Exploring  the hub genes and potential mechanisms of complement system-related genes in Parkinson disease: based on transcriptome sequencing and Mendelian randomization

doi:10.21203/rs.3.rs-4704744/v1

Download PDF

Research Article

Exploring the hub genes and potential mechanisms of complement system-related genes in Parkinson disease: based on transcriptome sequencing and Mendelian randomization

https://doi.org/10.21203/rs.3.rs-4704744/v1

This work is licensed under a CC BY 4.0 License

You are reading this latest preprint version

An accurate diagnosis of Parkinson's disease (PD) remains challenging and the exact cause of the disease is unclean. The aims are to identify hub genes associated with the complement system in PD and to explore their underlying molecular mechanisms. Initially, differentially expressed genes (DEGs) and key module genes related to PD were mined through differential expression analysis and WGCNA. Then, differentially expressed CSRGs (DE-CSRGs) were obtained by intersecting the DEGs, key module genes and CSRGs. Subsequently, MR analysis was executed to identify genes causally associated with PD. Based on genes with significant MR results, the expression level and diagnostic performance verification were achieved to yield hub genes. Functional enrichment and immune infiltration analyses were accomplished to insight into the pathogenesis of PD. qRT-PCR was employed to evaluate the expression levels of hub genes. After MR analysis and related verification, CD93, CTSS, PRKCD and TLR2 were finally identified as hub genes. Enrichment analysis indicated that the main enriched pathways for hub genes. Immune infiltration analysis found that the hub genes showed significant correlation with a variety of immune cells (such as myeloid-derived suppressor cell and macrophage). In the qRT-PCR results, the expression levels of CTSS, PRKCD and TLR2 were consistent with those we obtained from public databases. Hence, we mined four hub genes associated with complement system in PD which provided novel perspectives for the diagnosis and treatment of PD.

Parkinson’s disease

Complement system

MR analysis

Hub genes

Molecular mechanisms

Parkinson’s disease (PD) is a very complex and common progressive neurodegenerative disorder, the clinical manifestations mainly include motor symptoms of bradykinesia, rest tremor, and rigidity as well as changes in posture and gait and non-motor symptoms such as olfactory dysfunction, constipation, depression and rapid eye movement (REM) sleep behavior disorder (Armstrong and Okun 2020; Vos et al. 2017; Berg et al. 2013). PD affects more than 6 million individuals worldwide, including over 3.6 million in China (Qi et al. 2021), and its prevalence is expected to double over the next generation (Tolosa et al. 2021). There is no specific biomarker for PD, and the diagnosis of PD is primarily based on history and physical examination. Unfortunately, there is currently no known cure or preventive treatment to halt the progression of the disease. The mechanism of PD is complex and involves many molecular mechanisms such as gene mutations, oxidative stress, α-synuclein, ferroptosis, mitochondrial abnormalities, and neuroinflammatory responses (X et al. 2023), but the exact pathogenesis of PD and accurate diagnosis still puzzle us. Since the disease was first reported in 1817 (Parkinson, 2002), we have gained some understanding of PD, but there are still many problems that plague us. Therefore, it is necessary to further explore new potential mechanisms or therapeutic targets to provide ideas for optimizing the clinical diagnosis and treatment of PD.

Since McGeer et al. first reported evidence of neuroinflammation in PD in 1988, increasing evidence supports neuroinflammation as one of the key factors of PD, being associated with its onset and progression (Pajares et al. 2020; Hammond, Marsh, and Stevens 2019; Araújo et al. 2022). Although inflammation in PD is widely accepted, we are still unsure how neuroinflammation occurs. Complement system as a critical effector of both adaptive and innate immunity, plays a crucial role in the neuroinflammatory response. In addition to immune function, the complement system have many functions in the central nervous system (CNS) : promote the proliferation, differentiation, and migration of neural progenitor cells and eliminate excess neurotransmitters, aging proteins, and synapses (Dalakas, Alexopoulos, and Spaeth 2020; Ratajczak and Ratajczak 2021). Many studies have confirmed that dysregulation of complement system in CNS was related to many diseases, such as neurodevelopmental defects, cerebrovascular, chronic neurodegenerative, neurotrauma and other neurological diseases (Morgan 2018; Lee, Coulthard, and Woodruff 2019; Ratajczak and Ratajczak 2021). The potential mechanisms and therapeutic targets of the complement system in PD have attracted more and more attention from researchers. It has been reported that changes in proteins and cytokines related to the complement system are related to PD (Khosousi et al. 2023; Ma et al. 2021; L. Zhang et al. 2022; H.-C. Wu et al. 2021; Goldknopf et al. 2006).

It is widely understood that gene regulation controls the expression of cytokines and proteins. Therefore, we speculate that complement system-related genes (CSRG) may also affect the inflammatory response in PD and participate in the pathological response process of PD. Nevertheless, there is a dearth of research exploring the connection between PD and CSRGs at the genetic level. Mendelian randomization (MR) is a statistical framework that uses genetic variation as a statistical tool for the effect of exposure on an outcome of interest, minimizing selection bias (Davey Smith and Ebrahim 2003). MR has been widely used in research on PD (Domenighetti et al. 2022; Domínguez-Baleón et al. 2021; Gan-Or and Wood 2021; Li et al. 2022). But there are few studies using MR methods to study the relationship between the CSRGs and PD.

Therefore, in this study, we intend to use transcriptome sequencing and MR analysis to explore the possibility of CSRGs as potential hub genes in PD. Further explore the potential regulatory mechanisms of these hub genes in PD to provide new references for the clinical diagnosis and treatment of PD.

2.1 Raw data acquisition

Two PD-related datasets (GSE100054 and GSE165082) were downloaded from Gene Expression Omnibus (GEO) database (http://www.ncbi.nlm.nih.gov/geo/). Specifically, the GSE100054 (GPL23126) included 10 PD samples and 9 control samples, and the GSE165082 (GPL11154) included 12 PD samples and 14 control samples. The sample type in both datasets was peripheral blood. Furthermore, complement system-related genes (CSRGs) were collected in the Human Biology Pathway Unification (https://pathcards.genecards.org/), HUGO Gene Nomenclature Committee (https://www.gennames.org) and Molecular Signatures Database (MSigDB, http://software.broadInstitute.org/gsea/msigdb/index.jsp) using “complement” as the keyword, resulting 414 CSRGs were obtained by eliminating duplicated genes.For Mendelian randomization (MR) analysis, PD-related genome-wide association study (GWAS) dataset (ieu-b-7) was acquired from Integrated Epidemiology Unit (IEU) Open GWAS database (https://gwas.mrcieu.ac.uk/), including 17,891,936 single nucleotide polymorphisms (SNPs) from 482,730 samples.

2.2 Identification of differentially expressed gene (DEGs)

In GSE100054 dataset, differential expression analysis was carried out to generate DEGs between PD and control groups via “limma” package (version 1.38.0) (Ritchie et al. 2015), with the screening criteria of |log₂FoldChange (FC)| > 0.5, P < 0.05. To demonstrate the DEGs, the volcano map and heat map were plotted via “ggplot2” package (version 3.4.1) (Gustavsson et al. 2022) and “ComplexHeatmap” package (version 2.14.0) (Gu, Eils, and Schlesner 2016), respectively.

2.3 Weighted gene co-expression network analysis (WGCNA)

To screen for key module genes associated with PD traits, WGCNA analysis was performed using the expression data of GSE100054 dataset via ‘WGCNA’ package (version 1.71) (Langfelder and Horvath 2008). Firstly, sample clustering analysis was performed to remove outlier samples. Secondly, the best soft threshold was selected to ensure that the interactions between genes were maximally consistent with a scale-free distribution. Subsequently, the phylogenetic tree of genes was created based on the similarity of the genes, followed by a dynamic tree-cutting algorithm to separate the modules with a minimum module size of 50. Finally, gene modules with significant correlation with PD (P < 0.05) were used as key modules for subsequent analysis, and the genes in key modules were defined as key module genes.

2.4 Identification of the DE-CSRGs and functional enrichment analysis

The differentially expressed CSRGs (DE-CSRGs) were obtained by intersecting the DEGs, key module genes and CSRGs using the “ggvenn” package (version 0.1.9)(Ito and Murphy 2013). In addition, to better understand the potential biological roles of DE-CSRGs, Gene Ontology (GO) annotation and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analyses were performed using the “clusterProfiler” package (version 4.7.1.3) (T. Wu et al. 2021), with P_adj < 0.05 as threshold.

2.5 Screening of instrumental variables (IVs) in MR analysis

In this study, MR analysis was implemented with DE-CSRGs as the exposure factors and PD as the outcome to identify DE-CSRGs causally associated with PD. Firstly, the expression quantitative trait Loci (eQTL) data for DE-CSRGs were derived from IEU-OpenGWAS database (https://gwas.mrcieu.ac.uk/).

Three fundamental assumptions underpinned MR design: (1) IVs were strongly correlated with exposure factors. (2) IVs were independent from confounding factor. (3) IVs could only influence outcome through exposure factors. We screened the IVs based on the above three assumptions by the following steps. The extract_instruments function of "TwoSampleMR" package (version 0.5.6) (Hemani et al. 2018) was used to read the exposure factors and screen the IVs, as well as SNPs significantly associated with the exposure factors were screened as IVs with the threshold of P < 5 × 10^− 8. Then, IVs with linkage disequilibrium (LD) were removed (r² = 0.001 and kb = 50). Secondly, based on the GWAS data of outcome, the IVs associated with the PD were removed, as well as harmonisation of effect alleles and effect sizes were accomplished through harmonise_data function of “TwoSampleMR” package (version 0.5.6) (Hemani et al. 2018).

2.6 MR analysis

The “TwoSampleMR” (version 0.5.6) (Hemani et al. 2018) was used in combination with five algorithms, MR Egger (Bowden, Davey Smith, and Burgess 2015), weighted median (Bowden et al. 2016), inverse variance weighted (IVW) (Burgess et al. 2015), simple mode (Hemani et al. 2018) and weighted mode (Hartwig, Davey Smith, and Bowden 2017), to perform MR analysis in order to determine the causal relationship between DE-CSRGs and PD (P < 0.05). Meanwhile, odds ratio (OR) > 1 reflected DE-CSRGs as a risk factor for PD, while OR < 1 revealed DE-CSRGs as a protective factor for PD. The results were visualized using scatter plots, forest plots, and funnel plots. Subsequently, in order to evaluate the reliability of the MR analysis, the sensitivity analysis were performed, mainly including heterogeneity test, horizontal pleiotropy test and leave-one-out (LOO) analysis. DE-CSRGs causally associated with PD that underwent univariate MR analysis and sensitivity analysis were defined as candidate genes. Finally, multivariate MR analyses of the candidate genes were performed to investigate the causal relationship between them when the candidate genes acted together.

2.7 Identification of hub genes and correlation analysis

In order to further screen hub genes, expression validation was performed, as well as gene with significantly different (PD vs. control) and consistently expressed in GSE100054 and GSE165082 datasets were screened as hub genes. The results were visualized using box plots. The ability of hub genes in diagnosing PD was evaluated by receiver operating characteristic (ROC) curves and the area under curve (AUC) value was calculated (AUC > 0.7),which was accomplished via “pROC” package (version 1.18.0) (Robin et al. 2011). Furthermore, to determine the ability of genes to predict disease, the artificial neural network (ANN) was constructed and its effectiveness was judged by plotting ROC curve. Further, Spearman correlation analysis was completed to reveal the relationship between hub genes.

2.8 Chromosome localisation analysis and enrichment analysis of hub genes

The distribution of hub genes on chromosomes was visualized by the “Circos” (version 1.2.2) (Rasche and Hiltemann 2020). In addition, in order to elucidate the biological pathways involving hub genes, we executed gene set enrichment analysis (GSEA). Firstly, the samples of GSE100054 were divided into high- and low-expression groups based on the expression levels of hub genes, and the differential expression analysis was performed using the “limma” (version 1.38.0) (Ritchie et al. 2015), followed by sorting the differentially expressed genes in ascending order according to log₂FC. Then, based on the KEGG background gene set (c2.cp.v2023.1.Hs.symbols.gmt) in MSigDB (https://www.gsea-msigdb.org/), and the “clusterProfiler” (version 4.7.1.3) (T. Wu et al. 2021) was used to perform the GSEA with |NES| > 1 and P < 0.05 as the thresholds. Meanwhile, the GeneMANIA (http://genemania.org) was used to predict functionally similar genes of hub genes and construct the co-expression network among them.

2.9 Immune infiltration analysis

The ssGSEA analysis was performed by “GSVA” (version 1.42.0) (Hänzelmann, Castelo, and Guinney 2013) to calculate the enrichment fraction of 28 immune infiltrating cells in samples of GSE100054 dataset. Then, differences in immune cell infiltration between PD and control groups were analyzed using Wilcoxon-test (P < 0.05). Correlations between immune infiltrating cells were analyzed using the “corrplot” (version 0.92) (Z et al. 2022). Subsequently, the Spearman-test was used to explore the correlation between hub genes and differential immune infiltrating cells, which was considered significant when |cor|> 0.4 and P < 0.05 were satisfied.

2.10 Construction of regulate network and drug predictive

In this study, the upstream transcription factors (TFs) of hub genes were predicted applying the ENCODE database (https://www.encodeproject.org/). The miRTarBase (https://mirtarbase.cuhk.edu.cn/) and TarBase (https://www.tarbase.com/) were used to predict miRNAs that might regulate hub genes. Immediately thereafter, upstream lncRNAs of miRNAs were collected through the miRNet database (https://www.mirnet.ca/). Finally, the TF-mRNA network and the lncRNA-miRNA-mRNA network were constructed by “Cytoscape” (version 3.10.0) (Reimand et al. 2019).

In addition, DrugBank (https://go.drugbank.com/) and DGIdb (https://dgidb.org/) databases were used to predict potential drugs for hub genes in the treatment of PD, respectively. The hub genes and drug correspondences were visualized by the “Cytoscape” (version 3.10.0) (Reimand et al. 2019)

2.11 Quantitative real-time polymerase chain reaction (qRT-PCR)

Totally 10 matching blood samples were collected healthy subjects and PD patients in the Department of Neurology of the Second Affiliated Hospital of Chengdu Medical College (China National Nuclear Corporation 416 Hospital). Samples were collected from healthy control group and PD patient group. We rigorously select the PD patients who are closely matched for age, disease progression, and treatment regimen. Furthermore, all participants were diagnosed with PD only and without any other neurological or comorbid conditions. The details of these cases were shown in Supplementary Table 1. Total RNA was extracted from blood samples using the Trizol reagent, and the RNA concentration was measured after dissolving in DEPC water. The RNA was reverse transcribed by SureScript-First-strand-cDNA-synthesis-kit (Servicebio, China) to obtain cDNA. Afterwards, the qPCR reaction was performed by CFX96 real-time quantitative fluorescence PCR instrument (Bio-Rad, China) according to the following protocol: 95°C, 1min pre-denaturation; 95°C, 20s denaturation, 55°C, 20s annealing and 72°C, 30s extension, with the first three steps performed for 40 cycles. The GAPDH was employed as an internal reference gene, and the relative expression of hub genes was calculated using 2-^ΔΔCt method. Three replicates were set up for all system analysis. The primer sequences were shown in Table 1 .

Table 1

Primer sequence for RT-q PCR
Genes	Primer sequence (5’ − 3’)
CD93	F: TGATTGCTCTCACAGCCCAG
CD93	R: CACAGAGGGGTTCCATGACC
CTSS	F: TTTGTGTGCTCTTGGTGTGC
CTSS	R: GGGAACTCTCAGGGAACTCATC
PRKCD	F: AGCCAGGACTAAGGACAAGC
PRKCD	R: ATCTTCACGGCACAGAAGGG
TLR2	F: CCCAGGAGCTCTTAGTGACC
TLR2	R: CTGTGAGCCCTGAGGGAATG
GAPDH	F: CGAAGGTGGAGTCAACGGATTT
GAPDH	R: ATGGGTGGAATCATATTGGAAC

This study was approved by the Ethics Committee of the Second Affiliated Hospital of Chengdu Medical College (China National Nuclear Corporation 416 Hospital) (Approval No: YJ-2023-037 ).

2.12 Statistical analysis

All analyses were carried out in R language. If not specified above, the difference was considered statistically significant when P < 0.05.

3.1 Totally 58 DE-CSRGs were acquired

A total of 1,186 DEGs were acquired between PD and control groups from GSE100054, including 784 up-regulated and 402 down-regulated genes (Fig. 1A-B). Then, the sample clustering analysis showed no outlier samples in GSE100054 dataset (Fig. 1C). Based on the scale-free topology criterion with R² = 0.85, the soft-threshold was chosen as 12 (Fig. 1D). The adjacency matrix was converted into a TOM matrix, which was used to show the similarity between nodes by considering the weighted correlation. Next, eight modules were obtained through dynamic tree clipping (Fig. 1E), and the heat map found that the MEblue (Cor = 0.58, P = 0.01) and MEyellow (Cor = 0.55, P = 0.01) had significant positive correlation with PD. The two modules were selected as key modules, containing 2,139 key module genes (Fig. 1F). Finally, the 58 DE-CSRGs were obtained by taking the intersection of 1,186 DEGs, 2,139 key module genes and 414 CSRGs (Fig. 1G).

3.2 DE-CSRGs regulated PD through multiple biological pathways

Totally 58 DE-CSRGs were analyzed for functional enrichment. The GO showed that a total of 551 items were enriched, including 441 biological processes (BPs), 55 cellular components (CCs) and 55 molecular functions (MFs). In terms of BP, DE-CSRGs were significantly associated with blood coagulation, coagulation, hemostasis, wound healing and regulation of body fluid levels. In terms of CC, DE-CSRGs were significantly enriched in ficolin-1-rich granule, secretory granules lumen, tertiary granule, cytoplasmic vesicle lumen and vesicle lumen. In terms of MF, DE-CSRGs were significantly related to integrin binding, extracellular matrix binding, RAGE receptor binding, complement receptor activity and Toll-like receptor binding (Fig. 2A). The KEGG showed that 186 related pathways were obtained, of which 5 were most significant (P.adj < 0.05), including complement and coagulation cascades, neutrophil extracellular trap formation, malaria, platelet activation and tuberculosis (Fig. 2B).

3.3 CD93, TLR2, CTSS, PRKCD, and CR1L were causally associated with PD

After screening, 38 DE-CSRGs of the 58 DE-CSRGs had three or more SNPs, thus these 38 DE-CSRGs were incorporated into the subsequent MR analysis. Five algorithms were applied to perform univariate MR analysis of the 38 DE-CSRGs as exposure factors and PD as an outcome, respectively. The results showed that CD93 (P = 0.005, OR = 1.045), TLR2 (P = 0.022, OR = 1.178), CTSS (P = 0.010, OR = 0.970), PRKCD (P < 0.001, OR = 1.333), ITGAM (P < 0.001, OR = 1.479) and CR1L (P < 0.001, OR = 0.85) were significant causal relationship with PD. According to the OR value, CD93, TLR2, PRKCD, and ITGAM were risk factors for PD, whereas CTSS and CR1L were protective factors. Details about the MR analysis on significant causal exposure factors and PD were shown in Table 2. Scatter plots and forest plots further support our conclusions above, and funnel plots showed that MR analysis conform to Mendel's second law of random grouping (Supplementary Figs. 1–3).

Table 2

MR analysis on significant causal exposure factors and PD
exposure	outcome	gene	nSNP	pval	OR
eqtl-a-ENSG00000125810	Parkinson's disease\|\| id:ieu-b-7	CD93	52	0.005213979	1.044739
eqtl-a-ENSG00000137462	Parkinson's disease \|\| id:ieu-b-7	TLR2	4	0.021604242	1.178478
eqtl-a-ENSG00000163131	Parkinson's disease \|\| id:ieu-b-7	CTSS	49	0.010088317	0.970339
eqtl-a-ENSG00000163932	Parkinson's disease \|\| id:ieu-b-7	PRKCD	16	0.000000000531	1.332504
eqtl-a-ENSG00000169896	Parkinson's disease \|\| id:ieu-b-7	ITGAM	13	0.000184669	1.47872
eqtl-a-ENSG00000197721	Parkinson's disease \|\| id:ieu-b-7	CR1L	18	0.00000427	0.851668
OR value, 1 as the boundary, greater than 1 as risk factor, less than 1 as safety factor; Abbreviations: MR- Mendelian randomization; SNP - single nucleotide polymorphism

Furthermore, the sensitivity analysis was performed on the MR analysis results. The Cochran's Q-test results showed that no heterogeneity was found in the MR analyses of CD93, TLR2, CTSS, PRKCD, and CR1L (P > 0.05), whereas there was heterogeneity in the MR analysis of ITGAM (P < 0.05) (Table 3). Meanwhile, the above five genes were all passed the pleiotropy test (P > 0.05), means that there was no horizontal pleiotropy (Table 4). In the LOO analysis, the effect of the remaining SNPs on the outcome didn’t change significantly by gradually eliminating each SNP, indicating that the results of MR analysis were reliable and stable (Supplementary Fig. 4). Therefore, CD93, TLR2, CTSS, PRKCD, and CR1L were selected as candidate genes for subsequent analyses.

Table 3

MR heterogeneity test and analysis
exposure	outcome	Symbol	Q	Q_df	Q_pval
eqtl-a-ENSG00000125810	Parkinson's disease \|\| id:ieu-b-7	CD93	41.6124	50	0.79479
eqtl-a-ENSG00000137462	Parkinson's disease \|\| id:ieu-b-7	TLR2	1.358926	3	0.71519
eqtl-a-ENSG00000163131	Parkinson's disease \|\| id:ieu-b-7	CTSS	55.60518	45	0.133602
eqtl-a-ENSG00000163932	Parkinson's disease \|\| id:ieu-b-7	PRKCD	17.67404	14	0.222032
eqtl-a-ENSG00000169896	Parkinson's disease \|\| id:ieu-b-7	ITGAM	85.27118	12	<0.001
eqtl-a-ENSG00000197721	Parkinson's disease \|\| id:ieu-b-7	CR1L	6.066947	15	0.978601
Q_pval value greater than 0.05 indicates that there is no heterogeneity among the samples

Table 4

The horizontal pleiotropy test for five candidate exposure factors
exposure	outcome	Symbol	egger_intercept	se	p
eqtl-a-ENSG00000125810	Parkinson's disease \|\| id:ieu-b-7	CD93	0.004353	0.007369	0.557445
eqtl-a-ENSG00000137462	Parkinson's disease \|\| id:ieu-b-7	TLR2	0.043354	0.041348	0.40442
eqtl-a-ENSG00000163131	Parkinson's disease \|\| id:ieu-b-7	CTSS	-0.00091	0.0103	0.929632
eqtl-a-ENSG00000163932	Parkinson's disease \|\| id:ieu-b-7	PRKCD	0.022947	0.022314	0.322539
eqtl-a-ENSG00000197721	Parkinson's disease \|\| id:ieu-b-7	CR1L	-0.00894	0.027131	0.746666
p>0.05 indicates that there was no horizontal pleiotropy.

3.4 Diagnostic value of CD93, CTSS, PRKCD and TLR2

The expression of candidate genes was verified in GSE100054 and GSE165082 datasets. The boxplot showed that the the expression of CD93, CTSS, PRKCD and TLR2 was consistent in the two datasets and significantly different between PD and control groups (Figs. 3A-3B). Meanwhile, in the ROC curves, the AUC values of CD93, CTSS, PRKCD, and TLR2 were all greater than 0.7 with respect to two datasets (Figs. 3C-3D), showing that the four candidate genes had better predictive power in PD. Therefore, CD93, CTSS, PRKCD, and TLR2 were chosen as the hub genes for this study. Furthermore, based on these four hub genes, we constructed an ANN model in GSE100054 dataset to further diagnose PD (Fig. 3E), and the ROC curve results indicated that the AUC value of ANN model was 0.922, implying that the diagnostic effect of the ANN model was excellent (Fig. 3F).

Figure. 3 Evaluation of genes expression and diagnostic performance. A The box line diagram of the expression levels of candidate genes in GSE100054. B The box line diagram of the expression levels of candidate genes in GSE165082. C The ROC curves of the four candidate genes in GSE100054. D The ROC curves of the four candidate genes in GSE165082. E The ANN model of the four hub gens in GSE100054. F The ROC curve of the AUC value of ANN model.

3.5 Prediction of relevant pathways and functions of four hub genes

Chromosome localisation results showed that CTSS, PRKCD, TLR2 and CD93 were located in chromosomes 1, 3, 4 and 20, respectively (Fig. 4A). Simultaneously, the four hub genes were significantly positively correlated with each other (cor > 0.4, P < 0.05), among which there was the strongest positive correlation between PRKCD and CD93 (cor = 0.858) (Fig. 4B).

We explored the biological pathways enriched for four hub genes through GSEA. The results revealed that all four hub genes regulated a variety of immune responses, inflammatory factors, cytokines, and complement system-related pathways, including “signaling by interleukins”, “Toll-like receptor cascades”, “MHC class II antigen presentation”, “complement system”, “TYROBP causal network in microglia”, “IRAK-4 deficiency TLR2/4”, “cytokine-cytokine receptor interaction”, etc (Figs. 5A-5D). These analyses revealed these pathways were closely correlated with the pathogenesis of PD.

Furthermore, GeneMANIA results showed that the co-expression network and potential functions of four hub genes and their 20 reciprocal genes, and they were primarily linked to interleukin-6 production, pattern recognition receptor signaling pathway, regulation of tumor necrosis factor production, regulation of tumor necrosis factor superfamily cytokine production, and toll-like receptor signaling pathways (Fig. 5E)

3.6 CD93, CTSS, PRKCD and TLR2 were significantly correlated with immune cells

The enrichment scores of 28 immune infiltrating cells for all samples in GSE100054 were calculated, as shown in Fig. 6A, in which the enrichment score of myeloid-derived suppressor cell (MDSC) was the highest. Boxplots were drawn between PD and controls samples, which indicated significant differences (P < 0.05) in 17 immune cells, including activated dendritic cell, CD56^bright natural killer cell and macrophage, etc (Fig. 6B). In addition, the heat map was drawn to analyze the correlation between differential immune cells, as shown that MDSC and macrophage had the highest correlation with a correlation coefficient of 0.911, and activated CD8 + T cell and activated dendritic cell were most negative correlation with a correlation coefficient of -0.712 (Fig. 6C). Finally, the correlation between four hub genes and 17 differential immune cells was explored, showing that all four hub genes presented significant positive correlations with regulatory T (Treg) cells, macrophage, and activated dendritic cells, as well as significant negative correlations with activated CD8 T cells (Figs. 6D-6G).

3.7 The hub gene was controlled via several ceRNAs and TFs in PD

Totally 5 TFs of regulatory hub genes were predicted based on ENCODE database, of which four TFs (NFKB1, HIF1A, SP1 and RELA) were regulators of TLR2, while only one TF (SPI1) was a regulator of CTSS, and TF-mRNA network was shown in Fig. 7A In the samples of GSE100054 dataset, HIF1A was had higher expression, whereas SPI1 had lower expression (Fig. 7B). Additionally, totally 16 potential miRNAs regulating hub genes were predicted through miRTarBase and TarBase databases, immediately followed by 309 upstream lncRNAs of miRNAs were collected through miRNet. Subsequently, the lncRNA-miRNA-mRNA network was constructed (Fig. 7C), containing 4 mRNAs, 16 miRNAs and 309 lncRNAs. The multiple relationship pairs were found in network, such as lncRNAs (CCDC18-AS1, IL6R-AS1, BLACAT1, PINK1-AS, etc.) could simultaneously regulate two hub genes (CTSS and TLR2) via both hsa-mir-106b-5p, lncRNAs (GAS5, MIR29B2CHG, etc.) could regulate CD93 via hsa-mir-29a-3p, and lncRNAs (LINC01128, TNFRSF14-AS1, etc.) could simultaneously regulate PRKCD via miRNAs (hsa-mir-15a-5p, hsa-mir-16-5p and hsa-mir-195-5p).

In the DrugBank database, totally 6 drugs were eventually searched, including Copper and Hyaluronic acid targeting CD93, Adapalene and Tuberculin purified protein derivative targeting TLR2, Fostamatinib targeting both CTSS and PRKCD, as well as Ingenol mebutate likewise targeting PRKCD (Fig. 7D). Furthermore, based on DGIdb database, it was found 35 drugs that could potentially affect the hub genens. Of these, 1 drug (Petesicatib), 4 drugs (DiaPep277, Tomaralimab, ChEMBL 1836411, and Resveratrol hexanoic acid) and 30 drugs (Enoxolone, Rottlerin, Fasudil, etc.) targeting CTSS, TLR2 and PRKCD were predicted, respectively, unfortunately, our analyses did not yield predictions for drugs targeting CRISPLD2 (Fig. 7E).

3.8 Experimental validation of hub genes expression through qRT-PCR

Ultimately, we validated the four hub genes by qRT-PCR, and the result revealed that the mRNA expression of CTSS, PRKCD and TLR2 was significantly elevated in the PD samples (P < 0.05) (Figs. 8A-8C), which was consistent with the results from public databases, implicating that CTSS, PRKCD and TLR2 had regulatory role in PD. However, CD93 expression was not significantly different between PD and control samples (P > 0.05) (Figs. 8D), probably because of the small sample size, and we will perform further experiments to verify it afterwards.

PD is a prevalent neurodegenerative disease worldwide. However, the exact pathological mechanism of PD remains unclear and its treatment remains a big challenge. The complement system, which is a component of our immune mechanism, acts as a double-edged sword. Currently, the complement system has become an important direction in the research of neurodegenerative diseases. In this study, we used bioinformatics methods to screen CSRGs that may be related to PD, finally 4 genes (CD93, TLR2, CTSS, PRKCD) passed further verification, which were identified as the potential hub gens for PD.

Previous studies have reported on the relevant mechanisms of these hub genes, some of which involve biological mechanisms in PD. Such as Jae-Won Jeon et al. found that sCD93 can induce the production of IL-6, IL-1β and TNF-α and play a key role in stimulating monocytes and enhances TLR responses during inflammatory processes (Jeon et al. 2010; Bohlson et al. 2005). Besides, CTSS is mainly expressed in antigen-presenting cells (APCs) and has an important impact on MHC-II maturation, antigen processing and presentation (Hsing and Rudensky 2005). Then, Richard Gordon and his colleagues indicated that PRKCD encodes protein kinase C delta (PKCδ) plays a crucial role in microglial activation responses (Gordon et al. 2016; 2012). In addition, TLR2 is expressed in microglia and neurons and can participate in inflammatory responses through multiple pathways, such as the activation of microglia and the secretion of inflammatory factors such as IL-6 and TNF (Flynn et al. 2019; Da Rocha Sobrinho et al. 2021; Dzamko et al. 2017). It has been demonstrated that TLR-mediated neuroinflammation may contribute to dopaminergic nerve loss in PD patients. And TLR-2 has been linked to α-synuclein aggregation in PD, that suggested that TLR-2 was related to the disease progression of PD (Ho et al. 2022; He et al. 2023; Heidari, Yazdanpanah, and Rezaei 2022). CD93, PRKCD, and TLR2 may affect the pathophysiological process of PD by producing IL-6 or TNF or regulating TLR signaling pathways (Gordon et al. 2012; Flynn et al. 2019; Samidurai et al. 2020). In our research, the results of GSEA showed that all four hub genes regulated a variety of immune responses, inflammatory factors, cytokines, and complement system-related pathways, including “signaling by interleukins”, “Toll-like receptor cascades”, “MHC class II antigen presentation”, “IRAK-4 deficiency TLR2/4” etc. Further GeneMANIA analysis revealed that the hub gens were mostly engaged in IL-6 production, pattern recognition receptor signaling pathways, TNF regulation, TNF superfamily cytokine production, and Toll-like receptor signaling pathways. At the same time, chromosome mapping results showed that these four hub genes were significantly positively correlated with each other. Therefore, given the correlation of these four genes, we cautiously speculate that CD93, CTSS, and PRKCD may regulate the immune inflammatory response through TLR-2 and thereby affect the biological process of PD. However, a large number of related experiments are needed in the future to explore the specific mechanisms of these genes in PD.

To investigate the immune microenvironment in PD patients, immune infiltration analysis was conducted. Correlation analysis between immune infiltrating cells revealed a predominantly positive correlation among most cells, which may indicate that the inflammation observed in PD may extend beyond the CNS or be the result of a systemic inflammatory response. The relationship between these hub gens and the immune cells has rarely been studied in PD, but their relationship has been explored in other diseases. For example, Dheilly et al. reported CTSS regulates antigen processing and CD4⁺ and CD8⁺ T cell-mediated immune responses in Non-Hodgkin Lymphoma (Dheilly et al. 2020). Jeon et al. found the sCD93 induces differentiation of monocytes in rheumatoid arthritis (Jeon et al. 2010). And Beyer et al. used RNA sequencing to discover that CD93 is one of the surface markers of M2 macrophages (Beyer et al. 2012). Jiang et al. found that CD93 expression is positively correlated with the infiltration levels of B cells, CD8⁺T cells, CD4⁺T cells, macrophages, neutrophils and dendritic cells, and their results, and their results suggested that CD93 may be a predictor of clinical prognosis in hepatocellular carcinoma (Jiang et al. 2023). And Xu et al. reported the SPP1, PRKCD, and SLC11A1 genes have the strongest correlation with macrophages (Xu et al. 2021). TLR2 has been widely studied among these hub genes. Numerous research results suggest that TLR2 can affect the activity of various immune cells and participate in various inflammatory processes (Da Rocha Sobrinho et al. 2021; Ho et al. 2022; Heidari, Yazdanpanah, and Rezaei 2022; C.-C. Wu et al. 2016; Jafarzadeh et al. 2019; E. Zhang et al. 2019). These studies all suggest that these hub genes are related to immune cells or affect the function of immune cells, so we believe that they have immune regulatory functions in PD. However, the precise mechanism underlying the relationship between each gene and these immune cells in the neuroinflammation process of PD remains unclear and is worthy of further investigation.

MR analysis of the relationship between candidate genes and PD found that CD93, TLR2, PRKCD are risk factors for PD. As mentioned above, CD93, TLR2, and PRKCD can promote the inflammatory response in PD by activating inflammatory cells or stimulating related immune cells (Jeon et al. 2010; Gordon et al. 2016; E. Zhang et al. 2019; Griffiths et al. 2018). While, CTSS is a protective factor for PD. The lysosomal family plays a central role in clearing damaged cellular components and degrading and recycling misfolded proteins(such as α-synapsin, β-peptides) (Smyth et al. 2022; Yadati et al. 2020). Therefore, maintaining neuronal health relies heavily on lysosomal function. As a member of the lysosome family, we speculate that CTSS may play a protective role in PD by eliminating these abnormally aggregated proteins (Nübling et al. 2017). However, the specific mechanism requires further study. Thereafter, our qRT-PCR analysis showed that the expression of CTSS, PRKCD, and TLR2 mRNA was higher in PD patients than in healthy controls. It is suggested that these hub genes may have regulatory effects on PD. However, CD93 mRNA expression was not significantly higher than in healthy controls. The small sample size used in our study due to strict screening criteria may also be one of the reasons for the results.

Neuroinflammation plays a significant role in neurodegenerative diseases, with a specific focus on PD in this study. It is noteworthy that the hub genes identified in this research are also associated with other diseases like Alzheimer's disease (AD) and multiple sclerosis (MS). It is reported that inhibiting TLR2 can reduce the inflammatory response in diseases such as AD and MS (McDonald et al. 2016; Rangasamy et al. 2018; Wasko et al. 2019; Rubio-Araiz et al. 2018; Wasko, Nichols, and Clark 2020). Zhou et al. reported that noise exposure can increase hippocampal cell apoptosis in AD mice by upregulating CTSS (Zhou et al. 2023). In addition, studies have shown that PRKCD is associated with amyloid-β-induced neurodegeneration (Park et al. 2020). Our results may provide some new evidence for inflammatory responses in neurodegenerative diseases.

We also tried to explore the regulatory network of these hub genes and made drug predictions based on these hub genes. Our results found four related TFs for TLR2 and one related TF for CTSS. These TFs regulate the expression of these genes to a certain extent. Meanwhile, it is believed that miRNA dysregulation is one of the important causes of PD. To date, dozens of miRNAs have been found in the circulation of PD patients which were involved in various possible mechanisms of PD (Nies et al. 2021; Wang 2023; Ye et al. 2016; Ueta et al. 2023). Our study may contribute novel data to the existing miRNA database. Additionally, the Drug Bank and DGIdb databases were consulted to identify drugs targeting the hub genes. Some drugs have been reported to have potential therapeutic effects in PD (Tatenhorst et al. 2016; Maria Miranda Santos 2012). It is hoped that drugs predicted by the hub genes may bring opportunities for PD treatment in the future. However, this study still has some limitations. For instance, we have not conducted additional clinical and in vitro experiments to validate the mechanisms associated with these key genes. The sample size selected in this study is limited and may not represent the entire population. In the future, we will continue to focus on and investigate these genes and their related mechanisms

In summary, this study used a series of methods such as differential expression analysis, WGCNA, GSEA, MR and immune infiltration analysis to identify 4 hub genes(CD93, TLR2, CTSS and PRKCD)related to PD disease and CSRGs. We further analyzed the biological mechanisms, immune infiltration, regulatory network construction, and drug prediction of the hub genes. It is hoped that these results can provide new directions and some new ideas for future research.

PD -Parkinson's disease；MR- Mendelian randomization；CNS -central nervous system; CSRGs - complement system-related genes；DEGs - differentially expressed genes; DE-CSRGs -differentially expressed complement system-related genes; ROC -receiver operating characteristic; GEO- Gene Expression Omnibus; IEU - Integrated Epidemiology Unit; GWAS -genome-wide association study; SNPs - single nucleotide polymorphisms; WGCNA -Weighted gene co-expression network analysis ; GO- Gene Ontology; KEGG - Kyoto Encyclopedia of Genes and Genomes; eQTL- expression quantitative trait Loci; IVs - instrumental variables; LD - linkage disequilibrium; IVW -inverse variance weighted; OR- odds ratio; LOO - leave-one-out; AUC - area under curve ; ANN - artificial neural network; GSEA- gene set enrichment analysis; TFs - transcription factors; MFs - molecular functions; BPs- biological processes; CCs - cellular components; MDSC -myeloid-derived suppressor cell; Treg- regulatory T cell; ; APCs- antigen-presenting cells; PKCδ - protein kinase C delta.

Funding the Chengdu Health Commission Foundation (2022139).

Conflict of Interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Ethical approval This study was approved by the Ethics Committee of the Second Affiliated Hospital of Chengdu Medical College (China National Nuclear Corporation 416 Hospital) (Approval No: YJ-2023-037, Additional file 1: Ethical approval).

Informed consent All subjects provided informed consent.

Author contribution Xin Wang: Writing – original draft, Funding acquisition, Formal analysis, Data curation. Gaoming Yang: Formal analysis, Data curation, Methodology, Yali Lai: Formal analysis, Data curation. Yuanyuan Li: Formal analysis, Methodology. Xindong Liu: Writing – review & editing, Supervision, Conceptualization.

Data Availability Statement (Required) Data will be made available on request.

Acknowledgements We sincerely appreciated the data provided by GEO database.

Araújo, Bruna, Rita Caridade-Silva, Carla Soares-Guedes, Joana Martins-Macedo, Eduardo D. Gomes, Susana Monteiro, and Fábio G. Teixeira. 2022. “Neuroinflammation and Parkinson’s Disease—From Neurodegeneration to Therapeutic Opportunities.” Cells 11 (18): 2908. https://doi.org/10.3390/cells11182908.
Armstrong, Melissa J., and Michael S. Okun. 2020. “Diagnosis and Treatment of Parkinson Disease: A Review.” JAMA 323 (6): 548. https://doi.org/10.1001/jama.2019.22360.
Berg, Daniela, Anthony E Lang, Ronald B Postuma, Walter Maetzler, Guenther Deuschl, Thomas Gasser, Andrew Siderowf, et al. 2013. “Changing the Research Criteria for the Diagnosis of Parkinson’s Disease: Obstacles and Opportunities.” The Lancet Neurology 12 (5): 514–24. https://doi.org/10.1016/S1474-4422(13)70047-4.
Beyer, Marc, Michael R. Mallmann, Jia Xue, Andrea Staratschek-Jox, Daniela Vorholt, Wolfgang Krebs, Daniel Sommer, et al. 2012. “High-Resolution Transcriptome of Human Macrophages.” PLoS ONE 7 (9): e45466. https://doi.org/10.1371/journal.pone.0045466.
Bohlson, Suzanne S., Richard Silva, Maria I. Fonseca, and Andrea J. Tenner. 2005. “CD93 Is Rapidly Shed from the Surface of Human Myeloid Cells and the Soluble Form Is Detected in Human Plasma.” The Journal of Immunology 175 (2): 1239–47. https://doi.org/10.4049/jimmunol.175.2.1239.
Bowden, Jack, George Davey Smith, and Stephen Burgess. 2015. “Mendelian Randomization with Invalid Instruments: Effect Estimation and Bias Detection through Egger Regression.” International Journal of Epidemiology 44 (2): 512–25. https://doi.org/10.1093/ije/dyv080.
Bowden, Jack, George Davey Smith, Philip C. Haycock, and Stephen Burgess. 2016. “Consistent Estimation in Mendelian Randomization with Some Invalid Instruments Using a Weighted Median Estimator.” Genetic Epidemiology 40 (4): 304–14. https://doi.org/10.1002/gepi.21965.
Burgess, Stephen, Robert A. Scott, Nicholas J. Timpson, George Davey Smith, and Simon G. Thompson. 2015. “Using Published Data in Mendelian Randomization: A Blueprint for Efficient Identification of Causal Risk Factors.” European Journal of Epidemiology 30 (7): 543–52. https://doi.org/10.1007/s10654-015-0011-z.
Da Rocha Sobrinho, Hermínio Maurício, Rodrigo Saar Gomes, Delson José Da Silva, Valéria Bernadete Leite Quixabeira, Leo A. B. Joosten, Cristina Ribeiro De Barros Cardoso, and Fátima Ribeiro‐Dias. 2021. “Toll‐like Receptor 10 Controls TLR2‐induced Cytokine Production in Monocytes from Patients with Parkinson’s Disease.” Journal of Neuroscience Research 99 (10): 2511–24. https://doi.org/10.1002/jnr.24916.
Dalakas, Marinos C., Harry Alexopoulos, and Peter J. Spaeth. 2020. “Complement in Neurological Disorders and Emerging Complement-Targeted Therapeutics.” Nature Reviews Neurology 16 (11): 601–17. https://doi.org/10.1038/s41582-020-0400-0.
Davey Smith, George, and Shah Ebrahim. 2003. “‘Mendelian Randomization’: Can Genetic Epidemiology Contribute to Understanding Environmental Determinants of Disease?*.” International Journal of Epidemiology 32 (1): 1–22. https://doi.org/10.1093/ije/dyg070.
Dheilly, Elie, Elena Battistello, Natalya Katanayeva, Stephanie Sungalee, Justine Michaux, Gerben Duns, Sarah Wehrle, et al. 2020. “Cathepsin S Regulates Antigen Processing and T Cell Activity in Non-Hodgkin Lymphoma.” Cancer Cell 37 (5): 674-689.e12. https://doi.org/10.1016/j.ccell.2020.03.016.
Domenighetti, Cloé, Pierre‐Emmanuel Sugier, Ashwin Ashok Kumar Sreelatha, Claudia Schulte, Sandeep Grover, Océane Mohamed, Berta Portugal, et al. 2022. “Dairy Intake and Parkinson’s Disease: A Mendelian Randomization Study.” Movement Disorders 37 (4): 857–64. https://doi.org/10.1002/mds.28902.
Domínguez-Baleón, Carmen, Jue-Sheng Ong, Clemens R. Scherzer, Miguel E. Rentería, and Xianjun Dong. 2021. “Understanding the Effect of Smoking and Drinking Behavior on Parkinson’s Disease Risk: A Mendelian Randomization Study.” Scientific Reports 11 (July):13980. https://doi.org/10.1038/s41598-021-93105-y.
Dzamko, Nicolas, Amanda Gysbers, Gayathri Perera, Anita Bahar, Amrita Shankar, Jianqun Gao, YuHong Fu, and Glenda M. Halliday. 2017. “Toll-like Receptor 2 Is Increased in Neurons in Parkinson’s Disease Brain and May Contribute to Alpha-Synuclein Pathology.” Acta Neuropathologica 133 (2): 303–19. https://doi.org/10.1007/s00401-016-1648-8.
Flynn, Charlotte M., Yvonne Garbers, Juliane Lokau, Daniela Wesch, Dominik M. Schulte, Matthias Laudes, Wolfgang Lieb, Samadhi Aparicio-Siegmund, and Christoph Garbers. 2019. “Activation of Toll-like Receptor 2 (TLR2) Induces Interleukin-6 Trans-Signaling.” Scientific Reports 9 (May):7306. https://doi.org/10.1038/s41598-019-43617-5.
Gan‐Or, Ziv, and Nicholas W. Wood. 2021. “Mendelian Randomization Studies: A Path to Better Understand Sex and Gender Differences in Parkinson’s Disease?” Movement Disorders 36 (10): 2220–22. https://doi.org/10.1002/mds.28765.
Goldknopf, Ira L., Essam A. Sheta, Jennifer Bryson, Brian Folsom, Chris Wilson, Jeff Duty, Albert A. Yen, and Stanley H. Appel. 2006. “Complement C3c and Related Protein Biomarkers in Amyotrophic Lateral Sclerosis and Parkinson’s Disease.” Biochemical and Biophysical Research Communications 342 (4): 1034–39. https://doi.org/10.1016/j.bbrc.2006.02.051.
Gordon, Richard, Vellareddy Anantharam, Anumantha G Kanthasamy, and Arthi Kanthasamy. 2012. “Proteolytic Activation of Proapoptotic Kinase Protein Kinase Cδ by Tumor Necrosis Factor α Death Receptor Signaling in Dopaminergic Neurons during Neuroinflammation.” Journal of Neuroinflammation 9 (April):82. https://doi.org/10.1186/1742-2094-9-82.
Gordon, Richard, Neeraj Singh, Vivek Lawana, Anamitra Ghosh, Dilshan S. Harischandra, Huajun Jin, Colleen Hogan, et al. 2016. “Protein Kinase Cδ Upregulation in Microglia Drives Neuroinflammatory Responses and Dopaminergic Neurodegeneration in Experimental Models of Parkinson’s Disease.” Neurobiology of Disease 93 (September):96–114. https://doi.org/10.1016/j.nbd.2016.04.008.
Griffiths, Mark R., Marina Botto, Bryan Paul Morgan, James W. Neal, and Philippe Gasque. 2018. “CD93 Regulates Central Nervous System Inflammation in Two Mouse Models of Autoimmune Encephalomyelitis.” Immunology 155 (3): 346–55. https://doi.org/10.1111/imm.12974.
Gu, Zuguang, Roland Eils, and Matthias Schlesner. 2016. “Complex Heatmaps Reveal Patterns and Correlations in Multidimensional Genomic Data.” Bioinformatics 32 (18): 2847–49. https://doi.org/10.1093/bioinformatics/btw313.
Gustavsson, Emil K, David Zhang, Regina H Reynolds, Sonia Garcia-Ruiz, and Mina Ryten. 2022. “Ggtranscript: An R Package for the Visualization and Interpretation of Transcript Isoforms Using Ggplot2.” Bioinformatics 38 (15): 3844–46. https://doi.org/10.1093/bioinformatics/btac409.
Hammond, Timothy R., Samuel E. Marsh, and Beth Stevens. 2019. “Immune Signaling in Neurodegeneration.” Immunity 50 (4): 955–74. https://doi.org/10.1016/j.immuni.2019.03.016.
Hänzelmann, Sonja, Robert Castelo, and Justin Guinney. 2013. “GSVA: Gene Set Variation Analysis for Microarray and RNA-Seq Data.” BMC Bioinformatics 14 (January):7. https://doi.org/10.1186/1471-2105-14-7.
Hartwig, Fernando Pires, George Davey Smith, and Jack Bowden. 2017. “Robust Inference in Summary Data Mendelian Randomization via the Zero Modal Pleiotropy Assumption.” International Journal of Epidemiology 46 (6): 1985–98. https://doi.org/10.1093/ije/dyx102.
He, Yongtao, Jiayin Zhao, Hongtian Dong, Xiaoshuang Zhang, Yufei Duan, Yuanyuan Ma, Mei Yu, Jian Fei, and Fang Huang. 2023. “TLR2 Deficiency Is Beneficial at the Late Phase in MPTP-Induced Parkinson’ Disease Mice.” Life Sciences 333 (November):122171. https://doi.org/10.1016/j.lfs.2023.122171.
Heidari, Arash, Niloufar Yazdanpanah, and Nima Rezaei. 2022. “The Role of Toll-like Receptors and Neuroinflammation in Parkinson’s Disease.” Journal of Neuroinflammation 19 (June):135. https://doi.org/10.1186/s12974-022-02496-w.
Hemani, Gibran, Jie Zheng, Benjamin Elsworth, Kaitlin H Wade, Valeriia Haberland, Denis Baird, Charles Laurin, et al. 2018. “The MR-Base Platform Supports Systematic Causal Inference across the Human Phenome.” eLife 7 (May):e34408. https://doi.org/10.7554/eLife.34408.
Ho, Dong-Hwan, Daleum Nam, Mikyoung Seo, Sung-Woo Park, Wongi Seol, and Ilhong Son. 2022. “LRRK2 Inhibition Mitigates the Neuroinflammation Caused by TLR2-Specific α-Synuclein and Alleviates Neuroinflammation-Derived Dopaminergic Neuronal Loss.” Cells 11 (5): 861. https://doi.org/10.3390/cells11050861.
Hsing, Lianne C., and Alexander Y. Rudensky. 2005. “The Lysosomal Cysteine Proteases in MHC Class II Antigen Presentation.” Immunological Reviews 207 (1): 229–41. https://doi.org/10.1111/j.0105-2896.2005.00310.x.
Ito, K., and D. Murphy. 2013. “Application of Ggplot2 to Pharmacometric Graphics.” CPT: Pharmacometrics & Systems Pharmacology 2 (10): e79. https://doi.org/10.1038/psp.2013.56.
Jafarzadeh, Abdollah, Maryam Nemati, Hossein Khorramdelazad, and Abbas Mirshafiey. 2019. “The Toll-like Receptor 2 (TLR2)-Related Immunopathological Responses in the Multiple Sclerosis and Experimental Autoimmune Encephalomyelitis.” Iranian Journal of Allergy, Asthma and Immunology, June, 230–50. https://doi.org/10.18502/ijaai.v18i3.1117.
Jeon, Jae-Won, Joon-Goo Jung, Eui-Cheol Shin, Hye In Choi, Ho Youn Kim, Mi-La Cho, Sun-Wha Kim, et al. 2010. “Soluble CD93 Induces Differentiation of Monocytes and Enhances TLR Responses.” The Journal of Immunology 185 (8): 4921–27. https://doi.org/10.4049/jimmunol.0904011.
Jiang, Qianwei, Jing Kuai, Zhongyi Jiang, Weitao Que, Pusen Wang, Wenxin Huang, Wei Ding, and Lin Zhong. 2023. “CD93 Overexpresses in Liver Hepatocellular Carcinoma and Represents a Potential Immunotherapy Target.” Frontiers in Immunology 14 (July):1158360. https://doi.org/10.3389/fimmu.2023.1158360.
Khosousi, Shervin, Abdul Hye, Latha Velayudhan, Björn Bloth, Panagiota Tsitsi, Ioanna Markaki, and Per Svenningsson. 2023. “Complement System Changes in Blood in Parkinson’s Disease and Progressive Supranuclear Palsy/Corticobasal Syndrome.” Parkinsonism & Related Disorders 108 (March):105313. https://doi.org/10.1016/j.parkreldis.2023.105313.
Langfelder, Peter, and Steve Horvath. 2008. “WGCNA: An R Package for Weighted Correlation Network Analysis.” BMC Bioinformatics 9 (December):559. https://doi.org/10.1186/1471-2105-9-559.
Lee, John D., Liam G. Coulthard, and Trent M. Woodruff. 2019. “Complement Dysregulation in the Central Nervous System during Development and Disease.” Seminars in Immunology 45 (October):101340. https://doi.org/10.1016/j.smim.2019.101340.
Li, Chunyu, Junyu Lin, Tianmi Yang, and Huifang Shang. 2022. “Green Tea Intake and Parkinson’s Disease Progression: A Mendelian Randomization Study.” Frontiers in Nutrition 9 (May):848223. https://doi.org/10.3389/fnut.2022.848223.
Ma, Shi-Xun, Bo Am Seo, Donghoon Kim, Yulan Xiong, Seung-Hwan Kwon, Saurav Brahmachari, Sangjune Kim, et al. 2021. “Complement and Coagulation Cascades Are Potentially Involved in Dopaminergic Neurodegeneration in α-Synuclein-Based Mouse Models of Parkinson’s Disease.” Journal of Proteome Research 20 (7): 3428–43. https://doi.org/10.1021/acs.jproteome.0c01002.
Maria Miranda Santos, Claudia. 2012. “New Agents Promote Neuroprotection in Parkinson’s Disease Models.” CNS & Neurological Disorders - Drug Targets 11 (4): 410–18. https://doi.org/10.2174/187152712800792820.
McDonald, Claire L., Edel Hennessy, Ana Rubio-Araiz, Brian Keogh, William McCormack, Peter McGuirk, Mary Reilly, and Marina A. Lynch. 2016. “Inhibiting TLR2 Activation Attenuates Amyloid Accumulation and Glial Activation in a Mouse Model of Alzheimer’s Disease.” Brain, Behavior, and Immunity 58 (November):191–200. https://doi.org/10.1016/j.bbi.2016.07.143.
Morgan, B. Paul. 2018. “Complement in the Pathogenesis of Alzheimer’s Disease.” Seminars in Immunopathology 40 (1): 113–24. https://doi.org/10.1007/s00281-017-0662-9.
Nies, Yong Hui, Nor Haliza Mohamad Najib, Wei Ling Lim, Mohd Amir Kamaruzzaman, Mohamad Fairuz Yahaya, and Seong Lin Teoh. 2021. “MicroRNA Dysregulation in Parkinson’s Disease: A Narrative Review.” Frontiers in Neuroscience 15 (April):660379. https://doi.org/10.3389/fnins.2021.660379.
Nübling, Georg, M. Schuberth, K. Feldmer, A. Giese, L. M. Holdt, D. Teupser, and S. Lorenzl. 2017. “Cathepsin S Increases Tau Oligomer Formation through Limited Cleavage, but Only IL-6, Not Cathespin S Serum Levels Correlate with Disease Severity in the Neurodegenerative Tauopathy Progressive Supranuclear Palsy.” Experimental Brain Research 235 (8): 2407–12. https://doi.org/10.1007/s00221-017-4978-4.
Pajares, Marta, Ana I. Rojo, Gina Manda, Lisardo Boscá, and Antonio Cuadrado. 2020. “Inflammation in Parkinson’s Disease: Mechanisms and Therapeutic Implications.” Cells 9 (7): 1687. https://doi.org/10.3390/cells9071687.
Park, Young Ho, Angela Hodges, Shannon L. Risacher, Kuang Lin, Jae-Won Jang, Soyeon Ahn, SangYun Kim, et al. 2020. “Dysregulated Fc Gamma Receptor-Mediated Phagocytosis Pathway in Alzheimer’s Disease: Network-Based Gene Expression Analysis.” Neurobiology of Aging 88 (April):24–32. https://doi.org/10.1016/j.neurobiolaging.2019.12.001.
Qi, Shige, Peng Yin, Linhong Wang, Ming Qu, Ge Lin Kan, Han Zhang, Qingjun Zhang, et al. 2021. “Prevalence of Parkinson’s Disease: A Community‐Based Study in China.” Movement Disorders 36 (12): 2940–44. https://doi.org/10.1002/mds.28762.
Rangasamy, Suresh B., Malabendu Jana, Avik Roy, Grant T. Corbett, Madhuchhanda Kundu, Sujyoti Chandra, Susanta Mondal, et al. 2018. “Selective Disruption of TLR2-MyD88 Interaction Inhibits Inflammation and Attenuates Alzheimer’s Pathology.” Journal of Clinical Investigation 128 (10): 4297–4312. https://doi.org/10.1172/JCI96209.
Rasche, Helena, and Saskia Hiltemann. 2020. “Galactic Circos: User-Friendly Circos Plots within the Galaxy Platform.” GigaScience 9 (6): giaa065. https://doi.org/10.1093/gigascience/giaa065.
Ratajczak, Mariusz Z., and Janina Ratajczak. 2021. “Innate Immunity Communicates Using the Language of Extracellular Microvesicles.” Stem Cell Reviews and Reports 17 (2): 502–10. https://doi.org/10.1007/s12015-021-10138-6.
Reimand, Jüri, Ruth Isser, Veronique Voisin, Mike Kucera, Christian Tannus-Lopes, Asha Rostamianfar, Lina Wadi, et al. 2019. “Pathway Enrichment Analysis and Visualization of Omics Data Using g:Profiler, GSEA, Cytoscape and EnrichmentMap.” Nature Protocols 14 (2): 482–517. https://doi.org/10.1038/s41596-018-0103-9.
Ritchie, Matthew E., Belinda Phipson, Di Wu, Yifang Hu, Charity W. Law, Wei Shi, and Gordon K. Smyth. 2015. “Limma Powers Differential Expression Analyses for RNA-Sequencing and Microarray Studies.” Nucleic Acids Research 43 (7): e47. https://doi.org/10.1093/nar/gkv007.
Robin, Xavier, Natacha Turck, Alexandre Hainard, Natalia Tiberti, Frédérique Lisacek, Jean-Charles Sanchez, and Markus Müller. 2011. “pROC: An Open-Source Package for R and S+ to Analyze and Compare ROC Curves.” BMC Bioinformatics 12 (March):77. https://doi.org/10.1186/1471-2105-12-77.
Rubio-Araiz, Ana, Orla M. Finucane, Samuel Keogh, and Marina A. Lynch. 2018. “Anti-TLR2 Antibody Triggers Oxidative Phosphorylation in Microglia and Increases Phagocytosis of β-Amyloid.” Journal of Neuroinflammation 15 (1): 247. https://doi.org/10.1186/s12974-018-1281-7.
Samidurai, Manikandan, Prashant Tarale, Chelva Janarthanam, Crystal Gomez Estrada, Richard Gordon, Gary Zenitsky, Huajun Jin, Vellareddy Anantharam, Anumantha G. Kanthasamy, and Arthi Kanthasamy. 2020. “Tumor Necrosis Factor-Like Weak Inducer of Apoptosis (TWEAK) Enhances Activation of STAT3/NLRC4 Inflammasome Signaling Axis through PKCδ in Astrocytes: Implications for Parkinson’s Disease.” Cells 9 (8): 1831. https://doi.org/10.3390/cells9081831.
Smyth, Peter, Jutharat Sasiwachirangkul, Rich Williams, and Christopher J. Scott. 2022. “Cathepsin S (CTSS) Activity in Health and Disease - A Treasure Trove of Untapped Clinical Potential.” Molecular Aspects of Medicine 88 (December):101106. https://doi.org/10.1016/j.mam.2022.101106.
Tatenhorst, Lars, Katrin Eckermann, Vivian Dambeck, Luis Fonseca-Ornelas, Hagen Walle, Tomás Lopes da Fonseca, Jan C. Koch, et al. 2016. “Fasudil Attenuates Aggregation of α-Synuclein in Models of Parkinson’s Disease.” Acta Neuropathologica Communications 4 (April):39. https://doi.org/10.1186/s40478-016-0310-y.
Tolosa, Eduardo, Alicia Garrido, Sonja W. Scholz, and Werner Poewe. 2021. “Challenges in the Diagnosis of Parkinson’s Disease.” The Lancet. Neurology 20 (5): 385–97. https://doi.org/10.1016/S1474-4422(21)00030-2.
Ueta, Mayumi, Hiromi Nishigaki, Seitaro Komai, Katsura Mizushima, Risa Tamagawa-Mineoka, Yuji Naito, Norito Katoh, Chie Sotozono, and Shigeru Kinoshita. 2023. “Positive Regulation of Innate Immune Response by miRNA-Let-7a-5p.” Frontiers in Genetics 13 (January):1025539. https://doi.org/10.3389/fgene.2022.1025539.
Vos, Theo, Amanuel Alemu Abajobir, Kalkidan Hassen Abate, Cristiana Abbafati, Kaja M Abbas, Foad Abd-Allah, Rizwan Suliankatchi Abdulkader, et al. 2017. “Global, Regional, and National Incidence, Prevalence, and Years Lived with Disability for 328 Diseases and Injuries for 195 Countries, 1990–2016: A Systematic Analysis for the Global Burden of Disease Study 2016.” The Lancet 390 (10100): 1211–59. https://doi.org/10.1016/S0140-6736(17)32154-2.
Wang, Hsiuying. 2023. “microRNA Pathological Mechanisms between Parkinson’s Disease, Alzheimer’s Disease, Glaucoma and Macular Degeneration.” Expert Reviews in Molecular Medicine 25:e24. https://doi.org/10.1017/erm.2023.19.
Wasko, Nicholas J., Meghan Horne Kulak, Debayon Paul, Alexandra M. Nicaise, Stephen T. Yeung, Frank C. Nichols, Kamal M. Khanna, Stephen Crocker, Joel S. Pachter, and Robert B. Clark. 2019. “Systemic TLR2 Tolerance Enhances Central Nervous System Remyelination.” Journal of Neuroinflammation 16 (July):158. https://doi.org/10.1186/s12974-019-1540-2.
Wasko, Nicholas J., Frank Nichols, and Robert B. Clark. 2020. “Multiple Sclerosis, the Microbiome, TLR2, and the Hygiene Hypothesis.” Autoimmunity Reviews 19 (1): 102430. https://doi.org/10.1016/j.autrev.2019.102430.
Wu, Chiao-Chieh, Shih-Jen Liu, Hsin-Wei Chen, Kuan-Yin Shen, and Chih-Hsiang Leng. 2016. “A Toll-like Receptor 2 Agonist-Fused Antigen Enhanced Antitumor Immunity by Increasing Antigen Presentation and the CD8 Memory T Cells Population.” Oncotarget 7 (21): 30804–19. https://doi.org/10.18632/oncotarget.9001.
Wu, Hsiu-Chuan, Kuo-Hsuan Chang, Mu-Chun Chiang, and Chiung-Mei Chen. 2021. “Alterations of Plasma Galectin-3 and C3 Levels in Patients with Parkinson’s Disease.” Brain Sciences 11 (11): 1515. https://doi.org/10.3390/brainsci11111515.
Wu, Tianzhi, Erqiang Hu, Shuangbin Xu, Meijun Chen, Pingfan Guo, Zehan Dai, Tingze Feng, et al. 2021. “clusterProfiler 4.0: A Universal Enrichment Tool for Interpreting Omics Data.” Innovation (Cambridge (Mass.)) 2 (3): 100141. https://doi.org/10.1016/j.xinn.2021.100141.
X, Dong-Chen, Yong C, Yang X, Chen-Yu S, and Li-Hua P. 2023. “Signaling Pathways in Parkinson’s Disease: Molecular Mechanisms and Therapeutic Interventions.” Signal Transduction and Targeted Therapy 8 (1). https://doi.org/10.1038/s41392-023-01353-3.
Xu, Dafeng, Yu Wang, Jincai Wu, Yuliang Zhang, Zhehao Liu, Yonghai Chen, and Jinfang Zheng. 2021. “Systematic Characterization of Novel Immune Gene Signatures Predicts Prognostic Factors in Hepatocellular Carcinoma.” Frontiers in Cell and Developmental Biology 9 (September):686664. https://doi.org/10.3389/fcell.2021.686664.
Yadati, Tulasi, Tom Houben, Albert Bitorina, and Ronit Shiri-Sverdlov. 2020. “The Ins and Outs of Cathepsins: Physiological Function and Role in Disease Management.” Cells 9 (7): 1679. https://doi.org/10.3390/cells9071679.
Ye, Eun-Ah, Li Liu, Youde Jiang, Jenny Jan, Subhash Gaddipati, Susmit Suvas, and Jena J. Steinle. 2016. “miR-15a/16 Reduces Retinal Leukostasis through Decreased pro-Inflammatory Signaling.” Journal of Neuroinflammation 13 (December):305. https://doi.org/10.1186/s12974-016-0771-8.
Z, Liu, Wang L, Xing Q, Liu X, Hu Y, Li W, Yan Q, Liu R, and Huang N. 2022. “Identification of GLS as a Cuproptosis-Related Diagnosis Gene in Acute Myocardial Infarction.” Frontiers in Cardiovascular Medicine 9 (November). https://doi.org/10.3389/fcvm.2022.1016081.
Zhang, Ejuan, Zhiyong Ma, Qian Li, Hu Yan, Jia Liu, Weimin Wu, Jiabao Guo, et al. 2019. “TLR2 Stimulation Increases Cellular Metabolism in CD8+ T Cells and Thereby Enhances CD8+ T Cell Activation, Function, and Antiviral Activity.” The Journal of Immunology 203 (11): 2872–86. https://doi.org/10.4049/jimmunol.1900065.
Zhang, Lei, Yiye Shao, Chunlei Tang, Zhen Liu, Dingzhong Tang, Canfang Hu, Xuelin Liang, Zhaoyang Hu, and Guojun Luo. 2022. “Identification of Novel Biomarkers in Platelets for Diagnosing Parkinson’s Disease.” European Neurology 85 (2): 122–31. https://doi.org/10.1159/000520102.
Zhou, Zan, Wen-jun Jiang, Li Li, and Jun-qiang Si. 2023. “Noise Exposure Increase Apoptosis in the Hippocampus of AD Mice through the Upregulation of CTSS.” Biochemical and Biophysical Research Communications 681 (November):283–90. https://doi.org/10.1016/j.bbrc.2023.09.076.

No competing interests reported.

SupplementaryInfo.docx

Download PDF

Reviewers agreed at journal
14 Jul, 2024
Reviewers invited by journal
12 Jul, 2024
Editor assigned by journal
12 Jul, 2024
Submission checks completed at journal
12 Jul, 2024
First submitted to journal
08 Jul, 2024

You are reading this latest preprint version

Exploring the hub genes and potential mechanisms of complement system-related genes in Parkinson disease: based on transcriptome sequencing and Mendelian randomization

Status:

Version 1

Abstract

Figures

1. Introduction

2. Materials and methods