Novel therapeutic targets for major depressive disorder related to oxidative stress identified by integrative multi-omics and multi-trait study

Novel therapeutic targets for major depressive disorder related to oxidative stress identified by integrative multi-omics and multi-trait study | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article Novel therapeutic targets for major depressive disorder related to oxidative stress identified by integrative multi-omics and multi-trait study Gang Zhu, Xiaojun Shao, Yuan Wang, Yuan Wang, Guangming Liang, and 4 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-4641375/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 19 Oct, 2024 Read the published version in Translational Psychiatry → Version 1 posted 12 You are reading this latest preprint version Abstract Oxidative stress (OS) is strongly implicated in the pathophysiology of major depressive disorder (MDD) but the molecular mechanisms remain largely unknown. The purpose of this study is to identify genes related to both OS and MDD, and further to evaluate the utility of these genes as diagnostic markers and potential treatment targets. We searched datasets related to MDD from the Gene Expression Omnibus (GEO) database for differentially expressed genes (DEGs) also related to OS according to GeneCards. Bioinformatics analyses and machine learning algorithms were used to identify hub genes mediating OS–MDD interactions. A summary data-based Mendelian randomization (SMR) approach was employed to identify possible causal genes for MDD from blood tissue eQLT data. These investigations identified 32 genes mediating OS–MDD interactions, while SMR analysis identified KCNE1 (OR = 1.057, 95%CI = 1.013–1.102, P = 0.010), MAPK3 (OR = 1.023, 95%CI = 1.004–1.043, P = 0.020), and STIP1 (OR = 0.792, 95%CI = 0.641–0.979, P = 0.031) as OS-related causal genes for MDD. These genes may thus serve as useful diagnostic markers and potential therapeutic targets. Health sciences/Diseases/Psychiatric disorders/Depression Health sciences/Biomarkers/Diagnostic markers Oxidative stress Depression Machine learning Integrative Omics Mendelian randomization Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Figure 8 Figure 9 Figure 10 Introduction Major depressive disorder (MDD) is a common psychiatric illness characterized by persistent depressed mood accompanied by heterogenous cognitive, behavioral, and physical symptoms( 1 ). According to the World Health Organization, approximately 280 million people worldwide are currently afflicted with MDD( 2 ). A nationally representative cross-sectional survey from China (the China Mental Health Survey) revealed a weighted lifetime prevalence of 3.4% and weighted 12-month prevalence of 2.1%( 3 ), with substantially higher weighted lifetime prevalence and weighted 12-month prevalence among adult females (8.0% and 4.2%, respectively) compared to males (5.7% and 3.0%)( 4 ). It is generally accepted that genetic, biological, psychosocial, and personality traits all contribute to MDD risk, termed the diversified disease hypothesis of MDD( 5 ). Pathological processes implicated in MDD include glutamatergic excitotoxicity, brain-derived neurotrophic factor/ tyrosine receptor kinase B signaling insufficiency, neuroinflammation, and gut microbiota–brain axis disturbance( 6 , 7 ). However, there are no effective treatments based on these mechanisms, implying complex multidimensional pathogenesis. Cellular metabolism and various signaling mechanisms result in the formation of highly reactive free radicals, including reactive oxygen species (OS) and reactive nitrogen species (RNS). Under normal physiological conditions, these species are neutralized by endogenous antioxidants; however, an imbalance between ROS or RNS production and scavenging will result in oxidative stress (OS), which leads to the oxidative damage of cellular lipids, proteins, and genomic DNA( 8 ), while severe OS can trigger necrotic or apoptotic cell death. Oxidative stress and ensuing cell death is strongly implicated in neurodegenerative diseases such as the Alzheimer’s disease and Parkinson’s disease( 9 , 10 ), and recent studies have suggested that OS also contributes to the pathophysiology and treatment response of MDD( 11 , 12 ). For instance, a positive correlation was found between and the amplitude of low-frequency fluctuations in key MDD-associated brain regions, such as the thalamus, anterior cingulate gyrus, and superior frontal gyrus, and plasma concentrations of the antioxidant enzymes superoxide dismutase (SOD) and glutathione reductase( 13 ). Furthermore, depression severity and working memory impairment were associated with higher plasma concentrations of malondialdehyde (MDA), an indicator of lipid peroxidation from oxidative stress, among recurrent MDD patients both before and after antidepressant treatment( 14 ). Reduced SOD activity, lower levels of the non-enzyme antioxidant glutathione (GSH), and elevated lipid peroxidation products have been proposed as reliable biomarkers for MDD( 15 , 16 ). In experiment animals as well, chronic unpredictable mild stress can induce depression-like symptoms and concomitant abnormalities in redox balance across brain subregions, including decreased GSH and SOD activities and higher levels of ROS, MDA, and carbonyl in prefrontal cortex and hippocampus( 17 ). Despite a growing number of studies implicating OS in MDD, there are still discrepancies( 11 , 18 ). An early meta-analysis of 23 studies with 4980 individuals found increased OS among MDD patients compared to matched individuals without depression, but some included studies found lower antioxidant biomarker concentrations (MDA and 8-oxo-2′-deoxyguanosine) in MDD( 19 ). A recent meta-analysis also concluded that OS status was more severe in MDD, but found no significant differences in catalase, SOD, glutathione peroxidase, and uric acid between MDD patients and healthy controls( 20 ).Therefore, the precise contributions of OS to MDD pathogenesis remained unclear. To address this issue, we conducted a comprehensive multi-omics analysis of the association between OS and MDD and identified multiple differentially expressed genes (DEGs) linking the two conditions. Machine learning models trained using these DEGs reliably distinguished patients from controls in an independent dataset. Thus, these genes may serve as feasible biomarkers or treatment targets. Furthermore, we utilized a multi-omics summary data-based Mendelian randomization (SMR) approach to identify putative causal OS-related genes for MDD. Methods and materials An overview of the analytical methods is illustrated in Fig. 1 . No ethics committee approval was required for this summary-level study. The gene array data of three study cohorts were extracted from the Gene Expression Omnibus (GEO) database: GSE32280 (platform: GPL570)( 21 ), GSE39653 (platform: GPL10558)( 22 ) and GSE98793 (platform: GPL570)( 23 ). In total these datasets included gene expression profiles for 165 MDD patients and 96 health controls. Genes related to OS were identified from the GeneCards database ( https://www.genecards.org ) using the keyword “oxidative stress”. Those with a relevance score ≥ 7 were included according to previous methods( 24 ). Identification of differentially expressed genes (DEGs) associated with MDD and OS We identified genes common among the disease datasets GSE32280, GSE39653, GSE98793, and GeneCards using the “cbind ()” function of R, and then removed batch bias and performed log2(X + 1) normalization using the R sva function. These standardized gene expression data formed the basis for all further analysis. To analyze the effect of OS-related gene expression level on MDD risk, the R limma package was used to identify DEGs between disease and normal control samples, with the absolution value of log2-fold change (log2FC) > 1 and P < 0.05 set as thresholds. These DEGs were visualized by volcano plot and a heat map was constructed using the R packages heatmaps (version 1.0.12) and ggplot2 (version 3.3.2). Screening of hub genes mediating crosstalk between MDD and OS Univariate logistic regression and multivariate logistic regression analyses of the GSE39653 dataset were used to identify hub genes within the MDD- and OS-related gene set (Hub MDD-OS DEGs) with P < 0.05 considered statistically significant. This was followed by dimensionality reduction analysis. Identification of key modules and functional genes To reveal biological relationships among key genes and gene clusters related to OS and MDD, Gene Ontology (GO) enrichment analysis was conducted using the R package clusterProfiler. Genes were assessed for molecular function (MF), cell component (CC), and biological process (BP) annotations with a False Discovery Rate (FDR)-corrected P < 0.05 considered statistically significant. The R package cowplot was then used to conduct Spearman correlation analyses between key OS- and MDD-related genes as well as to draw heat maps, scatter plots, and correlation curves. Again, P < 0.05 was considered statistically significant. Key genes were then mapped to chromosomal locations using the R package RCircos based on data from the R package and downloaded from the ENSEMBL database. The open-source STRING database was then used to build a protein–protein interaction (PPI) network( 25 ), which was subsequently visualized using Cytoscape. In-depth investigation of key genes Key genes mediating MDD–OS interactions were further screened and evaluated using 6 machine learning algorithms, Bagged Trees, Bayesian, Random Forest, Wrapper (Bpruta), Learning Vector Quantification (LQV), and 1000 iterations 10-fold cross-validation Least Absolute Shrinkage and Selection Operator (LASSO). A DEG identified as a feature gene by at least 5 algorithms based on classification performance was considered an important Hub DEG for MDD–OS interactions. The interactions among these key intersecting genes were then visualized using the R package “Upset” application. To explore the associations of individual hub gene pairs in the GSE39653 and GSE9873 datasets, we conducted Spearman’s correlation analyses and plotting of heat maps, scatter plots, and correlation curves using cowplot. Finally, RCircos was used to draw the chromosome localization map for display. Based on these analyses, we constructed a nomogram based on the GSE39653 dataset using the R package rms. A nomogram model was then constructed based on the hub genes to predict the prevalence of OS-associated MDD. Calibration curve analysis was performed to evaluate the accuracy and resolution of the nomogram. Finally, receiver operating characteristic (ROC) curves were constructed to evaluate the accuracy of the diagnostic model for the onset of OS-associated MDD using the R package pROC. Thorough composition validation techniques were employed to confirm model performance for distinguishing MDD from control samples. We also developed a classification prediction model based on expression analysis of blood utilizing a combination of the aforementioned 6 machine learning algorithms. The performance of each model was evaluated by calculating the area under the ROC curve (AUC), followed by visual representation of the results (predictive genes) using heat maps. Optimal model performance was assessed using calibration curve and decision curve analysis (DCA). The R ConsensusClusterPlus package was used to identify subpopulations with distinct molecular phenotypes based on hub genes for MDD–OS interactions( 26 ). Principal co-ordinates analysis (PCoA) was then conducted to verify consistency of clustering. The R package ggpubr was used to draw box plots with sample cluster labels as groups, and the differences between groups were evaluated for statistical significance by the Wilcoxon rank sum test. Immuno-infiltration analysis The relative abundances of specific infiltrating immune cells were estimated using CIBERSORT ( https://cibersort.stanford.edu/)(27) , an analytical tool designed to reveal the distribution levels of LM22 immune cells based on gene expression profiles. Distinct enrichment fractions of immune cells were then compared using the Wilcox test. We further performed quantitative Single Sample Gene Set Enrichment Analysis (ssGSEA) to calculate the abundance of immune cells associated with MDD–OS interactions. Finally, differences in immune cell infiltration were visualized using ggplot2 with P < 0.05 set as the threshold for significance. SMR analysis We gathered genome-wide association study (GWAS) summary statistics of 170,756 MDD cases and 329,443 controls (8,483,301 genetic variants in total) from the Psychiatric Genomics Consortium ( https://www.med.unc.edu/pgc ), one of the largest, most innovative, and productive platforms in psychiatry( 28 ). Single nucleotide polymorphisms (SNPs) associated with expression quantitative trait loci (eQTLs) were selected as instrumental variables (IVs) to infer direct causal effects of gene expression or protein levels on MDD. Whole-blood eQTL summary statistics for 15,882 genes were obtained from 31,684 individuals in the eQTLGen database( 29 ). The current study focused only on cis-eQTLs within 1-Mb from the start or end of the gene. Statistical analysis All statistical analyses were performed using R 4.2.0. Continuous variables were compared between two groups using the Wilcoxon rank sum test and among three or more groups using the Kruskal–Wallis test. Categorical variables were compared by the chi-square test or Fisher's exact test. Associations between immune cell abundance and gene expression levels were evaluated using Spearman's correlation tests. A P < 0.05 was considered statistically significant for all tests. Causal inferences between GWAS and cis-eQTLs were evaluated using the SMR multi-tool. The top associated cis-QTLs were selected by considering a window centered around the corresponding gene (± 1000 kb) and surpassing a P-value threshold of 5.0 ×10 − 8 . All SNPs with allele frequency differences larger than the specified threshold (0.2 in the current study) between datasets, including the LD reference sample, the QTL summary data, and the outcome summary data, were excluded. All SMR analysis were implemented using SMR v1.3.1 and included SNPs as instruments, key genes for MDD–OS interactions as exposures, and MDD as the outcome (SMR P < 0.05, cis-eQTLs, and GWAS P 10 to minimize weak instrument bias( 30 ). The heterogeneity in the dependent instrument (HEIDI) test was applied using SMR v1.3.1 to distinguish pleiotropy from linkage. All instruments with P-HEIDI < 0.01 (indicating significant heterogeneity) were omitted from the analysis. Results Identification of DEGs between MDD cases and controls The GEO datasets GSE32280, GSE39653, and GSE98793 were combined using the R “cbind ()” function and the batch bias mitigated using the R “sva” package (see Fig. 2 and Supplementary Table 1). In total, 596 DEGs were identified between MDD patients and healthy controls (HCs) using Limma, including 294 upregulated genes and 302 downregulated genes in MDD. The magnitudes of differential expression for significant DEGs (P < 0.05) are shown as a volcano map in Fig. 3 A and as a heatmap in Fig. 3 B and details in Supplementary Table 2. These DEGs were enriched in GO CC terms ‘specific granule’, ‘specific granule membrane’, ‘secretory granule membrane’, ‘specific granule lumen’, and ‘tertiary granule’, BP terms ‘receptor signaling pathway via JAK-STAT and STAT’, ‘placenta development’, and ‘carbohydrate catabolic process’, and MF terms ‘identical protein binding’, ‘immune receptor activity’, ‘cytokine receptor activity’, ‘phospholipase activity’, and ‘1-phosphatidylinositol-3-kinase regulator activity’ (Fig. 3 C, D and Supplementary Table 3). To identify MDD-associated genes also related to OS, these DEGs were searched against the 817 genes in GeneCards with relevance score ≥ 7 for OS, yielding 38 potential MDD–OS interaction or crosstalk genes (Supplementary Table 4). The Venn diagram of these overlapping DEGs is shown in Fig. 4 A and the chromosomal positions in Fig. 4 B. Among these 38 DEGs, AMD , ALPP , CAMK2G , DDAH1 , KCNE1 , LEP , MAPK3 , IL10 , PINK1 , and SLC2A1 were significantly upregulated in the MDD–OS samples (Fig. 4 C). More details were shown in Supplementary Table 5. Identification of key MDD–OS interaction genes via machine learning Machine learning algorithms were trained to deduce the associations between the 38 MDD–OS DEGs and MDD pathogenesis using the GSE39653 dataset, and model performance was validated using the GSE98793 dataset. Intersection of results yielded by the Bagged Trees algorithm (Fig. 5 A), Bayesian algorithm (Fig. 5 B), Random Forest algorithm (Fig. 5 C), Wrapper algorithm (Fig. 5 D), LQV (Fig. 5 E) and 1000 times 10-fold cross-validation LASSO Logistic model (Fig. 5 F) identified 32 of these genes as closely related to the pathogenesis of MDD and OS (Fig. 5 G): ADM, AKR1C3, ALPP, CAMK2G, CREBBP, DDAH1, DNM1L, F5, FKBP5 , GADD45A , GATB , GDF15 , HSP90AA1 , HSP90AB1 , HSP90B1 , IL10 , INSR , KCNE1 (encoding Potassium Voltage-Gated Channel Subfamily E Regulatory Subunit 1), KLF2 , LEP , MAP2K1 , MAPK3 (encoding mitogen-activated protein kinase 3), MGST1 , PLA2G7 , PLAU , PTK2B , RETN , SLC2A1 , STIP1 (encoding stress-induced phosphoprotein 1), TNF , UGT1A1 , and VDR . Subsequently, we constructed a PPI network with these 32 genes as hubs using the STRING database. Based on the criteria |Logfc| < 0.05 and P value < 0.05, this 32-node PPI network included 341 edges, indicating complex multilevel interactions between OS and MDD pathogenesis. A diagnostic model was then constructed based on the GSE39653 dataset (Fig. 6 A), and discrimination was verified on both the training set GSE39653 and validation set GSE98793. The recall curve and Hosmer − Lemoeshow goodness-of-fit test results were equal to 1 for the training set (GSE39653), indicating a low probability of type I error, that prediction results were close to the real data, and that the calibration degree of the model was high (Fig. 6 B). The AUC of the diagnostic model was also 1.00 for the training set GSE39653 (Fig. 6 C). The C-index of the diagnostic model was 0.99 for the training set (GSE39653) and 1 for the validation set (GSE98793) (Fig. 6 D), and the recall curve and Hosmer − Lemeshow test result P values were equal to 1.00 (Fig. 6 E). Finally, the AUC of the diagnostic model was 0.876 for the validation set (GSE98793) (Fig. 6 F). Associations of key genes with MDD subtypes The physical, behavioral, and cognitive symptoms of MDD can vary substantially among patients, so we examined the potential functions of these 32 key genes in distinct MDD phenotypes by cluster analysis using ConsensusClusterPlus. This analysis yielded two patient clusters in the GSE39653 dataset, Cluster A and Cluster B, based on DEG profiles (Fig. 8 A). The slope of consistency index for different classifications (Fig. 7 B) further confirmed that two was the optimal cluster number for this cohort, while the scree plot in Fig. 8 C revealed the inflection point for best classification. We further validated the stability of this MDD patient subtyping in GSE39653 by PCoA (Fig. 7 D). To explore differences in pathogenesis between the two subtypes, the total DEGs for Cluster A and Cluster B were first identified. In total, 3560 DEGs were found, including 1367 expressed at higher levels in Cluster A and 1223 expressed at higher levels in Cluster B (Fig. 7 E). Then, GO enrichment analysis was conducted to reveal differential enrichment of biological functions. Enriched BP terms in Cluster A included ‘leukocyte degranulation’, ‘aerobic electron transport chain’, and ‘ATP synthesis coupled electron transport’, while enriched CC terms included ‘respiratory chain complex’ and ‘ficolin-1-rich granule lumen’, and enriched MF terms included ‘antioxidant activity’ and ‘oxidoreduction-driven active transmembrane transporter activity’ (Fig. 7 F and Supplementary Table 6). Analysis of immune characteristics Immune cell infiltration is a major driver of OS in the brain, suggesting an important contribution to OS-related MDD pathogenesis. We employed Cibersort analysis to calculate the infiltration status of 22 immune cells in the GSE39653 dataset, and the relative abundance of each immune cell type was then compared by the Wilcoxon sign rank test (Fig. 8 A). The CIBERSORT algorithm was also used to compare immune cell infiltration profiles between Cluster A and Cluster B. The results revealed that dendritic cells and activated mast cells were highly abundant in Cluster B (Fig. 8 B), and that the relative abundances of most immune cell subsets differed between clusters (Fig. 8 C). Moreover, the ssGSEA algorithm indicated greater infiltration of activated CD8 + T cells, effector memory CD8 + T cells, regulatory T cells, Type 1 T helper cell, eosinophils, macrophages, and monocytes in MDD samples of Cluster A (Fig. 9 A) (P < 0.05). Furthermore, abundances were correlated with the expression levels of the 32 most important DEGs (all P < 0.05) (Fig. 9 B). SMR analysis of characteristic genes We also performed SMR analysis to evaluate the association strengths of these 32 key genes with MDD (with P-smr < 0.05 set as the threshold for statistical significance). Results revealed an association between elevated expression of KCNE1 and MDD odds (OR = 1.057, 95% CI = 1.013–1.102, P-smr = 0.010) (Fig. 10 A, B, and G). Similarly, elevated MAPK3 expression was associated with greater MDD odds (OR = 1.023, 95% CI = 1.004–1.043, P-smr = 0.020) (Fig. 10 C, D, and G), while upregulation of STIP1 was associated with reduced MDD odds (OR = 0.792, 95% CI = 0.641–0.979, P-smr = 0.031) (Fig. 10 E, F, and G). Supplementary Table 7 showed the SMR association between expression of gene MAPK3 , KCNE1 , STIP1 and MDD. Sensitivity analysis The F-statistic of all SNPs included in the analysis ranged from 29.855 to 3394.048, indicating a powerful instrumental variable–exposure association (threshold set at 10) (Supplementary Table 7). Results of the HEIDI test further suggested that all observed associations were not due to linkage (p > 0.01). Discussion To our best knowledge, this is the first study on the contributions of oxidative stress-related genes to MDD pathogenesis using integrated multi-omics, machine learning, infiltrated immune cell profiling, genome-wide association, and summary data-based Mendelian randomization analysis. We identified 38 genes differentially expressed between MDD patients and controls that were also associated with OS, of which 32 were deemed important to the influence of OS on MDD pathogenesis (MDD–OS interaction genes) in training and validation cohorts by 6 separate machine learning algorithms. Further screening of blood tissue expression profiles by SMR analysis identified KCNE1, MAPK3 , and STIP1 as key linkage genes between OS and MDD. These DEGs may thus be convenient biomarkers for MDD as well as potential treatment targets. Neuroinflammation in strongly implicated in MDD as evidenced by elevated inflammatory marker concentrations, infiltrating immune cell numbers, and antibody titers( 31 ). These inflammatory processes both generate and are promoted by ROS and RNS (hence the OS–inflammation interaction is also known as “evil twins of aging”), further implicating OS in MDD pathogenesis( 32 ). Elevated ROS production leads to GSH depletion, oxidative damage, and ultimately enhanced inflammation( 33 ). Excessive ROS can promote the expression of proinflammatory cytokines through several pathways, including activation of promoting protein-1 and nuclear factor kappa-B (NFκB), increased histone acetylation, and activation of caspase-1 and NOD-like receptor thermal protein domain associated protein 3( 34 – 36 ). Inflammatory reactions induce expression and release of peroxiredoxin 2, which in turn stimulates macrophages to release pro-inflammatory tumor necrosis factor-α (TNF-α)( 37 ). Elevated TNF-α had been detected in serum and in multiple brain subregions (including the anterior cingulate cortex, prefrontal cortex, and hippocampus) of MDD patients( 38 , 39 ). Neuroinflammation also promotes the kynurenine pathway and ensuing quinolinic acid generation by activating indoleamine 2,3-dioxygenase (IDO), tryptophan-2,3-dioxygenase (TDO), and kynurenine 3-mono-oxygenase, which induces mitochondrial damage and results in further ROS production, glutamate release, N-methyl-D-aspartic acid (NMDA) receptor activation, Ca 2+ influx, and mitochondrial calcium overload, the end result of which is loss of mitochondrial membrane potential, reduced ATP generation, and accelerated ROS generation( 34 , 40 , 41 ). In addition, IDO and TDO activation may reduce 5-HT biosynthesis, and 5-HT insufficiency is widely believed to result in low mood( 42 ). These relationships also appear to be bidirectional, such that OS can promote neuroinflammation and vice versa in MDD. In accord with previous studies, we found that multiple immune-inflammatory gene pathways were activated in MDD compared to controls. These genes may in turn mediate the reciprocal exacerbation of OS generation and neuroinflammation leading to MDD. The OS–MDD crosstalk genes identified in this study are primarily involved in immune cell function, including activated CD8 + T cells, effector memory CD8 + T cells, regulatory T cells, type 1 T helper cells, eosinophils, macrophages, and monocytes, further supporting shared immune-inflammatory mechanisms in OS and MDD. We also conducted SMR analysis to identify new causal genes for MDD as such genes may be prime drug targets. Upregulation of KCNE1 and MAPK3 were found to increase MDD risk, potentially by promoting pathogenic mechanisms involving OS. The MAPK3 product extracellular signal-regulated kinase 1 (ERK1) regulates cell proliferation, differentiation, and cell cycle progression among other vital processes( 43 ). It has been reported that ERK signaling is significantly downregulated in the prefrontal cortex and hippocampus of both human patients and animal models of chronic depression( 44 – 46 ). The ERK1/2 isoforms are the most thoroughly investigated and well characterized isoforms in the central nervous system( 42 , 47 , 48 ), and both have been found to promote OS via ROS production and to amplify the inflammatory response through activation of the stress-responsive transcription factor NFκB( 49 , 50 ). At present, most studies on the role of MAPK3 in MDD have focused on the brain, while few studies have investigated expression changes in more accessible blood samples. Moreover, most studies have focused on ERK1/2, but few specifically on ERK1. We found higher MAPK3 expression in the blood tissue of MDD patients compared to controls, consistent with previous findings. One prospective case–control study reported that a MAPK3 SNP enhanced interferon-α-induced depression, possibly by increasing the propensity for glutamate dysregulation( 51 ). A bioinformatics analysis identified 5 genes including MAPK3 as key modulators of post-stroke depression risk, disease biomarkers, and therapeutic targets of acupuncture( 52 ). Others have found significant associations of MAPK3 with schizophrenia, and a recent genome-wide Mendelian randomization analysis identified MAPK3 as a potential drug target for schizophrenia treatment( 53 ), in line with previous studiess( 54 , 55 ). Based on these and our own findings, we speculate that MAPK3 may be a critical mediator of OS effects on MDD pathogenesis and thus a promising therapeutic target. However, in our present study, MAPK3 appeared to make only a limited contribution (OR = 1.023, 95% CI = 1.004–1.043). Nonetheless, the contributions of MAPK3 to OS and MDD warrant further exploration. In contrast to MAPKs, few studies have examined the genetic association of KCNE1 with MDD, although McCaffery and colleagues proposed that KCNE1 is associated with longer-term changes in depressive symptoms( 56 ). The KCNE family proteins are regulatory subunits of voltage-gated K(+) channels( 57 ), and are implicated in multiple arrhythmogenic cardiac myocardium diseases( 58 ). The KCNE1 subunit regulates the neuronal membrane potential through modulation of K(+) channels, including KCNQ channels( 59 ). Further, the KCNQ channel modulator retigabine has been shown to improve depressive symptoms, suggesting therapeutic potential for MDD( 60 ). Another study also included KCNE1 expression in a diagnostic model for MDD( 61 ), although no causal association was suggested. In the current study, preliminary genomic analysis indicated that KCNE1 was upregulated in MDD and positioned as a linker gene between MDD and OS, while according to SMR analysis, KCNE1 upregulation increases the risk of MDD. We speculate that drugs targeting KCNE1 could show therapeutic efficacy against MDD. These same genomics analyses also revealed downregulation of STIP1 , which encodes a co-chaperone that interacts with heat-shock proteins 70 and 90, in the blood tissue of MDD patients, in accord with previous reports( 62 , 63 ). Further, SMR identified STIP1 as a protective target against MDD (OR = 0.792, 95% CI = 0.641–0.979). Thus, activation of STIP1 expression may be a useful therapeutic strategy against MDD. In addition to acting as a chaperone, extracellular STIP1 acts as a trophic factor to engage PrP C , thereby enhancing neuritogenesis and neuronal survival( 64 , 65 ). Studies have also implicated STIP1 in functional recovery after stroke and regulation of Aβ peptide toxicity in Alzheimer's disease models. Moreover, a GWAS analysis identified a STIP1 polymorphism as a potential risk factor for attention-deficit disorder( 66 ). Mice with elevated STIP1 levels (up to nearly fivefold) showed no neuropathology, anxiety-like behaviors, depression-like behaviors, spatial memory deficits, or attention deficits( 67 ), suggesting that STIP1 augmentation may be a feasible strategy for antidepressant treatment; however, the detailed underlying mechanisms remain unclarified. Study limitations: This study has several limitations. First, it is possible that differences in gene expression between MDD patients and controls reflect the influences of factors such as age, sex, smoking, medications, and other health conditions. Second, we only focused on the cis-regions of OS and MDD genes, despite the possibility that trans-eQTL SNPs (SNPs > 5 Mb from the gene) may have a widespread impact on regulatory networks. Finally, functional experiments are still needed to confirm the importance of these DEGs in MDD pathogenesis through OS-dependent or OS-independent pathways. Conclusion This integrative multi-omics and multi-trait study identified numerous genes linking OS to MDD pathogenesis, including three genes, KCNE1, MAPK3 , and STIP1 , causally associated with MDD. These gene in particular could serve as diagnostic markers and drug targets for MDD treatment. In addition, the dozens of other genes identify may provide clues to novel pathological mechanisms for MDD. Declarations Declaration of interests All authors declare no competing interests. Data sharing Original data of major depressive depression could be retrieved from References( 21 – 23 , 28 ). The raw data of oxidative stress could be obtained from the GeneCards database ( https://www.genecards.org ). The data analyzed in this paper were publicly available and the specific information and links of source data of all Figures and Tables are showed in Supplementary files. Funding This work was supported by grants awarded by the Major Project of the Department of Science & Technology, Liaoning Province (2019JH8/10300019). Code availability R Code for the current analysis was available at Supplementary material- “Source code. R”. Please contact the corresponding author if you would like to see any data that are not included in the Article or the Appendix. Contributors GZ and XJS conceived and planned the study. XJS, YW, ZLG, GML, XTZ, LJ, MM and LD selected the articles and extracted data. XJS and YW did the statistical analyses. XJS wrote the report. GZ had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. References American Psychiatric Association D, Association AP. Diagnostic and statistical manual of mental disorders: DSM-5: American psychiatric association Washington, DC; 2013. Herrman H, Patel V, Kieling C, Berk M, Buchweitz C, Cuijpers P, et al. Time for united action on depression: a Lancet-World Psychiatric Association Commission. Lancet. 2022;399(10328):957–1022. Huang Y, Wang Y, Wang H, Liu Z, Yu X, Yan J, et al. Prevalence of mental disorders in China: a cross-sectional epidemiological study. Lancet Psychiatry. 2019;6(3):211–24. Lu J, Xu X, Huang Y, Li T, Ma C, Xu G, et al. Prevalence of depressive disorders and treatment in China: a cross-sectional epidemiological study. Lancet Psychiatry. 2021;8(11):981–90. Shao X, Zhu G. Associations Among Monoamine Neurotransmitter Pathways, Personality Traits, and Major Depressive Disorder. Front Psychiatry. 2020;11:381. Hu S, Sheng WS, Ehrlich LC, Peterson PK, Chao CC. Cytokine effects on glutamate uptake by human astrocytes. Neuroimmunomodulation. 2000;7(3):153–9. Zheng P, Zeng B, Zhou C, Liu M, Fang Z, Xu X, et al. Gut microbiome remodeling induces depressive-like behaviors through a pathway mediated by the host's metabolism. Mol Psychiatry. 2016;21(6):786–96. Camini FC, da Silva Caetano CC, Almeida LT, de Brito Magalhães CL. Implications of oxidative stress on viral pathogenesis. Arch Virol. 2017;162(4):907–17. Bai R, Guo J, Ye XY, Xie Y, Xie T. Oxidative stress: The core pathogenesis and mechanism of Alzheimer's disease. Ageing Res Rev. 2022;77:101619. Dong-Chen X, Yong C, Yang X, Chen-Yu S, Li-Hua P. Signaling pathways in Parkinson's disease: molecular mechanisms and therapeutic interventions. Signal Transduct Target Ther. 2023;8(1):73. Jiménez-Fernández S, Gurpegui M, Díaz-Atienza F, Pérez-Costillas L, Gerstenberg M, Correll CU. Oxidative stress and antioxidant parameters in patients with major depressive disorder compared to healthy controls before and after antidepressant treatment: results from a meta-analysis. J Clin Psychiatry. 2015;76(12):1658–67. Wang H, Jin M, Xie M, Yang Y, Xue F, Li W, et al. Protective role of antioxidant supplementation for depression and anxiety: A meta-analysis of randomized clinical trials. J Affect Disord. 2023;323:264–79. Liu Y, Zhang B, Zhou Y, Li M, Gao Y, Qin W, et al. Plasma oxidative stress marker levels related to functional brain abnormalities in first-episode drug-naive major depressive disorder. Psychiatry Res. 2024;333:115742. Talarowska M, Gałecki P, Maes M, Gardner A, Chamielec M, Orzechowska A, et al. Malondialdehyde plasma concentration correlates with declarative and working memory in patients with recurrent depressive disorder. Molecular biology reports. 2012;39:5359–66. Lindqvist D, Dhabhar FS, James SJ, Hough CM, Jain FA, Bersani FS, et al. Oxidative stress, inflammation and treatment response in major depression. Psychoneuroendocrinology. 2017;76:197–205. Camkurt MA, Fındıklı E, İzci F, Kurutaş EB, Tuman TC. Evaluation of malondialdehyde, superoxide dismutase and catalase activity and their diagnostic value in drug naïve, first episode, non-smoker major depression patients and healthy controls. Psychiatry Res. 2016;238:81–5. Parul, Mishra A, Singh S, Singh S, Tiwari V, Chaturvedi S, et al. Chronic unpredictable stress negatively regulates hippocampal neurogenesis and promote anxious depression-like behavior via upregulating apoptosis and inflammatory signals in adult rats. Brain Res Bull. 2021;172:164–79. Liu T, Zhong S, Liao X, Chen J, He T, Lai S, et al. A Meta-Analysis of Oxidative Stress Markers in Depression. PLoS One. 2015;10(10):e0138904. Palta P, Samuel LJ, Miller ER, 3rd, Szanton SL. Depression and oxidative stress: results from a meta-analysis of observational studies. Psychosom Med. 2014;76(1):12–9. Jiménez-Fernández S, Gurpegui M, Garrote-Rojas D, Gutiérrez-Rojas L, Carretero MD, Correll CU. Oxidative stress parameters and antioxidants in adults with unipolar or bipolar depression versus healthy controls: Systematic review and meta-analysis. J Affect Disord. 2022;314:211–21. Yi Z, Li Z, Yu S, Yuan C, Hong W, Wang Z, et al. Blood-based gene expression profiles models for classification of subsyndromal symptomatic depression and major depressive disorder. PLoS One. 2012;7(2):e31283. Savitz J, Frank MB, Victor T, Bebak M, Marino JH, Bellgowan PS, et al. Inflammation and neurological disease-related genes are differentially expressed in depressed patients with mood disorders and correlate with morphometric and functional imaging abnormalities. Brain Behav Immun. 2013;31:161–71. Leday GGR, Vértes PE, Richardson S, Greene JR, Regan T, Khan S, et al. Replicable and Coupled Changes in Innate and Adaptive Immune Gene Expression in Two Case-Control Studies of Blood Microarrays in Major Depressive Disorder. Biol Psychiatry. 2018;83(1):70–80. Xu S, Li X, Zhang S, Qi C, Zhang Z, Ma R, et al. Oxidative stress gene expression, DNA methylation, and gut microbiota interaction trigger Crohn's disease: a multi-omics Mendelian randomization study. BMC Med. 2023;21(1):179. Wagner RA, Fischer MJ. The string-to-string correction problem. Journal of the ACM (JACM). 1974;21(1):168–73. Wilkerson MD, Hayes DN. ConsensusClusterPlus: a class discovery tool with confidence assessments and item tracking. Bioinformatics. 2010;26(12):1572–3. Chen B, Khodadoust MS, Liu CL, Newman AM, Alizadeh AA. Profiling Tumor Infiltrating Immune Cells with CIBERSORT. Methods Mol Biol. 2018;1711:243–59. Howard DM, Adams MJ, Clarke TK, Hafferty JD, Gibson J, Shirali M, et al. Genome-wide meta-analysis of depression identifies 102 independent variants and highlights the importance of the prefrontal brain regions. Nat Neurosci. 2019;22(3):343–52. Võsa U, Claringbould A, Westra HJ, Bonder MJ, Deelen P, Zeng B, et al. Large-scale cis- and trans-eQTL analyses identify thousands of genetic loci and polygenic scores that regulate blood gene expression. Nat Genet. 2021;53(9):1300–10. Burgess S, Thompson SG. Avoiding bias from weak instruments in Mendelian randomization studies. Int J Epidemiol. 2011;40(3):755–64. Beurel E, Toups M, Nemeroff CB. The Bidirectional Relationship of Depression and Inflammation: Double Trouble. Neuron. 2020;107(2):234–56. Joseph J, Shukitt-Hale B, Casadesus G, Fisher D. Oxidative stress and inflammation in brain aging: nutritional considerations. Neurochemical research. 2005;30:927–35. Ghezzi P. Role of glutathione in immunity and inflammation in the lung. International journal of general medicine. 2011:105–13. de Oliveira LG, Angelo YdS, Iglesias AH, Peron JPS. Unraveling the link between mitochondrial dynamics and neuroinflammation. Frontiers in Immunology. 2021;12:624919. Adcock IM, Cosio B, Tsaprouni L, Barnes PJ, Ito K. Redox regulation of histone deacetylases and glucocorticoid-mediated inhibition of the inflammatory response. Antioxidants & redox signaling. 2005;7(1–2):144–52. Rahman I, Marwick J, Kirkham P. Redox modulation of chromatin remodeling: impact on histone acetylation and deacetylation, NF-κB and pro-inflammatory gene expression. Biochemical pharmacology. 2004;68(6):1255–67. Behl T, Rana T, Alotaibi GH, Shamsuzzaman M, Naqvi M, Sehgal A, et al. Polyphenols inhibiting MAPK signalling pathway mediated oxidative stress and inflammation in depression. Biomedicine & Pharmacotherapy. 2022;146:112545. Gritti D, Delvecchio G, Ferro A, Bressi C, Brambilla P. Neuroinflammation in major depressive disorder: a review of PET imaging studies examining the 18-kDa translocator protein. Journal of affective disorders. 2021;292:642–51. Enache D, Pariante CM, Mondelli V. Markers of central inflammation in major depressive disorder: a systematic review and meta-analysis of studies examining cerebrospinal fluid, positron emission tomography and post-mortem brain tissue. Brain, behavior, and immunity. 2019;81:24–40. Moncrieff J, Cooper RE, Stockmann T, Amendola S, Hengartner MP, Horowitz MA. The serotonin theory of depression: a systematic umbrella review of the evidence. Molecular psychiatry. 2023;28(8):3243–56. Visentin APV, Colombo R, Scotton E, Fracasso DS, da Rosa AR, Branco CS, et al. Targeting inflammatory-mitochondrial response in major depression: current evidence and further challenges. Oxidative medicine and cellular longevity. 2020;2020. Catena-Dell’Osso M, Rotella F, Dell’Osso A, Fagiolini A, Marazziti D. Inflammation, serotonin and major depression. Current Drug Targets. 2013;14(5):571–7. Roskoski R, Jr. ERK1/2 MAP kinases: structure, function, and regulation. Pharmacol Res. 2012;66(2):105–43. Wang JQ, Mao L. The ERK Pathway: Molecular Mechanisms and Treatment of Depression. Mol Neurobiol. 2019;56(9):6197–205. Yuan S, Jiang X, Zhou X, Zhang Y, Teng T, Xie P. Inosine alleviates depression-like behavior and increases the activity of the ERK-CREB signaling in adolescent male rats. Neuroreport. 2018;29(14):1223–9. Yuan P, Zhou R, Wang Y, Li X, Li J, Chen G, et al. Altered levels of extracellular signal-regulated kinase signaling proteins in postmortem frontal cortex of individuals with mood disorders and schizophrenia. J Affect Disord. 2010;124(1–2):164–9. Wang JQ, Fibuch EE, Mao L. Regulation of mitogen-activated protein kinases by glutamate receptors. J Neurochem. 2007;100(1):1–11. Li SY, Rong PJ, Gao GJ, Zhang Y, Wang JY, Wang Y, et al. [Auricular electroacupuncture improves depression possibly by promoting hippocampal Raf/ERK/RSK/CREB signaling in chronic unpredictable mild stress induced depression rats]. Zhen Ci Yan Jiu. 2019;44(8):554–9. Chiu FL, Lin JK. Tomatidine inhibits iNOS and COX-2 through suppression of NF-kappaB and JNK pathways in LPS-stimulated mouse macrophages. FEBS Lett. 2008;582(16):2407–12. Veres B, Radnai B, Gallyas F, Jr., Varbiro G, Berente Z, Osz E, et al. Regulation of kinase cascades and transcription factors by a poly(ADP-ribose) polymerase-1 inhibitor, 4-hydroxyquinazoline, in lipopolysaccharide-induced inflammation in mice. J Pharmacol Exp Ther. 2004;310(1):247–55. Cheng SW, Li JX, Chien YC, Chang JP, Shityakov S, Huang SY, et al. Genetic Variations of Ionotropic Glutamate Receptor Pathways on Interferon-α-induced Depression in Patients with Hepatitis C Viral Infection. Brain Behav Immun. 2021;93:16–22. Li M, Ding R, Yang X, Ran D. Study on Biomarkers Related to the Treatment of Post-Stroke Depression and Alternative Medical Treatment Methods. Neuropsychiatr Dis Treat. 2022;18:1861–73. Liu J, Cheng Y, Li M, Zhang Z, Li T, Luo XJ. Genome-wide Mendelian randomization identifies actionable novel drug targets for psychiatric disorders. Neuropsychopharmacology. 2023;48(2):270–80. Chang H, Cai X, Li HJ, Liu WP, Zhao LJ, Zhang CY, et al. Functional Genomics Identify a Regulatory Risk Variation rs4420550 in the 16p11.2 Schizophrenia-Associated Locus. Biol Psychiatry. 2021;89(3):246–55. Ma C, Gu C, Huo Y, Li X, Luo XJ. The integrated landscape of causal genes and pathways in schizophrenia. Transl Psychiatry. 2018;8(1):67. McCaffery JM, Papandonatos GD, Faulconbridge LF, Erar B, Peter I, Wagenknecht LE, et al. Genetic Predictors of Depressive Symptoms in the Look AHEAD Trial. Psychosom Med. 2015;77(9):982–92. Abbott GW. KCNE1 and KCNE3: The yin and yang of voltage-gated K(+) channel regulation. Gene. 2016;576(1 Pt 1):1–13. Erlandsdotter LM, Giammarino L, Halili A, Nikesjö J, Gréen H, Odening KE, et al. Long-QT mutations in KCNE1 modulate the 17β-estradiol response of Kv7.1/KCNE1. Sci Adv. 2023;9(11):eade7109. Melman YF, Um SY, Krumerman A, Kagan A, McDonald TV. KCNE1 binds to the KCNQ1 pore to regulate potassium channel activity. Neuron. 2004;42(6):927–37. Costi S, Han MH, Murrough JW. The Potential of KCNQ Potassium Channel Openers as Novel Antidepressants. CNS Drugs. 2022;36(3):207–16. Huang S, Li Y, Shen J, Liang W, Li C. Identification of a diagnostic model and molecular subtypes of major depressive disorder based on endoplasmic reticulum stress-related genes. Front Psychiatry. 2023;14:1168516. Feng J, Zhou Q, Gao W, Wu Y, Mu R. Seeking for potential pathogenic genes of major depressive disorder in the Gene Expression Omnibus database. Asia Pac Psychiatry. 2020;12(1):e12379. Martins-de-Souza D, Guest PC, Harris LW, Vanattou-Saifoudine N, Webster MJ, Rahmoune H, et al. Identification of proteomic signatures associated with depression and psychotic depression in post-mortem brains from major depression patients. Transl Psychiatry. 2012;2(3):e87. Roffé M, Beraldo FH, Bester R, Nunziante M, Bach C, Mancini G, et al. Prion protein interaction with stress-inducible protein 1 enhances neuronal protein synthesis via mTOR. Proc Natl Acad Sci U S A. 2010;107(29):13147–52. Beraldo FH, Arantes CP, Santos TG, Queiroz NG, Young K, Rylett RJ, et al. Role of alpha7 nicotinic acetylcholine receptor in calcium signaling induced by prion protein interaction with stress-inducible protein 1. J Biol Chem. 2010;285(47):36542–50. Mick E, McGough J, Loo S, Doyle AE, Wozniak J, Wilens TE, et al. Genome-wide association study of the child behavior checklist dysregulation profile. J Am Acad Child Adolesc Psychiatry. 2011;50(8):807 – 17.e8. Beraldo FH, Thomas A, Kolisnyk B, Hirata PH, De Jaeger X, Martyn AC, et al. Hyperactivity and attention deficits in mice with decreased levels of stress-inducible phosphoprotein 1 (STIP1). Dis Model Mech. 2015;8(11):1457–66. Additional Declarations The authors have declared there is NO conflict of interest to disclose Supplementary Files SupplementaryTable1.xlsx SupplementaryTable2.xlsx SupplementaryTable3.EnrichmentanalysisofkeyMDD.xlsx Supplementarytable4.OsrelatedgeneswitharelatedscoreXXX7.xlsx SupplementaryTable5.Inputofchromosomelocate.xlsx SupplementaryTable6.EnrichmentGOanalysisinclusterA.xlsx Supplementarytable7.SMRassociationbetweenexpressionofgeneMAPK3KCNE1STIP1andMDD.xlsx Cite Share Download PDF Status: Published Journal Publication published 19 Oct, 2024 Read the published version in Translational Psychiatry → Version 1 posted Editorial decision: revise 19 Aug, 2024 Review # 2 received at journal 26 Jul, 2024 Review # 1 received at journal 15 Jul, 2024 Review # 3 received at journal 14 Jul, 2024 Reviewer # 3 agreed at journal 09 Jul, 2024 Reviewer # 2 agreed at journal 09 Jul, 2024 Reviewer # 1 agreed at journal 08 Jul, 2024 Reviewers invited by journal 08 Jul, 2024 Submission checks completed at journal 28 Jun, 2024 First submitted to journal 27 Jun, 2024 Unknown event 27 Jun, 2024 Editor assigned by journal 26 Jun, 2024 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-4641375","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":324382921,"identity":"63b35488-6830-44a5-abd9-722219530cc3","order_by":0,"name":"Gang Zhu","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAArUlEQVRIiWNgGAWjYFACHgaGDxCWAfFaGGeQrIWZhyQt8v5nj0nb1GxLbGBv3ibBUHOHsBbDA+fSpHOO3U5s4DlWJsFw7BkRWhp7zKRzG4BaJHLMJBgbDhOhpZnHTNoSpEX+DZFa5NmAWhjBtvAQqcWAh8fYsufYbeM2nrRii4RjxNjSf8bwxo+a27L97Ic33vhQQ4wtB6AMNhCRQFgD0JYGYlSNglEwCkbByAYABLQ1Hdic6VkAAAAASUVORK5CYII=","orcid":"https://orcid.org/0000-0001-6967-8336","institution":"The First Affiliated Hospital of China Medical University","correspondingAuthor":true,"prefix":"","firstName":"Gang","middleName":"","lastName":"Zhu","suffix":""},{"id":324382922,"identity":"6fb4394b-0318-48a7-aa35-1f054135a8ae","order_by":1,"name":"Xiaojun Shao","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Xiaojun","middleName":"","lastName":"Shao","suffix":""},{"id":324382923,"identity":"511e15d3-953e-48bb-a148-9c431e914469","order_by":2,"name":"Yuan Wang","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Yuan","middleName":"","lastName":"Wang","suffix":""},{"id":324382924,"identity":"0ee0022d-8ae3-4c93-ba68-02c49566cf10","order_by":3,"name":"Yuan Wang","email":"","orcid":"","institution":"The First Affiliated Hospital of China Medical University","correspondingAuthor":false,"prefix":"","firstName":"Yuan","middleName":"","lastName":"Wang","suffix":""},{"id":324382925,"identity":"3f20735c-8986-43d4-b7a3-c31c4e488e3f","order_by":4,"name":"Guangming Liang","email":"","orcid":"","institution":"The First Affiliated Hospital of China Medical University","correspondingAuthor":false,"prefix":"","firstName":"Guangming","middleName":"","lastName":"Liang","suffix":""},{"id":324382926,"identity":"a1ec20fd-2c8d-4e3e-aedd-1d6370a0e025","order_by":5,"name":"Xiaotong Zhu","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Xiaotong","middleName":"","lastName":"Zhu","suffix":""},{"id":324382927,"identity":"459f314b-0c38-4611-abb1-0e161ccbbade","order_by":6,"name":"Lu Liu","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Lu","middleName":"","lastName":"Liu","suffix":""},{"id":324382928,"identity":"d27bada1-7c6f-4c06-afcd-97f0a2e4befe","order_by":7,"name":"Ming Meng","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Ming","middleName":"","lastName":"Meng","suffix":""},{"id":324382929,"identity":"fae81bfc-6bbe-4b14-8a4e-b277615ce3fe","order_by":8,"name":"Li Duan","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Li","middleName":"","lastName":"Duan","suffix":""}],"badges":[],"createdAt":"2024-06-26 09:00:45","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-4641375/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-4641375/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1038/s41398-024-03126-0","type":"published","date":"2024-10-19T04:00:00+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":61558594,"identity":"2a51db8c-6ae5-49f2-84bf-cb26f2e36aa0","added_by":"auto","created_at":"2024-08-01 08:08:17","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":617761,"visible":true,"origin":"","legend":"\u003cp\u003eFlowchart of the study. MDD, Major depressive depression；OS, Oxidative stress; DEG, Differentially expressed gene; PPI, Protein-Protein Interaction; GO, Gene ontology；GSEA, Single Gene Set Enrichment Analysis; ssGSEA, Single Sample Gene Set Enrichment Analysis; SMR, Summary data-based Mendelian randomization.\u003c/p\u003e","description":"","filename":"floatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-4641375/v1/e0129c027d0577dd93efd38a.png"},{"id":61558596,"identity":"c0f36045-2214-4df9-a925-1e8cd32cf889","added_by":"auto","created_at":"2024-08-01 08:08:17","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":104676,"visible":true,"origin":"","legend":"\u003cp\u003eGene expression distribution and batch effect correction for all MDD samples compared to healthy controls.\u003c/p\u003e","description":"","filename":"floatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-4641375/v1/dbfa159105b881de546744d1.png"},{"id":61558610,"identity":"978252f1-4c4d-4cf4-8607-6e731b150ec9","added_by":"auto","created_at":"2024-08-01 08:08:18","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":697486,"visible":true,"origin":"","legend":"\u003cp\u003eGene Ontogeny enrichment analysis of differentially expressed genes (DEGs) between MDD patients and healthy controls.\u003c/p\u003e","description":"","filename":"floatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-4641375/v1/3b0a32f403aa2b0860a7b804.png"},{"id":61558609,"identity":"854344e7-9dd1-46e6-ae05-52ea7782cbf1","added_by":"auto","created_at":"2024-08-01 08:08:18","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":233709,"visible":true,"origin":"","legend":"\u003cp\u003eIdentification of MDD-related and OS-related DEGs.\u003c/p\u003e","description":"","filename":"floatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-4641375/v1/b16f5ad11244dce4aaa1da01.png"},{"id":61558601,"identity":"97f4eb03-1efc-4991-b718-579d5f4f754e","added_by":"auto","created_at":"2024-08-01 08:08:17","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":305303,"visible":true,"origin":"","legend":"\u003cp\u003eMachine learning analysis of MDD-associated differentially expressed genes (DEGs) also involved in OS from the GSE39653 dataset. A-F. Results for the Bagged Tress algorithm (A), Bayesian algorithm (B), Random Forest algorithm (C), Wrapper algorithm (D), LQV algorithm (E), and 1000-times Lasso-Logistic algorithm (F). G. Genes identified as important for MDD–OS interactions by all 6 machine learning algorithms.\u003c/p\u003e","description":"","filename":"floatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-4641375/v1/c619b80b4f6724ebd0d578d7.png"},{"id":61559178,"identity":"45e14d19-0286-48a5-b274-4e274783c70f","added_by":"auto","created_at":"2024-08-01 08:16:17","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":176563,"visible":true,"origin":"","legend":"\u003cp\u003eNomogram model for MDD and OS-related genes from training dataset GSE39653 and validation dataset GSE98793.\u003c/p\u003e","description":"","filename":"floatimage6.png","url":"https://assets-eu.researchsquare.com/files/rs-4641375/v1/b608cd81da243d0014e1f925.png"},{"id":61558606,"identity":"794323ce-875f-4a77-8039-cb3f8e101341","added_by":"auto","created_at":"2024-08-01 08:08:18","extension":"png","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":312590,"visible":true,"origin":"","legend":"\u003cp\u003eUnsupervised cluster analysis of MDD patients in dataset GSE39653 revealing two distinct molecular phenotypes. A. Consensus clustering matrix with k = 2 as the optimal cluster number. B. Cumulative distribution function. C. Scree plot of cluster analysis. D. PCoA cluster validation analysis. E. Differential gene expression analysis of MDD-OS subtypes. F. GO enrichment analysis of highly expressed genes in ClusterA.\u003c/p\u003e","description":"","filename":"floatimage7.png","url":"https://assets-eu.researchsquare.com/files/rs-4641375/v1/4b2db5abba1649242a9d6d01.png"},{"id":61559180,"identity":"436beea7-bb6b-436a-a3b8-d9dc7b83f1cf","added_by":"auto","created_at":"2024-08-01 08:16:17","extension":"png","order_by":8,"title":"Figure 8","display":"","copyAsset":false,"role":"figure","size":227762,"visible":true,"origin":"","legend":"\u003cp\u003eImmune cell infiltration analysis of the GSE39653 dataset using the CIBERSORT algorithm.\u003c/p\u003e","description":"","filename":"floatimage8.png","url":"https://assets-eu.researchsquare.com/files/rs-4641375/v1/048611e24e43d9b0866edd1f.png"},{"id":61558603,"identity":"e9c689fa-a4a5-41c0-be9b-3acd0fef6e4c","added_by":"auto","created_at":"2024-08-01 08:08:17","extension":"png","order_by":9,"title":"Figure 9","display":"","copyAsset":false,"role":"figure","size":581676,"visible":true,"origin":"","legend":"\u003cp\u003eImmune cell infiltration analysis of the GSE39653 dataset using the ssGSEA algorithm.\u003c/p\u003e","description":"","filename":"floatimage9.png","url":"https://assets-eu.researchsquare.com/files/rs-4641375/v1/4debd871dc228672910ee5df.png"},{"id":61558599,"identity":"833d0b16-2dc8-4a79-a01e-a075c9299479","added_by":"auto","created_at":"2024-08-01 08:08:17","extension":"png","order_by":10,"title":"Figure 10","display":"","copyAsset":false,"role":"figure","size":404602,"visible":true,"origin":"","legend":"\u003cp\u003eSummary data-based Mendelian randomization associations between MDD and expression levels of \u003cem\u003eKCNE1\u003c/em\u003e, \u003cem\u003eMAPK3\u003c/em\u003e, and \u003cem\u003eSTIP1\u003c/em\u003e.\u003c/p\u003e","description":"","filename":"floatimage10.png","url":"https://assets-eu.researchsquare.com/files/rs-4641375/v1/60c3e28d5ae759225ad4c8bc.png"},{"id":67039533,"identity":"ddc5faed-5096-46af-9bf9-5fac4858a6f7","added_by":"auto","created_at":"2024-10-20 07:06:53","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":4087930,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-4641375/v1/4863f428-5c65-4201-919a-4422e49f5a7a.pdf"},{"id":61558608,"identity":"83bf890c-caf7-4e2a-b10a-ab03267337ff","added_by":"auto","created_at":"2024-08-01 08:08:18","extension":"xlsx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":47348524,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryTable1.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-4641375/v1/e3f13ea7d6b84042cf93d1c2.xlsx"},{"id":61558595,"identity":"e8a6c377-4252-4e95-b177-5c4d78fe7193","added_by":"auto","created_at":"2024-08-01 08:08:17","extension":"xlsx","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":1678447,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryTable2.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-4641375/v1/4c9d1ae89a7bc8538cac9a6b.xlsx"},{"id":61559176,"identity":"c802a382-f20b-42b9-b7f2-be3066997c5e","added_by":"auto","created_at":"2024-08-01 08:16:17","extension":"xlsx","order_by":3,"title":"","display":"","copyAsset":false,"role":"supplement","size":371874,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryTable3.EnrichmentanalysisofkeyMDD.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-4641375/v1/de3567712521b9d4068b7386.xlsx"},{"id":61559761,"identity":"0c134cae-3a5d-4906-b4ab-4e87651bf7b0","added_by":"auto","created_at":"2024-08-01 08:24:17","extension":"xlsx","order_by":4,"title":"","display":"","copyAsset":false,"role":"supplement","size":21427,"visible":true,"origin":"","legend":"","description":"","filename":"Supplementarytable4.OsrelatedgeneswitharelatedscoreXXX7.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-4641375/v1/03a18fd6154fc29efaf418e8.xlsx"},{"id":61559179,"identity":"6d47d9ec-d20b-4c3f-ade9-2d1c5aece340","added_by":"auto","created_at":"2024-08-01 08:16:17","extension":"xlsx","order_by":5,"title":"","display":"","copyAsset":false,"role":"supplement","size":8468744,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryTable5.Inputofchromosomelocate.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-4641375/v1/fbb07b2fb9b0edb162d11f13.xlsx"},{"id":61559182,"identity":"25a9e1b9-1cc8-4ebf-b6b6-b572f85b7d96","added_by":"auto","created_at":"2024-08-01 08:16:18","extension":"xlsx","order_by":6,"title":"","display":"","copyAsset":false,"role":"supplement","size":15517,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryTable6.EnrichmentGOanalysisinclusterA.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-4641375/v1/735e6a43ad932ed6366016c9.xlsx"},{"id":61559177,"identity":"cd0814ab-d438-4c68-a611-2ec4833e86ea","added_by":"auto","created_at":"2024-08-01 08:16:17","extension":"xlsx","order_by":7,"title":"","display":"","copyAsset":false,"role":"supplement","size":11193,"visible":true,"origin":"","legend":"","description":"","filename":"Supplementarytable7.SMRassociationbetweenexpressionofgeneMAPK3KCNE1STIP1andMDD.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-4641375/v1/248c4af5ca1ff856c713f2e1.xlsx"}],"financialInterests":"The authors have declared there is \u003cb\u003eNO\u003c/b\u003e conflict of interest to disclose","formattedTitle":"Novel therapeutic targets for major depressive disorder related to oxidative stress identified by integrative multi-omics and multi-trait study","fulltext":[{"header":"Introduction","content":"\u003cp\u003eMajor depressive disorder (MDD) is a common psychiatric illness characterized by persistent depressed mood accompanied by heterogenous cognitive, behavioral, and physical symptoms(\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e). According to the World Health Organization, approximately 280\u0026nbsp;million people worldwide are currently afflicted with MDD(\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e). A nationally representative cross-sectional survey from China (the China Mental Health Survey) revealed a weighted lifetime prevalence of 3.4% and weighted 12-month prevalence of 2.1%(\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e), with substantially higher weighted lifetime prevalence and weighted 12-month prevalence among adult females (8.0% and 4.2%, respectively) compared to males (5.7% and 3.0%)(\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e). It is generally accepted that genetic, biological, psychosocial, and personality traits all contribute to MDD risk, termed the diversified disease hypothesis of MDD(\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e). Pathological processes implicated in MDD include glutamatergic excitotoxicity, brain-derived neurotrophic factor/ tyrosine receptor kinase B signaling insufficiency, neuroinflammation, and gut microbiota\u0026ndash;brain axis disturbance(\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e, \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e). However, there are no effective treatments based on these mechanisms, implying complex multidimensional pathogenesis.\u003c/p\u003e \u003cp\u003eCellular metabolism and various signaling mechanisms result in the formation of highly reactive free radicals, including reactive oxygen species (OS) and reactive nitrogen species (RNS). Under normal physiological conditions, these species are neutralized by endogenous antioxidants; however, an imbalance between ROS or RNS production and scavenging will result in oxidative stress (OS), which leads to the oxidative damage of cellular lipids, proteins, and genomic DNA(\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e), while severe OS can trigger necrotic or apoptotic cell death. Oxidative stress and ensuing cell death is strongly implicated in neurodegenerative diseases such as the Alzheimer\u0026rsquo;s disease and Parkinson\u0026rsquo;s disease(\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e, \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e), and recent studies have suggested that OS also contributes to the pathophysiology and treatment response of MDD(\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e, \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e). For instance, a positive correlation was found between and the amplitude of low-frequency fluctuations in key MDD-associated brain regions, such as the thalamus, anterior cingulate gyrus, and superior frontal gyrus, and plasma concentrations of the antioxidant enzymes superoxide dismutase (SOD) and glutathione reductase(\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e). Furthermore, depression severity and working memory impairment were associated with higher plasma concentrations of malondialdehyde (MDA), an indicator of lipid peroxidation from oxidative stress, among recurrent MDD patients both before and after antidepressant treatment(\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e). Reduced SOD activity, lower levels of the non-enzyme antioxidant glutathione (GSH), and elevated lipid peroxidation products have been proposed as reliable biomarkers for MDD(\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e, \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e). In experiment animals as well, chronic unpredictable mild stress can induce depression-like symptoms and concomitant abnormalities in redox balance across brain subregions, including decreased GSH and SOD activities and higher levels of ROS, MDA, and carbonyl in prefrontal cortex and hippocampus(\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e). Despite a growing number of studies implicating OS in MDD, there are still discrepancies(\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e, \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e). An early meta-analysis of 23 studies with 4980 individuals found increased OS among MDD patients compared to matched individuals without depression, but some included studies found lower antioxidant biomarker concentrations (MDA and 8-oxo-2\u0026prime;-deoxyguanosine) in MDD(\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e). A recent meta-analysis also concluded that OS status was more severe in MDD, but found no significant differences in catalase, SOD, glutathione peroxidase, and uric acid between MDD patients and healthy controls(\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e).Therefore, the precise contributions of OS to MDD pathogenesis remained unclear.\u003c/p\u003e \u003cp\u003eTo address this issue, we conducted a comprehensive multi-omics analysis of the association between OS and MDD and identified multiple differentially expressed genes (DEGs) linking the two conditions. Machine learning models trained using these DEGs reliably distinguished patients from controls in an independent dataset. Thus, these genes may serve as feasible biomarkers or treatment targets. Furthermore, we utilized a multi-omics summary data-based Mendelian randomization (SMR) approach to identify putative causal OS-related genes for MDD.\u003c/p\u003e"},{"header":"Methods and materials","content":"\u003cp\u003eAn overview of the analytical methods is illustrated in Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eNo ethics committee approval was required for this summary-level study. The gene array data of three study cohorts were extracted from the Gene Expression Omnibus (GEO) database: GSE32280 (platform: GPL570)(\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e), GSE39653 (platform: GPL10558)(\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e) and GSE98793 (platform: GPL570)(\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e). In total these datasets included gene expression profiles for 165 MDD patients and 96 health controls. Genes related to OS were identified from the GeneCards database (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.genecards.org\u003c/span\u003e\u003cspan address=\"https://www.genecards.org\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e) using the keyword \u0026ldquo;oxidative stress\u0026rdquo;. Those with a relevance score\u0026thinsp;\u0026ge;\u0026thinsp;7 were included according to previous methods(\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e).\u003c/p\u003e \u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003eIdentification of differentially expressed genes (DEGs) associated with MDD and OS\u003c/h2\u003e \u003cp\u003eWe identified genes common among the disease datasets GSE32280, GSE39653, GSE98793, and GeneCards using the \u0026ldquo;cbind ()\u0026rdquo; function of R, and then removed batch bias and performed log2(X\u0026thinsp;+\u0026thinsp;1) normalization using the R sva function. These standardized gene expression data formed the basis for all further analysis. To analyze the effect of OS-related gene expression level on MDD risk, the R limma package was used to identify DEGs between disease and normal control samples, with the absolution value of log2-fold change (log2FC)\u0026thinsp;\u0026gt;\u0026thinsp;1 and P\u0026thinsp;\u0026lt;\u0026thinsp;0.05 set as thresholds. These DEGs were visualized by volcano plot and a heat map was constructed using the R packages heatmaps (version 1.0.12) and ggplot2 (version 3.3.2).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec4\" class=\"Section2\"\u003e \u003ch2\u003eScreening of hub genes mediating crosstalk between MDD and OS\u003c/h2\u003e \u003cp\u003eUnivariate logistic regression and multivariate logistic regression analyses of the GSE39653 dataset were used to identify hub genes within the MDD- and OS-related gene set (Hub MDD-OS DEGs) with P\u0026thinsp;\u0026lt;\u0026thinsp;0.05 considered statistically significant. This was followed by dimensionality reduction analysis.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec5\" class=\"Section2\"\u003e \u003ch2\u003eIdentification of key modules and functional genes\u003c/h2\u003e \u003cp\u003eTo reveal biological relationships among key genes and gene clusters related to OS and MDD, Gene Ontology (GO) enrichment analysis was conducted using the R package clusterProfiler. Genes were assessed for molecular function (MF), cell component (CC), and biological process (BP) annotations with a False Discovery Rate (FDR)-corrected P\u0026thinsp;\u0026lt;\u0026thinsp;0.05 considered statistically significant. The R package cowplot was then used to conduct Spearman correlation analyses between key OS- and MDD-related genes as well as to draw heat maps, scatter plots, and correlation curves. Again, P\u0026thinsp;\u0026lt;\u0026thinsp;0.05 was considered statistically significant. Key genes were then mapped to chromosomal locations using the R package RCircos based on data from the R package and downloaded from the ENSEMBL database. The open-source STRING database was then used to build a protein\u0026ndash;protein interaction (PPI) network(\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e), which was subsequently visualized using Cytoscape.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec6\" class=\"Section2\"\u003e \u003ch2\u003eIn-depth investigation of key genes\u003c/h2\u003e \u003cp\u003eKey genes mediating MDD\u0026ndash;OS interactions were further screened and evaluated using 6 machine learning algorithms, Bagged Trees, Bayesian, Random Forest, Wrapper (Bpruta), Learning Vector Quantification (LQV), and 1000 iterations 10-fold cross-validation Least Absolute Shrinkage and Selection Operator (LASSO). A DEG identified as a feature gene by at least 5 algorithms based on classification performance was considered an important Hub DEG for MDD\u0026ndash;OS interactions. The interactions among these key intersecting genes were then visualized using the R package \u0026ldquo;Upset\u0026rdquo; application. To explore the associations of individual hub gene pairs in the GSE39653 and GSE9873 datasets, we conducted Spearman\u0026rsquo;s correlation analyses and plotting of heat maps, scatter plots, and correlation curves using cowplot. Finally, RCircos was used to draw the chromosome localization map for display.\u003c/p\u003e \u003cp\u003eBased on these analyses, we constructed a nomogram based on the GSE39653 dataset using the R package rms. A nomogram model was then constructed based on the hub genes to predict the prevalence of OS-associated MDD. Calibration curve analysis was performed to evaluate the accuracy and resolution of the nomogram. Finally, receiver operating characteristic (ROC) curves were constructed to evaluate the accuracy of the diagnostic model for the onset of OS-associated MDD using the R package pROC. Thorough composition validation techniques were employed to confirm model performance for distinguishing MDD from control samples.\u003c/p\u003e \u003cp\u003eWe also developed a classification prediction model based on expression analysis of blood utilizing a combination of the aforementioned 6 machine learning algorithms. The performance of each model was evaluated by calculating the area under the ROC curve (AUC), followed by visual representation of the results (predictive genes) using heat maps. Optimal model performance was assessed using calibration curve and decision curve analysis (DCA).\u003c/p\u003e \u003cp\u003eThe R ConsensusClusterPlus package was used to identify subpopulations with distinct molecular phenotypes based on hub genes for MDD\u0026ndash;OS interactions(\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e). Principal co-ordinates analysis (PCoA) was then conducted to verify consistency of clustering. The R package ggpubr was used to draw box plots with sample cluster labels as groups, and the differences between groups were evaluated for statistical significance by the Wilcoxon rank sum test.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec7\" class=\"Section2\"\u003e \u003ch2\u003eImmuno-infiltration analysis\u003c/h2\u003e \u003cp\u003eThe relative abundances of specific infiltrating immune cells were estimated using CIBERSORT (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://cibersort.stanford.edu/)(27)\u003c/span\u003e\u003cspan address=\"https://cibersort.stanford.edu/)(27)\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e, an analytical tool designed to reveal the distribution levels of LM22 immune cells based on gene expression profiles. Distinct enrichment fractions of immune cells were then compared using the Wilcox test. We further performed quantitative Single Sample Gene Set Enrichment Analysis (ssGSEA) to calculate the abundance of immune cells associated with MDD\u0026ndash;OS interactions. Finally, differences in immune cell infiltration were visualized using ggplot2 with P\u0026thinsp;\u0026lt;\u0026thinsp;0.05 set as the threshold for significance.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003eSMR analysis\u003c/h2\u003e \u003cp\u003eWe gathered genome-wide association study (GWAS) summary statistics of 170,756 MDD cases and 329,443 controls (8,483,301 genetic variants in total) from the Psychiatric Genomics Consortium (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.med.unc.edu/pgc\u003c/span\u003e\u003cspan address=\"https://www.med.unc.edu/pgc\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e), one of the largest, most innovative, and productive platforms in psychiatry(\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e). Single nucleotide polymorphisms (SNPs) associated with expression quantitative trait loci (eQTLs) were selected as instrumental variables (IVs) to infer direct causal effects of gene expression or protein levels on MDD. Whole-blood eQTL summary statistics for 15,882 genes were obtained from 31,684 individuals in the eQTLGen database(\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e). The current study focused only on cis-eQTLs within 1-Mb from the start or end of the gene.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec9\" class=\"Section2\"\u003e \u003ch2\u003eStatistical analysis\u003c/h2\u003e \u003cp\u003eAll statistical analyses were performed using R 4.2.0. Continuous variables were compared between two groups using the Wilcoxon rank sum test and among three or more groups using the Kruskal\u0026ndash;Wallis test. Categorical variables were compared by the chi-square test or Fisher's exact test. Associations between immune cell abundance and gene expression levels were evaluated using Spearman's correlation tests. A P\u0026thinsp;\u0026lt;\u0026thinsp;0.05 was considered statistically significant for all tests.\u003c/p\u003e \u003cp\u003eCausal inferences between GWAS and cis-eQTLs were evaluated using the SMR multi-tool. The top associated cis-QTLs were selected by considering a window centered around the corresponding gene (\u0026plusmn;\u0026thinsp;1000 kb) and surpassing a P-value threshold of 5.0 \u0026times;10\u003csup\u003e\u0026minus;\u0026thinsp;8\u003c/sup\u003e. All SNPs with allele frequency differences larger than the specified threshold (0.2 in the current study) between datasets, including the LD reference sample, the QTL summary data, and the outcome summary data, were excluded. All SMR analysis were implemented using SMR v1.3.1 and included SNPs as instruments, key genes for MDD\u0026ndash;OS interactions as exposures, and MDD as the outcome (SMR P\u0026thinsp;\u0026lt;\u0026thinsp;0.05, cis-eQTLs, and GWAS P\u0026thinsp;\u0026lt;\u0026thinsp;1 \u0026times; 10\u003csup\u003e\u0026minus;\u0026thinsp;5\u003c/sup\u003e).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec10\" class=\"Section2\"\u003e \u003ch2\u003eSensitivity analysis\u003c/h2\u003e \u003cp\u003eThe strengths of SNPs used as instruments were assessed using the F-statistic, and we included only SNPs with an F-statistic\u0026thinsp;\u0026gt;\u0026thinsp;10 to minimize weak instrument bias(\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e). The heterogeneity in the dependent instrument (HEIDI) test was applied using SMR v1.3.1 to distinguish pleiotropy from linkage. All instruments with P-HEIDI\u0026thinsp;\u0026lt;\u0026thinsp;0.01 (indicating significant heterogeneity) were omitted from the analysis.\u003c/p\u003e \u003c/div\u003e"},{"header":"Results","content":"\u003cdiv id=\"Sec12\" class=\"Section2\"\u003e \u003ch2\u003eIdentification of DEGs between MDD cases and controls\u003c/h2\u003e \u003cp\u003eThe GEO datasets GSE32280, GSE39653, and GSE98793 were combined using the R \u0026ldquo;cbind ()\u0026rdquo; function and the batch bias mitigated using the R \u0026ldquo;sva\u0026rdquo; package (see Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e and Supplementary Table\u0026nbsp;1). In total, 596 DEGs were identified between MDD patients and healthy controls (HCs) using Limma, including 294 upregulated genes and 302 downregulated genes in MDD. The magnitudes of differential expression for significant DEGs (P\u0026thinsp;\u0026lt;\u0026thinsp;0.05) are shown as a volcano map in Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eA and as a heatmap in Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eB and details in Supplementary Table\u0026nbsp;2. These DEGs were enriched in GO CC terms \u0026lsquo;specific granule\u0026rsquo;, \u0026lsquo;specific granule membrane\u0026rsquo;, \u0026lsquo;secretory granule membrane\u0026rsquo;, \u0026lsquo;specific granule lumen\u0026rsquo;, and \u0026lsquo;tertiary granule\u0026rsquo;, BP terms \u0026lsquo;receptor signaling pathway via JAK-STAT and STAT\u0026rsquo;, \u0026lsquo;placenta development\u0026rsquo;, and \u0026lsquo;carbohydrate catabolic process\u0026rsquo;, and MF terms \u0026lsquo;identical protein binding\u0026rsquo;, \u0026lsquo;immune receptor activity\u0026rsquo;, \u0026lsquo;cytokine receptor activity\u0026rsquo;, \u0026lsquo;phospholipase activity\u0026rsquo;, and \u0026lsquo;1-phosphatidylinositol-3-kinase regulator activity\u0026rsquo; (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eC, D and Supplementary Table\u0026nbsp;3).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eTo identify MDD-associated genes also related to OS, these DEGs were searched against the 817 genes in GeneCards with relevance score\u0026thinsp;\u0026ge;\u0026thinsp;7 for OS, yielding 38 potential MDD\u0026ndash;OS interaction or crosstalk genes (Supplementary Table\u0026nbsp;4). The Venn diagram of these overlapping DEGs is shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eA and the chromosomal positions in Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eB. Among these 38 DEGs, \u003cem\u003eAMD\u003c/em\u003e, \u003cem\u003eALPP\u003c/em\u003e, \u003cem\u003eCAMK2G\u003c/em\u003e, \u003cem\u003eDDAH1\u003c/em\u003e, \u003cem\u003eKCNE1\u003c/em\u003e, \u003cem\u003eLEP\u003c/em\u003e, \u003cem\u003eMAPK3\u003c/em\u003e, \u003cem\u003eIL10\u003c/em\u003e, \u003cem\u003ePINK1\u003c/em\u003e, and \u003cem\u003eSLC2A1\u003c/em\u003e were significantly upregulated in the MDD\u0026ndash;OS samples (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eC). More details were shown in Supplementary Table\u0026nbsp;5.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec13\" class=\"Section2\"\u003e \u003ch2\u003eIdentification of key MDD\u0026ndash;OS interaction genes via machine learning\u003c/h2\u003e \u003cp\u003eMachine learning algorithms were trained to deduce the associations between the 38 MDD\u0026ndash;OS DEGs and MDD pathogenesis using the GSE39653 dataset, and model performance was validated using the GSE98793 dataset. Intersection of results yielded by the Bagged Trees algorithm (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eA), Bayesian algorithm (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eB), Random Forest algorithm (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eC), Wrapper algorithm (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eD), LQV (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eE) and 1000 times 10-fold cross-validation LASSO Logistic model (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eF) identified 32 of these genes as closely related to the pathogenesis of MDD and OS (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eG): \u003cem\u003eADM, AKR1C3, ALPP, CAMK2G, CREBBP, DDAH1, DNM1L, F5, FKBP5\u003c/em\u003e, \u003cem\u003eGADD45A\u003c/em\u003e, \u003cem\u003eGATB\u003c/em\u003e, \u003cem\u003eGDF15\u003c/em\u003e, \u003cem\u003eHSP90AA1\u003c/em\u003e, \u003cem\u003eHSP90AB1\u003c/em\u003e, \u003cem\u003eHSP90B1\u003c/em\u003e, \u003cem\u003eIL10\u003c/em\u003e, \u003cem\u003eINSR\u003c/em\u003e, \u003cem\u003eKCNE1\u003c/em\u003e (encoding Potassium Voltage-Gated Channel Subfamily E Regulatory Subunit 1), \u003cem\u003eKLF2\u003c/em\u003e, \u003cem\u003eLEP\u003c/em\u003e, \u003cem\u003eMAP2K1\u003c/em\u003e, \u003cem\u003eMAPK3\u003c/em\u003e (encoding mitogen-activated protein kinase 3), \u003cem\u003eMGST1\u003c/em\u003e, \u003cem\u003ePLA2G7\u003c/em\u003e, \u003cem\u003ePLAU\u003c/em\u003e, \u003cem\u003ePTK2B\u003c/em\u003e, \u003cem\u003eRETN\u003c/em\u003e, \u003cem\u003eSLC2A1\u003c/em\u003e, \u003cem\u003eSTIP1\u003c/em\u003e (encoding stress-induced phosphoprotein 1), \u003cem\u003eTNF\u003c/em\u003e, \u003cem\u003eUGT1A1\u003c/em\u003e, and \u003cem\u003eVDR\u003c/em\u003e. Subsequently, we constructed a PPI network with these 32 genes as hubs using the STRING database. Based on the criteria |Logfc| \u0026lt; 0.05 and P value\u0026thinsp;\u0026lt;\u0026thinsp;0.05, this 32-node PPI network included 341 edges, indicating complex multilevel interactions between OS and MDD pathogenesis.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eA diagnostic model was then constructed based on the GSE39653 dataset (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eA), and discrimination was verified on both the training set GSE39653 and validation set GSE98793. The recall curve and Hosmer\u0026thinsp;\u0026minus;\u0026thinsp;Lemoeshow goodness-of-fit test results were equal to 1 for the training set (GSE39653), indicating a low probability of type I error, that prediction results were close to the real data, and that the calibration degree of the model was high (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eB). The AUC of the diagnostic model was also 1.00 for the training set GSE39653 (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eC). The C-index of the diagnostic model was 0.99 for the training set (GSE39653) and 1 for the validation set (GSE98793) (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eD), and the recall curve and Hosmer\u0026thinsp;\u0026minus;\u0026thinsp;Lemeshow test result P values were equal to 1.00 (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eE). Finally, the AUC of the diagnostic model was 0.876 for the validation set (GSE98793) (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eF).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec14\" class=\"Section2\"\u003e \u003ch2\u003eAssociations of key genes with MDD subtypes\u003c/h2\u003e \u003cp\u003eThe physical, behavioral, and cognitive symptoms of MDD can vary substantially among patients, so we examined the potential functions of these 32 key genes in distinct MDD phenotypes by cluster analysis using ConsensusClusterPlus. This analysis yielded two patient clusters in the GSE39653 dataset, Cluster A and Cluster B, based on DEG profiles (Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003eA). The slope of consistency index for different classifications (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003eB) further confirmed that two was the optimal cluster number for this cohort, while the scree plot in Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003eC revealed the inflection point for best classification. We further validated the stability of this MDD patient subtyping in GSE39653 by PCoA (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003eD).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eTo explore differences in pathogenesis between the two subtypes, the total DEGs for Cluster A and Cluster B were first identified. In total, 3560 DEGs were found, including 1367 expressed at higher levels in Cluster A and 1223 expressed at higher levels in Cluster B (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003eE). Then, GO enrichment analysis was conducted to reveal differential enrichment of biological functions. Enriched BP terms in Cluster A included \u0026lsquo;leukocyte degranulation\u0026rsquo;, \u0026lsquo;aerobic electron transport chain\u0026rsquo;, and \u0026lsquo;ATP synthesis coupled electron transport\u0026rsquo;, while enriched CC terms included \u0026lsquo;respiratory chain complex\u0026rsquo; and \u0026lsquo;ficolin-1-rich granule lumen\u0026rsquo;, and enriched MF terms included \u0026lsquo;antioxidant activity\u0026rsquo; and \u0026lsquo;oxidoreduction-driven active transmembrane transporter activity\u0026rsquo; (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003eF and Supplementary Table\u0026nbsp;6).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec15\" class=\"Section2\"\u003e \u003ch2\u003eAnalysis of immune characteristics\u003c/h2\u003e \u003cp\u003eImmune cell infiltration is a major driver of OS in the brain, suggesting an important contribution to OS-related MDD pathogenesis. We employed Cibersort analysis to calculate the infiltration status of 22 immune cells in the GSE39653 dataset, and the relative abundance of each immune cell type was then compared by the Wilcoxon sign rank test (Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003eA). The CIBERSORT algorithm was also used to compare immune cell infiltration profiles between Cluster A and Cluster B. The results revealed that dendritic cells and activated mast cells were highly abundant in Cluster B (Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003eB), and that the relative abundances of most immune cell subsets differed between clusters (Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003eC). Moreover, the ssGSEA algorithm indicated greater infiltration of activated CD8\u0026thinsp;+\u0026thinsp;T cells, effector memory CD8\u0026thinsp;+\u0026thinsp;T cells, regulatory T cells, Type 1 T helper cell, eosinophils, macrophages, and monocytes in MDD samples of Cluster A (Fig.\u0026nbsp;\u003cspan refid=\"Fig9\" class=\"InternalRef\"\u003e9\u003c/span\u003eA) (P\u0026thinsp;\u0026lt;\u0026thinsp;0.05). Furthermore, abundances were correlated with the expression levels of the 32 most important DEGs (all P\u0026thinsp;\u0026lt;\u0026thinsp;0.05) (Fig.\u0026nbsp;\u003cspan refid=\"Fig9\" class=\"InternalRef\"\u003e9\u003c/span\u003eB).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec16\" class=\"Section2\"\u003e \u003ch2\u003eSMR analysis of characteristic genes\u003c/h2\u003e \u003cp\u003eWe also performed SMR analysis to evaluate the association strengths of these 32 key genes with MDD (with P-smr\u0026thinsp;\u0026lt;\u0026thinsp;0.05 set as the threshold for statistical significance). Results revealed an association between elevated expression of \u003cem\u003eKCNE1\u003c/em\u003e and MDD odds (OR\u0026thinsp;=\u0026thinsp;1.057, 95% CI\u0026thinsp;=\u0026thinsp;1.013\u0026ndash;1.102, P-smr\u0026thinsp;=\u0026thinsp;0.010) (Fig.\u0026nbsp;\u003cspan refid=\"Fig10\" class=\"InternalRef\"\u003e10\u003c/span\u003eA, B, and G). Similarly, elevated \u003cem\u003eMAPK3\u003c/em\u003e expression was associated with greater MDD odds (OR\u0026thinsp;=\u0026thinsp;1.023, 95% CI\u0026thinsp;=\u0026thinsp;1.004\u0026ndash;1.043, P-smr\u0026thinsp;=\u0026thinsp;0.020) (Fig.\u0026nbsp;\u003cspan refid=\"Fig10\" class=\"InternalRef\"\u003e10\u003c/span\u003eC, D, and G), while upregulation of \u003cem\u003eSTIP1\u003c/em\u003e was associated with reduced MDD odds (OR\u0026thinsp;=\u0026thinsp;0.792, 95% CI\u0026thinsp;=\u0026thinsp;0.641\u0026ndash;0.979, P-smr\u0026thinsp;=\u0026thinsp;0.031) (Fig.\u0026nbsp;\u003cspan refid=\"Fig10\" class=\"InternalRef\"\u003e10\u003c/span\u003eE, F, and G). Supplementary Table\u0026nbsp;7 showed the SMR association between expression of gene \u003cem\u003eMAPK3\u003c/em\u003e, \u003cem\u003eKCNE1\u003c/em\u003e, \u003cem\u003eSTIP1\u003c/em\u003e and MDD.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec17\" class=\"Section2\"\u003e \u003ch2\u003eSensitivity analysis\u003c/h2\u003e \u003cp\u003eThe F-statistic of all SNPs included in the analysis ranged from 29.855 to 3394.048, indicating a powerful instrumental variable\u0026ndash;exposure association (threshold set at 10) (Supplementary Table\u0026nbsp;7). Results of the HEIDI test further suggested that all observed associations were not due to linkage (p\u0026thinsp;\u0026gt;\u0026thinsp;0.01).\u003c/p\u003e \u003c/div\u003e"},{"header":"Discussion","content":"\u003cp\u003eTo our best knowledge, this is the first study on the contributions of oxidative stress-related genes to MDD pathogenesis using integrated multi-omics, machine learning, infiltrated immune cell profiling, genome-wide association, and summary data-based Mendelian randomization analysis. We identified 38 genes differentially expressed between MDD patients and controls that were also associated with OS, of which 32 were deemed important to the influence of OS on MDD pathogenesis (MDD\u0026ndash;OS interaction genes) in training and validation cohorts by 6 separate machine learning algorithms. Further screening of blood tissue expression profiles by SMR analysis identified \u003cem\u003eKCNE1, MAPK3\u003c/em\u003e, and \u003cem\u003eSTIP1\u003c/em\u003e as key linkage genes between OS and MDD. These DEGs may thus be convenient biomarkers for MDD as well as potential treatment targets.\u003c/p\u003e \u003cp\u003eNeuroinflammation in strongly implicated in MDD as evidenced by elevated inflammatory marker concentrations, infiltrating immune cell numbers, and antibody titers(\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e). These inflammatory processes both generate and are promoted by ROS and RNS (hence the OS\u0026ndash;inflammation interaction is also known as \u0026ldquo;evil twins of aging\u0026rdquo;), further implicating OS in MDD pathogenesis(\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e). Elevated ROS production leads to GSH depletion, oxidative damage, and ultimately enhanced inflammation(\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e). Excessive ROS can promote the expression of proinflammatory cytokines through several pathways, including activation of promoting protein-1 and nuclear factor kappa-B (NFκB), increased histone acetylation, and activation of caspase-1 and NOD-like receptor thermal protein domain associated protein 3(\u003cspan additionalcitationids=\"CR35\" citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e). Inflammatory reactions induce expression and release of peroxiredoxin 2, which in turn stimulates macrophages to release pro-inflammatory tumor necrosis factor-α (TNF-α)(\u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e). Elevated TNF-α had been detected in serum and in multiple brain subregions (including the anterior cingulate cortex, prefrontal cortex, and hippocampus) of MDD patients(\u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e, \u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e). Neuroinflammation also promotes the kynurenine pathway and ensuing quinolinic acid generation by activating indoleamine 2,3-dioxygenase (IDO), tryptophan-2,3-dioxygenase (TDO), and kynurenine 3-mono-oxygenase, which induces mitochondrial damage and results in further ROS production, glutamate release, N-methyl-D-aspartic acid (NMDA) receptor activation, Ca\u003csup\u003e2+\u003c/sup\u003e influx, and mitochondrial calcium overload, the end result of which is loss of mitochondrial membrane potential, reduced ATP generation, and accelerated ROS generation(\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e, \u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e40\u003c/span\u003e, \u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e41\u003c/span\u003e). In addition, IDO and TDO activation may reduce 5-HT biosynthesis, and 5-HT insufficiency is widely believed to result in low mood(\u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e42\u003c/span\u003e). These relationships also appear to be bidirectional, such that OS can promote neuroinflammation and vice versa in MDD. In accord with previous studies, we found that multiple immune-inflammatory gene pathways were activated in MDD compared to controls. These genes may in turn mediate the reciprocal exacerbation of OS generation and neuroinflammation leading to MDD. The OS\u0026ndash;MDD crosstalk genes identified in this study are primarily involved in immune cell function, including activated CD8\u0026thinsp;+\u0026thinsp;T cells, effector memory CD8\u0026thinsp;+\u0026thinsp;T cells, regulatory T cells, type 1 T helper cells, eosinophils, macrophages, and monocytes, further supporting shared immune-inflammatory mechanisms in OS and MDD.\u003c/p\u003e \u003cp\u003eWe also conducted SMR analysis to identify new causal genes for MDD as such genes may be prime drug targets. Upregulation of \u003cem\u003eKCNE1\u003c/em\u003e and \u003cem\u003eMAPK3\u003c/em\u003e were found to increase MDD risk, potentially by promoting pathogenic mechanisms involving OS. The \u003cem\u003eMAPK3\u003c/em\u003e product extracellular signal-regulated kinase 1 (ERK1) regulates cell proliferation, differentiation, and cell cycle progression among other vital processes(\u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e43\u003c/span\u003e). It has been reported that ERK signaling is significantly downregulated in the prefrontal cortex and hippocampus of both human patients and animal models of chronic depression(\u003cspan additionalcitationids=\"CR45\" citationid=\"CR44\" class=\"CitationRef\"\u003e44\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e46\u003c/span\u003e). The ERK1/2 isoforms are the most thoroughly investigated and well characterized isoforms in the central nervous system(\u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e42\u003c/span\u003e, \u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e47\u003c/span\u003e, \u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e48\u003c/span\u003e), and both have been found to promote OS via ROS production and to amplify the inflammatory response through activation of the stress-responsive transcription factor NFκB(\u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e49\u003c/span\u003e, \u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e50\u003c/span\u003e). At present, most studies on the role of \u003cem\u003eMAPK3\u003c/em\u003e in MDD have focused on the brain, while few studies have investigated expression changes in more accessible blood samples. Moreover, most studies have focused on ERK1/2, but few specifically on ERK1. We found higher \u003cem\u003eMAPK3\u003c/em\u003e expression in the blood tissue of MDD patients compared to controls, consistent with previous findings. One prospective case\u0026ndash;control study reported that a \u003cem\u003eMAPK3\u003c/em\u003e SNP enhanced interferon-α-induced depression, possibly by increasing the propensity for glutamate dysregulation(\u003cspan citationid=\"CR51\" class=\"CitationRef\"\u003e51\u003c/span\u003e). A bioinformatics analysis identified 5 genes including \u003cem\u003eMAPK3\u003c/em\u003e as key modulators of post-stroke depression risk, disease biomarkers, and therapeutic targets of acupuncture(\u003cspan citationid=\"CR52\" class=\"CitationRef\"\u003e52\u003c/span\u003e). Others have found significant associations of \u003cem\u003eMAPK3\u003c/em\u003e with schizophrenia, and a recent genome-wide Mendelian randomization analysis identified \u003cem\u003eMAPK3\u003c/em\u003e as a potential drug target for schizophrenia treatment(\u003cspan citationid=\"CR53\" class=\"CitationRef\"\u003e53\u003c/span\u003e), in line with previous studiess(\u003cspan citationid=\"CR54\" class=\"CitationRef\"\u003e54\u003c/span\u003e, \u003cspan citationid=\"CR55\" class=\"CitationRef\"\u003e55\u003c/span\u003e). Based on these and our own findings, we speculate that \u003cem\u003eMAPK3\u003c/em\u003e may be a critical mediator of OS effects on MDD pathogenesis and thus a promising therapeutic target. However, in our present study, \u003cem\u003eMAPK3\u003c/em\u003e appeared to make only a limited contribution (OR\u0026thinsp;=\u0026thinsp;1.023, 95% CI\u0026thinsp;=\u0026thinsp;1.004\u0026ndash;1.043). Nonetheless, the contributions of \u003cem\u003eMAPK3\u003c/em\u003e to OS and MDD warrant further exploration.\u003c/p\u003e \u003cp\u003eIn contrast to MAPKs, few studies have examined the genetic association of \u003cem\u003eKCNE1\u003c/em\u003e with MDD, although McCaffery and colleagues proposed that \u003cem\u003eKCNE1\u003c/em\u003e is associated with longer-term changes in depressive symptoms(\u003cspan citationid=\"CR56\" class=\"CitationRef\"\u003e56\u003c/span\u003e). The KCNE family proteins are regulatory subunits of voltage-gated K(+) channels(\u003cspan citationid=\"CR57\" class=\"CitationRef\"\u003e57\u003c/span\u003e), and are implicated in multiple arrhythmogenic cardiac myocardium diseases(\u003cspan citationid=\"CR58\" class=\"CitationRef\"\u003e58\u003c/span\u003e). The \u003cem\u003eKCNE1\u003c/em\u003e subunit regulates the neuronal membrane potential through modulation of K(+) channels, including KCNQ channels(\u003cspan citationid=\"CR59\" class=\"CitationRef\"\u003e59\u003c/span\u003e). Further, the KCNQ channel modulator retigabine has been shown to improve depressive symptoms, suggesting therapeutic potential for MDD(\u003cspan citationid=\"CR60\" class=\"CitationRef\"\u003e60\u003c/span\u003e). Another study also included \u003cem\u003eKCNE1\u003c/em\u003e expression in a diagnostic model for MDD(\u003cspan citationid=\"CR61\" class=\"CitationRef\"\u003e61\u003c/span\u003e), although no causal association was suggested. In the current study, preliminary genomic analysis indicated that \u003cem\u003eKCNE1\u003c/em\u003e was upregulated in MDD and positioned as a linker gene between MDD and OS, while according to SMR analysis, \u003cem\u003eKCNE1\u003c/em\u003e upregulation increases the risk of MDD. We speculate that drugs targeting \u003cem\u003eKCNE1\u003c/em\u003e could show therapeutic efficacy against MDD.\u003c/p\u003e \u003cp\u003eThese same genomics analyses also revealed downregulation of \u003cem\u003eSTIP1\u003c/em\u003e, which encodes a co-chaperone that interacts with heat-shock proteins 70 and 90, in the blood tissue of MDD patients, in accord with previous reports(\u003cspan citationid=\"CR62\" class=\"CitationRef\"\u003e62\u003c/span\u003e, \u003cspan citationid=\"CR63\" class=\"CitationRef\"\u003e63\u003c/span\u003e). Further, SMR identified \u003cem\u003eSTIP1\u003c/em\u003e as a protective target against MDD (OR\u0026thinsp;=\u0026thinsp;0.792, 95% CI\u0026thinsp;=\u0026thinsp;0.641\u0026ndash;0.979). Thus, activation of \u003cem\u003eSTIP1\u003c/em\u003e expression may be a useful therapeutic strategy against MDD. In addition to acting as a chaperone, extracellular \u003cem\u003eSTIP1\u003c/em\u003e acts as a trophic factor to engage PrP\u003csup\u003eC\u003c/sup\u003e, thereby enhancing neuritogenesis and neuronal survival(\u003cspan citationid=\"CR64\" class=\"CitationRef\"\u003e64\u003c/span\u003e, \u003cspan citationid=\"CR65\" class=\"CitationRef\"\u003e65\u003c/span\u003e). Studies have also implicated \u003cem\u003eSTIP1\u003c/em\u003e in functional recovery after stroke and regulation of Aβ peptide toxicity in Alzheimer's disease models. Moreover, a GWAS analysis identified a \u003cem\u003eSTIP1\u003c/em\u003e polymorphism as a potential risk factor for attention-deficit disorder(\u003cspan citationid=\"CR66\" class=\"CitationRef\"\u003e66\u003c/span\u003e). Mice with elevated \u003cem\u003eSTIP1\u003c/em\u003e levels (up to nearly fivefold) showed no neuropathology, anxiety-like behaviors, depression-like behaviors, spatial memory deficits, or attention deficits(\u003cspan citationid=\"CR67\" class=\"CitationRef\"\u003e67\u003c/span\u003e), suggesting that \u003cem\u003eSTIP1\u003c/em\u003e augmentation may be a feasible strategy for antidepressant treatment; however, the detailed underlying mechanisms remain unclarified.\u003c/p\u003e \u003cdiv id=\"Sec19\" class=\"Section2\"\u003e \u003ch2\u003eStudy limitations:\u003c/h2\u003e \u003cp\u003eThis study has several limitations. First, it is possible that differences in gene expression between MDD patients and controls reflect the influences of factors such as age, sex, smoking, medications, and other health conditions. Second, we only focused on the cis-regions of OS and MDD genes, despite the possibility that trans-eQTL SNPs (SNPs\u0026thinsp;\u0026gt;\u0026thinsp;5 Mb from the gene) may have a widespread impact on regulatory networks. Finally, functional experiments are still needed to confirm the importance of these DEGs in MDD pathogenesis through OS-dependent or OS-independent pathways.\u003c/p\u003e \u003c/div\u003e"},{"header":"Conclusion","content":"\u003cp\u003eThis integrative multi-omics and multi-trait study identified numerous genes linking OS to MDD pathogenesis, including three genes, \u003cem\u003eKCNE1, MAPK3\u003c/em\u003e, and \u003cem\u003eSTIP1\u003c/em\u003e, causally associated with MDD. These gene in particular could serve as diagnostic markers and drug targets for MDD treatment. In addition, the dozens of other genes identify may provide clues to novel pathological mechanisms for MDD.\u003c/p\u003e "},{"header":"Declarations","content":"\u003cp\u003e \u003ch2\u003eDeclaration of interests\u003c/h2\u003e \u003cp\u003eAll authors declare no competing interests.\u003c/p\u003e \u003c/p\u003e\u003cp\u003e \u003ch2\u003eData sharing\u003c/h2\u003e \u003cp\u003eOriginal data of major depressive depression could be retrieved from References(\u003cspan additionalcitationids=\"CR22\" citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e, \u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e). The raw data of oxidative stress could be obtained from the GeneCards database (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.genecards.org\u003c/span\u003e\u003cspan address=\"https://www.genecards.org\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e). The data analyzed in this paper were publicly available and the specific information and links of source data of all Figures and Tables are showed in Supplementary files.\u003c/p\u003e \u003c/p\u003e\u003ch2\u003eFunding\u003c/h2\u003e \u003cp\u003eThis work was supported by grants awarded by the Major Project of the Department of Science \u0026amp; Technology, Liaoning Province (2019JH8/10300019).\u003c/p\u003e\u003ch2\u003eCode availability\u003c/h2\u003e \u003cp\u003eR Code for the current analysis was available at Supplementary material- \u0026ldquo;Source code. R\u0026rdquo;. Please contact the corresponding author if you would like to see any data that are not included in the Article or the Appendix.\u003c/p\u003e\u003cdiv id=\"Sec21\" class=\"Section2\"\u003e \u003ch2\u003eContributors\u003c/h2\u003e \u003cp\u003eGZ and XJS conceived and planned the study. XJS, YW, ZLG, GML, XTZ, LJ, MM and LD selected the articles and extracted data. XJS and YW did the statistical analyses. XJS wrote the report. GZ had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.\u003c/p\u003e \u003c/div\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eAmerican Psychiatric Association D, Association AP. Diagnostic and statistical manual of mental disorders: DSM-5: American psychiatric association Washington, DC; 2013.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHerrman H, Patel V, Kieling C, Berk M, Buchweitz C, Cuijpers P, et al. Time for united action on depression: a Lancet-World Psychiatric Association Commission. Lancet. 2022;399(10328):957\u0026ndash;1022.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHuang Y, Wang Y, Wang H, Liu Z, Yu X, Yan J, et al. Prevalence of mental disorders in China: a cross-sectional epidemiological study. Lancet Psychiatry. 2019;6(3):211\u0026ndash;24.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLu J, Xu X, Huang Y, Li T, Ma C, Xu G, et al. Prevalence of depressive disorders and treatment in China: a cross-sectional epidemiological study. Lancet Psychiatry. 2021;8(11):981\u0026ndash;90.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eShao X, Zhu G. Associations Among Monoamine Neurotransmitter Pathways, Personality Traits, and Major Depressive Disorder. Front Psychiatry. 2020;11:381.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHu S, Sheng WS, Ehrlich LC, Peterson PK, Chao CC. Cytokine effects on glutamate uptake by human astrocytes. Neuroimmunomodulation. 2000;7(3):153\u0026ndash;9.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZheng P, Zeng B, Zhou C, Liu M, Fang Z, Xu X, et al. Gut microbiome remodeling induces depressive-like behaviors through a pathway mediated by the host's metabolism. Mol Psychiatry. 2016;21(6):786\u0026ndash;96.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCamini FC, da Silva Caetano CC, Almeida LT, de Brito Magalh\u0026atilde;es CL. Implications of oxidative stress on viral pathogenesis. Arch Virol. 2017;162(4):907\u0026ndash;17.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBai R, Guo J, Ye XY, Xie Y, Xie T. Oxidative stress: The core pathogenesis and mechanism of Alzheimer's disease. Ageing Res Rev. 2022;77:101619.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDong-Chen X, Yong C, Yang X, Chen-Yu S, Li-Hua P. Signaling pathways in Parkinson's disease: molecular mechanisms and therapeutic interventions. Signal Transduct Target Ther. 2023;8(1):73.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJim\u0026eacute;nez-Fern\u0026aacute;ndez S, Gurpegui M, D\u0026iacute;az-Atienza F, P\u0026eacute;rez-Costillas L, Gerstenberg M, Correll CU. Oxidative stress and antioxidant parameters in patients with major depressive disorder compared to healthy controls before and after antidepressant treatment: results from a meta-analysis. J Clin Psychiatry. 2015;76(12):1658\u0026ndash;67.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWang H, Jin M, Xie M, Yang Y, Xue F, Li W, et al. Protective role of antioxidant supplementation for depression and anxiety: A meta-analysis of randomized clinical trials. J Affect Disord. 2023;323:264\u0026ndash;79.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLiu Y, Zhang B, Zhou Y, Li M, Gao Y, Qin W, et al. Plasma oxidative stress marker levels related to functional brain abnormalities in first-episode drug-naive major depressive disorder. Psychiatry Res. 2024;333:115742.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTalarowska M, Gałecki P, Maes M, Gardner A, Chamielec M, Orzechowska A, et al. Malondialdehyde plasma concentration correlates with declarative and working memory in patients with recurrent depressive disorder. Molecular biology reports. 2012;39:5359\u0026ndash;66.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLindqvist D, Dhabhar FS, James SJ, Hough CM, Jain FA, Bersani FS, et al. Oxidative stress, inflammation and treatment response in major depression. Psychoneuroendocrinology. 2017;76:197\u0026ndash;205.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCamkurt MA, Fındıklı E, İzci F, Kurutaş EB, Tuman TC. Evaluation of malondialdehyde, superoxide dismutase and catalase activity and their diagnostic value in drug na\u0026iuml;ve, first episode, non-smoker major depression patients and healthy controls. Psychiatry Res. 2016;238:81\u0026ndash;5.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eParul, Mishra A, Singh S, Singh S, Tiwari V, Chaturvedi S, et al. Chronic unpredictable stress negatively regulates hippocampal neurogenesis and promote anxious depression-like behavior via upregulating apoptosis and inflammatory signals in adult rats. Brain Res Bull. 2021;172:164\u0026ndash;79.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLiu T, Zhong S, Liao X, Chen J, He T, Lai S, et al. A Meta-Analysis of Oxidative Stress Markers in Depression. PLoS One. 2015;10(10):e0138904.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePalta P, Samuel LJ, Miller ER, 3rd, Szanton SL. Depression and oxidative stress: results from a meta-analysis of observational studies. Psychosom Med. 2014;76(1):12\u0026ndash;9.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJim\u0026eacute;nez-Fern\u0026aacute;ndez S, Gurpegui M, Garrote-Rojas D, Guti\u0026eacute;rrez-Rojas L, Carretero MD, Correll CU. Oxidative stress parameters and antioxidants in adults with unipolar or bipolar depression versus healthy controls: Systematic review and meta-analysis. J Affect Disord. 2022;314:211\u0026ndash;21.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYi Z, Li Z, Yu S, Yuan C, Hong W, Wang Z, et al. Blood-based gene expression profiles models for classification of subsyndromal symptomatic depression and major depressive disorder. PLoS One. 2012;7(2):e31283.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSavitz J, Frank MB, Victor T, Bebak M, Marino JH, Bellgowan PS, et al. Inflammation and neurological disease-related genes are differentially expressed in depressed patients with mood disorders and correlate with morphometric and functional imaging abnormalities. Brain Behav Immun. 2013;31:161\u0026ndash;71.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLeday GGR, V\u0026eacute;rtes PE, Richardson S, Greene JR, Regan T, Khan S, et al. Replicable and Coupled Changes in Innate and Adaptive Immune Gene Expression in Two Case-Control Studies of Blood Microarrays in Major Depressive Disorder. Biol Psychiatry. 2018;83(1):70\u0026ndash;80.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eXu S, Li X, Zhang S, Qi C, Zhang Z, Ma R, et al. Oxidative stress gene expression, DNA methylation, and gut microbiota interaction trigger Crohn's disease: a multi-omics Mendelian randomization study. BMC Med. 2023;21(1):179.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWagner RA, Fischer MJ. The string-to-string correction problem. Journal of the ACM (JACM). 1974;21(1):168\u0026ndash;73.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWilkerson MD, Hayes DN. ConsensusClusterPlus: a class discovery tool with confidence assessments and item tracking. Bioinformatics. 2010;26(12):1572\u0026ndash;3.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChen B, Khodadoust MS, Liu CL, Newman AM, Alizadeh AA. Profiling Tumor Infiltrating Immune Cells with CIBERSORT. Methods Mol Biol. 2018;1711:243\u0026ndash;59.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHoward DM, Adams MJ, Clarke TK, Hafferty JD, Gibson J, Shirali M, et al. Genome-wide meta-analysis of depression identifies 102 independent variants and highlights the importance of the prefrontal brain regions. Nat Neurosci. 2019;22(3):343\u0026ndash;52.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eV\u0026otilde;sa U, Claringbould A, Westra HJ, Bonder MJ, Deelen P, Zeng B, et al. Large-scale cis- and trans-eQTL analyses identify thousands of genetic loci and polygenic scores that regulate blood gene expression. Nat Genet. 2021;53(9):1300\u0026ndash;10.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBurgess S, Thompson SG. Avoiding bias from weak instruments in Mendelian randomization studies. Int J Epidemiol. 2011;40(3):755\u0026ndash;64.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBeurel E, Toups M, Nemeroff CB. The Bidirectional Relationship of Depression and Inflammation: Double Trouble. Neuron. 2020;107(2):234\u0026ndash;56.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJoseph J, Shukitt-Hale B, Casadesus G, Fisher D. Oxidative stress and inflammation in brain aging: nutritional considerations. Neurochemical research. 2005;30:927\u0026ndash;35.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGhezzi P. Role of glutathione in immunity and inflammation in the lung. International journal of general medicine. 2011:105\u0026ndash;13.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ede Oliveira LG, Angelo YdS, Iglesias AH, Peron JPS. Unraveling the link between mitochondrial dynamics and neuroinflammation. Frontiers in Immunology. 2021;12:624919.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAdcock IM, Cosio B, Tsaprouni L, Barnes PJ, Ito K. Redox regulation of histone deacetylases and glucocorticoid-mediated inhibition of the inflammatory response. Antioxidants \u0026amp; redox signaling. 2005;7(1\u0026ndash;2):144\u0026ndash;52.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRahman I, Marwick J, Kirkham P. Redox modulation of chromatin remodeling: impact on histone acetylation and deacetylation, NF-κB and pro-inflammatory gene expression. Biochemical pharmacology. 2004;68(6):1255\u0026ndash;67.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBehl T, Rana T, Alotaibi GH, Shamsuzzaman M, Naqvi M, Sehgal A, et al. Polyphenols inhibiting MAPK signalling pathway mediated oxidative stress and inflammation in depression. Biomedicine \u0026amp; Pharmacotherapy. 2022;146:112545.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGritti D, Delvecchio G, Ferro A, Bressi C, Brambilla P. Neuroinflammation in major depressive disorder: a review of PET imaging studies examining the 18-kDa translocator protein. Journal of affective disorders. 2021;292:642\u0026ndash;51.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eEnache D, Pariante CM, Mondelli V. Markers of central inflammation in major depressive disorder: a systematic review and meta-analysis of studies examining cerebrospinal fluid, positron emission tomography and post-mortem brain tissue. Brain, behavior, and immunity. 2019;81:24\u0026ndash;40.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMoncrieff J, Cooper RE, Stockmann T, Amendola S, Hengartner MP, Horowitz MA. The serotonin theory of depression: a systematic umbrella review of the evidence. Molecular psychiatry. 2023;28(8):3243\u0026ndash;56.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eVisentin APV, Colombo R, Scotton E, Fracasso DS, da Rosa AR, Branco CS, et al. Targeting inflammatory-mitochondrial response in major depression: current evidence and further challenges. Oxidative medicine and cellular longevity. 2020;2020.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCatena-Dell\u0026rsquo;Osso M, Rotella F, Dell\u0026rsquo;Osso A, Fagiolini A, Marazziti D. Inflammation, serotonin and major depression. Current Drug Targets. 2013;14(5):571\u0026ndash;7.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRoskoski R, Jr. ERK1/2 MAP kinases: structure, function, and regulation. Pharmacol Res. 2012;66(2):105\u0026ndash;43.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWang JQ, Mao L. The ERK Pathway: Molecular Mechanisms and Treatment of Depression. Mol Neurobiol. 2019;56(9):6197\u0026ndash;205.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYuan S, Jiang X, Zhou X, Zhang Y, Teng T, Xie P. Inosine alleviates depression-like behavior and increases the activity of the ERK-CREB signaling in adolescent male rats. Neuroreport. 2018;29(14):1223\u0026ndash;9.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYuan P, Zhou R, Wang Y, Li X, Li J, Chen G, et al. Altered levels of extracellular signal-regulated kinase signaling proteins in postmortem frontal cortex of individuals with mood disorders and schizophrenia. J Affect Disord. 2010;124(1\u0026ndash;2):164\u0026ndash;9.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWang JQ, Fibuch EE, Mao L. Regulation of mitogen-activated protein kinases by glutamate receptors. J Neurochem. 2007;100(1):1\u0026ndash;11.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLi SY, Rong PJ, Gao GJ, Zhang Y, Wang JY, Wang Y, et al. [Auricular electroacupuncture improves depression possibly by promoting hippocampal Raf/ERK/RSK/CREB signaling in chronic unpredictable mild stress induced depression rats]. Zhen Ci Yan Jiu. 2019;44(8):554\u0026ndash;9.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChiu FL, Lin JK. Tomatidine inhibits iNOS and COX-2 through suppression of NF-kappaB and JNK pathways in LPS-stimulated mouse macrophages. FEBS Lett. 2008;582(16):2407\u0026ndash;12.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eVeres B, Radnai B, Gallyas F, Jr., Varbiro G, Berente Z, Osz E, et al. Regulation of kinase cascades and transcription factors by a poly(ADP-ribose) polymerase-1 inhibitor, 4-hydroxyquinazoline, in lipopolysaccharide-induced inflammation in mice. J Pharmacol Exp Ther. 2004;310(1):247\u0026ndash;55.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCheng SW, Li JX, Chien YC, Chang JP, Shityakov S, Huang SY, et al. Genetic Variations of Ionotropic Glutamate Receptor Pathways on Interferon-α-induced Depression in Patients with Hepatitis C Viral Infection. Brain Behav Immun. 2021;93:16\u0026ndash;22.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLi M, Ding R, Yang X, Ran D. Study on Biomarkers Related to the Treatment of Post-Stroke Depression and Alternative Medical Treatment Methods. Neuropsychiatr Dis Treat. 2022;18:1861\u0026ndash;73.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLiu J, Cheng Y, Li M, Zhang Z, Li T, Luo XJ. Genome-wide Mendelian randomization identifies actionable novel drug targets for psychiatric disorders. Neuropsychopharmacology. 2023;48(2):270\u0026ndash;80.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChang H, Cai X, Li HJ, Liu WP, Zhao LJ, Zhang CY, et al. Functional Genomics Identify a Regulatory Risk Variation rs4420550 in the 16p11.2 Schizophrenia-Associated Locus. Biol Psychiatry. 2021;89(3):246\u0026ndash;55.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMa C, Gu C, Huo Y, Li X, Luo XJ. The integrated landscape of causal genes and pathways in schizophrenia. Transl Psychiatry. 2018;8(1):67.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMcCaffery JM, Papandonatos GD, Faulconbridge LF, Erar B, Peter I, Wagenknecht LE, et al. Genetic Predictors of Depressive Symptoms in the Look AHEAD Trial. Psychosom Med. 2015;77(9):982\u0026ndash;92.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAbbott GW. KCNE1 and KCNE3: The yin and yang of voltage-gated K(+) channel regulation. Gene. 2016;576(1 Pt 1):1\u0026ndash;13.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eErlandsdotter LM, Giammarino L, Halili A, Nikesj\u0026ouml; J, Gr\u0026eacute;en H, Odening KE, et al. Long-QT mutations in KCNE1 modulate the 17β-estradiol response of Kv7.1/KCNE1. Sci Adv. 2023;9(11):eade7109.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMelman YF, Um SY, Krumerman A, Kagan A, McDonald TV. KCNE1 binds to the KCNQ1 pore to regulate potassium channel activity. Neuron. 2004;42(6):927\u0026ndash;37.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCosti S, Han MH, Murrough JW. The Potential of KCNQ Potassium Channel Openers as Novel Antidepressants. CNS Drugs. 2022;36(3):207\u0026ndash;16.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHuang S, Li Y, Shen J, Liang W, Li C. Identification of a diagnostic model and molecular subtypes of major depressive disorder based on endoplasmic reticulum stress-related genes. Front Psychiatry. 2023;14:1168516.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFeng J, Zhou Q, Gao W, Wu Y, Mu R. Seeking for potential pathogenic genes of major depressive disorder in the Gene Expression Omnibus database. Asia Pac Psychiatry. 2020;12(1):e12379.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMartins-de-Souza D, Guest PC, Harris LW, Vanattou-Saifoudine N, Webster MJ, Rahmoune H, et al. Identification of proteomic signatures associated with depression and psychotic depression in post-mortem brains from major depression patients. Transl Psychiatry. 2012;2(3):e87.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRoff\u0026eacute; M, Beraldo FH, Bester R, Nunziante M, Bach C, Mancini G, et al. Prion protein interaction with stress-inducible protein 1 enhances neuronal protein synthesis via mTOR. Proc Natl Acad Sci U S A. 2010;107(29):13147\u0026ndash;52.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBeraldo FH, Arantes CP, Santos TG, Queiroz NG, Young K, Rylett RJ, et al. Role of alpha7 nicotinic acetylcholine receptor in calcium signaling induced by prion protein interaction with stress-inducible protein 1. J Biol Chem. 2010;285(47):36542\u0026ndash;50.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMick E, McGough J, Loo S, Doyle AE, Wozniak J, Wilens TE, et al. Genome-wide association study of the child behavior checklist dysregulation profile. J Am Acad Child Adolesc Psychiatry. 2011;50(8):807\u0026thinsp;\u0026ndash;\u0026thinsp;17.e8.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBeraldo FH, Thomas A, Kolisnyk B, Hirata PH, De Jaeger X, Martyn AC, et al. Hyperactivity and attention deficits in mice with decreased levels of stress-inducible phosphoprotein 1 (STIP1). Dis Model Mech. 2015;8(11):1457\u0026ndash;66.\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"translational-psychiatry","isNatureJournal":false,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"tp","sideBox":"Learn more about [Translational Psychiatry](http://www.nature.com/tp/)","snPcode":"41398","submissionUrl":"https://mts-tp.nature.com/cgi-bin/main.plex","title":"Translational Psychiatry","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"ejp","reportingPortfolio":"Nature AJ","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"Oxidative stress, Depression, Machine learning, Integrative Omics, Mendelian randomization","lastPublishedDoi":"10.21203/rs.3.rs-4641375/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-4641375/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eOxidative stress (OS) is strongly implicated in the pathophysiology of major depressive disorder (MDD) but the molecular mechanisms remain largely unknown. The purpose of this study is to identify genes related to both OS and MDD, and further to evaluate the utility of these genes as diagnostic markers and potential treatment targets. We searched datasets related to MDD from the Gene Expression Omnibus (GEO) database for differentially expressed genes (DEGs) also related to OS according to GeneCards. Bioinformatics analyses and machine learning algorithms were used to identify hub genes mediating OS\u0026ndash;MDD interactions. A summary data-based Mendelian randomization (SMR) approach was employed to identify possible causal genes for MDD from blood tissue eQLT data. These investigations identified 32 genes mediating OS\u0026ndash;MDD interactions, while SMR analysis identified \u003cem\u003eKCNE1\u003c/em\u003e (OR\u0026thinsp;=\u0026thinsp;1.057, 95%CI\u0026thinsp;=\u0026thinsp;1.013\u0026ndash;1.102, P\u0026thinsp;=\u0026thinsp;0.010), \u003cem\u003eMAPK3\u003c/em\u003e (OR\u0026thinsp;=\u0026thinsp;1.023, 95%CI\u0026thinsp;=\u0026thinsp;1.004\u0026ndash;1.043, P\u0026thinsp;=\u0026thinsp;0.020), and \u003cem\u003eSTIP1\u003c/em\u003e (OR\u0026thinsp;=\u0026thinsp;0.792, 95%CI\u0026thinsp;=\u0026thinsp;0.641\u0026ndash;0.979, P\u0026thinsp;=\u0026thinsp;0.031) as OS-related causal genes for MDD. These genes may thus serve as useful diagnostic markers and potential therapeutic targets.\u003c/p\u003e","manuscriptTitle":"Novel therapeutic targets for major depressive disorder related to oxidative stress identified by integrative multi-omics and multi-trait study","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2024-08-01 08:08:12","doi":"10.21203/rs.3.rs-4641375/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"revise","date":"2024-08-19T14:36:15+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"This content is not available.","date":"2024-07-26T06:01:19+00:00","index":2,"fulltext":"This content is not available."},{"type":"editorInvitedReview","content":"This content is not available.","date":"2024-07-16T01:21:03+00:00","index":1,"fulltext":"This content is not available."},{"type":"editorInvitedReview","content":"This content is not available.","date":"2024-07-14T16:20:03+00:00","index":3,"fulltext":"This content is not available."},{"type":"reviewerAgreed","content":"This content is not available.","date":"2024-07-10T02:13:25+00:00","index":3,"fulltext":"This content is not available."},{"type":"reviewerAgreed","content":"This content is not available.","date":"2024-07-09T13:45:45+00:00","index":2,"fulltext":"This content is not available."},{"type":"reviewerAgreed","content":"This content is not available.","date":"2024-07-08T23:19:58+00:00","index":1,"fulltext":"This content is not available."},{"type":"reviewersInvited","content":"","date":"2024-07-08T21:19:39+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2024-06-28T09:19:26+00:00","index":"","fulltext":""},{"type":"submitted","content":"Translational Psychiatry","date":"2024-06-27T21:24:35+00:00","index":"","fulltext":""},{"type":"checksFailed","content":"","date":"2024-06-27T15:10:22+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2024-06-26T08:58:15+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"translational-psychiatry","isNatureJournal":false,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"tp","sideBox":"Learn more about [Translational Psychiatry](http://www.nature.com/tp/)","snPcode":"41398","submissionUrl":"https://mts-tp.nature.com/cgi-bin/main.plex","title":"Translational Psychiatry","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"ejp","reportingPortfolio":"Nature AJ","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"8b0951f8-b268-43bf-9aa4-9fff3f9df6f3","owner":[],"postedDate":"August 1st, 2024","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[{"id":34310364,"name":"Health sciences/Diseases/Psychiatric disorders/Depression"},{"id":34310365,"name":"Health sciences/Biomarkers/Diagnostic markers"}],"tags":[],"updatedAt":"2024-10-20T07:06:45+00:00","versionOfRecord":{"articleIdentity":"rs-4641375","link":"https://doi.org/10.1038/s41398-024-03126-0","journal":{"identity":"translational-psychiatry","isVorOnly":false,"title":"Translational Psychiatry"},"publishedOn":"2024-10-19 04:00:00","publishedOnDateReadable":"October 19th, 2024"},"versionCreatedAt":"2024-08-01 08:08:12","video":"","vorDoi":"10.1038/s41398-024-03126-0","vorDoiUrl":"https://doi.org/10.1038/s41398-024-03126-0","workflowStages":[]},"version":"v1","identity":"rs-4641375","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-4641375","identity":"rs-4641375","version":["v1"]},"buildId":"qtupq5eGEP_6zYnWcrvyt","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}