Causal relationship between inflammatory cytokines and posttraumatic stress disorder: a Mendelian randomization study and potential mechanism analysis

ABSTRACT

Background: Post-traumatic stress disorder (PTSD) is a complex condition linked to inflammation. The causality between inflammatory cytokines and PTSD risk remains unclear.

Methods: We conducted a bidirectional two-sample Mendelian randomization (MR) study using genome-wide association study (GWAS) data from 41 inflammatory cytokines and PTSD. Additional analyses included differential gene expression, protein–protein interaction, and functional enrichment to explore underlying mechanisms.

Results: MR analysis indicated that higher levels of stem cell factor (SCF) and interleukin-4 (IL-4) are associated with a reduced risk of PTSD. Genes POGZ and LRIG2 were identified as mediators, implicated in the TGF-beta signalling pathway.

Conclusion: Our findings suggest a protective role of certain cytokines against PTSD and highlight potential molecular mediators. This knowledge could inform future therapeutic strategies for PTSD.

HIGHLIGHTS

• This study identifies a protective link between specific inflammatory cytokines, namely stem cell factor (SCF) and interleukin-4 (IL-4), and reduced risk of posttraumatic stress disorder (PTSD), highlighting new avenues for PTSD prevention.

• Through detailed gene analysis, POGZ and LRIG2 genes are shown to be key mediators in the interaction between cytokines and PTSD risk, offering potential therapeutic targets.

• Findings emphasize the role of TGF-beta signalling in PTSD, with implications for developing targeted treatments addressing both inflammation and neuroprotective pathways.

1. Introduction

Posttraumatic stress disorder (PTSD) is a psychiatric disorder that arises from exposure to life-threatening or sexually traumatic events, manifesting in four core symptom clusters, i.e. re-experiencing the trauma, avoidance of triggers, persistent negative alterations in cognition and mood, and hyperarousal, which significantly impair occupational and social functioning (Miao et al., 2018; Pai et al., 2017). Epidemiological studies have found that the lifetime occurrence of PTSD among adults in the general population of the USA and Canada is estimated to be between 6.1% and 9.2% (Goldstein et al., 2016; Koenen et al., 2017). Population subgroups exposed to higher levels of trauma consistently exhibit elevated PTSD rates. Notably, increased prevalence has been observed among socially disadvantaged individuals, youth, women, military personnel, police officers, firefighters, and first responders to disasters and mass traumatic events (Karam et al., 2014). Current treatment options for PTSD include psychological therapies and pharmacological interventions. To date, the U.S. Food and Drug Administration (FDA) has approved only two selective serotonin reuptake inhibitors (SSRIs), sertraline (Zoloft) and paroxetine (Paxil), which primarily alleviate symptoms such as anxiety and depression but do not directly target the core symptoms of PTSD (e.g. intrusive memories, hyperarousal) (Schnurr et al., 2024). The National Institute for Health and Care Excellence (NICE) guidelines advocate for the use of trauma-focused cognitive behavioural therapy (Tf-CBT) or eye movement desensitization and reprocessing (EMDR) as the preferred psychological therapies for treating PTSD (Tait et al., 2024). Early identification and proactive management of PTSD are crucial, especially in light of the limited pharmacological treatments that effectively target its underlying pathophysiology. There is an urgent need to explore new biological mechanisms to guide the development of more precise and targeted therapies, emphasizing the critical importance of clarifying the causal connections between inflammatory pathways and PTSD risk.

The development of PTSD is believed to involve a multifaceted interplay of genetic susceptibility, trauma exposure, and the characteristics and intensity of the trauma that initiates the disorder (Hiscox et al., 2021; McLaughlin et al., 2010). Although evidence supports this perspective, the precise mechanisms by which these factors interact are not yet fully understood. In turn, these dynamics can be influenced by protective factors that may be personal or social in nature (Panagou & MacBeth, 2022). PTSD is linked to considerable changes in critical biochemical pathways, especially those involving the functioning of the hypothalamic–pituitary–adrenal (HPA) axis, immune system regulation, systemic inflammation, and oxidative stress (Dell'Oste et al., 2023; Pan et al., 2018; Schiavone et al., 2013). Given that these alterations could heighten the risk of various medical conditions, some scholars consider PTSD to be a systemic disorder (Krantz et al., 2022). Growing research supports the involvement of inflammation and immune dysregulation in the development of PTSD, with pro-inflammatory cytokines potentially playing a significant role in the clinical manifestations of the disorder, notably through their interactions with pathways like NF-κB and P38MAPK (Dell'Oste et al., 2023). Specifically, changes in interleukin-6 (IL-6), interleukin-1β (IL-1β), tumour necrosis factor-α (TNF-α), and interferon-γ (INF-γ) have been linked to disruptions in synaptic plasticity and neuroinflammatory pathways, which underpin the functional and cognitive disturbances associated with PTSD (Levin & Godukhin, 2017). Conversely, inflammatory cytokines can also play a protective role by enhancing neuroplasticity and facilitating healing processes that help reduce the effects of the disorder (Sun et al., 2022). Inflammatory factors act as a ‘double-edged sword’ in mental health disorders, including PTSD, and have been extensively studied. However, the causal link between these inflammatory factors and the risk of PTSD has been investigated but continues to be unclear.

To more thoroughly investigate the correlation between PTSD and inflammatory cytokines, there is a need for more rigorous research. Pretrauma prospective studies are considered the gold standard for determining causality in PTSD research, as they allow for the assessment of neurobiological factors prior to trauma exposure and their changes post-exposure. However, conducting such studies requires significant financial and logistical resources, making them challenging to implement on a large scale. Traditional observational studies, however, frequently encounter challenges such as confounding factors and reverse causality, which can compromise the reliability of their findings. Recently, Mendelian Randomization (MR) has become a widely used and powerful tool for causal inference, utilizing genetic variations as instrumental variables (IVs) to investigate causal relationships between exposures and outcomes (Davies et al., 2018; Hu et al., 2022). The primary advantage of MR lies in its ability to mitigate confounding and reverse causality, as genetic variants are randomly assigned at conception. This random assignment closely resembles a randomized controlled trial (RCT), enabling causal inferences without the need for experimental intervention. Consequently, MR minimizes biases commonly found in traditional epidemiological studies, providing a robust alternative to randomized clinical trials. Therefore, in this study, we conducted a bidirectional MR analysis to investigate the relationship between inflammatory factors and PTSD. Additionally, we employed bioinformatics approaches to investigate the potential mechanisms through which inflammatory factors that are causally linked to PTSD influence its development.

2. Method

2.1. Study design

The design of the present study was guided by the Report List of Mendelian Randomization-Enhanced Epidemiological Observational Studies (STROBE-MR) (Skrivankova et al., 2021) (Table S1). We carried out a two-sample MR study utilizing data from 42 publicly available genome-wide association studies (GWAS) – 41 inflammatory cytokines for exposure and one PTSD for outcomes. Subsequently, a reverse MR analysis was conducted using PTSD as the exposure and 41 inflammatory factors as the outcomes to explore the impact of PTSD on these inflammatory markers. To minimize population stratification bias, these cohorts were exclusively composed of individuals of European descent. To guarantee the accuracy of causal deductions from MR analysis, it is crucial that the IVs adhere to three essential criteria (Figure 1A): (1) the relevance criterion, ensuring the single nucleotide polymorphisms (SNPs) robustly correlate with the exposure; (2) the independence criterion, confirming the SNPs are not connected with any confounding factors; and (3) the exclusion-restriction criterion, asserting that the SNPs influence the outcome exclusively through their interaction with the exposure. Figure 1B provides a detailed illustration of the MR framework. All data utilized in this research were derived from studies that obtained subject consent and ethical approval. Consequently, our study does not require further ethical clearance from the institutional review board.

Figure 1.

Schematic representation of the Mendelian randomization (MR) analysis assumptions (A) and workflow (B). SNPs, single nucleotide polymorphisms; GWAS, genome-wide association study (GWAS); MR, Mendelian randomization; p, p-value; F, F-statistic.

2.2. GWAS data source

In the current study, data on PTSD was collected from the FinnGen database, a distinctive project that merges genomic data with extensive digital health records from Finnish residents aged 18 and older. PTSD diagnoses in the FinnGen GWAS dataset were classified based on DSM-5 criteria. The GWAS dataset for PTSD encompasses a total of 199,213 samples, which includes 1103 patients and 198,110 healthy controls, along with 1,6380,382 SNPs. GWAS data on inflammatory factors, covering 41 cytokines and growth factors, were acquired from a study involving 8293 individuals. These datasets were sourced from the IEU OPEN GWAS database, and a comprehensive description of these GWAS datasets is available in the supplementary documents (Table S2).

2.3. Transcriptomic data acquisition

Transcriptomic datasets related to PTSD were sourced from the National Centre for Biotechnology Information Gene Expression Omnibus (GEO, https://www.ncbi.nlm.nih.gov/geo/) by searching with the terms ‘Posttraumatic stress disorder’ AND ‘Homo sapiens.’ The selection criteria for the datasets were as follows: (1) inclusion of both PTSD patients and non-PTSD control samples; (2) utilization of RNA-Seq as the sequencing method; (3) a minimum sample size of 15; (4) use of human blood as the test specimen; (5) collection of data samples from a European demographic; (6) negligible age differences between the PTSD and control groups; (7) unrestricted access to the data for downloading and compatibility with all analytical methods used in the study. Using these selection criteria, we accessed mRNA expression profiles from dataset GSE97356, which comprises 324 World Trade Center responders, categorized into 201 who never had PTSD (non-PTSD), 81 current sufferers, and 42 past sufferers (Kuan et al., 2017, 2021). The profiling was performed on the GPL11154 Illumina HiSeq 2000 platform (Homo sapiens). For differential gene expression analysis, the PTSD group (81 current and 42 past PTSD cases) was compared against the non-PTSD group (201 individuals who never had PTSD)

2.4. IVs selection

Initially, we attempted to identify SNPs associated with each inflammatory cytokine by applying the standard GWAS significance threshold of p < 5 × 10⁻⁸, commonly used in MR studies to ensure genome-wide significance (Zhang et al., 2023). However, this approach failed to yield an adequate number of IVs. As a result, we relaxed the threshold to p < 5 × 10⁻⁶ to guarantee a sufficient number of IVs for dependable outcomes. Secondly, to guarantee statistical independence among SNPs, we conducted a linkage disequilibrium (LD) analysis using data from the European-based 1000 Genome Projects, setting the parameters to an R² of less than 0.001 and a clumping distance of 10,000 kb. Thirdly, we maintained the principle that the chosen SNPs should demonstrate consistent allele effects on both the exposure and outcome variables. As a result, we eliminated any palindromic SNPs without A/T or C/G polymorphisms from our selection of IVs. Ultimately, we determined the robustness of each SNP by calculating the F-statistic, which assesses the strength of the genetic influence on the trait using the formula: F = R² (N – 2)/(1 – R²), where R² indicates the proportion of variance in the trait explained by the SNP and N represents the sample size in the GWAS analyzing SNPs associated with the trait (Burgess & Thompson, 2011). We calculated R² values using the equation R² = 2 × EAF × (1 – EAF) × β², where effect allele frequency (EAF) denotes the effect allele frequency and β the estimated effect of the SNP on the trait. SNPs with an F-statistic below 10 were excluded, since an F-statistic above 10 is considered indicative of sufficient strength to validate the SNPs as IVs. To address possible correlations between SNPs and confounding variables, we utilized LDtrait Tool (https://ldlink.nih.gov/?tab=ldtrait) to scrutinize all included IVs. We then excluded any SNPs that exhibited associations with confounding factors. In the reverse MR analysis, we used PTSD as the exposure and applied the same criteria to select IVs.

2.5. MR analysis

In the present study, we implemented several statistical techniques, such as Inverse Variance Weighted (IVW), MR Egger, Weighted Median, Weighted Mode, and Simple Mode, to explore the causal connections between inflammatory cytokines and PTSD while assessing the potential influence of pleiotropy. The IVW method, serving as the principal technique, was employed to estimate the potential directional causality between inflammatory cytokines and PTSD, effectively providing reliable causal estimates even in the absence of directional pleiotropy and ensuring robust and accurate conclusions in their relationship analysis. Reverse MR analysis was performed to ascertain the direction of causality, utilizing a method similar to forward MR, but with PTSD as the exposure and 41 inflammatory cytokines as the outcomes. For this analysis, SNPs associated with PTSD were carefully chosen to serve as the IVs.

All statistical analyses were performed using the ‘TwoSampleMR’ (Version 0.6.4) and ‘MR-PRESSO’ packages within RStudio (Version: 2023.06.1 + 524). The ‘TwoSampleMR’ package facilitated initial MR estimates, leveraging its array of tools designed for data harmonization and various analytical processes specific to Mendelian Randomization. The ‘MR-PRESSO’ package was used to detect and adjust for potential pleiotropic effects in the IVs employed. After extracting the IVs with the TwoSampleMR R package, the exposure and outcome data were synchronized using the ‘harmonise’ function within the package. This harmonization ensures that the impact of each SNP on both the exposure and the outcome is associated with the same allele, thus improving the precision and consistency of the results.

2.6. Heterogeneity and pleiotropy

Sensitivity analysis includes tests for heterogeneity and pleiotropy, which are conducted primarily from three distinct perspectives. Heterogeneity testing targets the detection of variations across diverse IVs. To this end, the Q-test method is utilized to assess heterogeneity within both the MR-IVW and MR-Egger models, conducting evaluations on the IVs through a range of statistical analysis techniques. When a p-value exceeds 0.05, it indicates no significant heterogeneity. However, if heterogeneity is observed among the IVs, a random effects model grounded in the IVW methodology is applied to determine the influence of exposure on the outcome.

Pleiotropy testing primarily investigates whether IVs demonstrate horizontal pleiotropy, specifically whether SNPs serving as IVs have associations with other unrelated variables. For assessing horizontal pleiotropy, the MR-Egger method is commonly applied, where a significant deviation of the intercept term from zero (p < .05) signals its presence. Furthermore, the MR-PRESSO method is employed to identify and exclude outlier SNPs, thus correcting horizontal pleiotropy and strengthening the validity of causal inferences between exposure and outcome. Leave-one-out sensitivity analysis is conducted by methodically excluding each individual SNP locus through the leave-one-out technique. MR analysis is then performed with the remaining SNP loci to evaluate whether any specific SNP locus contributes bias to the results.

2.7. GEO data processing and differential gene screening

Differentially expressed genes (DEGs) from dataset GSE97356 were identified using the ‘limma’ R package, with genes showing a log fold change (FC) > 0 classified as upregulated and those with log FC < 0 as downregulated, all selected based on p-values of less than 0.05. The visualization of these DEGs was accomplished through volcano plots created with the ‘ggplot2’ R package.

2.8. Gene annotation near SNP and protein–protein interaction (PPI) analysis

We utilized bioinformatics approaches in RStudio and queried online databases to identify genes at genetic loci commonly linked to both inflammatory factors and PTSD, aiming to explore the mechanisms through which these factors may influence PTSD development. Initially, the ‘vautils’ R package was employed to pinpoint candidate genes associated with SNPs common to both inflammatory factors and PTSD. Furthermore, we used Venn diagrams to identify and illustrate the overlaps between DEGs from the GSE97356 dataset and genes adjacent to loci frequently associated with both inflammatory factors and PTSD. Utilizing the STRING (https://string-db.org) online database, we constructed a human PPI network featuring genes that overlap between the DEGs identified in the GSE97356 dataset and genes adjacent to loci commonly associated with both inflammatory factors and PTSD, specifically focusing on association data pertinent to H. sapiens. In addition, we explored 50 genes that interact within this context. To ensure the reliability of the interactions depicted in the network, we implemented a minimum interaction score cutoff of 0.400. To delve deeper into the interactions and functional relationships between the overlapping genes and the 50 interacting genes, we employed the sophisticated gene function prediction tool, GeneMANIA (https://genemania.org/). By inputting the names of these 10 central genes into the GeneMANIA platform and selecting the species ‘Homo sapiens,’ an interaction network diagram was automatically generated, reflecting their established interactions and functional linkages.

2.9. Functional and Pathway enrichment analysis

Gene Ontology (GO) enrichment analysis and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis were employed to elucidate the biological functions and pathways linked to both the overlapping genes and the 50 genes they interact with. The ‘clusterProfiler’ R package facilitated the GO enrichment analysis and KEGG pathway enrichment analysis, with p-values adjusted for multiple testing using the Benjamini-Hochberg method and set at less than 0.05 to denote significant enrichment. For further analysis, the top p-values for biological process (BP), molecular function (MF), and cellular component (CC) were selected from the GO analysis. Additionally, the top 10 pathways, based on p-values, were identified for KEGG pathway enrichment analysis, which included creating a Bubble Chart.

2.10. Drug sensitivity analysis

The RNAactDrug (Dong et al., 2020) (http://bio-bigdata.hrbmu.edu.cn/RNAactDrug/) database serves as an extensive platform for exploring connections between drug sensitivity and RNA molecules, such as mRNAs, lncRNAs, and miRNAs, across four molecular dimensions: expression, copy number variation, mutation, and methylation. This analysis integrates data from three major pharmacogenomic databases: GDSC, CellMiner, and CCLE. From this resource, we derived correlations between mRNA and drug sensitivity. Within the RNAactDrug database, we selected data derived from the CCLE and set parameters with an FDR < 0.001 and a p value < 0.001 to investigate the expression of overlapping genes and their potential drug sensitivity. The results of the drug sensitivity analysis were displayed using a bubble chart. Finally, we utilized ChEMBL (https://www.ebi.ac.uk/chembl/) to explore the molecular structures of potential sensitive drugs.

3. Results

3.1. IVs used for analysis

The primary objective of this study is to explore the potential association between inflammatory cytokines and PTSD. Following the application of genome-wide significance criteria and the exclusion of LD, we identified 452 SNPs from each inflammatory cytokine dataset to serve as IVs for subsequent analysis. Ultimately, we compiled summary data for the IVs linked to the inflammatory cytokines and employed the F-statistic to evaluate their robustness. Significantly, the F-statistics for all SNPs used as IVs ranged from 11.16 to 788.95, indicating a strong correlation between the selected SNPs and cytokine levels and the absence of weak instrument bias, as detailed in Table S3. We pinpointed the relevant SNP information for the outcome variable by verifying each SNP's robust association with the exposure across the respective datasets (Table S4).

3.2. Effect of inflammatory cytokines on PTSD risk

We gathered data on 41 inflammatory cytokines as exposure factors and PTSD as the outcome, successfully clarifying the causal link between them (Figure 2, Table S5). Genetic predictors of systemic inflammatory regulators have been found to be associated with PTSD, as demonstrated by the ensuing results. Using IVW methods, we found that elevated levels of stem cell factor (SCF) (OR = 0.673, 95% CI = 0.479–0.946, p = .023) and interleukin-4 (IL-4) (OR = 0.688, 95% CI = 0.483–0.931, p = 0.039) are associated with a reduced risk of PTSD. The relationships remain uniform across MR Egger, Weighted Median, Weighted Mode, and Simple Mode methods, showing comparable estimates in both direction and magnitude of the causal effects (Figure 3A,B). The leave-one-out analysis indicates that all SNPs consistently fall on the same side of zero, demonstrating that no single SNP drives the observed causality (Figure 3C–D).

Figure 2.

A forest plot illustrates the associations between genetically predicted factors of 41 inflammatory cytokines as exposures and the risk of developing Posttraumatic Stress Disorder (PTSD).

Figure 3.

Investigating the causal relationships between stem cell factor (SCF), interleukin-4 (IL-4), and the risk of developing Posttraumatic Stress Disorder (PTSD). (A) A scatter plot depicting the causal relationships between SCF and PTSD. (B) A scatter plot depicting the causal relationships between IL-4 and PTSD. (C) Leave-one-out plot of the causal relationship between SCF and PTSD. (D) Leave-one-out plot of the causal relationship between SCF and PTSD.

3.3. Reverse MR analysis

Following the application of genome-wide significance criteria and the exclusion of LD, we identified nine SNPs from PTSD dataset to serve as IVs for subsequent reverse MR analysis (Table S6). Reverse MR analysis, leveraging five different statistical techniques, was conducted to investigate the potential causal link between PTSD and 41 inflammatory cytokines, as outlined in Table S7 and Figure 4. Within the study, the Interleukin-2 receptor alpha subunit (IL2rα) (OR = 0.907, 95% CI = 0.826–0.996, p = .042) and Growth regulated oncogene-α (GRO-α) (OR = 0.912, 95% CI = 0.837–0.993, p = .034) emerged as factors negatively correlated with PTSD risk. Consistency in the relationships is maintained across various methods including MR Egger, Weighted Median, Weighted Mode, and Simple Mode, with causal estimates exhibiting similar directions and magnitudes (Figure 5A,B). Additionally, leave-one-out analysis reveals that all SNPs uniformly align on the same side of zero, confirming that the causality observed is not influenced by any individual SNP (Figure 5C,D).

Figure 4.

A forest plot illustrates the associations between genetically predicted factors of Posttraumatic Stress Disorder (PTSD) as exposures and the risk of developing 41 inflammatory cytokines.

Figure 5.

Investigating the causal relationships between Posttraumatic Stress Disorder (PTSD) and the risk of developing Interleukin-2 receptor alpha subunit (IL2rα), Growth regulated oncogene-α (GRO-α). (A) A scatter plot depicting the causal relationships between PTSD and IL2rα. (B) A scatter plot depicting the causal relationships between PTSD and GRO-α. (C) Leave-one-out plot of the causal relationship between PTSD and IL2rα. (D) Leave-one-out plot of the causal relationship between PTSD and GRO-α.

3.4. Sensitivity analysis

In order to verify the accuracy of our MR findings, assessments for both heterogeneity and pleiotropy were conducted (Tables S8 and S9). To strengthen the reliability of our results, these assessments were undertaken to explore the impact of genetic variations on multiple traits or confounding factors that might affect the observed associations. Initially, evidence from the MR-Egger Intercept test suggested that horizontal pleiotropy was unlikely (p > .05). Furthermore, the MR-Egger and IVW heterogeneity tests showed no significant heterogeneity among the genetic variants involved in the analysis (p > .05). Ultimately, the MR-PRESSO approach detected no significant outliers influencing the study outcomes.

3.5. Differential gene expression between PTSD and Non-PTSD groups

After screening, 440 DEGs were identified from the GSE97356 dataset (Table S10). Volcano plots were used to illustrate the genes with differing expression levels between the PTSD and Non-PTSD groups (Figure 6A). In these plots, 273 genes displayed in red were significantly upregulated in the PTSD group, while another 167 shown in blue were markedly downregulated compared to the Non-PTSD group. Additionally, 269 genes are located near the IVs for inflammatory cytokines (SCF and IL-4) with positive results in forward MR analysis (Table S11). Subsequently, we created a Venn diagram to identify overlaps between these DEGs and genes situated near SNPs associated with inflammatory cytokines linked to PTSD risk (Figure 6B). The diagram revealed two common genes, including POGZ and LRIG2, between these groups. This result suggests that these two common genes may play a critical role in how inflammatory cytokines affect PTSD risk. Consequently, we conducted a series of in-depth studies on these two genes.

Figure 6.

Identification of differentially expressed genes (DEGs) and functional enrichment analysis. (A) A volcano plot of DEGs shows the upregulated genes as red dots, the downregulated genes as green dots, and the genes with no significant expression difference as black dots. (B) Venn diagram analysis identifies and illustrates the overlaps between DEGs from the GSE97356 dataset and genes located near loci commonly associated with both inflammatory factors and Posttraumatic Stress Disorder (PTSD). (C) Protein – protein interaction (PPI) network of POGZ and LRIG2 and their interacting genes based on STRING. (D) Co-expression network of POGZ and LRIG2 and related genes based on GeneMANIA.

3.6. PPI analysis on POGZ and LRIG2

To better understand the potential biological functions of these two common genes, we input the proteins corresponding to these candidate genes into the STRING database to ascertain their protein–protein interaction relationships and identify 50 functionally similar proteins. The PPI network, featuring 52 nodes and 98 edges, demonstrated significant connectivity with a p-value of less than 5.57e-08, as shown in Figure 6C. Using the GeneMANIA database, a detailed network mapping the gene interactions among 52 genes was constructed to analyse their biological functions (Figure 6D). This network highlighted physical interactions accounting for 62.56%, co-expression representing 23.28%, and shared protein domains making up 2.02%. The analysis further indicated that these genes are associated with heterochromatin, the ubiquitin ligase complex, and the cullin-RING ubiquitin ligase complex.

3.7. Functional enrichment analysis

Using the ‘clusterProfiler’ R package, we conducted GO and KEGG analyses to explore the biological relevance of these 52 genes (Figure 7A). GO analysis revealed enrichment in processes related to neuronal regeneration, transcriptional regulation, and chromatin organization. KEGG pathway analysis highlighted associations with immune response (herpes simplex virus 1 infection, bacterial invasion of epithelial cells), signalling pathways (TGF-beta, apelin), and cellular processes (cell cycle, axon guidance) (Figure 7B). These findings provide functional insights into the potential roles of these genes in PTSD-related mechanisms.

Figure 7.

Bubble chart of functional and pathway enrichment for POGZ and LRIG2 and related genes. (A) Gene Ontology (GO) functional enrichment bubble chart. (B) Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment bubble chart.

3.8. Drug sensitivity analysis on POGZ and LRIG2

We utilized the RNAactDrug database to identify potential therapeutic drugs for PTSD, targeting gene expressions of POGZ and LRIG2. Initial searches focused on POGZ mRNA expression, revealing that the sensitivity of seven drugs was significantly inversely correlated with POGZ expression levels (FDR <0.05, Spearman.stat < 0). Subsequent searches for LRIG2 mRNA expression identified two drugs whose sensitivity was significantly inversely correlated with LRIG2 levels (FDR <0.05, Spearman.stat < 0). Bubble charts were used to visualize the correlations between these drugs and their targets (Figure 8A). Additionally, we utilized the ChEMBL database to explore the molecular structures and formulas of these drugs (Figure 8B-8F).

Figure 8.

Drug sensitivity analysis and molecular structure presentation for POGZ and LRIG2. (A) Drug sensitivity analysis for POGZ and LRIG2. Triangles represent potential drugs related to POGZ, while circles represent potential drugs related to LRIG2. (B-H) Drug molecular structures for POGZ and LRIG2 sensitive drugs.

4. Discussion

Traumatic injuries profoundly affect the human experience, often with repercussions that surpass mere physical harm. Following these events, individuals frequently experience psychological aftermaths such as PTSD and depression, which can profoundly hinder their recovery and daily reintegration (Robles et al., 2024). PTSD, along with many other mental health conditions, exhibits significant heterogeneity and is characterized by distinct underlying dimensions that may each arise from unique causes (Galatzer-Levy & Bryant, 2013). Currently, treatment outcomes for PTSD remain inadequate, underscoring the importance of identifying underlying factors to enhance our understanding and treatment of its pathogenesis (Maercker et al., 2022). While pharmacological interventions for PTSD remain limited, psychological treatments such as EMDR and Prolonged Exposure (PE) therapy have shown high efficacy in treating PTSD (Daniëls et al., 2025; Rentinck et al., 2025). Both EMDR and PE are considered first-line treatments for PTSD and have been found to significantly reduce symptoms in a variety of patient populations (Hoppen et al., 2025). Despite the success of these treatments, challenges such as high dropout rates and the need for greater accessibility remain. The current research primarily focuses on symptom management, but there is still much to be desired in terms of a comprehensive treatment approach that targets the core pathophysiological mechanisms of PTSD. This underscores the importance of identifying underlying biological factors that can inform the development of pharmacological treatments that complement psychological therapies like EMDR and PE. Inflammation has emerged as a critical factor in the onset and progression of PTSD, playing a central role in regulating nerve damage and conduction (Ursini & Punzi, 2021). Elevated peripheral inflammatory cytokines can breach the blood–brain barrier, inducing neuroinflammation that potentially disrupts brain architecture and impairs cognitive functions such as memory and learning, as well as synaptic plasticity and neurogenesis, increasing the susceptibility to PTSD (Quinones et al., 2020). Although there is an initial understanding of the relationship between inflammatory factors and PTSD, further research is needed to clarify specific inflammatory pathways, potential genetic susceptibilities, and the variations across different ethnicities and genders in this process.

Inflammation is the body's intricate biological reaction to pathogens, damaged cells, and detrimental stimuli, involving immune cells, blood vessels, and molecular mediators, serving as a protective mechanism to eradicate the source of cellular injury and facilitate the restoration of affected tissues (Ferrero-Miliani et al., 2007). Traditionally considered immune-privileged and only susceptible during certain diseases or injuries, the brain is now understood to be influenced by peripheral pro-inflammatory cytokines, which can reach neural tissues through several pathways, including penetration of the blood–brain barrier (Al-Ghraiybah et al., 2022; Woodburn et al., 2021). Neuroinflammation, a form of chronic inflammation distinct from acute central nervous system inflammation, involves the persistent activation of nerve cells and the recruitment of other immune cells to the brain, a process intimately linked to the progression of neurodegenerative diseases (Ashwal et al., 2021; Jurcau & Simion, 2021). Similar to neurodegenerative diseases, neuroinflammation significantly affects neurological damage and degenerative pathways linked to psychological disorders including PTSD (Lee et al., 2022). PTSD-related neuroinflammation has been associated with elevated inflammatory mediators, activation of glial cells, and the recruitment of leukocytes, as observed in both preclinical and clinical studies (Eraly et al., 2014; Sun et al., 2021).

Sun Y et.al, in a detailed review of existing research, reveal that patients with PTSD demonstrate a disrupted immune equilibrium, marked by elevated plasma concentrations of pro-inflammatory cytokines such as IL-6 and TNF-α, along with increased levels of immune-activating Th1 and inflammatory Th17 cells in the blood, pointing to a pro-inflammatory condition (Sun et al., 2021). This finding is consistent with our own research; through MR studies, we observed a positive correlation between IL-6 (OR = 1.117, 95% CI = 0.711–1.755, p = .632) and TNF-α (OR = 1.098, 95% CI = 0.697–1.729, p = .686) levels and the incidence of PTSD. This suggests that elevated IL-6 and TNF-α may increase the risk of developing PTSD. Although the p-values for both are greater than 0.05, indicating statistical non-significance, the ORs are both greater than 1, which suggests a trend that warrants further investigation. In combat-related PTSD, the levels of the inhibitory cytokines IL-4 and IL-10 were significantly reduced in affected patients (Wang et al., 2016). Notably, reductions in the levels of inhibitory cytokines IL-4 and IL-10 were more pronounced in saliva than in plasma, suggesting a distinct alteration in the regulatory cytokine profile and emphasizing the potential of saliva as a diagnostic medium for assessing neuroimmune changes associated with PTSD (Wang et al., 2016). In the current study, we observed that levels of IL-4 (OR = 0.688, 95% CI = 0.483–0.931, p = .039) and IL-10 (OR = 0.898, 95% CI = 0.716–1.125, p = .349) were negatively associated with PTSD. Additionally, our studies have identified that SCF (OR = 0.673, 95% CI = 0.479–0.946, p = .023) is associated with a decreased risk of developing PTSD. SCF, also known as kit ligand or c-kit ligand, is a vital cytokine involved in hematopoiesis, the process of blood cell formation. Additionally, SCF significantly contributes to the reduction of inflammation by interacting with mast cells and eosinophils, which are key players in the inflammatory response (Ptaschinski et al., 2023).

In addition to inflammation and immune system dysregulation contributing to the development of PTSD, differential gene expression can also trigger its onset. Our research also revealed that 273 genes were significantly upregulated in the PTSD group, while another 167 genes were notably downregulated compared to the non-PTSD group. Additionally, we identified two DEGs associated with IL-4 and SCF, namely POGZ and LRIG2. This suggests that these two genes, POGZ and LRIG2, may play a critical role in how IL-4 and SCF influence the risk of developing PTSD. It is important to highlight that gene expression analysis was performed as a secondary exploratory investigation, utilizing a single transcriptomic dataset (GSE97356) to offer preliminary insights into potential transcriptional alterations associated with PTSD. Additionally, we discovered that POGZ, LRIG2, and their associated genes are involved in the TGF-beta signalling pathway, according to KEGG analysis. TGF-β signalling plays a crucial role in the pathophysiology of PTSD by influencing several key biological processes. It modulates neuroinflammation, which can result from chronic stress, leading to structural and functional changes in the brain (Cekanaviciute et al., 2014; Patel et al., 2017). Additionally, TGF-β supports neuroprotective processes by promoting neuron survival and maintaining blood–brain barrier integrity, while also influencing neuroplasticity, which is crucial for adapting to traumatic experiences (Dobolyi et al., 2012; Hao et al., 2024). Interaction of TGF-β with the hypothalamic–pituitary–adrenal (HPA) axis may influence cortisol levels and stress resilience, key aspects of PTSD (Bangsgaard et al., 2017). Moreover, TGF-β regulates immune function, which is often altered in PTSD, further affecting disease progression through immune system modulation (Bauché & Marie, 2017; Schmidt-Weber & Blaser, 2004). Collectively, these mechanisms underscore the multifaceted role of TGF-β in PTSD, positioning it as a potential target for therapeutic intervention. To enhance our understanding of the relationship between inflammatory factors and the development of PTSD, we created a schematic diagram that details the potential pathways through which SCF and IL-4 may influence PTSD formation (Figure 9).

Figure 9.

The potential mechanisms by which stem cell factor (SCF) and interleukin-4 (IL-4) influence the development of Posttraumatic Stress Disorder (PTSD).

The role of inflammatory cytokines in PTSD has been widely studied, yet findings across studies remain inconsistent (Passos et al., 2015; Speer et al., 2018). Some research has reported elevated levels of pro-inflammatory cytokines in PTSD patients, while others have found weak or non-significant associations (Hori & Kim, 2019; Rohleder, 2019). These inconsistencies may be due to differences in study designs, PTSD populations, symptom heterogeneity, cytokine measurement techniques, and confounding factors such as comorbid conditions and medication use. Our study, utilizing MR, mitigates many of these limitations by reducing bias from confounding and reverse causation. While our findings indicate only a few cytokines with significant associations to PTSD, this suggests that the causal influence of inflammatory cytokines may be more limited than previously hypothesized or relevant only to specific subtypes of PTSD. Future research should focus on stratified analyses based on PTSD symptom profiles and integrate multi-omics approaches to further elucidate these relationships.

Our MR analysis, which examined a broad range of inflammatory cytokines, identified a limited number of significant associations with PTSD. This does not necessarily refute the hypothesis that cytokines contribute to PTSD risk but rather reflects the complexity of PTSD and the methodological constraints of MR studies. First, MR relies on genetic variants as IVs, and the strength of these IVs depends on the statistical power of the underlying GWAS data. Some cytokines may indeed influence PTSD risk, but if their genetic determinants are weak, significant associations may not be detected. Second, PTSD is a heterogeneous disorder with multiple biological pathways contributing to its pathophysiology. It is possible that cytokine dysregulation plays a more significant role in specific subgroups of PTSD rather than uniformly across all affected individuals. Third, we applied multiple MR methods, including IVW as our primary approach, supported by MR Egger, Weighted Median, Weighted Mode, and Simple Mode, to ensure the robustness of our findings. The consistency of results across these methods suggests a degree of reliability, though further studies with larger datasets and refined genetic instruments are needed to validate our conclusions. Additionally, while MR is a robust method for causal inference, it relies on genetic associations derived from cross-sectional GWAS data. Furthermore, the 42 inflammatory factor datasets were based on self-reported questionnaires, which may introduce biases and false positives, inherently limiting the ability to establish direct causal relationships. Another limitation of this study is the reliance on a single cohort of first responders, which may not fully represent the wider PTSD population. Moreover, the relatively small number of PTSD cases within this sample restricts the generalizability of our findings. Therefore, although our results offer valuable insights, they should be regarded as preliminary, and additional research with larger, more diverse cohorts is necessary to confirm our conclusions.

5. Conclusion

The present study employed MR analysis to investigate the causal relationship between inflammatory cytokines and PTSD. Our findings suggest that elevated levels of SCF and IL-4 are associated with a reduced risk of PTSD. We identified two DEGs, POGZ and LRIG2, which may play crucial roles in how IL-4 and SCF influence PTSD risk. These genes are involved in the TGF-beta signalling pathway, which is critical in PTSD pathophysiology. Future studies incorporating larger-scale transcriptomic datasets or prospective multi-omic analyses will be essential to validate these findings and further explore their biological relevance in PTSD.

Glossary

Abbreviations: PTSD: posttraumatic stress disorder; MR: Mendelian randomization; GWAS: genome-wide association study; PPI: Protein–protein interaction; SCF, stem cell factor; IL-4: interleukin-4; HPA: hypothalamic–pituitary–adrenal; IL-6: interleukin-6; IL-1β: interleukin-1β; TNF-α: tumour necrosis factor-α; INF-γ: interferon-γ; IVs: instrumental variables; GO: gene ontology; KEGG: Kyoto Encyclopedia of Genes and Genomes; BP: biological process; MF: molecular function; CC: cellular component

Disclosure statement

No potential conflict of interest was reported by the author(s).

Authors’ contributions

All authors made substantial contributions to conception and design, acquisition of data, or analysis and interpretation of data; took part in drafting the article or revising it critically for important intellectual content; agreed to submit to the current journal; gave final approval of the version to be published; and agree to be accountable for all aspects of the work.

Data availability statement

The datasets used and analysed in the present study are available from the corresponding authors on reasonable request. The datasets generated and/or analysed during the current study are available in GWAS (https://gwas.mrcieu.ac.uk/) database.

Supplemental Material

Supplemental data for this article can be accessed online at https://doi.org/10.1080/20008066.2025.2494480.