Exploring genetic causal relationships between spinal cord injury and glioma: a Mendelian randomization study

Abstract

Background: Gliomas and spinal cord injuries represent significant health challenges with potential shared genetic underpinnings. Understanding the causal genetic relationships between these conditions could provide valuable insights for targeted therapeutic interventions. This study aimed to investigate potential causal genetic associations between spinal cord injury and glioma using Mendelian Randomization approaches.

Methods: We employed Mendelian Randomization (MR) to examine potential genetic associations between spinal cord injury and glioma. Four SNPs (rs1358980, rs217992, rs789990, and rs158541) were used as instrumental variables, identified from the FinnGen R11 release's "finngen_R11_C3_GBM_EXALLC" dataset. We applied three MR statistical approaches: MR Egger regression, Inverse Variance Weighted (IVW), and Weighted mode. Additionally, we analyzed gene expression patterns using RNA-sequencing data from TCGA and GEO databases, performed machine learning-based risk stratification, and validated our findings using single-cell RNA sequencing data from glioma patient tissues (GSE131928).

Results: Forest plot analyses revealed that while individual SNPs did not show significant effects on spinal cord injury (confidence intervals crossing zero), different MR methods yielded varying results. The MR Egger method demonstrated a positive correlation trend between glioma-associated genetic factors and spinal cord injury risk, while other methods showed more gradual effects. The MR analysis with the finngen_R11_C3_GBM_EXALLC genetic instrument yielded odds ratios close to 1.000 across all statistical methods (MR Egger: OR = 1.001, 95% CI 0.997-1.004, p = 0.759; IVW: OR = 1.000, 95% CI 1.000-1.000, p = 0.634), suggesting no significant causal relationship. Heterogeneity test results indicated moderate heterogeneity. Additionally, risk stratification analysis revealed significant differences in immune cell infiltration, gene expression patterns, and survival outcomes between high-risk and low-risk groups.

Conclusion: Our comprehensive analysis using Mendelian randomization provides evidence of complex genetic relationships between glioma and spinal cord injury.

Keywords: Genetic association; Glioma; Mendelian randomization; Risk stratification; SNPs; Spinal cord injury.

Fig. 1

Genetic Association Studies on the Impact of SNPs in Spinal Cord Injury and Glioma. A–D The analysis of multiple SNP loci revealed that while individual SNPs did not show significant effects on spinal cord injury (as their confidence intervals crossed the zero line), the overall trend appeared negative. The MR Egger method showed heterogeneity in the data and demonstrated a clearer negative correlation compared to other statistical methods like weighted median and simple mode

Fig. 2

The Mendelian randomization analysis investigating the potential causal relationship between glioma and spinal cord injury. All three statistical approaches (MR Egger, MR IVW, and Weighted mode) produced odds ratios very close to 1.000 with non-significant p-values (0.759, 0.634, and 0.710, respectively), suggesting no significant causal relationship between glioma and spinal cord injury

Fig. 3

Mendelian randomization analysis reveals potential causal relationship between glioma and spinal cord injury. The forest plot analyses of specific genetic variant loci (rs1358980, rs217992, rs789990, and rs158541) confirmed that individual SNP effects were not significant. Interestingly, the relationship between SNP effects on glioma and spinal cord injury showed a positive correlation trend with the MR Egger method, while weighted median and simple mode methods showed more gradual effects

Fig. 4

The relationship between risk stratification and multiple biomarkers and scores A shows correlations between immune cell types and risk scores. B Indicates higher immune cell scores in high-risk groups compared to low-risk groups. C Displays gene expression differences between high and low-risk groups, with significant variations. D Demonstrates that high-risk groups have higher stromal, immune, and ESTIMATE scores, suggesting a more active tumor microenvironment

Fig. 5

The impact of different clusters and risk stratification on survival rate. A Clustering analysis identifies two distinct groups (C1 and C2). B, C Principal component analysis (PCA) confirms clear separation between clusters and risk levels. D Cluster C1 is associated with low risk, while C2 is linked to high risk. E Survival analysis indicates that Cluster C1 has significantly better survival outcomes than C2 (p < 0.001)

Fig. 6

Differences in gene expression between different risk groups. A High-risk groups show significantly higher expression of various immune-related genes compared to low-risk groups. B A heatmap displays distinct gene expression patterns across different clusters, highlighting differences between high and low-risk groups

Fig. 7

Gene expression is linked to clinical characteristics and biological networks. A A heatmap illustrating gene expression patterns across different clinical and demographic factors. B A correlation plot highlighting relationships between genes, with strong correlations among certain genes. C A network diagram showing interactions between genes, indicating functional connections

Fig. 8

Clustering and cell type composition in glioma samples using single-cell RNA sequencing data. A, B Visualize cell clusters in glioma, identifying various cell types, including malignant and immune cells. C Shows the proportion of different cell types across patients, highlighting variability. D Illustrates the distribution of major cell types, with MES-like malignant and Mono/Macro being predominant

Fig. 9

The expression levels of specific genes across cell clusters in glioma samples using t-SNE plots. The results show expression patterns of six genes (DHFR, GART, IDH1, OGDHL, SHMT2, SUCLG2) across cell clusters. Each gene has varying levels of expression in different clusters, indicating diverse roles in the tumor microenvironment

Fig. 10

Hallmark gene sets that are differentially expressed across various cell lineages in glioma. A Down-regulated hallmark gene sets across various cell types, indicating decreased activity in specific pathways. B Up-regulated hallmark gene sets, highlighting increased activity in certain pathways across different cell types

Fig. 11

The enrichment of hallmark pathways in different risk groups and their expression patterns in glioma. A High-risk groups are enriched with specific pathways, such as DNA repair and inflammatory response. B Low-risk groups show enrichment in pathways like metabolism and immune regulation. C, D UMAP plots display the distribution of hallmark pathways, with coagulation and complement pathways showing distinct expression patterns across clusters