Identification of Novel Susceptibility Genes for Glioma: A Multi-Omics and Transcriptome-Wide Association Study Integrating Cross-Tissue and Single-Tissue Analyses, Summary-Data-Based Mendelian Randomization and Bioinformatic Validation

Abstract

Background: Genome-wide association study (GWAS) have successfully identified a substantial number of susceptibility loci in gliomas. However, the underlying genetic mechanisms are complicated and have not been entirely elaborated.

Methods: Cross-tissue and single-tissue transcriptome-wide association studies (TWAS) were conducted based on summary statistics from 12496 cases, including 6,191 glioblastomas (GBM) and 5,819 low-grade gliomas (LGG), and 18,196 controls of large glioma GWAS and different genetic-expression matrix models of 49 human tissues in the Genotype-Tissue Expression (GTEx) project. The TWAS results were validated with different gene-based methodologies, including summary data-based Mendelian randomization, colocalization analysis, and multimarket analysis of genomic annotation. Additionally, bioinformatics analyses were performed to characterize the candidate genes.

Results: Evaluation of both the cross-tissue and single-tissue test identified five susceptibility genes (CDKN2B, HEATR3, STMN3, D2HGDH, and GALNT6) in known loci in addition to seven novel genes (GMPPB, CCDC88B, BAIAP2L2, FAIM, GPX1, RRM1 and SCFD1). The two tier 1 genes CDKN2B and HEATR3 were significantly associated with the risk of all glioma, while one tier 1 gene CDKN2B and two tier 3 genes: D2HGDH and STMN3 were specifically associated with the risk of LGG, and six tier 1 genes, including GMPPB, CDKN2B, CCDC88B, GALNT6, HEATR3and BAIAP2L2, were associated with the risk of GBM. Single-tissue analysis found that the tier 2 genes: SCFD1 and the tier 3 gene D2HGDH were negatively associated with the risk of all glioma, whereas the tier 1 gene RRM1, the tier 2 genes: GPX1, FAIM were positively associated with the risk of all glioma. Similarly, the predicted tier 3 gene D2HGDH was negatively associated with the risk of LGG. Furthermore, high expression of the tier 1 genes FAIM and GMPPB, and the tier 3 gene: GPX1 may increase the risk of GBM. Further validation of differential expression and survival analysis supported the cross-tissue and single-tissue TWAS results.

Conclusions: These results highlight the power of TWAS to identify and prioritize putative causal genes, thereby enhancing our understanding of the genetic architecture and complex causality of glioma.

Funding: This work was supported by Doctoral Workstation Climbing Project of Zhangzhou Hospital (PDA202306) and Natural Science Foundation of Fujian Province (No.2021J011412).

Declaration of Interest: The authors report no competing interests.

Ethical Approval: Included studies had been approved by corresponding ethical review committees and all participants signed the consent forms.

Keywords: Glioma, Cross-tissue transcriptome-wide association study, Summary-data-based Mendelian randomization, FUSION, Multimarker analysis of genomic annotation, Unified test for molecular signatures