Brain multi-omic Mendelian randomisation to identify novel drug targets for gliomagenesis

• Objective: This study employed a novel brain multi-omic Mendelian randomization (MR) approach, integrating brain-specific genetic, transcriptomic (eQTLs), and proteomic (pQTLs) data to identify genes and proteins causally associated with glioma risk.
• Key Findings: The analysis identified 25 causal genes and, more specifically, 13 causal proteins. The strongest evidence pointed to two promising drug targets for glioblastoma (GBM): genetically predicted lower levels of the protein FAM178B and higher levels of the protein MDM4 both increased risk.
• Significance: The work validates the use of brain-specific proteomic data in MR to enhance the prioritization of highly relevant molecular drug targets for neurological cancers.

PubMed: 39565278 DOI: 10.1093/hmg/ddae168 Overview generated by: Gemini 2.5 Flash, 28/11/2025

Key Findings: Multi-Omic Prioritization of Glioma Drug Targets

This study utilized a novel multi-omic Mendelian randomization (MR) approach, integrating genetic, transcriptomic (gene expression), and proteomic (protein expression) data specifically from human brain tissue, to prioritize genes and proteins causally implicated in the development of glioma and its subtypes (glioblastoma, GBM).

• Prioritized Causal Genes (Transcriptome): The MR analysis using brain-specific expression data (eQTLs) identified 25 genes whose genetically predicted expression was causally associated with overall glioma risk, including several novel candidate genes.
• Prioritized Causal Proteins (Proteome): A key advancement was the use of brain-specific proteomic data (pQTLs), which identified 13 proteins whose genetically predicted levels were causally associated with glioma risk. The most compelling findings included:
  • Increased risk of GBM associated with genetically predicted lower levels of the protein FAM178B (Odds Ratio (OR) of 0.81 per 1 SD decrease, \(P=1.92 \times 10^{-6}\)).
  • Increased risk of GBM associated with genetically predicted higher levels of the protein MDM4 (OR of 1.15 per 1 SD increase, \(P=2.87 \times 10^{-5}\)).
• Integration of Omics: The multi-omic approach successfully validated previously known targets (e.g., TP53, MDM4) and identified novel mechanisms. The strongest and most robust evidence for potential drug targets were the proteins FAM178B and MDM4, which showed consistent association with glioma subtypes.
• Tissue Specificity Confirmed: The results confirmed the importance of brain-specific molecular data, as the effects observed were often specific to brain tissue and not replicated when using non-brain tissues (e.g., whole blood).

Study Design and Methods

Study Design

This was a two-sample multi-omic Mendelian randomization (MR) study. It leveraged two distinct sets of instrumental variables: one for gene expression and one for protein expression, both derived from brain tissue, to infer causal relationships with glioma risk.

Data Sources and Instrumental Variables

1. Outcome GWAS Data (Glioma Risk): Summary statistics were used for overall glioma risk and the specific subtypes glioblastoma (GBM) and non-GBM glioma, sourced from large-scale consortia (up to 7,400 cases).
2. Exposure Data (Brain Multi-omics):
   • Transcriptomics (Gene Expression): Expression quantitative trait loci (eQTLs) were used from the GTEx consortium and other brain-specific datasets for gene expression in various brain regions.
   • Proteomics (Protein Expression): Protein quantitative trait loci (pQTLs) were used from two key human brain-derived proteomic datasets, providing genetic instruments for protein levels.
3. Instrument Selection: Genetic variants (SNPs) acting as both eQTLs or pQTLs were selected. Strict quality control, including linkage disequilibrium (LD) clumping, was applied to ensure the independence of the instruments.

Statistical Analysis

• Primary MR Method: The Inverse-Variance Weighted (IVW) method was used to calculate the causal effect estimates (Odds Ratios, ORs).
• Sensitivity Analyses: The following robust MR methods were applied to assess the validity of the instruments and detect pleiotropy:
  • MR-Egger regression (to test for balanced pleiotropy).
  • Weighted Median and Weighted Mode estimators (to provide consistent estimates even with invalid instruments).
  • MR-PRESSO (to detect and adjust for horizontal pleiotropy outliers).
• Colocalization Analysis: Bayesian colocalization (using the moloc method) was performed to ensure that the same causal variant drove both the molecular exposure (expression or protein level) and the disease outcome (glioma risk), strengthening the causal evidence.
• Drug Target Prioritization: The results were systematically compared against existing databases (DGIdb, Open Targets) to prioritize genes and proteins that are known to be targetable by existing drugs or are considered strong drug candidates.

Conclusions and Recommendations

The multi-omic MR pipeline successfully identified and prioritized several brain-specific molecular targets for gliomagenesis, offering substantial advantages over relying solely on genetic or expression data.

• Drug Development Focus: The strong evidence for a causal role of the proteins FAM178B and MDM4 in GBM risk, based on brain-derived proteomic data, provides highly compelling candidates for drug development and repurposing efforts.
• Methodological Advance: The study strongly recommends the use of brain-specific pQTL data in future MR studies of neurological diseases, as protein levels are often a more direct and clinically relevant measure of biological function than mRNA expression.
• Need for Functional Studies: While the MR analysis provides strong evidence for a causal link, the authors stress that follow-up functional validation studies are necessary to fully elucidate the exact molecular mechanisms by which these prioritized targets influence tumor risk.