Integrative Mendelian randomization and spatial transcriptomics analysis reveals causal links between plasma lipidome...

This study aimed to investigate the potential causal relationship between plasma lipidome and non-small cell lung cancer (NSCLC) using a two-sample Mendelian randomization (MR) approach based on large-scale genome-wide association study (GWAS) data.

A two-sample bidirectional MR analysis was conducted using summary-level data from the GeneRISK and FinnGen GWAS. Genetic instruments were selected for 179 plasma lipid traits. Causal effects were estimated using inverse variance weighting (IVW), MR-Egger, weighted median, and Mendelian Randomization Pleiotropy RESidual Sum and Outlier (MR-PRESSO) methods. Spatial transcriptomic profiling and machine learning-based gene prioritization were further employed to explore potential mechanistic links.

MR identified 30 lipid species significantly associated with NSCLC risk, of which 17 demonstrated consistent causal effects across multiple sensitivity analyses. A subset of phosphatidylcholine and phosphatidylethanolamine species exhibited inverse associations, indicating potential protective effects, while specific sterol esters, sphingomyelins, and triacylglycerols were positively associated with disease susceptibility. Plasmacytoma Variant Translocation 1 (PVT1) was first prioritized as a key lipid-associated gene by three machine learning algorithms and was subsequently validated by spatial transcriptomics, showing elevated expression in tumor-dense regions.

This study reveals a causal link between distinct plasma lipid species and NSCLC risk, emphasizing the role of lipid structural diversity. PVT1 may act as a critical mediator connecting lipid dysregulation to lung tumorigenesis, offering novel biomarker and therapeutic target potential.