Exploring Metabolites as Therapeutic Targets in Lung Cancer: Insights from Mendelian Randomization

Background： The primary objective of this study is to utilize Mendelian randomization analysis methods to investigate causal relationships between various metabolites and lung cancer. Our aims include not only validating direct associations between metabolites and lung cancer risk, but also exploring potential molecular mechanisms and regulatory pathways underlying these relationships. Through these efforts, we aim to provide new biomarkers and therapeutic targets for early diagnosis and personalized treatment strategies for lung cancer. Methods： This study aims to explore causal relationships between 1,091 metabolites and lung cancer (LC). The research design includes metabolite data from 8,299 individuals collected from the Canadian Longitudinal Study on Aging (CLSA), encompassing large-scale GWAS analyses. Summary statistics for LC come from GWAS covering 29,836 cases and 55,586 controls, as well as another dataset comprising 3,791 cases and 489,012 control individuals. In Mendelian randomization (MR) analysis, 753 metabolites were selected and studied using SNPs as instrumental variables, rigorously screened and confirmed for validity using F-statistics. Various MR methods were employed, including inverse variance weighting and Wald ratio, with significance determined by Bonferroni correction. Sensitivity analyses included Cochran's Q statistics and MR-Egger method to assess heterogeneity and pleiotropy among instrumental SNPs. Generalized summary data-based MR (GSMR) was used to validate causal relationships between LC and metabolites, with HEIDI method applied to exclude pleiotropic SNPs. Additionally, colocalization analysis explored shared regulatory genes of causal SNPs between LC and metabolites. Finally, SMR analysis further investigated relationships with gene expression. In summary, this study integrates multiple methods and large datasets to uncover potential associations and molecular mechanisms between various metabolites and LC. Results： In Mendelian randomization analysis, using IVW and Wald ratio methods, 6 metabolites significantly associated with lung cancer (LC) were identified: increased risk with 3-methylxanthine levels and X-18935 levels, decreased risk with Paraxanthine levels, Isovalerylcarnitine (C5) levels, Indolin-2-one levels, and 6-hydroxyindole sulfate levels. Additionally, 63 metabolites showed potential associations with LC. Validation analyses confirmed LC associations for 35 metabolites using external data. GSMR analysis validated LC associations for 32 metabolites, with consistent effect directions, including significant associations for Paraxanthine levels, X-18935 levels, and 6-hydroxyindole sulfate levels. Colocalization analysis revealed significant evidence of shared colocalization for 5 metabolites with LC, all associated with the same candidate causal SNP. Gene expression analysis demonstrated complex regulatory relationships between multiple genes (e.g., HIST1H4E and GATAD2A) and LC and its related metabolite levels. Conclusion： This study provides a comprehensive analysis of complex relationships between various metabolites and lung cancer, revealing their potential roles and regulatory mechanisms in disease development. By integrating different analytical and validation approaches, we offer important scientific insights for future personalized strategies in disease prevention and treatment. These findings not only contribute to discovering new biomarkers but also lay groundwork for targeted therapeutic approaches and personalized medical interventions.