核磁共振脂质代谢物与胰腺癌风险因果关系的双样本孟德尔随机化分析

A two-sample Mendelian randomization analysis of the causal relationship between NMR-based lipid metabolites and pancreatic cancer risk

摘要:

背景与目的 胰腺癌（PC）是一种预后极差的消化系统恶性肿瘤，其发生机制尚未完全明确。现有研究提示血浆代谢组学改变与PC发生可能存在关联，但传统观察性研究易受混杂因素和反向因果关系干扰，难以明确代谢物与PC的因果联系。本研究基于双样本孟德尔随机化（MR）方法，系统评估325种核磁共振代谢物与PC风险之间是否存在潜在的因果关系。方法 整合UK Biobank中325种核磁共振代谢物的GWAS数据与FinnGen胰腺癌GWAS数据，筛选与代谢物显著相关的单核苷酸多态性作为工具变量。以逆方差加权法为主，结合MR-Egger、加权中位数、加权模式、贝叶斯加权孟德尔随机化（BWMR）和约束最大似然法（cML）等方法进行验证，并开展多重敏感性分析以评估结果稳健性结果 共发现4项代谢物与胰腺癌风险存在显著因果关联。其中，中密度脂蛋白（IDL）中磷脂与总脂质的比值（GCST90445881）、小高密度脂蛋白（HDL）中磷脂/总脂质比值（GCST90446027），以及极大极低密度脂蛋白（VLDL）中游离胆固醇/总脂质比值（GCST90446151）升高均与PC风险降低相关；相反，乳糜微粒和极大VLDL中甘油三酯/总脂质比值（GCST90446157）升高则增加PC风险。多种敏感性分析均支持结果的稳健性。结论 本研究从遗传学角度揭示脂质代谢异常与PC风险之间的因果关系，提示磷脂及游离胆固醇比例升高具有保护作用，而甘油三酯水平升高则为危险因素。相关代谢特征有望成为PC早期诊断与干预的重要生物标志物，为临床风险评估和代谢靶向治疗提供新思路。

Abstract:

Background and Aims Pancreatic cancer (PC) is a highly lethal gastrointestinal malignancy with poorly understood pathogenesis. Previous studies suggest that alterations in plasma metabolomics may be associated with PC development; however, traditional observational studies are prone to confounding and reverse causation, making it difficult to establish causal relationships. This study employed a two-sample Mendelian randomization (MR) approach to systematically evaluate the potential causal relationship between 325 nuclear magnetic resonance (NMR) metabolites and PC risk.Methods Genome-wide association study (GWAS) data of 325 NMR metabolites from the UK Biobank were integrated with GWAS data of PC from FinnGen. Single nucleotide polymorphisms (SNPs) significantly associated with metabolites were selected as instrumental variables. The inverse variance weighted method served as the primary analysis, supplemented by MR-Egger regression, weighted median, weighted mode, Bayesian weighted Mendelian randomization (BWMR), and constrained maximum likelihood (cML) for validation. Multiple sensitivity analyses were performed to assess the robustness of the results.Results Four metabolites were identified to have significant causal associations with PC risk. Higher phospholipid-to-total lipid ratios in intermediate-density lipoproteins (IDL) (GCST90445881) and small high density lipoproteins (HDL) (GCST90446027), as well as higher free cholesterol-to-total lipid ratios in extremely large very-low-density lipoproteins (VLDL) (GCST90446151), were inversely associated with PC risk. Conversely, an elevated triglyceride-to-total lipid ratio in chylomicrons and extremely large VLDL (GCST90446157) was positively associated with increased PC risk. The findings were consistently supported by multiple sensitivity analyses.Conclusion This study provides genetic evidence linking lipid metabolism alterations to PC risk. Elevated phospholipid and free cholesterol ratios appear protective, whereas increased triglyceride levels act as risk factors. These metabolite profiles may serve as promising biomarkers for early diagnosis and intervention in PC, offering novel insights for risk assessment and potential metabolic-targeted therapies.