Ma Canchen, Shen Ziyuan, Tian Jing, Schooneveldt Yvette L, Giles Corey, Cicuttini Flavia, Jones Graeme, Meikle Peter J, Pan Feng
Menzies Institute for Medical Research, University of Tasmania, Private Bag 23, Hobart, TAS, 7000, Australia.
Department of Rheumatology, School of Medicine, Renji Hospital, Shanghai Jiaotong University, Shanghai, China.
Calcif Tissue Int. 2025 Jun 25;116(1):89. doi: 10.1007/s00223-025-01399-1.
To identify and validate lipid metabolites associated with bone mineral density (BMD) change and fracture risk through integrated Mendelian randomization (MR) and observational analyses. Two-sample MR analysis was first performed to uncover potential causal relationships between 32 lipid classes and 576 lipid species and BMD and fractures. Identified signatures were subsequently validated in an independent cohort (N = 492), where lipids, BMD, and fracture status were measured at two time points, 8 years apart. The false discovery rate method was employed to control multiple testing. Linear and log binomial mixed-effects models were used to analyze lipid associations with hip BMD and fracture risk, respectively. Two-sample MR revealed seven lipid classes causally associated with BMD and/or fractures, including acylcarnitine (AC), cholesteryl ester (CE), sphingomyelin (SM), phosphatidylinositol (PI), GM3 ganglioside (GM3), alkylphosphatidylcholine (PC(O)) and triacylglycerol (TG). Causal associations were found between 18 lipid species across these classes and BMD, and 10 lipid species were associated with fractures. Validation in an independent longitudinal cohort confirmed associations for total SM, SM(d18:1/16:0), SM(d18:2/24:0), and CE(18:3) with hip BMD change (β ranging from - 0.036 to - 0.012 g/cm, per log µM increase, p < 1.13 × 10). Total SM, total GM3, and SM(d18:2/18:1), SM(d18:2/22:0), SM(d18:2/17:0) were associated with an increased risk of fractures (RR ranging from 1.038 to 1.290 g/cm, per log µM increase, p < 5 × 10) over 8 years. Our findings suggest that alterations in lipid metabolism play a causal role in bone remodeling and fracture risk. This warrants further investigation into the mechanisms of lipid-mediated BMD changes and the potential for identifying patients at 'high risk' of osteoporotic fracture.
通过整合孟德尔随机化（MR）和观察性分析来识别和验证与骨密度（BMD）变化及骨折风险相关的脂质代谢物。首先进行两样本MR分析，以揭示32种脂质类别和576种脂质种类与BMD及骨折之间的潜在因果关系。随后在一个独立队列（N = 492）中对识别出的特征进行验证，在该队列中，分别在相隔8年的两个时间点测量脂质、BMD和骨折状态。采用错误发现率方法来控制多重检验。分别使用线性和对数二项混合效应模型分析脂质与髋部BMD及骨折风险的关联。两样本MR显示，七种脂质类别与BMD和/或骨折存在因果关联，包括酰基肉碱（AC）、胆固醇酯（CE）、鞘磷脂（SM）、磷脂酰肌醇（PI）、GM3神经节苷脂（GM3）、烷基磷脂酰胆碱（PC(O)）和三酰甘油（TG）。在这些类别中的18种脂质种类与BMD之间发现了因果关联，10种脂质种类与骨折有关。在一个独立的纵向队列中的验证证实，总SM、SM(d18:1/16:0)、SM(d18:2/24:0)和CE(18:3)与髋部BMD变化相关（每增加log μM，β范围为-0.036至-0.012 g/cm，p < 1.13×10）。总SM、总GM3以及SM(d18:2/18:1)、SM(d18:2/22:0)、SM(d18:2/17:0)与8年内骨折风险增加相关（每增加log μM，RR范围为1.038至1.290 g/cm，p < 5×10）。我们的研究结果表明，脂质代谢改变在骨重塑和骨折风险中起因果作用。这值得进一步研究脂质介导的BMD变化机制以及识别骨质疏松性骨折“高危”患者的潜力。
