代谢组学在抑郁症生物标志物研究中的应用进展

Progress in the Application of Metabolomics to Biomarker Studies of Depression

摘要: 抑郁症是一种常见且高度异质性的精神障碍，其发生发展涉及神经递质失衡、能量代谢异常、免疫炎症激活、氧化应激增强及肠道菌群失调等多种生物学过程。由于当前临床诊断仍主要依赖症状学评估，缺乏客观稳定的生物学指标，探索可重复、可转化的生物标志物已成为抑郁症研究的重要方向。代谢组学能够从整体层面刻画生物体内小分子代谢物的动态变化，为解析抑郁症相关代谢紊乱及筛选潜在标志物提供了有力工具。本文围绕抑郁症，系统综述了代谢组学在其生物标志物研究中的应用进展。首先，从氨基酸代谢、能量代谢和脂质代谢三个方面总结抑郁症的主要代谢异常特征，重点讨论色氨酸–犬尿氨酸通路、三羧酸循环及脂质代谢网络的改变。其次，从单一代谢物、多代谢物组合以及基于代谢通路的标志物三个层面归纳近年来的研究进展，并分析其在疾病识别、分型及疗效预测中的潜力。与此同时，本文进一步讨论了代谢组学研究中常见的方法学局限，包括样本来源差异、检测平台异质性、统计模型过拟合，以及抗抑郁药物使用、饮食结构、昼夜节律和共病状态等混杂因素对结果稳定性的影响。总体来看，抑郁症并非单一通路异常，而是多通路协同紊乱的系统性代谢失衡状态。相较于单一代谢物，多代谢物组合及通路层面的标志物具有更高的稳定性和诊断潜力。未来应加强前瞻性纵向研究、多时间点动态采样及大样本外部验证，并结合转录组、蛋白质组和微生物组等多组学数据，推动代谢组学标志物向临床转化。

Abstract: Depression is a common and highly heterogeneous psychiatric disorder involving multiple biological processes, including neurotransmitter imbalance, altered energy metabolism, immune-inflammatory activation, oxidative stress, and gut microbiota dysbiosis. Since current clinical diagnosis still relies mainly on symptom-based assessment and lacks objective and stable biological indicators, the identification of reproducible and clinically translatable biomarkers has become a major research priority. Metabolomics enables global profiling of dynamic small-molecule changes in biological systems and provides a powerful approach for characterizing metabolic disturbances and identifying potential biomarkers in depression. This review systematically summarizes recent advances in metabolomics-based biomarker studies in depression. First, the major metabolic abnormalities associated with depression are discussed from the perspectives of amino acid metabolism, energy metabolism, and lipid metabolism, with particular emphasis on the tryptophan-kynurenine pathway, the tricarboxylic acid cycle, and lipid metabolic networks. Second, recent progress in single-metabolite biomarkers, multi-metabolite panels, and pathway-based biomarkers is reviewed, together with their potential applications in diagnosis, stratification, and treatment response prediction. In addition, common methodological limitations in metabolomics studies are addressed, including heterogeneity in sample sources and analytical platforms, statistical overfitting, and the confounding effects of antidepressant exposure, diet, circadian rhythm, and comorbidities. Overall, depression should be viewed as a systemic metabolic dysregulation involving multiple disturbed pathways rather than a single metabolic defect. Compared with individual metabolites, multi-metabolite panels and pathway-level markers may provide greater robustness and diagnostic value. Future studies should prioritize prospective longitudinal designs, multi-time-point sampling, large-scale external validation, and multi-omics integration to facilitate the clinical translation of metabolomics-based biomarkers in depression.