A Big Data investigation of the influence of life-long metabolic factors that contribute to age-related eye diseases

有大量证据表明，老年性眼病（AREDs），特别是老年性黄斑变性（AMD）、青光眼和老年性白内障与代谢功能障碍有关。以前的AREDs代谢组学研究在不同的队列中使用一系列不同的代谢物评估技术进行了全代谢组分析。然而，许多队列以临床为基础，动力不足，使用各种平台，文献中几乎没有趋同。例如，一项对13项AMD研究的荟萃综述(1)揭示了广泛的潜在生物标志物，包括苯丙氨酸、腺苷、次黄嘌呤、酪氨酸、肌酸、柠檬酸、肉碱、脯氨酸和麦芽糖，并且在受影响的个体中，脂质、碳水化合物、氨基酸和核苷酸代谢可能失调。meta分析的研究结果缺乏一致性，因此需要进一步研究，以更高的可信度选择和验证AMD的代谢物生物标志物。在青光眼中，最近对18项代谢组学研究的回顾(2)表明，血浆中羟脯氨酸和蛋氨酸持续增加，而房水结果显示青光眼中丙氨酸、肌酸、甘氨酸和赖氨酸增加。作者认为，基于等离子体的研究结果不如使用水幽默的结果一致。由于血浆是大规模人群研究中唯一实用且相对无创的数据来源，因此需要进一步改进AREDs的血浆代谢组学研究。该项目的主管在2019年发表了TwinsUK对313种代谢物的研究(3)，发现维生素C代谢物o -甲基抗坏血酸酯与眼压（青光眼的主要危险因素）有关，并导致眼压降低，这突出了抗氧化剂在对抗光氧化应激中的作用。这项特殊的研究还通过使用随机森林机器学习来识别最具影响力的代谢物，以及孟德尔随机化来研究代谢物对IOP的因果关系，为该项目的方法学方法树立了榜样。迄今为止，与年龄相关的白内障受到的关注较少，很少有高通量代谢组学研究：一项基于血浆代谢组学分析的研究没有发现任何已发表的数据，除了一篇关于青光眼的文章，其中使用了性别匹配的白内障对照。虽然有一种有效的手术治疗方法，但揭示白内障患者的代谢特征将提高对分子机制的理解，如果干预可以延缓白内障，那么就可以显著减少手术量和节省成本。该项目的主管可以访问包括英国生物银行在内的四个超过50万人的大型队列的数据，这意味着该研究有能力识别新的关联，以进一步了解与年龄相关的眼病的病理生理学，潜在地检查因果关系，提高新疗法的前景，并使用机器学习模型使用代谢组学平台来预测那些有疾病风险的人。侯晓伟等（2020）黄斑变性代谢组学研究综述。唐莹等。（2022）原发性开角型青光眼的代谢组学研究：系统评价和meta分析。中国眼科杂志，61,13-13。Hysi PG等（2019）抗坏血酸代谢物参与普通人群的眼压控制。氧化还原生物学学报，20 (2):349-353

There is a body of evidence linking age-related eye diseases (AREDs), and in particular age-related macular degeneration (AMD), glaucoma and age-related cataract to metabolic dysfunction. Previous metabolomic studies of AREDs have performed metabolome-wide analyses in different cohorts with a range of different metabolite evaluation techniques. However, many cohorts were clinic-based, underpowered and used a variety of platforms and there has been little convergence in the literature. For example, a meta-review of 13 AMD studies(1) revealed that a broad range of potential biomarkers, including phenylalanine, adenosine, hypoxanthine, tyrosine, creatine, citrate, carnitine, proline and maltose, and that lipid, carbohydrate, amino acid and nucleotide metabolism may be dysregulated in affected individuals. The lack of convergence of results from the studies in the meta-analysis supports the need for further studies to select and validate metabolite biomarkers of AMD with higher confidence. In glaucoma, a recent review of 18 metabolomics studies(2) suggested that hydroxyproline and methionine were consistently increased in plasma, while results from aqueous humour revealed alanine, creatine, glycine and lysine increased in glaucoma. The authors argued that plasma-based findings were less consistent than those using aqueous humour. Since plasma is the only practical and relatively non-invasive source of data for large population studies, there is therefore need for further improvements in plasma metabolomics studies of AREDs. The supervisors of this project published a 2019 study(3) from TwinsUK of 313 metabolites and found a vitamin C metabolite O-methylascorbate is associated with and causally lowers IOP, the main risk factor for glaucoma, highlighting the power of antioxidants in combating photooxidative stress. This particular study also sets an example for the methodological approach in this project by using random forest machine learning to identify the most influential metabolites, and Mendelian randomization to investigate a causal effect of metabolites on IOP. Age-related cataract has to date received less attention, with few high-throughput metabolomic studies: a search for plasma-based metabolomic profiling returned no published data except for one article on glaucoma where sex-matched controls with cataract were used. While there is an effective surgical remedy, revealing metabolic signatures of cataract patients would improve understanding of molecular mechanisms and if an intervention could delay cataract, then significant surgical volume reduction and cost savings could follow. The supervisors of this project have access to data from four large cohorts of over 500,000 people, including the UK Biobank, meaning that this study has the power to identify novel associations to further understand the pathophysiology of age-related eye diseases, to potentially examine for causality raising the prospect of novel therapies, and to use machine learning models to predict those at risk of disease using metabolomics platforms. Hou XW et al. (2020) Metabolomics in Age-Related Macular Degeneration: A Systematic Review. Invest Ophthalmol Vis Sci. 61, 13-13 Tang Y et al. (2022) Metabolomics in Primary Open Angle Glaucoma: A Systematic Review and Meta-Analysis. Front Neurosci. 16, 835736 Hysi PG et al. (2019) Ascorbic acid metabolites are involved in intraocular pressure control in the general population. Redox Biol. 20, 349-353