Department of Neurology, Henry Ford Hospital, West Bloomfield, Michigan, USA; School of Medicine, Wayne State University, Detroit, Michigan, USA.
Mov Disord. 2013 Oct;28(12):1653-60. doi: 10.1002/mds.25555. Epub 2013 Jul 19.
Parkinson's disease (PD) biomarkers are needed to enhance therapeutics research and to understand PD pathogenesis. Methods that simultaneously measure hundreds of small molecular-weight compounds-metabolomic analysis-"fingerprint" disease-specific alterations in individual compounds or metabolic pathways. Beyond a nontargeted search for PD biomarkers, we hypothesized that PD cerebrospinal fluid would show increased formation of the excitotoxin 3-hydroxykynurenine and diminished concentration of the antioxidant glutathione. Cerebrospinal fluid was collected at <4 hours postmortem from 48 pathologically-verified PD subjects and 57 comparably-aged controls. Assays involved ultra-high-performance liquid and gas chromatography linked to mass spectrometry. We used univariate techniques to determine fold-changes in concentrations of biochemicals; false-discovery rates were calculated to exclude spurious findings. Data was modeled using a Support Vector Machine for analyzing compounds selected by Welch's t test. Classification accuracy was determined by cross-validation. Of 243 structurally-identified biochemicals,19 compounds differentiated PD from controls at a 20% false-discovery level. In PD, mean 3-hydroxykynurenine concentration was increased by one-third, and mean oxidized glutathione was decreased by 40% (for each, P < .01). Four of the 19 compounds differentiating PD from controls were N-acetylated amino acids, suggesting a generalized alteration in N-acetylation activity. The Support Vector Machine classification model distinguished between groups at 83% sensitivity and 91% specificity for the learning data, and at 65% and 79% from cross-validation. In this study, the first for metabolomic profiling of PD cerebrospinal fluid, we found several novel biomarkers and offer new directions for recognizing disease-specific biochemical indicators. The findings support involvement of excitotoxicity and oxidative stress in the pathogenesis of PD.
帕金森病 (PD) 生物标志物对于增强治疗学研究和理解 PD 发病机制非常重要。同时测量数百种小分子化合物的方法——代谢组学分析——“指纹”个体化合物或代谢途径的疾病特异性改变。除了针对 PD 生物标志物的非靶向搜索外，我们假设 PD 脑脊液中会显示出兴奋性毒素 3-羟基犬尿氨酸的形成增加和抗氧化剂谷胱甘肽的浓度降低。从 48 名经病理证实的 PD 患者和 57 名年龄相近的对照组中，在死后 <4 小时收集脑脊液。测定涉及超高效液相和气相色谱与质谱联用。我们使用单变量技术来确定生化物质浓度的变化倍数；使用错误发现率来排除虚假发现。使用支持向量机分析 Welch t 检验选择的化合物来建模数据。通过交叉验证确定分类准确性。在 243 种结构鉴定的生化物质中，有 19 种化合物在 20%的错误发现率下将 PD 与对照组区分开来。在 PD 中，3-羟基犬尿氨酸的平均浓度增加了三分之一，氧化型谷胱甘肽的平均浓度降低了 40%（均为 P<0.01）。从对照组中区分 PD 的 19 种化合物中有 4 种是 N-乙酰化氨基酸，这表明 N-乙酰化活性普遍改变。支持向量机分类模型在学习数据中的敏感性为 83%，特异性为 91%，交叉验证中的敏感性为 65%，特异性为 79%。在这项研究中，首次对 PD 脑脊液进行代谢组学分析，我们发现了几种新的生物标志物，并为识别疾病特异性生化指标提供了新的方向。这些发现支持兴奋性毒性和氧化应激在 PD 发病机制中的作用。
