Institute of Molecular Cardiovascular Research, University Hospital RWTH Aachen, Aachen, Germany.
Department of Internal Medicine, Nephrology, Rheumatology, Diabetology and Endocrinology, Helios Hospital Krefeld, Krefeld, Germany.
J Mol Med (Berl). 2019 Oct;97(10):1451-1463. doi: 10.1007/s00109-019-01823-8. Epub 2019 Aug 5.
Chronic kidney disease (CKD) may progress to end-stage renal disease (ESRD) at different pace. Early markers of disease progression could facilitate and improve patient management. However, conventional blood and urine chemistry have proven unable to predict the progression of disease at early stages. Therefore, we performed untargeted plasma peptidome analysis to select the peptides involved in progression, which are suitable for long prospective studies in future. The study consists of non-CKD (n = 66) and CKD (n = 106) patients with different stages. We performed plasma peptidomics on these subjects using chromatography and mass spectrometric approaches. Initially, we performed LC-ESI-MS and applied least absolute shrinkage and selection operator logistic regressions to select the peptides that are differentially expressed and we generated a peptidomic score for each subject. Later, we identified and sequenced the peptides with MALDI-MS/MS and also performed univariate and multivariate analyses with the clinical variables and peptidomic score to reveal their association with progression of renal disease. A logistic regression model selected 14 substances showing different concentrations according to renal function, of which seven substances were most likely occur in CKD patients. The peptidomic model had a global P value of < 0.01 with R of 0.466, and the area under the curve was 0.87 (95% CI, 0.8149-0.9186; P < 0.0001). The predicted score was significantly higher in CKD than in non-CKD patients (2.539 ± 0.2637 vs - 0.9382 ± 0.1691). The model was also able to predict stages of CKD: the Spearman correlation coefficient of the linear predictor with CKD stages was 0.83 with concordance indices of 0.899 (95% CI 0.863-0.927). In univariate analysis, the most consistent association of peptidomic score in CKD patients was with C-reactive protein, sodium level, and uric acid, which are unanticipated substances. Peptidomic analysis enabled to list some unanticipated substances that have not been extensively studied in the context of CKD but were associated with CKD progression, thus revealing interesting candidate markers or mediators of CKD of potential use in CKD progression management. KEY MESSAGES: • Conventional blood and urine chemistry have proven unable to predict the progression of disease at early stages of chronic kidney disease (CKD). • We performed untargeted plasma peptidome analysis to select the peptides involved in progression. • A logistic regression model selected 14 substances showing different concentrations according to renal function. • These peptides are unanticipated substances that have not been extensively studied in the context of CKD but were associated with CKD progression, thus revealing markers or mediators of CKD of potential use in CKD progression management.
慢性肾脏病（CKD）可能以不同的速度进展为终末期肾病（ESRD）。疾病进展的早期标志物可以促进和改善患者管理。然而，常规的血液和尿液化学已被证明无法在早期阶段预测疾病的进展。因此，我们进行了非靶向血浆肽组学分析，以选择与疾病进展相关的肽，这些肽适合未来进行长期前瞻性研究。该研究包括非 CKD（n=66）和 CKD（n=106）不同阶段的患者。我们使用色谱和质谱方法对这些受试者进行了血浆肽组学分析。最初，我们进行了 LC-ESI-MS 并应用最小绝对收缩和选择算子逻辑回归来选择差异表达的肽，为每个受试者生成肽组学评分。之后，我们使用 MALDI-MS/MS 鉴定和测序了肽，并对临床变量和肽组学评分进行了单变量和多变量分析，以揭示它们与肾脏疾病进展的关联。逻辑回归模型选择了 14 种根据肾功能显示不同浓度的物质，其中 7 种物质很可能出现在 CKD 患者中。肽组学模型的全局 P 值<0.01，R 为 0.466，曲线下面积为 0.87（95%CI，0.8149-0.9186；P<0.0001）。CKD 患者的预测评分明显高于非 CKD 患者（2.539±0.2637 与-0.9382±0.1691）。该模型还能够预测 CKD 阶段：线性预测因子与 CKD 阶段的斯皮尔曼相关系数为 0.83，一致性指数为 0.899（95%CI 0.863-0.927）。在单变量分析中，CKD 患者肽组学评分最一致的关联是与 C 反应蛋白、钠水平和尿酸，这是意外物质。肽组学分析能够列出一些在 CKD 背景下尚未广泛研究但与 CKD 进展相关的意外物质，从而揭示出潜在的 CKD 候选标志物或介质，可用于 CKD 进展管理。关键信息： • 常规血液和尿液化学已被证明无法在慢性肾脏病（CKD）的早期阶段预测疾病的进展。 • 我们进行了非靶向血浆肽组学分析以选择与疾病进展相关的肽。 • 逻辑回归模型根据肾功能选择了 14 种显示不同浓度的物质。 • 这些肽是意外物质，在 CKD 背景下尚未广泛研究，但与 CKD 进展相关，从而揭示了潜在的 CKD 标志物或介质，可用于 CKD 进展管理。
