[Analyzing urinary proteome patterns of metabolic syndrome patients with early renal injury by magnet bead separation and matrix-assisted laser desorption ionization time-of-flight mass spectrometry].

OBJECTIVE

To determine the potential urinary biomarkers of metabolic syndrome (MS) with early renal injury and establish diagnostic models by magnetic bead-based separation and matrix-assisted laser desorption ionization time-of-flight mass spectrometry (MALDI-TOF-MS).

METHODS

Participants were selected from the epidemiologic study on MS and renal involvement among residents in Pinggu district, Beijing. Eight-hour overnight urine samples were fractionated by means of magnetic bead-based weak cation exchange chromatography and subsequently analyzed with MALDI-TOF-MS. Wilcoxon test and random forests were used to screen differential protein peaks of MS patients with early renal injury, then combined with genetic algorithm and support vector machine, respectively, to establish diagnostic models.

RESULTS

Totally 54 cases of MS without renal injury and 46 cases of MS with early renal injury were enrolled. Totally twenty protein peaks were up-regulated in the urine of MS patients with early renal injury by Wilcoxon test (P < 0.05); random forests algorithm revealed twelve protein peaks up-regulated in the urine of MS patients with early renal injury (importance value of mean decrease in accuracy > 0.005). Genetic algorithm based model showed 82.6% sensitivity, 84.3% specificity, and 83.5% accuracy by a 10-fold cross-validation in identifying MS patients with early renal injury; correspondingly, the support vector machine based model reported 89.2% sensitivity, 81.1% specificity and 85.5% accuracy. Four protein peaks were included in two diagnostic models with mass-to-charge ratios of 2756.98, 3019.11, 9077.04, and 10 054.26.

CONCLUSIONS

The urinary proteome patterns of MS with early renal injury were successfully established with magnetic bead-based separation and MALDI-TOF-MS technology. A series of urinary differential expressing protein peaks were identified with bioinformatics tools. Diagnostic models combining cluster of protein peaks are capable of differentiating MS patients with early renal injury from those without renal injury. The different urine protein excretion patterns revealed in this study provide urinary candidate biomarkers of MS patients with early renal injury for future identification and biological roles investigation.