iTRAQ-based proteomic analysis of plasma reveals abnormalities in lipid metabolism proteins in chronic kidney disease-related atherosclerosis

Abstract

Patients with chronic kidney disease (CKD) have a considerably higher risk of death due to cardiovascular causes. Using an iTRAQ MS/MS approach, we investigated the alterations in plasma protein accumulation in patients with CKD and classical cardiovascular disease (CVD) without CKD. The proteomic analysis led to the identification of 130 differentially expressed proteins among CVD and CKD patients and healthy volunteers. Bioinformatics analysis revealed that 29 differentially expressed proteins were involved in lipid metabolism and atherosclerosis, 20 of which were apolipoproteins and constituents of high-density lipoprotein (HDL) and low-density lipoprotein (LDL). Although dyslipidemia is common in CKD patients, we found that significant changes in apolipoproteins were not strictly associated with changes in plasma lipid levels. A lack of correlation between apoB and LDL concentration and an inverse relationship of some proteins with the HDL level were revealed. An increased level of apolipoprotein AIV, adiponectin, or apolipoprotein C, despite their anti-atherogenic properties, was not associated with a decrease in cardiovascular event risk in CKD patients. The presence of the distinctive pattern of apolipoproteins demonstrated in this study may suggest that lipid abnormalities in CKD are characterized by more qualitative abnormalities and may be related to HDL function rather than HDL deficiency.

Figure 1. Workflow of the experimental strategy used in this study.

Plasma samples from six experimental groups were trypsin digested, labeled with isobaric tags, pooled and then purified and fractionated using SCX method. Quantitative proteomic analyses were simultaneously performed using ESI-nanoLC-MS/MS and MALDI-nanoLC-MS/MS and then obtained data were analyzed with three types of software: MaxQuant, ProteinScape and Proteome Discoverer. Only proteins identified by all software were found to have a differential accumulation level.

Figure 2. A Venn diagram comparing the results from the MaxQuant (MQ), Proteome Discoverer (PD), and ProteinScape (PS) software in ten iTRAQ experiments.

(A) A total of 1,038 unique proteins were identified, 229 of which were common between the approaches. (B) A total of 221 differentially expressed proteins were identified in all the experiments, 130 of which were common between the approaches.

Figure 3. Representative correlation plots of 115 and 121 reporter ion intensities from two technical (A) and two biological (B) experiments.

The Pearson correlation coefficient is provided for each plot.

Figure 4. PCA of the reporter ion intensities obtained from the plasma of HVs (blue), CKD1-2 (yellow), CKD3-4 (green), CKD5 (red), CVDI (black) and CVDII (pink) patients.

Calculations were performed with Perseus.

Figure 5

(A) Relative abundance of apoAIV in HVs, CKD1-2, CKD3-4, CKD5, CVDI and CVDII groups based on reporter ion intensities. (B) Relative abundance of apoAIV in experimental groups based on PRM analysis showing transition 480.60 m/z to 855.42 m/z. (C) ELISA measurements of apoAIV. Charts show mean and SD for all analyzed plasma samples. Anova and Student’s t-tests were completed and statistical significance is indicated (*P < 0.05, NS-non-significant).