Mass Spectrometry-Based Lipidomics Reveals Differential Changes in the Accumulated Lipid Classes in Chronic Kidney Disease

Abstract

Chronic kidney disease (CKD) is characterized by the progressive loss of functional nephrons. Although cardiovascular disease (CVD) complications and atherosclerosis are the leading causes of morbidity and mortality in CKD, the mechanism by which the progression of CVD accelerates remains unclear. To reveal the molecular mechanisms associated with atherosclerosis linked to CKD, we applied a shotgun lipidomics approach fortified with standard laboratory analytical methods and gas chromatography-mass spectrometry technique on selected lipid components and precursors to analyze the plasma lipidome in CKD and classical CVD patients. The MS-based lipidome profiling revealed the upregulation of triacylglycerols in CKD and downregulation of cholesterol/cholesteryl esters, sphingomyelins, phosphatidylcholines, phosphatidylethanolamines and ceramides as compared to CVD group and controls. We have further observed a decreased abundance of seven fatty acids in CKD with strong inter-correlation. In contrast, the level of glycerol was elevated in CKD in comparison to all analyzed groups. Our results revealed the putative existence of a functional causative link-the low cholesterol level correlated with lower estimated glomerular filtration rate and kidney dysfunction that supports the postulated "reverse epidemiology" theory and suggest that the lipidomic background of atherosclerosis-related to CKD is unique and might be associated with other cellular factors, i.e., inflammation.

Keywords: cardiovascular disease; chronic kidney disease; lipid profiling; lipidomics; mass spectrometry.

Figure 1

Unsupervised principal component analysis (PCA) performed on the total number of lipids identified in (A) stage 5 chronic kidney disease (CKD5) (red), cardiovascular disease (CVD) (blue), stage 1 and 2 chronic kidney disease (CKD1-2) (black), and healthy volunteers (HVs) (green) experimental groups. (B) To highlight the observed differences, only CKD5 (red) and CVD (blue) are marked. Other groups are presented in grey. (C) Hierarchical clustering analysis of identified lipid classes. PC, PC-O- phosphatidylcholines; PE, PE-O—phosphatidylethanolamines; PI—phosphatidylinositols; LPC—lysophosphatidylcholines; LPE—lysophosphatidylethanolamines; PS—phosphatidylserines; SM—sphingomyelins; Cer—ceramides; HexCer—hexosylceramides; TAG—triacylglycerols; DAG—diacylglycerols; CE—cholesterol and cholesteryl esters.

Figure 2

Differentially accumulated lipid classes sorted according to Anova/Kruskal–Wallis p-values: (A) triacyloglycerols (TAG). (B) cholesterol and cholesterol esters (CE). (C) lysophosphatidylcholines (LPE). (D) phosphatidylcholines (PC). (E) ceramides (CER). (F) phosphatidylethanolamines (PE-O). (G) sphingomyelins (SM). (H) phosphatidylserines (PS). (I) lysophosphatidylcholines (LPC). A box on a plot presents interquartile ranges and medians. The black dots correspond to the normalized accumulation of lipid classes in all samples. The notch indicates the 95% confidence interval around the median of each group. The mean level is indicated with a yellow diamond. Bars and asterisks show the results of U-Mann–Whitney/t-test: * p < 0.05, ** p < 0.01, *** p < 0.001; n = 25 in each group.

Figure 3

Results of the Pearson’s correlation analysis performed for all identified lipid classes and presented as a heat map with the corresponding hierarchical tree. Color saturation corresponds to the value of Pearson’s correlation coefficient.

Figure 4

Abundances of representative species for all classes of lipids. (A) TAG [52:2]. (B) Chol [20:3]. (C) LPE [20:4]. (D) PC [38:7]. (E) CER [36:3:0, OH]. (F) PE-O [38:4]. (G) SM [39:1]. (H) PS [36:4]. (I) LPC [16:0]. A box on plot presents the interquartile ranges and medians. The black dots correspond to the normalized accumulation of selected lipid species in all samples. The notch indicates the 95% confidence interval around the median for each group. The mean level is indicated with a yellow diamond. Bars and asterisks show the results of U-Mann–Whitney/t-test: * p < 0.05, ** p < 0.01, *** p < 0.001; n = 25 in each group.

Figure 5

LION-term enrichment analysis of CKD1-2 vs. CVD, CKD5 vs. HV, CKD5 vs. CVD and CKD1-2 vs. HVs presented in the “ranking mode”. The gray vertical lines indicate the threshold of –log10 False discovery rate (FDR)-corrected p-value 0.05 (q-value 1.3). Red bars present terms with FDR-corrected p-value < 0.01 (q > 2).

Figure 6

Network visualization of the most dysregulated lipid species found in CKD5 vs. CVD (left panel) and CKD5 vs. HV comparisons. Graphs express lipidomic pathways connections with clustering into individual lipid classes prepared using the Cytoscape software. Circles correspond to detected lipid species, the circle size articulates the significance according to Mann Whitney U test p-value, while the color gradient explains the degree of variations (red/blue) according to the calculated fold change.

Figure 7

The abundance of differentially accumulated fatty acids, cholesterol and glycerol revealed in GC-MS analysis. (A) Palmitoleic acid. (B) Palmitic acid. (C) Oleic acid. (D) Linoleic acid. (E) Arachidonic acid. (F) Stearic acid. (G) 2-hydroxybutyric acid. (H) Cholesterol. (I) Glycerol. A box on plot presents interquartile ranges and medians. The black dots represent the normalized accumulation of molecules in all samples. The notch indicates the 95% confidence interval around the median of each group. The mean level is indicated with a yellow diamond. Bars and asterisks show results of U-Mann–Whitney/t-test: * p < 0.05, ** p < 0.01, *** p < 0.001; n = 14 in each group.

Figure 8

Results of the Spearman’s rank correlation test performed for compounds identified in GC/MS analyses as differentially accumulated between analyzed experimental groups and estimated glomerular filtration rate (eGFR) and C-reactive protein (CRP) level presented as a heat map with the corresponding hierarchical tree. Color saturation corresponds to the value of Spearman’s correlation coefficient. Gly—glycerol, Chol—cholesterol, AA—arachidonic acid, SA- stearic acid, HA—hydroxybutyric acid, PA—palmitic acid, PeA—palmitoleic acid, OA—oleic acid, LA—linoleic acid.