Metabolomics Defines Complex Patterns of Dyslipidaemia in Juvenile-SLE Patients Associated with Inflammation and Potential Cardiovascular Disease Risk

Figures

• 

  Figure 1. Dyslipidaemia in JSLE is associated with reduced atheroprotective small HDL lipoproteins. (A–C) Comparison of 130 different lipid metabolites (Supplementary Table S1) in healthy controls (HCs, n = 32) versus juvenile-onset SLE (JSLE, n = 67) using the balanced random forest (BRF) machine learning model as described in the methods. Demographic variables including gender, age, and race were adjusted in the model. (A) The out of bag (OOB) error rate was 29.17%. (B) Reciever operator characteristic (ROC) curve analysis and 10-fold cross validation was used to validate the model providing an area under the curve (AUC) of 0.7613 and an accuracy of 73%. (C) The top 10 variables contributing to the BRF model are shown. The mean decrease in Gini measures the importance of each variable to the model: a higher score indicates a higher importance of the variable. (D,E) Individual scatter plots (D) and ROC plots (with AUC) (E) of the top 3 metabolites from the BRF predictive model comparing HCs (n = 32) to JSLE patients (n = 65). Unpaired t test. Mean. SEM. **** = p < 0.0001. Abbreviations: VLDL, very low-density lipoprotein; HDL, high density lipoprotein; XS, very small; S, small; M, medium;…

• 

  Figure 2. Atherogenic lipoproteins are associated with increased disease activity in JSLE. (A) Forest plot showing odds ratios (ORs) and 95% confidence intervals (CIs) of serum lipid metabolites between all JSLE patients (n = 65), and those with clinically active (n = 22) and inactive (n = 43) disease compared to healthy controls (HCs, n = 32) by logistic regression analysis adjusted for age, sex, and race. The concentration of total lipids (mmol/L), apolipoproteins (g/L), and lipoprotein measurements including particle size (nm), concentration, and lipid content (mmol/L) are displayed. Statistically significant differences denoted by solid black diamond; non-statistically significant differences denoted by open diamond. Abbreviations: Apo, apolipoprotein; VLDL, very low-density lipoprotein; IDL, intermediate density lipoprotein; LDL, low density lipoprotein; HDL, high density lipoprotein. HC vs. all JSLE, black line; HC vs. active JSLE, red line; HC vs. inactive JSLE, blue line.

• 

  Figure 3. Atherogenic lipoproteins correlate positively with markers of JSLE disease activity and are associated with markers of liver damage. (A) Venn diagram (http://www.biovenn.nl/, Accessed on 20 June 2020) displaying the proportional overlap of statistically significantly altered metabolites from (A) that overlap between the different group comparisons. Metabolites that are both top variables in the BRF model and are significantly altered by logistic regression analysis are underlined. Abbreviations: Apo, apolipoprotein; VLDL, very low-density lipoprotein; LDL, low density lipoprotein; HDL, high density lipoprotein; C, cholesterol; CE, cholesterol ester; D, diameter; FA, fatty acid; FC, free cholesterol; L, total lipids; P, particles; PL, phospholipids; FAw6, omega-6 fatty acid; LA, linoleic acid. Volcano plot showing (B) the correlation (r-value, Pearson’s) of systemic lupus erythematosus disease activity index (SLEDAI) with serum lipid metabolomic measures in JSLE patients (n = 65) and (C) the fold change of serum lipids in JSLE patients experiencing a clinical flare (previously inactive patients with an increase in SLEDAI by 4 points or more [21], mean increase of 7.2 points))…

• 

  Figure 4. Lymphocyte inflammatory lipid signalling platforms are associated with disease activity and circulating lipoproteins in JSLE. (A) Heat map displaying r values (Pearson’s) of the correlation between lipoprotein particles (VLDL, very low-density lipoprotein; IDL, intermediate density lipoprotein; LDL, low density lipoprotein; HDL, high density lipoprotein), including apolipoproteins (ApoB and ApoA1), and serum lipids associated with lipid rafts: SM (sphingomyelin), TotCho (total cholesterol), TotPG (total phosphoglycerides), and PC (phosphatidylcholine and other cholines). (B) Volcano plot displaying the correlation (r-value, Pearson’s) of CD4+ and CD8+ T-cell, CD19 B-cell, and CD14 monocyte lipid raft expression (cholera toxin B, CTB; mean fluorescence intensity, MFI, measured ex vivo by flow cytometry) with clinical measures of disease activity (systemic lupus erythematosus disease activity index; SLEDAI, erythrocyte sedimentation rate; ESR, dsDNA; anti-double-stranded DNA antibodies, C3; complement protein 3) in JSLE patients (n = 35). Log10 p values displayed. (C) Volcano plot showing the correlation (r-value, Pearson’s) of CD4+, CD8+ T-cell, CD19 B-cell, and CD14 monocyte…

• 

  Figure 5. Association between selected serum lipid metabolites, clinical measure of disease activity and lipid rafts in JSLE. Correlations between selected lipid metabolites, clinical factors, and lipid raft. Pearson correlation coefficients are represented as edges connecting the lipid metabolites section (green/pink/brown), clinical factor (grey) section and lipid raft section (blue). The width of the edges represents the size of the correlation. Only statistically significant correlations following adjustment for multiple comparisons (10% FDR) are shown. Red lines represent positive correlation and blue lines represent negative correlation. Abbreviations: VLDL, very low-density lipoprotein; HDL, high density lipoprotein; XS, very small; S, small; M, medium; C, cholesterol; CE, cholesterol ester; FC, free cholesterol; L, total lipids; P, particles; PL, phospholipids; LA, linoleic acid; BMI, Body Mass Index; SLEDAI, Systemic Lupus Erythematosus Disease Activity Index; dsDNA, anti-double-stranded-DNA antibodies; C3, complement component 3; LC, lymphocyte count.