LDL-containing immune complexes in the DCCT/EDIC cohort: associations with lipoprotein subclasses

Abstract

Immune complexes containing modified LDL (LDL-IC) and NMR-determined Total LDL particle concentrations are significantly associated with intima-media thickness (IMT). We analyzed the associations between concentrations of NMR-determined lipoprotein subclasses and LDL-IC in the DCCT/EDIC cohort. LDL-IC concentrations in women and men of the DCCT/EDIC cohort did not differ significantly and were positively associated with Total LDL particle concentrations in men and women (r=0.34, r=0.32, respectively; P<.01) and with Small LDL concentration (r=0.22, r=0.13, respectively; P<.01). In women, Large LDL concentrations were also associated with LDL-IC (r=0.20, P<.01) while in men, the association was more modest (r=0.11, P<.05). Thus, both Small and Large LDL are associated with LDL-IC formation. Based on the results from statistical mediation analyses, we concluded that plasma concentrations of LDL-IC may provide a physiological link between the statistically significant association of Total LDL particle concentration with carotid artery IMT in subjects with Type 1 diabetes. Furthermore, after adjusting for conventional risk factors, there was a decrease in LDL-IC concentration even in the presence of high Total LDL particle concentrations in those women with high concentrations of Large HDL, but the association was not evident in men. This suggests that the associations between Large HDL and Total LDL particle concentrations, and their associations with LDL-IC levels, differ by gender and suggest that LDL-IC partially mediate the contribution of Total LDL particle concentration to increased carotid IMT in diabetic men.

FIGURE 1

Estimated means and standard errors of NMR-determined Small, Large and Total LDL particle concentrations in Type 1 diabetes patients grouped according to the LDL-IC concentration separated by tertile: Tertile 1 (<131), Tertile 2 (131 to 217), and Tertile 3 (>217) µg/ml. All mean LDL particle concentrations were adjusted for conventional risk factors (age, prior DCCT treatment group, concurrent HbA1c level, waist-to-hip ratio, and treatment with lipid lowering medications). Separate ANCOVA models were run for each of the three LDL particle groups (Small, Large, and Total). Statistically significant comparisons after the Tukey-Kramer adjustment (alpha=0.05) are shown. * P < 0.001; ** P < 0.01; # P < 0.05

FIGURE 2

Influence of increasing Large HDL particle concentration on the association between Total LDL particle and LDL-IC cholesterol concentrations in men and women with Type 1 diabetes. The particle concentration of Large HDL averaged <6.2µM, 6.2–9.7µM, and >9.7µM for the Low, Medium, and High tertiles, respectively. The Total LDL particle concentration averaged <966nM, 966–1289nM, and >1289nM in the Low, Medium, and High tertiles, respectively. Estimated means and standard errors for the particle concentrations for each combination of LDL and HDL were adjusted using the conventional risk factors detailed in Figure 1. Statistically significant comparisons after the Tukey-Kramer adjustment (alpha=0.05) are discussed for each panel. All LDL and HDL groupings are listed as “Total LDL Tertile/Large HDL Tertile”, so the label ‘High/Low’ refers to the cell pertaining to a high Total LDL concentration with a low Large HDL concentration. Panel A: In men, the concentration of LDL-IC in the High/Low group was greater than both the Low/Low and Low/High groups (p=0.008 and p=0.043, respectively). The concentration of LDL-IC in the High/Medium group was also greater than that in the Low/Low group (p=0.024). Panel B: In women, the concentration of LDL-IC in the Medium/Medium group was significantly greater than both the Low/Medium and Low/High groups (p=0.022 and p=0.004, respectively). Also, the concentration of LDL-IC in the High/Low group was greater than in both the Low/Medium and Low/High groups (p=0.022 and p=0.006, respectively). Increasing Total LDL particle concentration was significantly associated (p<0.001) with LDL-IC concentration in both women and men when controlling for these factors but the association of LDL-IC with Large HDL failed to reach statistical significance. No significant interaction term of Total LDL particle concentration by Large HDL particle concentration was observed; however, this term was retained in the model.

FIGURE 3

Schematic depicting the potential mechanisms whereby increased concentrations of immune complexes formed with modified LDL may mediate the association between elevated Total LDL particle concentration in plasma and increased carotid artery intimal media thickness (IMT) used as a surrogate marker of macrovascular disease. Increased plasma Total LDL particle concentration is associated with IMT in the DCCT/EDIC cohort (14). We also have shown that LDL-IC concentration is independently associated with the 5-year progression of carotid IMT (13) in the same patients. We now demonstrate that increased Total LDL particle concentration, regardless whether increased concentrations of Small LDL or Large LDL concentration contributed, was positively associated with LDL-IC concentration in the DCCT/EDIC cohort of Type 1 diabetic patients. Furthermore, using statistical mediation analyses, we now demonstrate that the increase in LDL-IC concentration may mediate this previously observed, positive association between Total LDL particle concentration and IMT. LDL-IC may mediate this pathogenic metabolism through multiple mechanisms. We have shown that LDL-IC metabolised by phagocytic cells (30) can activate the cell (29) which results in increased release of matrix metalloproteinases and pro-inflammatory cytokines which can contribute to unstable plaques and vascular inflammation, respectively (4). LDL-IC metabolism by phagocytic cells also leads to cholesteryl ester accumulation in the cell (1) which may be sufficient to result in foam cell formation. Collectively, these processes may both contribute to macrovascular disease in diabetes.