Serum amyloid A facilitates the binding of high-density lipoprotein from mice injected with lipopolysaccharide to vascular proteoglycans

Abstract

Objective: Levels of serum amyloid A (SAA), an acute-phase protein carried on high-density lipoprotein (HDL), increase in inflammatory states and are associated with increased risk of cardiovascular disease. HDL colocalizes with vascular proteoglycans in atherosclerotic lesions. However, its major apolipoprotein, apolipoprotein A-I, has no proteoglycan-binding domains. Therefore, we investigated whether SAA, which has proteoglycan-binding domains, plays a role in HDL retention by proteoglycans.

Methods and results: HDL from control mice and mice deficient in both SAA1.1 and SAA2.1 (SAA knockout mice) injected with bacterial lipopolysaccharide (LPS) was studied. SAA mRNA expression in the liver and plasma levels of SAA increased dramatically in C57BL/6 mice after LPS administration, although HDL cholesterol did not change. Fast protein liquid chromatography analysis showed most of the SAA to be in HDL. Mass spectrometric analysis indicated that HDL from LPS-injected control mice had high levels of SAA1.1/2.1 and reduced levels of apolipoprotein A-I. HDL from LPS-injected control mice demonstrated high-affinity binding to biglycan relative to normal mouse HDL. In contrast, HDL from LPS-injected SAA knockout mice showed very little binding to biglycan, consistent with SAA facilitating the binding of HDL to vascular proteoglycans.

Conclusion: SAA enrichment of HDL under inflammatory conditions plays an important role in the binding of HDL to vascular proteoglycans.

Conflict of interest statement

Disclosures: There are no real or apparent conflicts of interest.

Figures

Figure 1. Effect of LPS-injection on plasma SAA and lipid levels

Plasma from control or LPS-injected mice were evaluated for SAA by ELISA(A). Plasma levels of cholesterol (B), HDL-cholesterol (C), and triglyceride (D) were measured enzymatically. Results represent the mean ± SEM; n=12, Control; n=11, LPS. *P<0.05, ***P<0.001 vs. Control.

Figure 2. MALDI-TOF analysis

HDL (d=1.063–1.210 g/mL) was isolated by ultracentrifugation from plasma of control and LPS-injected mice. HDL (20 μg protein) was separated by SDS-PAGE (10 – 20 % gradient gel) and the gel was stained with Coomassie Brilliant Blue. Each gel band corresponding to the apparent molecular weight of SAA1.1/2.1, SAA4, apoA-I and apoE was cut out, digested with trypsin, and the peptide digest was extracted for tandem mass spectrometric analysis by MALDI-TOF. The arrows indicate bands that were identified by MALDI-TOF and database searching that contained peptides unique to SAA1.1/2.1, SAA4, apoA-I, and apoE (Figure 2A). Albumin-depleted plasma samples (20 μl) from control (A) and LPS-injected (B) mice were separated by two dimensional electrophoresis (first dimension: IEF pH 3–10, second dimension: 10% SDS-PAGE) and the gel was stained with a silver stain. Selected spots from 2D gels were identified by in-gel tryptic digest and MALDI-TOF analysis. The small arrows indicate bands that were identified by tandem MS/MS MALDI-TOF and database searching that contained peptides unique to SAA1.1, SAA2.1, and apoA-I (Figure 2B). The spot designated as SAA2.1 was identified as such with a MASCOT MOWSE score of 374 (CI 100% - Figure 2C), and the adjacent spot was identified as SAA1.1 with a MOWSE score of 403 (CI 100% - Figure 2D). The asterisks indicate peaks corresponding to SAA peptides.

Figure 3. HDL from LPS-injected mice has high affinity for biglycan

(A) HDL-biglycan binding was evaluated by gel mobility shift assay at physiological pH. A constant amount of biglycan was incubated with no lipoproteins (0), increasing concentrations of human LDL (0.025, 0.05, 0.25 mg/mL), human HDL (3 mg/mL) or HDL from control and LPS-treated mice (3 mg/mL) and electrophoresed. Figure shown is representative of three independent experiments. (B) HDL-biglycan binding was quantified as % Bound using Opti-Quant software(Packard). Representative binding curves are shown.

Figure 4. HDL from LPS-injected SAA knockout mice does not bind biglycan

(A) HDL from wild-type and SAAKO mice treated with LPS were evaluated for SAA by Western immunoblot analysis. ApoE was removed using an antibody bound to magnetic beads, and resulting HDL lacking apoE (HDL-E) preparations were evaluated for apoE content by Western immunoblot analysis. (B) Binding of apoE-free HDL from LPS-injected wild-type and SAAKO mice to biglycan was evaluated at acidic pH. A constant amount of biglycan was incubated with no lipoproteins (0), human LDL (0.1 mg/mL), or increasing concentrations of HDL (0–1 mg/mL) from LPS-treated wild type mice, or from LPS-treated SAAKO mice and electrophoresed. (C) HDL-biglycan binding was quantified as % Bound using Opti-Quant software(Packard). Representative binding curves are shown.