LDL binding to cell receptors and extracellular matrix is proatherogenic in obesity but improves after bariatric surgery

Abstract

Obesity is a major global public health issue involving dyslipidemia, oxidative stress, inflammation, and increased risk of CVD. Weight loss reduces this risk, but the biochemical underpinnings are unclear. We explored how obesity and weight loss after bariatric surgery influence LDL interactions that trigger proatherogenic versus antiatherogenic processes. LDL was isolated from plasma of six patients with severe obesity before (basal) and 6-12 months after bariatric surgery (basal BMI = 42.7 kg/m²; 6-months and 12-months postoperative BMI = 34.1 and 30 kg/m²). Control LDL were from six healthy subjects (BMI = 22.6 kg/m²). LDL binding was quantified by ELISA; LDL size and charge were assessed by chromatography; LDL biochemical composition was determined. Compared to controls, basal LDL showed decreased nonatherogenic binding to LDL receptor, which improved postoperatively. Conversely, basal LDL showed increased binding to scavenger receptors LOX1 and CD36 and to glycosaminoglycans, fibronectin and collagen, which is proatherogenic. One year postoperatively, this binding decreased but remained elevated, consistent with elevated lipid peroxidation. Serum amyloid A and nonesterified fatty acids were elevated in basal and postoperative LDL, indicating obesity-associated inflammation. Aggregated and electronegative LDL remained elevated, suggesting proatherogenic processes. These results suggest that obesity-induced inflammation contributes to harmful LDL alterations that probably increase the risk of CVD. We conclude that in obesity, LDL interactions with cell receptors and extracellular matrix shift in a proatherogenic manner but are partially reversed upon postoperative weight loss. These results help explain why the risk of CVD increases in obesity but decreases upon weight loss.

Keywords: CD36 and LOX-1; Inflammation in obesity; LDL binding to LDL receptor; LDL retention by matrix proteins and glycosaminoglycans; Lipoprotein lipolysis; oxidation and aggregation.

Fig. 1

Inflammatory and proatherogenic markers in LDL from patients with obesity (basal, 6 months and one year after surgery) versus normolipidemic healthy controls. A: NEFA and (B) TBARS were measured using standard assays; (C) SAA and (D) Lp(a) were measured using ELISA kits. Each data point represents an average of five independent measurements. The bars show median ±SD; ns, not significant, ∗P < 0.05, ∗∗P < 0.01, and ∗∗∗P < 0.001. SAA, serum amyloid A; TBARS, thiobarbituric acid reactive substance.

Fig. 2

LDL binding to cell receptors measured by ELISA at pH 7.4. A–C: Validation assays for binding of control LDL to cell receptors; (A) LDLR, (B) CD36; (C) LOX-1. The receptors were immobilized on the wells and LDL binding was measured by ELISA using increasing concentration of control LDL that were either intact (for binding to LDLR) or have been oxidized by copper as described in Materials and methods (for binding to CD36 and LOX-1). The absorbance values were corrected for nonspecific binding using BSA-coated wells as a control. The binding was measured using increasing LDL concentrations at pH 7.4. D–F: Measurements of LDL binding to cell receptors; (D) LDLR, (E) CD36, (F) LOX-1. LDL protein concentration was 0.5 μg for LDLR, 5 μg for CD36, or 10 μg for LOX-1. Each data point represents an average of five independent measurements. The bars show median ±SD; ns, not significant, ∗P < 0.05, ∗∗P < 0.01, and ∗∗∗P < 0.001. LDLR, LDL receptor.

Fig. 3

LDL binding to heparin explored using heparin affinity chromatography. Dashed line indicates NaCl gradient. Normolipidemic LDL binds heparin and elutes at 30% NaCl (control, vertical line); unbound LDL elutes at 0% (0 M) NaCl; 100% corresponds to 1 M NaCl.

Fig. 4

LDL binding to matrix components measured by ELISA at pH 7.4. LDL (10 μg protein) was added to the matrix immobilized on the well plate. A: heparin, (B) heparan sulfate, (C) chondroitin sulfate, (D) hyaluronic acid, (E) biglycan, (F) perlecan, (G) fibronectin, and (H) collagen-IV. Well coating with sulfated GAGs was verified by colorimetric assay as described in Materials and methods. The assay was validated using control LDL (supplemental Fig. S2). The absorbance values were corrected for nonspecific binding using BSA-coated wells as a control. Each data point represents an average of five independent measurements. The bars show median ±SD; ns, not significant, ∗P < 0.05, ∗∗P < 0.01, and ∗∗∗P < 0.001. GAG, glycosaminoglycan.

Fig. 5

Size-exclusion chromatography profiles show particle size distribution in LDL. Profiles for single-donor LDL from six patients with severe obesity and six healthy controls are marked 1–6. Profiles of normolipidemic human HDL, LDL, and VLDL are shown for reference (top left).

Fig. 6

Effects of LDL oxidation and lipolysis in vitro on LDL binding to matrix components measured by ELISA at pH 7.4. Well coating with sulfated GAGs was verified by a colorimetric assay as described in Materials and methods. The absorbance values were corrected for nonspecific binding using BSA-coated wells as a control. Control LDL from lean single donors were modified in vitro by copper to generate oxidized LDL (ox.LDL) or by secretory PLA₂ to generate PLA₂-treated LDL (pla2.LDL) as described above in supplemental Methods. Lipolysis of PLA₂-treated LDL was monitored by NEFA levels as shown in supplemental Fig. S6. Next, modified LDL (10 μg protein) were added to the matrix component immobilized on the well plate, and the binding was measured to (A) heparin, (B) heparan sulfate, (C) chondroitin sulfate, (D) hyaluronic acid, (E) biglycan, (F) perlecan, (G) fibronectin, and (H) collagen type-IV. Nonmodified intact LDL (int.LDL) are shown as a control. Each data point represents an average of five independent measurements. The bars show median ±SD; ns, not significant, ∗P < 0.05, ∗∗P < 0.01, and ∗∗∗P<0.001. GAG, glycosaminoglycan; PLA₂, phospholipase A₂.