LDL-c/HDL-c Ratio and NADPH-Oxidase-2-Derived Oxidative Stress as Main Determinants of Microvascular Endothelial Function in Morbidly Obese Subjects

Abstract

The identification of obese subjects at higher risk for cardiovascular disease (CVD) is required. We aimed to characterize determinants of endothelial dysfunction, the initial step to CVD, in small omental arteries of visceral fat from obese subjects. The influences of analytical parameters and vascular oxidative stress mediated by NADPH-oxidase-2 (NOX2) on endothelial function were determined. Specimens were obtained from 51 obese subjects undergoing bariatric surgery and 14 non-obese subjects undergoing abdominal surgery. Obese subjects displayed reduced endothelial vasodilation to bradykinin (BK). Endothelial vasodilation (pEC₅₀ for BK) among obese subjects was significantly and negatively associated with low-density lipoprotein cholesterol (LDL-c)/high-density lipoprotein cholesterol (HDL-c) ratio (r = -0.510, p = 0.0001) in both women and men, while other metabolic parameters and comorbidities failed to predict endothelial function. The vascular expression of NOX2 was upregulated in obese subjects and was related to decreased endothelial vasodilation (r = -0.529, p = 0.0006, n = 38) and increased oxidative stress (r = 0.783, p = 0.0044, n = 11) in arterial segments. High LDL-c/HDL-c (>2) and high NOX2 (above median) were independently associated with reduced endothelial function, but the presence of both conditions was related to a further impairment. Concomitant elevated LDL-c/HDL-c ratio and high vascular expression of NOX2 would exacerbate endothelial impairment in obesity and could reveal a deleterious profile for cardiovascular outcomes among obese subjects.

Keywords: NADPH-oxidase; endothelial dysfunction; high-density lipoprotein cholesterol; human mesenteric small arteries; human obesity; low-density lipoprotein cholesterol.

Figure 1

Obesity is related to the impairment of endothelium-dependent vasodilation through endothelial hyperpolarization and nitric oxide-mediated relaxation in human mesenteric small arteries. Endothelium-dependent vasodilations induced by bradykinin (BK, 1 nM to 10 µM) in human mesenteric arteries (HMAs) contracted with the thromboxane analogue, U46619 (10–30 nM) (A,B), or 25–35 mM K⁺ © obtained from non-obese subjects (control, BMI < 30) and from subjects with morbid obesity (obese, BMI > 35). Panel (A) shows BK-induced responses in HMAs without treatment, while HMAs were treated with indomethacin (INDO, 10 µM) plus N^G-nitro-L-arginine methyl ester (L-NAME, 100 µM) in panel (B) and with INDO (10 µM) in panel (C). Data are expressed as the mean ± S.E.M. of the percentage of relaxation. n indicates the number of patients from whom the vessels were collected. *** indicates p < 0.001 vs. control by a two-factor ANOVA test.

Figure 2

Low-density lipoprotein cholesterol (LDL-c)/high-density lipoprotein cholesterol (HDL-c) ratio determines endothelial vasodilation in obese subjects. Upper panels show linear regressions of LDL-c/HDL-c ratio versus pEC₅₀ for bradykinin (BK) in subjects with morbid obesity (A–C). pEC₅₀ values for BK for each subject were obtained in human mesenteric arteries (HMAs) contracted with the thromboxane analogue, U46619 (10–30 nM) (A,B), or 25–35 mM K⁺ (C). Lower panels show complete BK-induced vasodilations in HMAs contracted with the thromboxane analogue, U46619 (10–30 nM) (D,E), or 25–35 mM K⁺ (F) obtained from obese subjects displaying LDL-c/HDL-c ratios equal to or below 2 (LDL-c/HDL-c ≤ 2) versus those with LDL-c/HDL-c above 2 (LDL-c/HDL-c > 2). Data in panels (A,D) were obtained in HMAs without treatment, while HMAs were treated with indomethacin (INDO, 10 µM) plus N^G-nitro-L-arginine methyl ester (L-NAME, 100 µM) in panels (B,E) and with INDO (10 µM) in panels (C,F). n indicates the number of patients from whom the determinations were obtained. Coefficients of determination and probability (p) values are indicated in upper panels. Significant associations are highlighted in bold plus italic. Data in lower panels are expressed as the mean ± S.E.M. of the percentage of relaxation. * indicates p < 0.05; *** p < 0.001 vs. control by a two-factor ANOVA test.

Figure 3

A negative association of low-density lipoprotein cholesterol (LDL-c)/high-density lipoprotein cholesterol (HDL-c) ratio with endothelial vasodilation is observed in both female and male obese subjects. Upper panels show vasodilations induced by bradykinin (BK, 1 nM to 10 µM) in HMAs contracted with the thromboxane analogue, U46619 (10–30 nM) (A,B), or 25–35 mM K+ (C) obtained from female (women) and male (men) obese subjects. Data in panel (A) were obtained in HMAs without treatment, while HMAs were treated with indomethacin (INDO, 10 µM) plus N^G-nitro-L-arginine methyl ester (L-NAME, 100 µM) in panel (B) and with INDO (10 µM) in panel (C). Data are expressed as the mean ± S.E.M. of the percentage of relaxation. Lower panels show linear regressions of LDL-c/HDL-c ratio versus pEC₅₀ for BK in female (D) and male (E) subjects with morbid obesity. pEC₅₀ values for BK for each subject were obtained in HMAs contracted with the thromboxane analogue, U46619. Coefficients of determination and probability (p) values are indicated. Significant associations are highlighted in bold plus italic.

Figure 4

The vascular expression of NADPH-oxidase-2 (NOX2) is related to increased oxidative stress and reduced endothelial vasodilation in obese subjects. Panels (A,B) show representative immunofluorescence detection of NOX2 protein in sections of small mesenteric arteries from a control and an obese subject, respectively. Magnification ×200. High immunoreactivity (green fluorescence) is detected in the arterial wall of the obese subject. Nuclei are stained with DAPI (blue). Panel (C) shows representative immunoblots for the detection of NOX2 and corresponding β-actin in mesenteric artery homogenates from non-obese (control, C) and obese (Ob) subjects as well as the quantification of the expression assays. Data are expressed as the mean ± S.E.M. of NOX2 band intensities normalized by respective β-actin band intensities. * indicates p < 0.05 vs. control by the Mann–Whitney U test. Panels (D–F) show linear regressions of NOX2 expression (NOX2/β-actin ratio) versus pEC₅₀ (D,E) or E_max for BK (F) in subjects with morbid obesity. BK-induced responses for each subject were obtained in HMAs contracted with the thromboxane analogue, U46619 (10–30 nM) (D,E), or 25–35 mM K⁺ (F). Data in panel (D) were obtained in HMAs without treatment, while HMAs were treated with indomethacin (INDO, 10 µM) plus N^G-nitro-L-arginine methyl ester (L-NAME, 100 µM) in panel (E) and with INDO (10 µM) in panel (F). Panels (G,H) are representative images showing the detection of dihydroethidium (DHE) fluorescence (red) and DAPI staining (blue) in HMA sections from an obese subject with low (G) and high (H) DHE/DAPI fluorescence ratios. Magnification ×400. Panel (I) shows the linear regression of NOX2 expression in HMA homogenates versus superoxide generation determined by DHE/DAPI fluorescence in HMA sections from patients with morbid obesity, while panel (J) shows the linear regression of DHE/DAPI fluorescence versus pEC₅₀ for BK in subjects with morbid obesity. pEC₅₀ values for BK for each subject were obtained in HMAs contracted with U46619 and without further treatment. Coefficients of determination (r²) and probability (p) values are indicated in panels (D–F,I,J). Significant associations are highlighted in bold plus italic. n always indicates the number of subjects from whom the tissues were collected for determinations.

Figure 5

The vascular expression of NADPH-oxidase-4 (NOX4) is not related to endothelial vasodilation in obese subjects. Panel (A) shows representative immunoblots for the detection of NOX4 and corresponding β-actin in mesenteric artery homogenates from non-obese (control, C) and obese (Ob) subjects as well as the quantification of the expression assays. Data are expressed as the mean ± S.E.M. of NOX4 band intensities normalized by respective β-actin band intensities. p by the Mann–Whitney U test is indicated. Panels (B–D) show linear regressions of NOX4 expression (NOX4/β-actin ratio) versus pEC₅₀ (B,C) or E_max for BK (D) in subjects with morbid obesity. BK-induced responses for each subject were obtained in HMAs contracted with the thromboxane analogue, U46619 (10–30 nM) (B,C), or 25–35 mM K⁺ (D). Data in panel (B) were obtained in HMAs without treatment, while HMAs were treated with indomethacin (INDO, 10 µM) plus N^G-nitro-L-arginine methyl ester (L-NAME, 100 µM) in panel (C) and with INDO (10 µM) in panel (D). Coefficients of determination (r²) and probability (p) values are indicated in panels (B–D). n always indicates the number of subjects from whom the tissues were collected for determinations.

Figure 6

The endothelial vasodilation of human mesenteric arteries (HMAs) from obese subjects is additively impaired by high LDL-c/HDL-c ratio and high vascular expression of NADPH-oxidase-2 (NOX2). Panel (A) shows vasodilations induced by bradykinin (BK, 1 nM to 10 µM) in HMAs contracted with the thromboxane analogue, U46619 (10–30 nM), obtained from obese subjects with low LDL-c/HDL-c ratio (≤2) and low vascular expression of NOX2 (below median) (low/low), from obese subjects with high LDL-c/HDL-c ratio (above 2) but low NOX2 expression (high LDL-c/HDL-c), from obese subjects with high vascular expression of NOX2 (above median) but low LDL-c/HDL-c ratio (high NOX2) and from obese subjects with both high LDL-c/HDL-c and NOX2 expression (high/high). Data are expressed as the mean ± S.E.M. of the percentage of relaxation. n indicates the number of subjects from whom the tissues were collected for determinations. ** indicates p < 0.01, *** p < 0.001 vs. low/low, ††† p < 0.001 vs. high/high by two-factor ANOVA test corrected by Bonferroni’s test. Right panels show mean ± S.E.M. of pEC₅₀ (B) and E_max (C) values for BK corresponding to each subject group. § p < 0.05, §§ p < 0.01 vs. low/low by Kruskal–Wallis followed by Dunn’s test.