Divergent effects of genetic and pharmacological inhibition of Nox2 NADPH oxidase on insulin resistance-related vascular damage

Abstract

Insulin resistance leads to excessive endothelial cell (EC) superoxide generation and accelerated atherosclerosis. The principal source of superoxide from the insulin-resistant endothelium is the Nox2 isoform of NADPH oxidase. Here we examine the therapeutic potential of Nox2 inhibition on superoxide generation in saphenous vein ECs (SVECs) from patients with advanced atherosclerosis and type 2 diabetes and on vascular function, vascular damage, and lipid deposition in apolipoprotein E-deficient (ApoE^-/-) mice with EC-specific insulin resistance (ESMIRO). To examine the effect of genetic inhibition of Nox2, ESMIRO mice deficient in ApoE^-/- and Nox2 (ESMIRO/ApoE^-/-/Nox2^-/y) were generated and compared with ESMIRO/ApoE^-/-/Nox2^+/y littermates. To examine the effect of pharmacological inhibition of Nox2, we administered gp91dstat or scrambled peptide to ESMIRO/ApoE^-/- mice. SVECs from diabetic patients had increased expression of Nox2 protein with concomitant increase in superoxide generation, which could be reduced by the Nox2 inhibitor gp91dstat. After 12 wk Western diet, ESMIRO/ApoE^-/-/Nox2^-/y mice had reduced EC superoxide generation and greater aortic relaxation to acetylcholine. ESMIRO/ApoE^-/-/Nox2^-/y mice developed more lipid deposition in the thoraco-abdominal aorta with multiple foci of elastin fragmentation at the level of the aortic sinus and greater expression of intercellular adhesion molecule-1 (ICAM-1). Gp91dstat reduced EC superoxide and lipid deposition in the thoraco-abdominal aorta of ESMIRO/ApoE^-/- mice without causing elastin fragmentation or increased ICAM-1 expression. These results demonstrate that insulin resistance is characterized by increased Nox2-derived vascular superoxide. Complete deletion of Nox2 in mice with EC insulin resistance exacerbates, whereas partial pharmacological Nox2 inhibition protects against, insulin resistance-induced vascular damage.

Keywords: Nox2; atherosclerosis; insulin resistance.

Fig. 1.

Saphenous vein endothelial cells (SVECs) from diabetic patients have increased expression of Nox2 and superoxide generation. A: increased NADPH-dependent superoxide generation in SVECs from patients with type 2 diabetes mellitus (No Diabetes n = 5; Diabetes, n = 3). B: increased expression of Nox2 protein in SVECs from patients with type 2 diabetes mellitus (No Diabetes n = 12; Diabetes, n = 7). C: the increased superoxide generated by SVECs from patients with type 2 diabetes is reduced by the specific Nox2 inhibitor gp91dstat. (No Diabetes n = 5, Diabetes n = 5). AU, arbitrary units; DM, diabetes mellitus. Data are expressed as mean ± SE; n = number of mice per genotype. *P < 0.05, **P < 0.01, ****P < 0.0001.

Fig. 2.

ESMIRO/ApoE^−/−/Nox2^−/y have reduced endothelial cell (EC) superoxide generation and greater aortic relaxation to acetylcholine. A: representative image showing Nox2 mRNA was undetectable in endothelial cells from ESMIRO/ApoE^−/−/Nox2^−/y mice [ESMIRO/ApoE^−/−/Nox2^−/y (n = 3) and ESMIRO/ApoE^−/−/Nox2^+/y mice (n = 4)]. B: reduced NADPH-dependent superoxide generation in endothelial cells from ESMIRO/ApoE^−/−/Nox2^−/y (n = 5) compared with ESMIRO/ApoE^−/−/Nox2^+/y mice (n = 5). C: in response to the endothelium-dependent vasorelaxant acetylcholine (Ach), aortic rings from ESMIRO/ApoE^−/−/Nox2^−/y mice (n = 15) had greater maximal relaxation than ESMIRO/ApoE^−/−/Nox2^+/y mice (n = 11). Both groups were fed a Western diet for 12 wk before the experiment. D: acetylcholine EC₅₀ values derived from C. E: no difference in response to the endothelium independent vasorelaxant sodium nitroprusside (SNP) in aortic rings from ESMIRO/ApoE^−/−/Nox2^−/y (n = 7) compared with ESMIRO/ApoE^−/−/Nox2^+/y mice (n = 5). Both groups were fed a Western diet for 12 wk before the experiment. ESMIRO, endothelium-specific mutant insulin receptor-overexpressing mice; ApoE, apolipoprotein E; Nox2, Nox2 isoform of NADPH oxidase; AU, arbitrary units. Data are expressed as mean ± SE; n = number of mice per genotype. *P < 0.05, **P < 0.01.

Fig. 3.

ESMIRO/ApoE^−/−/Nox2^−/y mice have increased lipid deposition in the thoraco-abdominal aorta and multiple foci of elastin fragmentation. A: greater lipid deposition in thoraco-abdominal aorta of ESMIRO/ApoE^−/−/Nox2^−/y (n = 21) compared with ESMIRO/ApoE^−/−/Nox^+/y mice (n = 15). Scale bar = 250 µm. B: no difference in atherosclerosis at the level of the aortic sinus of ESMIRO/ApoE^−/−/Nox2^−/y (n = 7) compared with ESMIRO/ApoE^−/−/Nox^+/y mice (n = 6). Scale bar = 500 µm. C: increased elastin fragmentation in aortic wall at the level of the aortic sinus in ESMIRO/ApoE^−/−/Nox2^−/y (n = 8) compared with ESMIRO/ApoE^−/−/Nox^+/y mice (n = 6). Scale bar = 500 µm. D: similar size populations of circulating leukocytes in ESMIRO/ApoE^−/−/Nox2^−/y (n = 13) compared with ESMIRO/ApoE^−/−/Nox2^+/y mice (n = 11). E: measurement of Nox4 NADPH oxidase, superoxide dismutase 2 (SOD2) and catalase mRNA level showed no difference between ESMIRO/ApoE^−/−/Nox2^−/y (n = 10) mice and ESMIRO/ApoE^−/−/Nox2^+/y mice (n = 9). F: measurement of mRNA level showed no difference in interleukin 1β (IL-1β), tumor necrosis factor-α (TNF-α), the chemokines CCL2 and CCR2, adhesion molecules intercellular adhesion molecule 1 (ICAM-1), and vascular cell adhesion protein 1 (VCAM-1) expression in ESMIRO/ApoE^−/−/Nox2^−/y (n = 8–10) compared with ESMIRO/ApoE^−/−/Nox^+/y mice (n = 8–9). ESMIRO, endothelium-specific mutant insulin receptor-overexpressing mice; ApoE, apolipoprotein E; Nox2, Nox2 isoform of NADPH oxidase; AU, arbitrary units. Data are expressed as mean ± SE; n = number of mice per genotype. *P < 0.05.

Fig. 4.

ESMIRO/ApoE^−/−/Nox2^−/y mice have greater expression of ICAM-1. A: serum IL-1β activity was similar in ESMIRO/ApoE^−/−/Nox2^−/y (n = 12) and ESMIRO/ApoE^−/−/Nox2^+/y. mice (n = 11). B and C: expression of VCAM-1 was reduced (B) whereas expression of ICAM-1 was increased (C) in aorta from ESMIRO/ApoE^−/−/Nox2^−/y (n = 9) compared with ESMIRO/ApoE^−/−/Nox2^+/y mice (n = 9). ESMIRO, endothelium-specific mutant insulin receptor-overexpressing mice; ApoE, apolipoprotein E; Nox2, Nox2 isoform of NADPH oxidase. Data are expressed as mean ± SE; n = number of mice per genotype. *P < 0.05.

Fig. 5.

The gp91dstat treatment has no effect on growth, random insulin, fasting triglyceride, and cholesterol levels and on insulin or glucose tolerance. A: no difference in growth was observed between gp91dstat (n = 24)-treated ESMIRO/ApoE^−/− mice compared with mice treated with scrambled peptide (n = 27). Arrow 1 denotes commencement of Western diet, and arrows 2 and 3 denote time of minipump implantation. B: fasting glucose was lower in gp91dstat-treated ESMIRO/ApoE^−/− (n = 22) compared ESMIRO/ApoE^−/− mice treated with scrambled peptide (n = 16). C: no difference was measured in random insulin concentration in gp91dstat-treated ESMIRO/ApoE^−/− mice (n = 12) compared with ESMIRO/ApoE^−/− mice treated with scrambled peptide (n = 15). D: scrambled peptide-treated ESMIRO/ApoE^−/− mice (n = 10) and gp91dstat-treated ESMIRO/ApoE^−/− mice (n = 10) had similar fasting triglycerides. E: scrambled peptide-treated ESMIRO/ApoE^−/− mice (n = 10) and gp91dstat-treated ESMIRO/ApoE^−/− mice (n = 10) had similar fasting total cholesterol. F and G: no difference was observed in insulin (scrambled n = 12 vs. gp91dstat n = 12; F) or glucose tolerance tests (scrambled n = 19 vs. gp91dstat n = 19; G) in gp91dstat-treated ESMIRO/ApoE^−/− mice compared with ESMIRO/ApoE^−/−mice treated with scrambled peptide. ESMIRO, endothelium-specific mutant insulin receptor-overexpressing mice; ApoE, apolipoprotein E. Data are expressed as mean ± SE; n = number of mice per genotype. *P < 0.05.

Fig. 6.

The gp91dstat treatment reduces endothelial cell (EC) superoxide generation and lipid deposition in the thoraco-abdominal aorta of ESMIRO/ApoE^−/− mice without causing elastin fragmentation or increased ICAM-1 expression. A: NADPH-dependent superoxide generation in pulmonary endothelial cells was reduced in gp91ds-tat-treated ESMIRO/ApoE^−/− mice (n = 7) compared with ESMIRO/ApoE^−/− mice (n = 8) treated with scrambled peptide. B: ESMIRO/ApoE^−/− mice treated with gp91dstat (n = 13) developed less atherosclerosis in the thoraco-abdominal aorta than ESMIRO/ApoE^−/− mice treated with scrambled peptide (n = 14). Scale bar = 250 µm. C: there was no difference in atherosclerosis at the level of the aortic sinus in gp91ds-tat-treated ESMIRO/ApoE^−/− mice (n = 6) compared with ESMIRO/ApoE^−/− mice (n = 6) treated with scrambled peptide. Scale bar = 500 µm. D: there was no difference in the number of defects in the aorta at the level of the sinus in ESMIRO/ApoE^−/− mice treated with gp91dstat (n = 6) compared with ESMIRO/ApoE^−/− mice treated with scrambled peptide (n = 7). Scale bar = 500 µm. E and F: VCAM-1 (E) and ICAM-1 (F) expression were not significantly different between ESMIRO/ApoE^−/− mice treated with gp91dstat (n = 5) or scrambled peptide (n = 5) although there was a tendency for ICAM-1 expression to be less in gp91dstat-treated mice. ESMIRO, endothelium-specific mutant insulin receptor-overexpressing mice; ApoE, apolipoprotein E; AU, arbitrary units. Data are expressed as mean ± SE; n = number of mice per genotype. *P < 0.05.