β3 Adrenergic Stimulation Restores Nitric Oxide/Redox Balance and Enhances Endothelial Function in Hyperglycemia

Abstract

Background: Perturbed balance between NO and O2 (•-). (ie, NO/redox imbalance) is central in the pathobiology of diabetes-induced vascular dysfunction. We examined whether stimulation of β3 adrenergic receptors (β3 ARs), coupled to endothelial nitric oxide synthase (eNOS) activation, would re-establish NO/redox balance, relieve oxidative inhibition of the membrane proteins eNOS and Na(+)-K(+) (NK) pump, and improve vascular function in a new animal model of hyperglycemia.

Methods and results: We established hyperglycemia in male White New Zealand rabbits by infusion of S961, a competitive high-affinity peptide inhibitor of the insulin receptor. Hyperglycemia impaired endothelium-dependent vasorelaxation by "uncoupling" of eNOS via glutathionylation (eNOS-GSS) that was dependent on NADPH oxidase activity. Accordingly, NO levels were lower while O2 (•-) levels were higher in hyperglycemic rabbits. Infusion of the β3 AR agonist CL316243 (CL) decreased eNOS-GSS, reduced O2 (•-), restored NO levels, and improved endothelium-dependent relaxation. CL decreased hyperglycemia-induced NADPH oxidase activation as suggested by co-immunoprecipitation experiments, and it increased eNOS co-immunoprecipitation with glutaredoxin-1, which may reflect promotion of eNOS de-glutathionylation by CL. Moreover, CL reversed hyperglycemia-induced glutathionylation of the β1 NK pump subunit that causes NK pump inhibition, and improved K(+)-induced vasorelaxation that reflects enhancement in NK pump activity. Lastly, eNOS-GSS was higher in vessels of diabetic patients and was reduced by CL, suggesting potential significance of the experimental findings in human diabetes.

Conclusions: β3 AR activation restored NO/redox balance and improved endothelial function in hyperglycemia. β3 AR agonists may confer protection against diabetes-induced vascular dysfunction.

Keywords: endothelial dysfunction; endothelial nitric oxide synthase; hyperglycemia; oxidative stress; β3 adrenergic receptors.

Figure 1

Effects of diabetogenic agents on redox modification of the Na⁺‐K⁺ pump. A, Effect of alloxan on β₁ Na⁺‐K⁺ pump glutathionylation (β₁‐GSS) in isolated rabbit cardiac myocytes loaded with biotinylated glutathione (BiOGEE technique) prior to alloxan exposure (n=5). B, β₁‐GSS in rabbits receiving a bolus injection of alloxan (100 mg/kg) with and without developing diabetes, as detected by immunoblots (IB) with anti‐glutathione and anti‐β₁ subunit antibodies in the β₁ subunit immunoprecipitate (IP). C, Effect of STZ on β₁‐GSS examined by BiOGEE technique (n=5). D, Effect of S961 on β₁‐GSS (n=6). α tubulin was used as loading control. Statistical significance is indicated by *. STZ indicates streptozocin.

Figure 2

Blood glucose levels and vascular function in S961‐induced hyperglycemia. A, Blood glucose levels of rabbits infused subcutaneously via osmotic mini‐pump with S961 (12 μg/kg per hour). (n=5). B, Mean blood glucose levels of control rabbits and rabbits with S961‐induced hyperglycemia over 7 days (n=10). C, Endothelium‐dependent vasorelaxation in aortic rings of control and hyperglycemic rabbits (n=6 control and 5 hyperglycemic rabbits, with 2–3 or 3–5 rings studied per rabbit in each group, respectively). These baseline data are re‐presented in Figure 3A. P=0.01 on 2‐way repeated‐measures ANOVA. D, SNP‐induced vasorelaxation in control and diabetic rabbits (n=6 rabbits and 2–3 rings per rabbit). P=0.69 on 2‐way repeated‐measures ANOVA. Relaxations are plotted as the percentage decrease in PE‐induced contraction against the concentration of ACh or SNP on a logarithmic scale. PE‐induced contractions were not different across different groups. Statistical significance is indicated by *. ACh indicates acetylcholine; PE, phenylephrine; SNP, sodium nitroprusside.

Figure 3

Effect of β₃ AR agonism on endothelial function, NO‐ and O₂ ^•− levels, and redox modification of eNOS in S961‐induced hyperglycemia. A, Effect of CL316243 (CL) on ACh‐induced relaxation in aorta of control and hyperglycemic rabbits. Number of rabbits per each group: 6 controls (2–3 rings/rabbit), 5 hyperglycemics (3–5 rings/rabbit), 5 CL‐treated controls (4–6 rings/rabbit), and 6 CL‐treated hyperglycemics (HG) (3–6 rings/rabbit). P=0.01 (control vs HG) and 0.005 (HG vs HG+CL) on 2‐way repeated‐measures ANOVA. B, O₂ ^•− levels in rabbit aorta by HPLC analysis of the specific‐ (2‐OH‐E⁺) and nonspecific product (E⁺) of DHE oxidation. Representative chromatograms from a control and a hyperglycemic rabbit are shown. [2‐OH‐E⁺]=mean values over protein concentration (pmol/mg per mL). n=5 control, 5 CL‐treated, 6 hyperglycemic‐ and 5 CL‐treated hyperglycemic rabbits. C, NO measurement by electron paramagnetic resonance of Fe(DETC)₂ in rabbit aorta. The amplitude of NO‐Fe(DETC)₂ signal was determined as the height between the top of the first low field signal (g₁=3291 G) and the valley of the third high field signal (g₂=3327 G) (up–down double arrows on the left). The trace from the standard spermine NONOate in dimethyl sulfoxide solution is shown. n=7 control‐, 4 CL‐treated‐, 5 hyperglycemic‐, and 5 CL‐treated hyperglycemic rabbits. D, Glutathione and eNOS immunoblot (IB) of eNOS immunoprecipitate (IP) from rabbit aorta. n=5. *P<0.05 vs control and **P<0.05 vs hyperglycemia. eNOS indicates endothelial nitric oxide synthase.

Figure 4

β₃ AR stimulation and redox regulation of the vascular Na⁺‐K⁺ pump in S961‐induced hyperglycemia. A, Effect of CL on K⁺‐induced relaxation in aortic rings of control‐ and hyperglycemic rabbits. n=7 control‐ (2–4 rings/rabbit), 5 hyperglycemic‐ (2 rings/rabbit), 4 CL‐treated‐ (4–6 rings/rabbit), 6 CL‐treated hyperglycemic‐ (4–6 rings/rabbit) rabbits. P=0.01 (HG vs control), 0.02 (CL vs control) and 0.001 (HG vs HG+CL) on 2‐way repeated‐measures ANOVA. B, Glutathionylated β₁ subunit detected by glutathione antibody technique. β₁ Na⁺‐K⁺ pump subunit IB of the glutathione IP is shown on top row. IB of β₁ subunit and α tubulin from raw homogenates of rabbit aorta are also shown. n=5. *P<0.05 vs control and **P<0.05 vs hyperglycemia. HG indicates hyperglycemia; IB, immunoblot; IP, immunoprecipitate.

Figure 5

Effect of β₃ AR agonism on sources of ROS and de‐glutathionylation enzymatic system in hyperglycemia. A and B, ACh‐induced relaxation in aortic rings from control and hyperglycemic rabbits with and without ex vivo incubation with gp91ds‐tat (5 μmol/L, 37°C, 1 hour) prior to vasomotor studies. n=5 control and 5 hyperglycemic rabbits with 3 rings pre‐incubated vs 3 rings not pre‐incubated with gp91ds‐tat (5 μmol/L, 37°C, 1 hour) in each rabbit. P=0.15 in controls and 0.01 in hyperglycemics on 2‐way repeated‐measures ANOVA. C and D, eNOS IB of glutathione IP from rabbit aorta with and without ex vivo exposure and α tubulin IB as loading control are shown. n=5. E, p47^phox and eNOS IB of p47^phox IP from rabbit aorta n=5. F, Glutaredoxine‐1 (Grx1), β₁ Na⁺‐K⁺ pump subunit (β₁), and eNOS IB of Grx1 IP from rabbit aorta. n=5. Grx1/Grx1 co‐immunoprecipitation was unaltered between the groups. The panels show the mean of signal for eNOS‐ and β₁ subunit coimmunoprecipitation with Grx1. *P<0.05 vs control and **P<0.05 vs hyperglycemia. β₃ AR indicates β₃ adrenergic receptors; ACh, acetylcholine; CL, CL316243; eNOS, endothelial nitric oxide synthase; GSS, glutathionylation; HG, hyperglycemia; IB, immunoblot; IP, immunoprecipitate; ROS, reactive oxygen species.

Figure 6

eNOS glutathionylation in vessels of diabetic humans and effects of β₃ AR stimulation in human endothelial cells and vascular tissue. A, Glutathione (GSH) and eNOS IB of eNOS IP from homogenates of the arteries from diabetic patients and their matched nondiabetic controls. There was a nonsignificant trend for increase in eNOS/eNOS co‐immunoprecipitation in diabetics (bottom bands). The values for eNOS‐GSS shown in the panel are the mean of densitometries normalized to the corresponding eNOS/eNOS bands. The median of the densitometries was 0.95 (IQR 0.8–1.0) in nondiabetics vs 1.87 (IQR 1.66–2.04) in diabetics. n=4. IP with a nonspecific IgG was used as negative control. B, The effect of CL (1 μmol/L, 37°C, 1 hour) on basal NO‐sensitive DAF‐FM fluorescence in HUVECs. The cell nuclei were counterstained with DAPI (blue). n=4. Cells from 2 to 3 randomly chosen areas of interest in each group were analyzed. C, Glutathione‐ and eNOS IB of eNOS IP from homogenates of the freshly harvested arteries from diabetic patients exposed to CL (1 μmol/L, 37°C, 1 hour) ex vivo. Mean densitometries are shown. The median of the densitometries was 1.0 (IQR 0.89–1.24) in −CL group vs 0.48 (IQR 0.36–0.66) in +CL group. n=4. IP with a nonspecific IgG was used as negative control. *P<0.05 vs control. β₃ AR indicates β₃ adrenergic receptors; CL, CL316243; DAF‐FM, 4‐amino‐5‐methylamino‐2',7'‐difluorofluorescein; DAPI, 4',6‐diamidino‐2‐phenylindole; DAPI,; eNOS, endothelial nitric oxide synthase; GSS, glutathionylation; HUVECs, human umbilical vein endothelial cells; IB, immunoblot; IP, immunoprecipitate; IQR, interquartile range.

Figure 7

Effects of CL on eNOS phosphorylation in human endothelial cells, and expression of nNOS in human endothelial cells or vessels. A, Immunoblot (IB) of peNOS at serine 114 and 1177 residues from eNOS immunoprecipitate of HUVECs with (+CL) and without (−CL) exposure to CL (n=4). B, Immunoblots of nNOS in HUVECs, human myocardium (positive control), and arterial segments from diabetic (DM) and nondiabetic (non DM) patients exposed to CL ex vivo. CL indicates CL316243; eNOS, endothelial nitric oxide synthase; peNOS, phosphorylated endothelial nitric oxide synthase; HUVECs, human umbilical vein endothelial cells; IB, immunoblot; IP, immunoprecipitate; nNOS, neuronal nitric oxide synthase.

Figure 8

β₃ AR stimulation, redox‐ and NO‐dependent signaling and endothelial function in hyperglycemia. A, NADPH oxidase (NOX) is activated in aorta of rabbits with S961‐induced hyperglycemia, increasing the levels of ROS. ROS can directly scavenge and decrease NO levels, increase eNOS glutathionylation (eNOS‐GSS) and, possibly, impair de‐glutathionylation by affecting the activity of glutaredoxin‐1 (Grx1) through yet unknown but likely redox‐dependent mechanism(s)26 (dashed line). Glutathionylation‐mediated eNOS uncoupling results in eNOS‐derived ROS generation, which might further increase eNOS‐GSS and reduce NO bioavailability in endothelial cells (ECs) and impair relaxation of vascular smooth muscle cells (VSMCs). B, The β₃ AR agonist CL316243 (CL) increases NO. NO effect in reducing ROS by direct quenching may also decrease eNOS‐derived ROS by reducing ROS‐induced eNOS‐GSS (green cross). Additionally, NO may decrease NOX activity (dashed line). An increase in Grx1/eNOS association by CL suggests enhanced eNOS de‐glutathionylation by Grx1. These changes induced by CL culminate in enhanced vasodilatation, which my in turn reduce blood pressure (dotted arrow). A decrease in blood pressure per se may contribute to these beneficial changes (dotted line). β₃ AR indicates β₃ adrenergic receptors; eNOS, endothelial nitric oxide synthase; GSS, glutathionylation; ROS, reactive oxygen species.