β₃-Adrenoceptor activation relieves oxidative inhibition of the cardiac Na⁺-K⁺ pump in hyperglycemia induced by insulin receptor blockade

Abstract

Dysregulated nitric oxide (NO)- and superoxide (O₂^·−)-dependent signaling contributes to the pathobiology of diabetes-induced cardiovascular complications. We examined if stimulation of β₃-adrenergic receptors (β₃-ARs), coupled to endothelial NO synthase (eNOS) activation, relieves oxidative inhibition of eNOS and the Na⁺-K⁺ pump induced by hyperglycemia. Hyperglycemia was established in male New Zealand White rabbits by infusion of the insulin receptor antagonist S961 for 7 days. Hyperglycemia increased tissue and blood indexes of oxidative stress. It induced glutathionylation of the Na⁺-K⁺ pump β₁-subunit in cardiac myocytes, an oxidative modification causing pump inhibition, and reduced the electrogenic pump current in voltage-clamped myocytes. Hyperglycemia also increased glutathionylation of eNOS, which causes its uncoupling, and increased coimmunoprecipitation of cytosolic p47^phox and membranous p22^phox NADPH oxidase subunits, consistent with NADPH oxidase activation. Blocking translocation of p47^phox to p22^phox with the gp91ds-tat peptide in cardiac myocytes ex vivo abolished the hyperglycemia-induced increase in glutathionylation of the Na⁺-K⁺ pump β₁-subunit and decrease in pump current. In vivo treatment with the β₃-AR agonist CL316243 for 3 days eliminated the increase in indexes of oxidative stress, decreased coimmunoprecipitation of p22^phox with p47^phox, abolished the hyperglycemia-induced increase in glutathionylation of eNOS and the Na⁺-K⁺ pump β₁-subunit, and abolished the decrease in pump current. CL316243 also increased coimmunoprecipitation of glutaredoxin-1 with the Na⁺-K⁺ pump β₁-subunit, which may reflect facilitation of deglutathionylation. In vivo β₃-AR activation relieves oxidative inhibition of key cardiac myocyte proteins in hyperglycemia and may be effective in targeting the deleterious cardiac effects of diabetes.

while a high prevalence of coronary artery disease predisposes patients with diabetes to ischemic cardiomyopathy, direct effects of diabetes on myocardial structure and function also contribute to an increased risk of heart failure (8, 37). Raised levels of reactive oxygen species (ROS), or “oxidative stress” due to increased activity of NADPH oxidase in cell membranes, contribute to the pathogenesis (28). The NADPH oxidase-derived ROS can quench nitric oxide (NO) and also uncouple endothelial NO synthase (eNOS), switching its function from NO to ROS generation (21). A “nitroso-redox imbalance” (51) then causes redox modifications of cell proteins and cell dysfunction.

The Na⁺-K⁺ pump is one of the important membrane proteins susceptible to redox modifications. We have shown that its activity is regulated by glutathionylation of Cys⁴⁶ on the β₁-subunit of the pump α/β-heterodimer (13, 17). Glutathionylation is a reversible oxidative modification whereby a disulfide bond is formed between glutathione (GSH) and protein cysteine residues. It has structural and functional effects on proteins reminiscent of phosphorylation and is involved in physiological and pathophysiological cell signaling (15).

As reviewed elsewhere (29), oxidative signaling and downstream glutathionylation and deglutathionylation of the Na⁺-K⁺ pump β₁-subunit play important roles in receptor-coupled inhibition and stimulation of the Na⁺-K⁺ pump. Within this scheme, glutathionylation and the functional equivalent of pump inhibition are mediated by activation of NADPH oxidase, while receptor-coupled deglutathionylation and pump stimulation involve activation of NOS. This raises the possibility that NADPH oxidase activation in diabetes uncouples eNOS and disturbs the “nitroso-redox balance” (51) to the extent that it may cause the sarcolemmal Na⁺-K⁺ pump inhibition that can occur with diabetes (24). Na⁺-K⁺ pump inhibition is expected to increase cytosolic Na⁺ levels, and such an increase has adverse effects in heart failure (1, 5, 35, 36). In support of such harmful effects, treatments that are beneficial in heart failure stimulate the Na⁺-K⁺ pump and, hence, Na⁺ export, while treatments that are harmful in clinical trials inhibit the pump, as we summarized previously (29).

The β₁-subunit in a proportion of Na⁺-K⁺ pumps in cardiac myocytes isolated from normal rabbits is glutathionylated at baseline, and we previously showed that in vitro exposure of the myocytes to β₃-adrenergic receptor (β₃-AR) agonists decreases glutathionylation and increases electrogenic Na⁺-K⁺ pump current (I_p) measured in voltage-clamped myocytes. These effects are NOS-dependent (9). In the present study we examine if in vivo treatment of rabbits with the β₃-AR agonist CL316243 (CL) can reverse hyperglycemia-induced pump inhibition through NO- and ROS-dependent signaling. We previously studied the effect of alloxan-induced type 1 diabetes on the Na⁺-K⁺ pump in cardiac myocytes (24). However, nonspecific intrinsic oxidant effects of alloxan (39) are undesirable when NO- and ROS-dependent signaling is examined. In this study we induced hyperglycemia in rabbits by infusion of S961.

S961 and the closely related S661 are synthetic peptides that competitively displace insulin from purified human, rat, and pig insulin receptors with affinities similar to the affinity of insulin itself and with a selectivity for the insulin receptor vs. the insulin-like growth factor 1 receptor that is higher than the selectivity for insulin. In a cellular assay, S961 blocks insulin action with an IC₅₀ of 1.3 nM, and in vivo intravenous bolus administration of S661 to rats transiently raises blood glucose nearly fivefold. The increase is reduced or eliminated in a dose-dependent manner by administration of insulin, consistent with competition for binding to the insulin receptor (41). We modified the protocol that induces transient hyperglycemia with bolus administration (41) to a protocol that induces sustained hyperglycemia by administration of S961 via osmotic mini-pumps.¹

METHODS

Animals, tissues, and cells.

Pharmacology and physiological effects on the heart mediated by the β₃-ARs are highly species-dependent, and since rabbits provide a relevant model for the human receptor (3), we used male New Zealand White rabbits (2.2–2.6 kg body wt). S961 (a gift from Novo Nordisk, Denmark), with the peptide sequence GSLDESFYDWFERQLGGGSGGSSLEEEWAQIQCEVWGRGCPSY and a disulfide bridge connecting the two cysteine residues (41), was infused for 7 days at a rate of 12 μg·kg⁻¹·h⁻¹ via osmotic mini-pumps (Alzet, Palo Alto, CA) implanted subcutaneously under general anesthesia. The selective β₃-AR agonist CL (Sigma-Aldrich, St. Louis, MO) was infused via mini-pumps at a rate of 40 μg·kg⁻¹·h⁻¹ for the last 3 days of the 7-day infusion period. Since infusion of vehicle alone had no effect on I_p in six myocytes from five rabbits, we did not use sham infusions in controls. Durations of S961 and CL infusions were limited by the amount of S961 we had available and the cost of CL.

Blood samples for analysis of systemic oxidative stress were obtained from the central ear artery of rabbits anesthetized by subcutaneous injection of ketamine hydrochloride (50 mg/kg) and xylazine hydrochloride (50 mg/kg), while blood glucose in a drop of blood from the marginal ear vein was measured using a glucometer and strips (Optium Xceed, Abbott Diabetes Care, Australia). Hearts were harvested after rabbits were killed by bolus intravenous injection of ketamine hydrochloride. Cardiac myocytes were isolated as previously described (25). Hearts from wild-type mice and mice with genetically deleted β₃-ARs were kindly provided by Professor Roger Summers' Laboratory (Monash University, Melbourne). Protocols for animal experiments were approved by the Animal Care and Ethics Committee of our institution.

Measurement of I_p.

We measured I_p in freshly isolated, voltage-clamped myocytes. Solutions and voltage-clamp protocol minimized nonpump membrane currents. Wide-tipped patch pipettes (4–5 μm) were filled with solutions containing (in mM) 5 HEPES, 2 MgATP, 5 EGTA, 70 potassium glutamate, 10 sodium glutamate, and 80 tetramethylammonium chloride and 0.01 l-arginine. l-Arginine was omitted in a subset of experiments. The solutions were titrated to pH 7.20 at 35°C with KOH. While we established the whole cell configuration, myocytes were superfused with a solution containing (in mM) 140 NaCl, 5.6 KCl, 2.16 CaCl₂, 1 MgCl₂, 10 glucose, 0.44 NaH₂PO₄, and 10 HEPES and titrated to pH 7.40 at 35°C with NaOH. We switched to a solution that was identical, except it was nominally Ca²⁺-free and contained 2 mM BaCl₂ and 0.2 mM CdCl₂. I_p was identified as the difference between holding current induced by Na⁺-K⁺ pump blockade before and after exposure to 100 μM ouabain.

Immunodetection of glutathionylated proteins and protein-protein association.

We detected glutathionylation of proteins in myocyte lysate by immunoprecipitation of glutathionylated proteins with an anti-GSH antibody (Virogen) using Protein A/G PLUS-Agarose beads (Santa Cruz Biotechnology), as previously described (13). The immunoprecipitate was immunoblotted with an antibody against the Na⁺-K⁺ pump β₁-subunit (Upstate Biotechnology) or an anti-eNOS antibody (Sigma-Aldrich) in Western blots. Glutathionylation of the Na⁺-K⁺ pump β₁-subunit was also detected by the biotinylated GSH ethyl ester technique in cardiac myocytes. The glutathionylated subfraction in lysate from myocytes that had been loaded with biotinylated GSH (500 μM, 1 h at room temperature) was precipitated with streptavidin-Sepharose beads (GE Healthcare Biosciences), as described elsewhere (13). Coimmunoprecipitation protocols were used to examine the association of proteins using antibodies against the α₁-subunit (Upstate Biotechnology), p47^phox and p22^phox subunits of NADPH oxidase, and glutaredoxin-1 (Grx1; Santa Cruz Biotechnology).

Measurement of lipid peroxidation in plasma.

Plasma lipid peroxidation levels were analyzed by the thiobarbituric acid-reactive substances (TBARS) assay, which reacts with malondialdehyde (MDA), using the OxiSelect TBARS assay kit (Cell Biolabs). Briefly, fresh plasma was mixed with butylated hydroxytoluene to arrest oxidation, mixed with sodium dodecyl sulfate lysis solution, and then incubated with thiobarbituric acid at 95°C for 1 h before centrifugation. The TBARS-MDA adduct was measured by fluorometry in the supernatant by absorption at 530 nm. Lipid peroxidation level was determined from a standard curve.

High-performance liquid chromatography analysis of dihydroethidium oxidation products.

High-performance liquid chromatography was used to separate the O₂^·−-dependent 2-hydroxyethidium (2-OH-E⁺) product from the nonspecific product ethidium (E⁺) following dihydroethidium oxidation (50) in rabbit aorta. Rabbit aorta was incubated with dihydroethidium (50 μM, 30 min at 37°C) in Krebs buffer containing diethylenetriaminepentaacetic acid (100 μM) in darkness. Aortic segments were lysed in methanol, and after protein precipitation by addition of HClO₄ (200 mM) in methanol at 4°C and centrifugation, the oxidation products in the supernatant (50 μl) were separated by high-performance liquid chromatography (Shimadzu). Products were quantified by fluorescence and electrochemical oxidation, and their mean value was normalized to protein concentration.

Statistical analysis.

Values are means ± SE. Student's t-test was used for comparison between two groups. For multiple comparisons, one-way analysis of variance was used, with Tukey's post hoc analysis for multiple comparisons. P < 0.05 was considered statistically significant.

RESULTS

Effect of S961 on blood glucose, insulin levels, and oxidative stress.

We first examined if transient hyperglycemia induced by a bolus injection of S661 in rats (41) can be reproduced by a bolus injection of S961 in rabbits. We administered S961 subcutaneously at 15, 150, or 300 μg/kg to a rabbit. The two larger doses induced transient hyperglycemia (Fig. 1A). We next examined the effect of continuous subcutaneous infusion of S961. The time course of blood glucose levels in a rabbit infused with S961 at 12 μg·kg⁻¹·h⁻¹ is shown in Fig. 1B. Additional higher and lower rates of infusion were used, and blood glucose levels were sampled daily over the following 7 days. Mean levels sampled in one rabbit for each rate of infusion increased with an increase in the rate of infusion (Fig. 1C).

Fig. 1.

Effects of S961 on blood glucose and plasma insulin levels. A: blood glucose after subcutaneous injection of S961 at 15 μg/kg (⧫), 150 μg/kg (■), or 300 μg/kg (●); n = 1 for each dose. B: blood glucose in a rabbit during subcutaneous infusion of S961 at 12 μg·kg⁻¹·h⁻¹ via osmotic mini-pump. C: blood glucose during 7 days of infusion of S961 at 5, 10, 12, and 18 μg·kg⁻¹·h⁻¹ or saline vehicle. Values are means ± SE of ≥7 levels sampled during the period of infusion in 1 rabbit at each rate of infusion. D: mean glucose levels in 7 control rabbits and 8 rabbits infused with S961 at 12 μg·kg⁻¹·h⁻¹. HG, hyperglycemia. E: plasma insulin levels before and after infusion of S961 at 12 μg·kg⁻¹·h⁻¹ for 7 days in 5 rabbits. *P < 0.05.

On the basis of results shown in Fig. 1C, we selected an infusion rate of 12 μg·kg⁻¹·h⁻¹ for the study. The mean blood glucose levels in a number of rabbits treated for 7 days with this rate of infusion compared with levels in controls are shown in Fig. 1D. The S961-induced increase in blood glucose levels was accompanied by an increase in plasma insulin levels (Fig. 1E), consistent with a counterregulatory response to the competitive insulin receptor blockade in in vitro studies (41). There were no effects of S961-induced hyperglycemia on serum Na⁺, K⁺, cholesterol, or triglycerides or on indexes of renal function (data not shown). The level of 2-OH-E⁺ in aorta (Fig. 2A) and the level of MDA in plasma (Fig. 2B) as indexes of oxidative stress were increased with S961-induced hyperglycemia relative to control.

Fig. 2.

Indexes of oxidative stress in tissue and blood. A: superoxide (O₂^·−) in rabbit aorta measured by HPLC analysis of the O₂^·−-specific product of dihydroethidium (DHE) oxidation, 2-hydroxyethidium (2-OH-E⁺). Levels are normalized to protein concentration of total aorta homogenate. Aorta was obtained from 4 control rabbits and 5 rabbits with S961-induced hyperglycemia. B: plasma malondialdehyde in 5 control rabbits and 4 rabbits in each of the other groups. CL and HG + CL groups represent rabbits treated with the β₃-adrenergic receptor (AR) agonist CL316243 (CL) on the last 3 days of a 7-day period of hyperglycemia before euthanization. *P < 0.05.

Na⁺-K⁺ pump current and glutathionylation of the Na⁺-K⁺ pump β₁-subunit.

Figure 3A illustrates the experimental protocol used to measure I_p in myocytes. Traces of currents in myocytes from a control rabbit and a rabbit treated with CL are shown. I_p for each myocyte was identified as the mean value of holding currents sampled with an electronic cursor before and after exposure to ouabain. Sampling of currents and criteria for their stability are described elsewhere (46). I_p was reduced in myocytes from S961-treated rabbits relative to I_p of myocytes from control rabbits. In vivo treatment with CL for 3 days reversed Na⁺-K⁺ pump inhibition (Fig. 3B). CL had no effect on plasma glucose, insulin levels, or heart rate, but it reduced mean arterial pressure in control and S961-treated rabbits (18) and reduced levels of MDA (Fig. 2B).

Fig. 3.

Effects of hyperglycemia and the β₃-AR agonist CL316243 on function and oxidative modification of the Na⁺-K⁺ pump. A: representative membrane currents of a myocyte from a control rabbit and a CL-treated rabbit without hyperglycemia. Pump current (I_p) was identified by the shift in holding current induced by ouabain. C_m, membrane capacitance. B: effects of S961-induced hyperglycemia and CL treatment on I_p. Number of experiments on myocytes from 5 different rabbits in each group is indicated within columns. Shaded columns indicate I_p obtained using l-arginine-free patch-pipette solutions. C, control. C: Na⁺-K⁺ pump β₁-subunit glutathionylation (β₁-GSS) detected by the biotinylated GSH ethyl ester technique. TL:β₁, expression of Na⁺-K⁺ pump β₁-subunits in myocyte total lysates. Glyceraldehyde 3-phosphate dehydrogenase (GAPDH) was used as a loading control. Immunoblots are from myocytes isolated from 5 rabbits in each group. D: Na⁺-K⁺ pump β₁-subunit glutathionylation in myocyte lysate as detected by β₁-subunit immunoblot (IB) of protein immunoprecipitate (IP) with anti-GSH antibody. Immunoblots are from 5 rabbits in each group. *P < 0.05.

In a subset of experiments, we omitted l-arginine from patch-pipette solutions used to voltage-clamp myocytes. l-Arginine deficiency uncouples NOS (52), and omission of l-arginine from pipette solutions reduces I_p relative to I_p in l-arginine-containing solutions (48). There was no significant difference between I_p measured in myocytes from control and S961-treated rabbits when intracellular l-arginine was depleted by use of l-arginine-free pipette solutions (Fig. 2B).

Expressions of Na⁺-K⁺ pump α₁- or β₁-subunits were not altered by S961 or treatment with CL (Fig. 3C; see Fig. 6), suggesting that S961 and CL affected pump turnover rate, rather than abundance. There is a causal relationship between Na⁺-K⁺ pump inhibition and glutathionylation of the β₁-subunit (13), and consistent with functional measurements summarized in Fig. 3B, infusion of S961 increased glutathionylation of the subunit, as detected by two independent techniques. Treatment with CL reversed the increase (Fig. 3, C and D).

Fig. 6.

Effects of hyperglycemia and treatment and CL316243 on NADPH oxidase. Top left: immunoblot of p22^phox and p47^phox NADPH oxidase subunits and the Na⁺-K⁺ pump α₁-subunit in total myocyte lysates. α-Tubulin was used as loading control. Top right: p47^phox, p22^phox and the Na⁺-K⁺ pump α₁-subunit immunoblot of p47^phox immunoprecipitate from myocyte protein. Bottom: mean densitometry for coimmunoprecipitation (Co-IP) experiments from 5 rabbits in each group. *P < 0.05.

Effect of the β₃-AR on glutathionylation of eNOS.

Glutathionylation mediates uncoupling of eNOS under conditions of oxidative stress (10), and we examined if infusion of S961 is associated with glutathionylation of eNOS. S961 induced an increase in glutathionylation of eNOS that was abolished by treatment with CL (Fig. 4A). To independently support a role for β₃-AR-dependent signaling in glutathionylation of eNOS, we examined glutathionylation of eNOS in the myocardium of mice lacking β₃-ARs (β₃-AR^−/−). The abundance of eNOS in β₃-AR^−/− mice and their wild-type littermates was similar, but the signal for glutathionylation of eNOS was consistent with an increase in eNOS in the myocardium from the β₃-AR^−/− mice relative to their wild-type littermates (Fig. 4B). Glutathionylation of the myocardial Na⁺-K⁺ pump β₁-subunit also is increased in β₃-AR^−/− mice (9), consistent with an effect of glutathionylation-dependent eNOS uncoupling on the Na⁺-K⁺ pump.

Fig. 4.

Effects of hyperglycemia and β₃-AR-coupled signaling on endothelial nitric oxide synthase (eNOS) glutathionylation. A: glutathionylation of eNOS in cardiac myocyte lysate as detected by an eNOS immunoblot of protein immunoprecipitate with anti-GSH antibody. Immunoblots are from 5 rabbits in each group. B: glutathionylation of eNOS in myocardium of mice lacking β₃-ARs (β₃-AR^−/−) and their wild-type (WT) littermates as detected by GSH immunoblot of protein immunoprecipitate with anti-eNOS antibody. Immunoblots and immunoprecipitates are from 3 mice in each group. *P < 0.05.

NADPH oxidase and effects of S961 and β₃-AR activation with CL.

In view of the role NADPH oxidase can have in uncoupling of eNOS (21), we examined effects of blocking NADPH oxidase on I_p and glutathionylation of the Na⁺-K⁺ pump β₁-subunit in myocytes isolated from control and S961-treated rabbits. In the functional studies, we blocked translocation of the cytosolic p47^phox subunit to the membranous p22^phox subunit necessary for NADPH oxidase activation by incubating myocytes with the gp91ds-tat peptide (5 μM) for 1 h at 37°C, and we included the peptide in patch-pipette solutions at the same concentration when we subsequently measured I_p in voltage-clamp experiments. The peptide induced a significant increase in I_p of myocytes from S961-treated rabbits (Fig. 5A). Exposure to the gp91ds-tat peptide had no effect on glutathionylation of the Na⁺-K⁺ pump β₁-subunit in myocytes from the control rabbits but decreased glutathionylation in myocytes from S961-treated rabbits (Fig. 5B), corresponding to the increase in I_p (Fig. 5A).

Fig. 5.

Sources of reactive oxygen species (ROS) in hyperglycemia. A: mean I_p in myocytes from control rabbits and rabbits with hyperglycemia incubated with gp91ds-tat peptide to inhibit NADPH oxidase. The peptide was also included in patch-pipette solutions. Numbers of voltage-clamp experiments are indicated within columns. B: Na⁺-K⁺ pump β₁-subunit glutathionylation detected by the biotinylated GSH technique in myocytes from control rabbits and rabbits with hyperglycemia incubated ex vivo with and without gp91ds-tat. Immunoblots of Na⁺-K⁺ pump β₁-subunits in myocyte total lysates are also shown. Three paired samples in each group were examined. *P < 0.05.

Since results in Fig. 5 suggest that increased constitutive NADPH oxidase activity contributed to the hyperglycemia-induced Na⁺-K⁺ pump inhibition, we examined coimmunoprecipitation of the p47^phox subunit with the membranous p22^phox subunit and, as an index of p47^phox subunit translocation to the membrane, coimmunoprecipitation of the p47^phox subunit with the Na⁺-K⁺ pump α₁-subunit. We determined coimmunoprecipitation in lysate of myocytes from control and S961-treated rabbits. Hyperglycemia had no effect on expression of p22^phox and p47^phox NADPH oxidase subunits or the Na⁺-K⁺ pump α₁-subunit (Fig. 6). However, it increased coimmunoprecipitation of p47^phox with p22^phox and p47^phox with the Na⁺-K⁺ pump α₁-subunit. In vivo treatment with CL reversed these changes (Fig. 6).

Effect of β₃-AR activation on coimmunoprecipitation of Grx1 with the Na⁺-K⁺ pump β₁-subunit.

Grx1 mediates deglutathionylation of proteins, and it coimmunoprecipitates with the Na⁺-K⁺ pump β-subunit in cardiac myocytes (7). Addition of recombinant Grx1 to patch-pipette solutions prevents an oxidation-induced decrease in I_p in voltage-clamped myocytes (13). We examined coimmunoprecipitation of Grx1 with the Na⁺-K⁺ pump β₁-subunit in lysate of myocytes from control and S961-treated rabbits. S961 induced a decrease in coimmunoprecipitation of Grx1 with the Na⁺-K⁺ pump β₁-subunit that was reversed by treatment with CL. Treatment of control rabbits with CL also increased coimmunoprecipitation of Grx1 with the Na⁺-K⁺ pump β₁-subunit (Fig. 7).

Fig. 7.

In vivo β₃-AR activation and association of Grx1 with the Na⁺-K⁺ pump in hyperglycemia. Top left: immunoblots of the Na⁺-K⁺ pump β₁-subunit and glutaredoxin-1 (Grx1) in total myocyte lysates. α-Tubulin was used as loading control. Top right: Na⁺-K⁺ pump β₁-subunit and Grx1 immunoblot of Na⁺-K⁺ pump β₁-subunit immunoprecipitate from myocyte protein. Bottom: densitometry for coimmunoprecipitation experiments from 5 rabbits in each group. *P < 0.05.

DISCUSSION

Animal models of hyperglycemia are usually associated with obesity and elevated triglyceride levels (26). Such elevated levels can influence lipid peroxides detected by the TBARS assay, and it is important that infusion of S961 induced stable hyperglycemia in this study but did not increase serum triglyceride levels relative to levels in control rabbits (18). Although the prevalence of hypertriglyceridemia is increased in type 2 diabetes (40), the pattern of hyperglycemia and normal triglyceride levels induced by S961 is similar to that in type 2 diabetes, because in most patients with type 2 diabetes, triglyceride levels are not elevated. A strong correlation between increased TBARS levels and markers of insulin resistance and plasma glucose in humans (45) suggests that the S961 model of hyperglycemia reproduces the oxidative stress that is a central feature in the pathophysiology of human diabetes. The β₃-AR agonist CL reversed the increase in TBARS levels, but not the hyperglycemia.

Infusion of S961 induced a decrease in I_p of myocytes studied ex vivo and a corresponding increase in Na⁺-K⁺ pump β₁-subunit glutathionylation. In vivo treatment with CL decreased Na⁺-K⁺ pump β₁-subunit glutathionylation and increased I_p of myocytes isolated from control and S961-treated rabbits. This indicates that the effects of CL on glutathionylation and I_p in the complex in vivo milieu, even in a disease model, are similar to previously reported effects of exposure of myocytes from normal rabbits to β₃-AR agonists in vitro (9). The effects of in vitro exposure to CL are mediated by a NOS-dependent pathway (9).

While it is firmly established that β₃-AR signaling in the heart is mediated by NOS, the specific isoform involved has varied between experimental models, as reviewed elsewhere (31). In this study, infusion of S961 was associated with an increase in glutathionylation of eNOS, which is expected to uncouple its function from synthesis of NO to synthesis of O₂^·− (10). This is consistent with the increase in vascular eNOS glutathionylation with streptozotocin-induced hyperglycemia in rats reported previously (42). Our finding in the present study of no effect of S961 when I_p was measured ex vivo using l-arginine-free patch-pipette solutions provides functional support for the role of NOS uncoupling in effects of S961-induced hyperglycemia on the Na⁺-K⁺ pump, because with use of the wide-tipped patch pipettes needed for effective perfusion of the intracellular compartment and accurate measurements of I_p (19), relatively low-molecular-weight intracellular substances, such as l-arginine, are expected to be markedly depleted. This facilitates uncoupling of NOS and, as observed experimentally, is expected to abolish the difference in I_p between myocytes from control and diabetic rabbits that is due to uncoupling of NOS developed in vivo. The parallel effect of treatment with CL to reverse the S961-induced increase in eNOS glutathionylation and the decrease in I_p suggests that glutathionylation of the eNOS isoform and Na⁺-K⁺ pump inhibition are interrelated.

Blocking translocation of the cytosolic p47^phox subunit to the membranous p22^phox subunit with the gp91ds-tat peptide reversed the decrease in I_p and the increase in glutathionylation of the Na⁺-K⁺ pump β₁-subunit in myocytes isolated from rabbits treated with S961. However, the peptide had no significant effect on I_p or β₁-subunit glutathionylation in myocytes from control rabbits. The lack of significant effect of gp91ds-tat on I_p in control myocytes, although shown here in only a small number of myocytes, is consistent with the same result in a previous report (47). S961 also increased coimmunoprecipitation of the p47^phox and p22^phox subunits in myocyte lysate. These results strongly implicate an increased activity of NADPH oxidase in the effects of S961-induced hyperglycemia and are consistent with the increased myocardial oxidative stress shown to occur in streptozotocin-induced diabetes (42). NADPH oxidase is also a major source of diabetes-induced oxidative stress in vascular tissue in experimental animal models (20), as well as in humans (22), suggesting that effects of hyperglycemia on the Na⁺-K⁺ pump reported here for the heart may be similar for blood vessels.

NADPH oxidase-derived ROS can uncouple eNOS, and, conversely, activation of NO production can downregulate the NOX2 isoform of NADPH oxidase (28). Cross talk between NOS and NADPH oxidase activity may contribute to the hyperglycemia-induced oxidative Na⁺-K⁺ pump inhibition in the present study. Reversal of the uncoupling is heat shock protein 90 (Hsp90)-dependent (23, 43), and we previously showed that receptor-coupled, NOS-dependent stimulation of I_p in vitro in cardiac myocytes, including stimulation mediated by the β₃-AR, is blocked by the Hsp90 inhibitor radicicol, implicating Hsp90 in the mechanism for stimulation (9, 49). Recoupling of eNOS function by in vivo treatment with CL in this study would have reduced the role of uncoupled eNOS as a source of ROS, but it may also have inhibited NADPH oxidase-dependent ROS synthesis, because we have found that activation of the canonical NO-dependent signaling pathway with the “soluble” guanylyl cyclase stimulator 3-(5′-hydroxymethyl-2′-furyl)-1-benzylindazole (widely known as YC-1) reverses NADPH oxidase-derived oxidative stress and oxidative pump inhibition in cardiac myocytes (11). Figure 8 provides a schematic summary of the effects of S961-induced hyperglycemia and CL on redox pathways and the Na⁺-K⁺ pump in cardiac myocytes.

Fig. 8.

Proposed effects of S961-induced hyperglycemia and β₃-AR stimulation on redox regulation of the Na⁺-K⁺ pump. A: hyperglycemia induces oxidative stress by activating NADPH oxidase. This in turn causes glutathionylation and uncoupling of eNOS, and the uncoupled eNOS generates O₂^·−, rather than nitric oxide (NO). Elevated O₂^·− promotes glutathionylation of the Na⁺-K⁺ pump β₁-subunit and pump inhibition. B: activation of the β₃-AR recouples eNOS, likely by promoting its association with heat shock protein 90 (Hsp90), and NO synthesized by eNOS suppresses hyperglycemia-induced NADPH oxidase activation via the canonical soluble guanylyl cyclase (sGC)-dependent pathway in steps that include dephosphorylation of the p47^phox subunit and its dissociation from the membranous subunit (dashed line). These changes reduce the forward reaction rate for β₁-GSS formation and restore Na⁺-K⁺ pump function.

Administration of S961 was associated with a decrease in coimmunoprecipitation of Grx1 with the Na⁺-K⁺ pump β₁-subunit, while treatment with CL increased coimmunoprecipitation in myocyte lysate with or without prior administration of S961. Activation of the cytosolic Grx1 is poorly understood but is believed to involve its translocation to target proteins (16). The level of coimmunoprecipitation of Grx1 with the Na⁺-K⁺ pump β₁-subunit in the different treatment groups was in the opposite direction of effects of the treatments on glutathionylation of the Na⁺-K⁺ pump β₁-subunit, indicating that abundance of target disulfide bonds is an unlikely primary determinant of translocation of Grx1. Grx1 itself has highly reactive cysteine residues at its active site and is inhibited under conditions of oxidative stress (16). In vitro studies have suggested that the main role of Grx1 is to reverse effects of oxidative stress after the stress has subsided, rather than limit effects during the oxidative stress (34). The coimmunoprecipitation patterns in this study are consistent with, but do not prove, that a similar scheme operates in vivo.

While it had been a widely held early view that the β₃-AR mediates harmful effects in heart failure, indirect evidence from human clinical data suggests that receptor activation might actually be useful, as reviewed elsewhere (38). Such indirect evidence includes the observation that use of thiazolidinediones (TZDs) to treat type 2 diabetes increases the risk of developing heart failure (27, 44) and that TZDs in low, clinically relevant concentrations rapidly downregulate the β₃-AR in various cell types maintained in culture (4). A similar downregulation in vivo would reduce β₃-AR-mediated signaling activated by endogenous catecholamines and perhaps contribute to TZD-induced cardiac dysfunction. This, in combination with the effect of CL to reverse effects of S961-induced hyperglycemia on glutathionylation-induced Na⁺-K⁺ pump inhibition in the present study, suggests that β₃-AR agonists might be useful in prevention and treatment of heart failure in diabetes.

Recent experimental evidence supports the efficacy of β₃-AR agonists in heart failure. After transverse aortic constriction, mice that genetically lack the receptor have enhanced NOS uncoupling, increased myocardial oxidative stress, exacerbation of pathological remodeling, and an increased mortality compared with wild-type mice (30). These results are consistent with a beneficial effect of treatment of mice with transverse aortic constriction with the β₃-AR agonist BRL37344 (33). It is also consistent with beneficial effects mediated by the β₃-AR that overexpression of the receptor in mice reduces the hypertrophic response to infusion of isoproterenol and ANG II through a NOS-mediated mechanism (6). Studies on larger animals also suggest that β₃-AR activation might be useful. Intravenous infusion of BRL37344 had acute beneficial hemodynamic effects in sheep with ischemic cardiomyopathy (9), and in a preliminary report we described how CL administered via osmotic mini-pumps reverses glutathionylation-induced inhibition of I_p in cardiac myocytes and improves clinical indexes of heart failure after coronary ligation in rabbits (14). A phase II study in human heart failure is in progress (12).

Overexpression of the β₃-AR in heart failure (32) may be a useful compensatory mechanism, as suggested by the recent studies on heart failure in animal models. Overexpression of the receptor, as indicated by studies on the diabetic rat heart (2), might have a similar protective role, and we suggest that early-phase human studies examining the effect of selective β₃-AR agonists on cardiac dysfunction in diabetes might be justified by the common and serious long-term complication of overt heart failure. A clinically effective selective β₃-AR agonist, mirabegron, is in use for the treatment of overactive bladder syndrome, and other compounds are being developed.

GRANTS

This work was supported by National Health and Medical Research Council (NHMRC) Project Grant 633252 and Heart Research Australia. K. Karimi Galougahi was supported by a scholarship from Heart Research Australia. C. C. Liu was supported by National Heart Foundation of Australia Fellowship PF 12S 6924. N. A. Fry was supported by research scholarships from AstraZeneca and the Beryl Raymer Trust. G. A. Figtree was supported by the University of Sydney Medical Foundation and a co-funded NHMRC/Heart Foundation Fellowship.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

K.K.G., C.-C.L., A.G., G.A.F., and H.H.R. developed the concept and designed the research; K.K.G., C.-C.L., A.G., N.A.F., and E.J.H. performed the experiments; K.K.G., C.-C.L., A.G., N.A.F., and E.J.H. analyzed the data; K.K.G., C.-C.L., A.G., N.A.F., E.J.H., G.A.F., and H.H.R. interpreted the results of the experiments; K.K.G. and C.-C.L. prepared the figures; K.K.G., C.-C.L., A.G., N.A.F., E.J.H., G.A.F., and H.H.R. edited and revised the manuscript; K.K.G., C.-C.L., A.G., N.A.F., E.J.H., G.A.F., and H.H.R. approved the final version of the manuscript; H.H.R. drafted the manuscript.