Upregulation of Arginase by H₂O₂ Impairs Endothelium-Dependent Nitric Oxide-Mediated Dilation of Coronary Arterioles

Reactive oxygen species (ROS) from mitochondria and other subcellular sources have been regarded as toxic byproducts of metabolism, especially when excessive production of ROS outstrips endogenous antioxidant defense mechanisms.¹ However, ROS are also known to influence the expression of a number of genes and signal transduction pathways² and are thought to act as subcellular messengers for certain growth factors.³ Interestingly, several cardiovascular diseases with diverse etiologies, such as atherosclerosis,⁴ hypertension,⁵ vascular complications in diabetes,⁶ and after ischemia/reperfusion injury⁷ are associated with the common hallmarks of increased oxidative stress and endothelial cell dysfunction.⁸ Although the molecular basis of endothelial dysfunction is not completely understood, numerous studies point to the reduction of nitric oxide (NO) biosynthesis and/or bioactivity as a major mechanism.⁹ However, the underlying cellular mechanisms contributing to the reduction of NO-mediated effects remain unclear.

Perfusion of coronary artery with H₂O₂ has recently been shown to impair vasodilation in response to NO-mediated agonists;¹⁰ however, the studies suggested that endothelial dysfunction caused by H₂O₂ was not mediated by the disruption of arginine-NO pathway.¹¹ In fact, NO synthase (NOS) activity and its expression in endothelial cells or in vascular tissues treated with H₂O₂ are not reduced¹¹ and are even increased in some studies.¹² These results imply that H₂O₂ may decrease NO-mediated functions via other mechanisms independent of NOS. Because a sufficient supply of substrate L-arginine is required for NO synthesis, it is possible that reduction of L-arginine availability is involved in the impairment of NO-mediated vasodilation by H₂O₂. Interestingly, recent studies have shown that arginase enzyme, which consumes L-arginine to form L-ornithine and urea, is expressed in the endothelium^13,14 and plays a counteracting role in the stimulated NO production^14–16 and in NO-mediated vasodilatory function in coronary microcirculation.¹⁴ In addition, there is substantial evidence that the expression of arginase is elevated in a variety of cells and tissues under the conditions with inflammation and oxidative stress.^17–20 It is plausible that the upregulation of arginase and its competition with NOS for their common substrate L-arginine may be involved in the microvascular dysfunction induced by H₂O₂. Because NO released from the endothelium plays a major role in the determination of coronary vasomotor activity,²¹ it is important to understand the direct effect of H₂O₂ on coronary arteriolar function and to elucidate the underlying mechanism responsible for the impairment of NO-mediated dilation in these microvessels. Herein, we tested the hypothesis that H₂O₂ specifically inhibits endothelium-dependent NO-mediated dilation of coronary arterioles by reducing L-arginine availability through upregulated arginase. Using an isolated vessel preparation, we examined the effect of H₂O₂ on vasodilatory function of coronary arterioles in response to various endothelium-dependent and endothelium-independent agonists. The role of arginase in influencing vasomotor function was addressed using pharmacological, molecular, and immunohistochemical tools.

Methods

Effect of H₂O₂ on Vasodilatory Function of Isolated Coronary Arterioles

The procedures followed were in accordance with guidelines set by the Laboratory Animal Care Committee at Texas A&M University. See the online-only data supplement for detailed description of methods (http://atvb.ahajournals.org). Pigs (Milberger Farms, Kurten, Tex) were anesthetized with pentobarbital (20 mg/kg) and the heart was quickly excised. Individual coronary arterioles (60 to 120 μm, in internal diameter in situ) were dissected from the subepicardium of left ventricle for in vitro study as previously described.²¹ Vessels were cannulated and pressurized to 60 cmH₂O intraluminal pressure. After development of stable basal tone, the effects of H₂O₂ on coronary arteriolar dilations mediated by different signaling mechanisms were examined before and after intraluminal incubation with H₂O₂ (100 μmol/L) for 60 minutes. Preliminary studies indicated that 60 minutes but not 30 minutes of exposure to 100 μmol/L H₂O₂ was sufficient to inhibit adenosine-induced vasodilation. First, to assess the signaling mechanisms, we used adenosine²² and ionomycin²³ as activators for NO-mediated vasodilation through receptor-dependent and receptor-independent mechanisms, respectively. Second, endothelium-dependent agonists bradykinin and arachidonic acid were used to stimulate vasodilation mediated by the cytochrome P450 monooxygenase²⁴ and the cyclooxygenase-derived prostanoid pathways.²⁴ Third, pinacidil²² and sodium nitroprusside²³ were used as the endothelium-independent agonists to assess the vasodilatory function in response to ATP-sensitive potassium (K_ATP) channel and guanylyl cyclase activation, respectively. Fourth, the involvement of NOS and cyclooxygenase pathways in vasodilation was examined before and after extraluminal incubation of the vessels with the specific inhibitors N^G-nitro-l-arginine methyl ester (L-NAME) (10 μmol/L, 30 minutes)²⁴ and indomethacin (10 μmol/L, 30 minutes),²⁴ respectively. Fifth, the role of endothelial cytochrome P450 monooxygenase pathway in vasodilation was examined by intraluminal incubation of the vessels with its inhibitor miconazole (30 μmol/L, 30 minutes).²⁵ Finally, to rule out time-dependent and nonspecific effects of H₂O₂, the vasodilatory responses were also examined in a separate series of experiments after a 60-minute intraluminal incubation of the vessels with vehicle (PSS).

Specificity of H₂O₂ in Impairing Vascular Function

The specificity of the effect of H₂O₂ on vasodilator responses was examined by intraluminal administration of H₂O₂ solution (100 μmol/L) containing catalase (1000 U/mL). To determine whether the impaired vascular function can be restored after H₂O₂ removal, in another group of vessels the agonist-induced vasodilations were initially studied in the presence of intraluminal H₂O₂ (60-minute incubation) and then re-examined at 30 minutes after replacing the intraluminal H₂O₂ with PSS. To evaluate whether superoxide anions or hydroxyl radicals contribute to the vascular dysfunction elicited by H₂O₂, coronary arteriolar vasodilation to agonists was examined before and after intraluminal administration of H₂O₂ solution containing cell permeable superoxide anion scavenger, polyethylene glycol (PEG)-superoxide dismutase (PEG-SOD) (100 U/mL), or hydroxyl radical production inhibitor, deferoxamine (100 μmol/L).^26,27

Role of L-Arginine and Arginase in Vascular Dysfunction

To determine whether the deficiency of L-arginine contributes to the impaired NO-mediated response, the adenosine-induced and ionomycin-induced vasodilations in the presence of H₂O₂ were further examined after extraluminal incubation of the vessels with L-arginine (3 mmol/L) for 30 minutes. In addition, to determine whether arginase plays a role in vascular dysfunction, the vessels were initially treated with H₂O₂ and then the vasodilations to adenosine and ionomycin were examined after intraluminal incubation of the vessels with arginase inhibitors α-difluoromethylornithine (DFMO) (0.4 mmol/L)¹⁴ or N^ω-hydroxy-nor-l-arginine (nor-NOHA) (0.1 mmol/L)²⁸ for 30 minutes.

RNA Isolation and Reverse-Transcription Polymerase Chain Reaction Analysis

Total RNA was isolated from porcine subepicardial coronary arterioles (3 to 4 vessels, &100 μm diameter, 2 to 3 mm length) after incubation with H₂O₂ (100 μmol/L) or vehicle for 60 minutes at 37°C, based on the protocols described previously.¹⁴ RNA isolated from liver tissue and kidney tissue were used as positive control for arginase I and arginase II, respectively.¹⁴ Using primers specific for arginase I, arginase II, endothelial NOS (eNOS), and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) genes, reverse-transcription polymerase chain reaction was conducted as delineated previously.¹⁴

Western Blot Analysis

Isolated coronary arterioles (4 to 5 vessels per sample, 60 to 120 μm diameter, 3 to 4 mm length) were incubated with H₂O₂ (100 μmol/L) or vehicle for 60 minutes at 37°C. The vessels were then homogenized and prepared for Western blot analysis, as described previously with slight modification.²⁹ Five micrograms of protein per lane were separated by 10% SDS-PAGE under reducing conditions, transferred onto a nitrocellulose membrane, and then allowed to react with a primary antibody for arginase I (1:1000; BD Transduction Laboratories, Lexington, Ky) or β-actin (1:1000; Ambion, Austin, Tex). The antigen-antibody complexes were revealed with horseradish peroxidase-labeled goat anti-rabbit IgG secondary antibody (Alpha Diagnostic International, San Antonio, Tex) by an enhanced chemiluminescence assay (Amersham Pharmacia, Piscataway, NJ).

Immunohistochemical Analysis

Isolated coronary arterioles (&100 to 150 μm in diameter) were pressurized and incubated with intraluminal H₂O₂ (100 μmol/L) or vehicle for 60 minutes. The vessels were removed from the cannulating pipettes and prepared for immunohistochemical analysis, as described previously.¹⁴ Sections (12-μm-thick) were immunolabeled with anti-arginase I antibody or anti-eNOS antibody (1:100, BD Transduction Laboratories) and observed by means of confocal microscopy, as described previously.¹⁴

Effect of Cycloheximide on Vascular Dysfunction

To determine the regulatory level of arginase activation by H₂O₂, the vessels were incubated with cycloheximide (CHX) (20 μg/mL, intraluminal incubation), a protein synthesis inhibitor, for 60 minutes and followed by the treatment of H₂O₂ (100 μmol/L) containing CHX for 60 minutes. The vasodilatory function was then evaluated by adenosine. Finally, the same vessels were prepared for immunohistochemical analysis of arginase I expression as described.

Data Analysis

Diameter changes in response to vasodilator agonists were normalized to the maximum diameter changes in response to 100 μmol/L sodium nitroprusside in an ethylenediaminetetraacetic acid (1 mmol/L) calcium-free PSS and expressed as a percentage of maximal dilation.²³ Statistical comparisons were performed by means of 2-way ANOVA or Student t test. P<0.05 was considered significant. Data are presented as mean±SEM (n=number of vessels).

Results

Effect of H₂O₂ on Endothelium-Dependent and Endothelium-Independent Vasodilations

All isolated coronary arterioles developed a similar level of basal tone (eg, constricted to 62±1% of their maximal diameter) and dilated to adenosine (Figure 1A) and ionomycin (Figure 1B) in a concentration-dependent manner. In the presence of NOS inhibitor L-NAME, the basal vascular tone was slightly increased but did not reach statistical significance (before L-NAME: 62±3% of maximal diameter; after L-NAME: 60±4% of maximal diameter); however, the dilation of these vessels to adenosine and ionomycin was significantly inhibited (Figure 1A and 1B).

• Download figure
• Download PowerPoint

In another set of experiments, the vasodilations to adenosine and ionomycin were examined before and after treating the vessels with intraluminal H₂O₂. The resting vascular tone was not altered by H₂O₂ (before H₂O₂: 62±1%; after H₂O₂: 63±1%), but the dilation to adenosine and ionomycin was significantly inhibited in the same manner as by L-NAME (Figure 1A and 1B). Subsequent administration of L-NAME to the H₂O₂-treated vessels did not further reduce the vasodilator responses (data not shown). Contrarily, H₂O₂ did not affect the endothelium-dependent vasodilation to a cytochrome P450 monooxygenase activator bradykinin (supplemental Figure I, available online at http://atvb.ahajournals.org). Activation of cyclooxygenase pathway by arachidonic acid (10 μmol/L) caused a 77±5% dilation of coronary arterioles; and this dilation was not altered by H₂O₂ (ie, 79±4% dilation, n=5; data not shown). Furthermore, H₂O₂ also had no effect on the vasodilation elicited by a smooth muscle K_ATP channel opener pinacidil (supplemental Figure I) or a guanylyl cyclase activator sodium nitroprusside (supplemental Figure I). To rule out the possible nonspecific endothelial deterioration during the experimental procedure, we examined the effect of luminal incubation of arterioles with a vehicle solution or a low concentration of H₂O₂ (10 μmol/L) for 60 minutes. As shown in supplemental Table I, dose-dependent dilations of adenosine and sodium nitroprusside were not altered by these treatments.

Effect of H₂O₂ Removal and ROS Scavengers on Vascular Dysfunction

The impaired vasodilations to adenosine were not restored after removing H₂O₂ from the lumen for 30 minutes (Figure 2A). In contrast, co-administration of H₂O₂ with catalase, but not superoxide scavenger PEG-SOD, prevented the inhibitory effect of H₂O₂ (Figure 2A). A similar result was observed in the vessels challenged with a receptor-independent NO-mediated vasodilator ionomycin (Figure 2B).

• Download figure
• Download PowerPoint

Role of L-Arginine and Arginase in Vascular Dysfunction

As shown in Figure 3, administration of L-arginine completely restored the H₂O₂-impaired vasodilation to adenosine (Figure 3A) and ionomycin (Figure 3B). Restoration of vasodilations to these agonists was also observed in the vessels treated with an arginase inhibitor DFMO (Figure 3A and 3B). Administration of another specific arginase inhibitor nor-NOHA also restored the impaired vasodilation to adenosine and ionomycin (n=3, data not shown). It should be noted that L-arginine did not enhance NO-mediated dilations to adenosine and ionomycin in control arterioles (n=4, data not shown) as demonstrated in our previous studies.^21,23,24

• Download figure
• Download PowerPoint

Role of Hydroxyl Radicals in Vascular Dysfunction

Because H₂O₂ can be converted to hydroxyl radical in vascular cells, we also examined the effect of deferoxamine, an inhibitor of hydroxyl radical formation, on vascular dysfunction. Deferoxamine prevented the H₂O₂-induced inhibitory effect on adenosine-induced vasodilation (Figure 4) but did not alter the adenosine-induced response of control vessels (Figure 4) or the sodium nitroprusside-induced vasodilation in the presence of H₂O₂ (supplemental Figure I).

• Download figure
• Download PowerPoint

Effect of H₂O₂ on Arginase and eNOS Expression

Reverse-transcription polymerase chain reaction studies showed that coronary arterioles express arginase I (liver tissue was used as a positive control) but not arginase II (kidney tissue was used as a positive control) (Figure 5A). Treating coronary arterioles with H₂O₂ for 60 minutes increased arginase I mRNA by &2-fold without altering eNOS expression (Figure 5B). At the protein level, immunoblotting showed that H₂O₂ treatment also stimulated a 2-fold increase in arginase I protein in arterioles (Figure 6A). For cellular localization of arginase, immunohistochemical analyses indicated that arginase I protein was expressed in the vascular wall with relatively low levels. Treating the vessels with H₂O₂ significantly increased arginase I expression mainly in endothelial cells (Figure 6B). This upregulation was not observed in the vessels pretreated with a protein synthesis inhibitor CHX (Figure 6C). However, the eNOS protein expression was not altered by H₂O₂ (Figure 6C). CHX also protected the adenosine-induced and ionomycin-induced vasodilation from the inhibitory effect of H₂O₂ (supplemental Figure II).

• Download figure
• Download PowerPoint

• Download figure
• Download PowerPoint

Discussion

Previous studies have shown that H₂O₂ can cause vasodilation of small porcine²⁵ and human³⁰ coronary arterioles when it is administered extraluminally. In de-endothelialized pig coronary artery rings, H₂O₂ caused transient contraction and a subsequent relaxation.^31,32 However, there are few studies examining the intraluminal effect of H₂O₂ on arteriolar function, despite the evidence showing that a substantial increase in H₂O₂ was detected in the luminal surface of the vessels subjected to oxidative stress.^33,34 To the best of our knowledge, there is limited information on the endogenous level of H₂O₂ in the intact vascular wall. However, a level from 2.5 μmol/L to 50 μmol/L has been reported in human plasma.^35–39 In general, it is regarded that H₂O₂ at the concentration <50 μmol/L exhibits limited cytotoxicity in many cell types.⁴⁰ It appears that endothelial cells are less susceptible to H₂O₂ because a relatively high concentration of H₂O₂ (ie, >200 μmol/L) is generally required to produce irreversible endothelial barrier dysfunction^41,42 and induce apoptosis.^41,43,44 In the context of neutrophil-endothelial interaction, the H₂O₂ released from activated neutrophils are capable of destroying endothelial cells,^45,46 suggesting a high level of H₂O₂ can be reached at the local circulation during inflammation. However, the direct effect of intraluminal H₂O₂ on vasomotor function has not been systematically examined.

A recent study on KCl precontracted pig coronary arteries indicated that NO-mediated relaxation was attenuated after luminal perfusion with 500 μmol/L H₂O₂;¹⁰ however, the underlying mechanism has not been fully investigated. At the microvascular levels, our present findings indicate that the intraluminal exposure of coronary arterioles to a sublethal level of H₂O₂ (100 μmol/L) leads to a selective impairment of NO-mediated vasodilation independent of endothelial receptors. There are several lines of evidence to support this contention. First, endothelium-dependent vasodilation to NO-mediated agonists adenosine (receptor-dependent)²⁹ and ionomycin (receptor-independent)²³ were inhibited by intraluminal H₂O₂ and L-NAME in an identical fashion. We have previously shown that adenosine-induced dilation in coronary arterioles is mediated by the activation of endothelial NO pathway and smooth muscle K_ATP channels.²² Because vasodilation in response to the activation of K_ATP channel by pinacidil was not altered, the impaired adenosine response appears to be caused by the selective action of H₂O₂ on endothelial NO pathways. Second, the dilations induced by bradykinin (cytochrome P450 pathway²⁴) and arachidonic acid (cyclooxygenase pathway²⁴) were unaltered. Third, the H₂O₂-treated vessels exhibited normal dilation to sodium nitroprusside, an NO donor, which activates smooth muscle guanylyl cyclase. Furthermore, H₂O₂-induced impairment is not caused by the time-dependent deterioration of NO-mediated function, because a 60-minute incubation of the vessels with either vehicle solution or a low concentration of H₂O₂ (10 μmol/L) did not affect the vasodilatory response to NOS activators and to the NO donor sodium nitroprusside. It should be noted that to avoid the potential confounding influences imposed on these microvessels caused by the prolonged experimental protocol, we chose a 60-minute incubation as the cutoff point. Therefore, it is not known whether 10 μmol/L H₂O₂ is sufficient to elicit endothelial dysfunction if a prolonged incubation (ie, >60 minutes) were allowed. Nevertheless, the adverse effect caused by 100 μmol/L H₂O₂ was not extended to the smooth muscle cells because there was no significant change in vascular tone by luminal H₂O₂ and the vessels exhibited normal response to endothelium-independent vasodilators sodium nitroprusside and pinacidil. It appears that vascular smooth muscle function was preserved and the endothelium plays an important role in protecting smooth muscle cells against luminal H₂O₂. This result is in agreement with the finding in large arteries that H₂O₂ does not readily get across the endothelium to exert its cytotoxicity.¹⁰ It is possible that a high level of catalase in the endothelium allows protection of the underlying smooth muscle cells.⁴⁷

Our findings on the improvement of NO-mediated dilation of H₂O₂-treated vessels by L-arginine suggested that a reduction in the availability of NOS precursor was involved in the vascular dysfunction. In terms of L-arginine metabolism, beside NOS isoforms, arginase is another major L-arginine consuming enzyme that converts L-arginine to L-ornithine and urea. To date, 2 arginase isoforms have been identified. Arginase I isoform is expressed most abundantly, but not exclusively, in the liver,⁴⁸ whereas arginase II is expressed in the kidney and many other extrahepatic tissues.⁴⁹ The main function of the hepatic arginase is for ammonia detoxification via the urea cycle.⁴⁸ However, the biological role of the extrahepatic arginase remains obscure. Nonetheless, our previous studies have shown that arginase I can modulate coronary arteriolar function by reducing NO production from NOS.¹⁴ It is plausible that upregulation of arginase in the H₂O₂-treated vessels causes a reduction of L-arginine availability to NOS and thus compromises NO-mediated vasodilation.

Indeed, we found that administration of arginase inhibitors, DFMO or nor-NOHA, effectively restored vasomotor function impaired by H₂O₂, suggesting the involvement of arginase in vascular dysfunction. It is worth noting that we¹⁴ and other laboratories⁵⁰ have previously shown that these inhibitors effectively reduce arginase activity without affecting NOS function. Interestingly, H₂O₂ appears to upregulate the gene and protein expression of arginase I in coronary arteriolar wall, especially in endothelial cells. At the present time, the mechanism underlying the upregulation of arginase remains unclear. However, the induction of arginase protein synthesis appears to be involved in the vascular dysfunction because administration of the protein synthesis inhibitor CHX before H₂O₂ exposure not only inhibited the increased arginase expression but also preserved the eNOS-dependent vasodilation. Although it is somewhat surprising that a significant arginase induction in coronary microvessels can be achieved within such a short period (ie, 60 minutes) of exposure to H₂O₂, previous studies have shown that pharmacological and pathophysiological stimulations can alter the expression of mRNA and/or protein within 60 minutes.^51–54 Interestingly, our recent studies demonstrated that vascular arginase I was upregulated leading to the impaired NO-mediated dilation in the porcine heart subjected to either chronic hypertension (8 weeks)⁵⁵ or an acute episode of ischemia-reperfusion.⁵⁶ Because ROS, including H₂O₂, play an important role in the vascular dysfunction in hypertension⁵ and ischemia-reperfusion injury,⁷ it is speculated that H₂O₂ may be the molecule that triggers the overexpression of vascular arginase and consequently leads to the impairment of NOS-mediated vascular function under these pathophysiological conditions.

In addition to the involvement of arginase, other potential mechanisms such as NOS expression and the production of ROS that could potentially influence NO-mediated vasodilation by H₂O₂ should be considered. Interestingly, H₂O₂ has been shown to increase, rather than decrease, NOS expression in both mRNA and protein levels in cultured endothelial cells.¹² However, these phenomena were not observed in our study in intact coronary arterioles since both mRNA and protein expressions in these vessels did not appear to be affected by H₂O₂. This discrepancy may be related to the differences in experimental model (cultured endothelium versus intact tissue) and/or incubation time (minutes versus hours) for H₂O₂ treatment. Nevertheless, in the present study it is unlikely that the reduced NO-mediated vasodilation by H₂O₂ is mediated by the alteration of NOS expression. Another possible route for reducing NO-mediated vasodilation is through the production of superoxide. It has been shown that H₂O₂ may lead to an increase in other ROS such as superoxide,⁵⁷ which can directly inactivate NO to form peroxynitrite leading to an increased cellular redox stress.⁹ However, treating the vessels with catalase, but not PEG-SOD (a cell-permeable superoxide scavenger), preserved the NO-mediated vascular function (Figure 2). It should be noted that the concentration of PEG-SOD used in the present study is sufficient to eliminate superoxide effect on coronary arterial function.⁵⁸ Furthermore, immunohistochemical studies with superoxide-sensitive dye (dihydroethidium) did not detect an increase in superoxide in the coronary arterioles after H₂O₂ treatment (100 μmol/L, n=3, data not shown). Thus, it is unlikely that superoxide plays a role in arteriolar dysfunction associated with H₂O₂. However, we found that treatment of the vessels with deferoxamine, an inhibitor of hydroxyl radical formation, prevented the H₂O₂-induced impairment of vasodilation to adenosine. The effect appeared to be specific because deferoxamine did not alter the vasodilator response of control vessels to adenosine. Because H₂O₂ can be rapidly converted to hydroxyl radical, these results suggest that formation of this ROS may contribute to the reduction of endothelium-dependent NO-mediated vasodilation. Interestingly, H₂O₂ and hydroxyl radicals have been shown to activate the p38 MAP kinase^59–61 and cAMP⁶² pathways and activation of both p38 and cAMP can cause arginase induction in some cells.^63,64 It is speculated that these ROS-induced signaling cascades may be involved in the upregulation of arginase expression.

In summary, we demonstrate that H₂O₂ inhibits endothelium-dependent NO-mediated dilation of coronary arterioles by upregulating arginase expression. Administration of L-arginine or inhibition of arginase activity restores the impaired vascular function. These results may suggest potential therapeutic interventions targeting L-arginine administration and/or inhibition of arginase induction/activity to improve compromised coronary arteriolar function during oxidative stress.

Footnotes