L-arginine and Arginase Products Potentiate Dexmedetomidine-induced Contractions in the Rat Aorta

DEXMEDETOMIDINE is an α₂-adrenoceptor agonist used as an intravenous sedative/anesthetic agent.¹  The popularity of the compound is based primarily on its bradycardic and sympatholytic effect when used clinically.^2,3  Besides causing sedation and anesthesia, dexmedetomidine also possesses clinically significant vascular effects; the systolic arterial blood pressure response to dexmedetomidine in patients can be quite different from that observed in healthy subjects—where most studies on the hemodynamic effects of the drug have been performed^4,5 —and can be unpredictable, with, for example, an increase during general anesthesia, but a decrease during regional anesthesia.⁶ In vivo, the compound protects the rat myocardium against ischemia/reperfusion injury.⁷ In vitro, dexmedetomidine contracts rat aortae without endothelium⁸  and potentiates the contraction to 40 mM KCl in human gastroepiploic and internal mammary arteries without endothelium.^9,10  Earlier work from the laboratory showed that in rat mesenteric arteries with endothelium, dexmedetomidine causes concentration-dependent, biphasic changes in tension with α₂-adrenergic relaxation followed by α₁-adrenergic endothelium-independent contraction; in aortic rings, the former effect of the compound is masked by the latter.¹¹  Indeed, the relaxation induced by dexmedetomidine is prevented by α₂-adrenergic antagonists, endothelium-dependent and mediated by nitric oxide, whereas the contractile phase is potentiated by the nitric oxide synthase inhibitor N^ω-nitro-l-arginine methyl ester hydrochloride (l-NAME) or by removal of the endothelium, but abolished by the α₁-adrenoceptor antagonist prazosin.¹¹  These findings suggest that under control conditions, the direct α₁-adrenergic effect of dexmedetomidine causing contraction of vascular smooth muscle is blunted by endothelium-derived nitric oxide.

l-arginine the precursor of endothelial nitric oxide,¹²  can reverse impairments of nitric oxide-mediated endothelium-dependent relaxations.¹³  Clinically, acute l-arginine supplementation can improve the release of nitric oxide in patients suffering from endothelial dysfunction.^14–16  Besides nitric oxide synthase, l-arginine is also a substrate for arginases, major enzymes of the urea cycle.¹⁷  In patients with vascular diseases with upregulated arginase activity, there is an associated decrease in nitric oxide-mediated dilatation.^18–20 

The present study was designed to investigate the vascular effects of dexmedetomidine under conditions of combined endothelial impairment and l-arginine supplementation, and to examine the hypothesis that arginases play a role in the observed effects. The results help to better understand the vasoconstrictor effects of dexmedetomidine in patients with endothelial dysfunction, particularly in stressing how the absence of preexisting constrictor tone, combined with endothelial dysfunction and/or l-arginine supplementation, affects the vascular response to dexmedetomidine.

All investigations were approved by the Committee for the Use of Laboratory Animals for Teaching and Research in the University of Hong Kong (Hong Kong, China), in accordance with the Guide for the Care and Use of Laboratory Animals by the National Research Council of United States. Ten weeks-old male Sprague Dawley rats were kept in a temperature-controlled room (21 ± 1°C) with a 12 h light dark cycle. They had free access to standard laboratory chow (LabDiet 5053, USA) and tap water.

Rats were anesthetized intraperitoneally with pentobarbital sodium (70 mg/kg; Ganes Chemicals Inc., USA). The thoracic aortae were isolated and transferred to oxygenated Krebs-Henseleit buffer (control solution) of the following composition: 120 mM NaCl, 25 mM NaHCO₃, 5.5 mM glucose, 4.8 mM KCl, 1.2 mM MgSO₄, 1.2 mM NaH₂PO₄, and 1.25 mM CaCl₂. The isolated blood vessels were dissected free of surrounding fat and connective tissue and cut into six to eight rings (2 to 3 mm in length) for isometric tension measurements. The ring segments were randomly assigned to different treatments and studied in parallel in conventional organ chambers. Each ring was exposed to a different treatment, and no single ring was exposed to more than one treatment.

In some aortae, the endothelium was removed by perfusion with 1 ml 0.5% Triton X-100 solution for 30 s before cutting the rings. Integrity of the smooth muscle layer after endothelial removal was confirmed by normal contractions to 60 mM KCl (data not shown).

The aortic rings were suspended in organ chambers filled with 5 ml control solution maintained at 37°C and aerated with 95% O₂ and 5% CO₂. Each ring was connected to a force transducer (AD Instruments, Australia) for isometric tension recording (PowerLab, AD Instruments). The rings were allowed to equilibrate for 1.5 h to reach their optimal resting tensions of approximately 2.5 g, as determined in preliminary experiments (data not shown).

After exposing the rings twice to 60 mM KCl to obtain a reference contraction, the presence or absence of endothelium was confirmed by adding 1 μM acetylcholine to rings contracted with 1 μM phenylephrine.²¹  The endothelium was considered viable when 80% or more relaxation to acetylcholine was obtained while it was regarded to be removed when relaxations to the muscarinic agonist were abolished. After determining endothelial viability, the rings were washed thoroughly, incubated with vehicle or different compounds including l-NAME (100 μM; Sigma-Aldrich, USA), prazosin (α₁-adrenoceptor antagonist, 1 μM; Sigma-Aldrich), rauwolscine (α₂-adrenoceptor antagonist, 100 nM; Sigma-Aldrich), l-arginine (1 mM; Sigma-Aldrich), (S)-(2-boronethyl)-l-cysteine hydrochloride (BEC, arginase inhibitor, 10 μM; Calbiochem, Germany), N^G-hydroxy-l-arginine (l-NOHA, arginase inhibitor, 10 μM; Calbiochem), urea (100 μM; Sigma-Aldrich) and ornithine (100 μM; Sigma-Aldrich) for 40 min. Cumulative concentrations of dexmedetomidine (10 nM to 10 μM; Abbott Laboratories, USA) were added to quiescent rings; the concentration of dexmedetomidine was increased when the reaction to the previous one had stabilized for at least three minutes, or after ten minutes in the absence of (further) increase in tension in response to the compound. Contractions to dexmedetomidine were expressed as percentages of the reference contraction to 60 mM KCl. At the end of each experiment, the viability of vascular smooth muscle was confirmed by the endothelium-independent relaxation caused by sodium nitroprusside (10 μM; Sigma-Aldrich) during contractions to phenylephrine.

Frozen sections of rat aortae were fixed with cold acetone for 10 min and then washed with water. After fixation, the sections were blocked with normal blocking serum for 30 min, incubated overnight at 4°C, and then at 37°C for one hour, with antibodies against von Willebrand factor (vWF,1:50; Sigma-Aldrich), arginase I (1:50; Santa Cruz, USA) or arginase II (1:50; Santa Cruz) in diluting buffer (phosphate buffered saline [PBS] + 0.01% (v/v) Triton X-100, 0.01% (v/v) Tween 20 and 0.1% (w/v) bovine serum albumin [BSA]). After incubation, the sections were washed with 0.01% Triton X in PBS and incubated with Alexa fluor 594 anti-rabbit IgG (1:50; Invitrogen, USA) and Alexa fluor 488 anti-goat IgG (1:50; Invitrogen) for two hours in the dark at room temperature. Antifade reagent with 4’,6-diamino-2-phenylindole dihydrochloride was added. The frozen sections were then examined under a fluorescence microscope (Eclipse TE300; Nikon, Japan).

Rat aortae were incubated with dexmedetomidine and other compounds for three hours, and the solution bathing the rat aortae was collected. Samples were assayed in duplicate. The amount of urea present was measured using a QuantiChrom Urea assay kit (DIUR-500), following the manufacturer’s (BioAssay Systems, USA) instructions. For each assay, working reagent was prepared freshly by mixing equal volumes of kit reagents A and B. Fifty μl of the samples and standards in serial dilutions were injected into separate wells of a 96-well plate; 200 μl of working reagent were added to each well. The plate was then put in the dark at room temperature for 50 min. The concentration of urea was determined by quantifying the optical density of each well at 450 nm using a microplate reader (MRX, Dynex Technology, USA).

The aortae used for the urea assay were collected and cut into small pieces. The samples were homogenized in lysis buffer (0.02 M Tris-HCl, 1% Triton X-100, 0.15 M NaCl, 1 mM ethylenediamine tetraacetic acid, 1 mM ethylene glycol tetraacetic acid, 2.5 mM sodium pyrophosphate, 1 mM β-glycerophosphate, 1 mM sodium orthovanadate) with 1:1,000 protease inhibitors cocktail. The mixture was centrifuged at 5,000 rpm for five minutes at 4°C. The supernatant was kept at –80°C until use. The protein concentrations of samples were measured using the Bradford assay. Protein samples were mixed with sodium dodecyl sulfate sample buffer and 5% bond breaker. The sample was boiled at 95°C for 10 min and subjected to 10% sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE). Proteins were transferred electrophoretically onto nitrocellulose membranes. The membranes were blocked with 0.1% Tween Tris-buffered saline (TTBS) containing 2% BSA for two hours at room temperature and incubated overnight with primary antibodies (1:300 arginase I or II) in 2% BSA TTBS at 4°C. They then were washed three times for 10 min in TTBS. The membranes were incubated with secondary anti-goat antibodies (1:3,000; Dako, Denmark) at room temperature for two hours, followed by three more washes with TTBS. The probed membranes were visualized by chemiluminescence using an ECL plus Western Detection System (GE Healthcare Life Science, USA), and subsequently exposed to X-ray film (Fuji Photo Film, Germany). To reprobe β-actin, membranes were stripped with TTBS (pH 2) for five minutes. After washing thoroughly, they were blocked and subsequently reprobed with monoclonal β-actin antibodies (1:3,000; Sigma-Aldrich). The optical densities of the protein bands were determined with the computerized program MultiAnalysis (Bio-Rad Laboratories, USA). Densitometric analysis was normalized to the immunoreactive β-actin band.

Contractions to dexmedetomidine are expressed as percentage of the reference contraction to 60 mM KCl obtained at the start of the experiment. The results are shown as mean ± SD, with N representing the number of rats. No blinding method was applied during these experiments; the present experiments were performed in parallel on preparations from the same aorta, thus scientifically permitting correct evaluation of the results without the need for “blind” analysis. Due to the limited number of aortic segments harvested, all combinations of treatment could not be studied at the same time. Some treatments were stopped with a sample size of four when significant differences were obtained. For the treatments that had a more fundamental role in comparison, they were repeated in each experiment so that the responses within the same artery could be compared. In order to avoid selection of data, all data obtained were compared. No missing, lost, or selected data have been excluded from the results. Data were analyzed and curve fitting was performed using the statistical program GraphPad Prism version 5.01 (GraphPad Software Inc., USA). To compare the effects of different inhibitors on dexmedetomidine-induced contractions, areas under the curve, half maximal effective concentrations (EC₅₀), and maximal effects (E_max) were calculated using the statistical program GraphPad Prism version 5.01. Since no significant changes in EC₅₀ to dexmedetomidine were observed (table 1), most of the data of the functional studies are shown as areas under the curve. Data were analyzed by Student’s t test for paired observations, one-way analysis of variances (ANOVA) followed by the Newman-Keuls post hoc test for multiple comparison, or two-way ANOVA followed by the Bonferroni post hoc test. An interim analysis was employed, but no further adjustments were made for multiple comparisons. P values less than 0.05 were considered to indicate statistically significant differences.

Dexmedetomidine did not cause significant increases in tension in control quiescent aortic rings with endothelium (fig. 1, left). However, the compound evoked concentration-dependent contractions in rings with endothelium when treated with the nitric oxide synthase inhibitor l-NAME (100 μM; E_max 50 ± 4%, N = 11) or in preparations without endothelium (E_max 74 ± 5%, N = 12); the response was significantly larger in the latter (fig. 1; table 1). In rings without endothelium, the contractions were significantly larger than those observed in preparations with endothelium treated with l-NAME and were significantly augmented further by the nitric oxide synthase inhibitor (E_max 101 ± 4%; N = 13; fig. 1; table 1). In preparations with endothelium, treated with l-NAME, the contractions to dexmedetomidine were abolished by prazosin (1 μM), but not significantly affected by rauwolscine (100 nM; fig. 2; table 1).

In rings with endothelium, incubation with l-arginine (1 mM) did not unmask contractions to dexmedetomidine, but did significantly augment the concentration-dependent contractions to the compound in preparations treated with l-NAME (100 μM) (E_max 75 ± 3%, N = 4; fig. 3; table 1). The arginase inhibitors BEC (10 μM) and l-NOHA (10 μM) significantly attenuated the contraction to dexmedetomidine of rings with endothelium incubated with l-arginine plus l-NAME (BEC: E_max 16 ± 4%, N = 9 and l-NOHA: E_max 18 ± 4%, N = 5; fig. 3; table 1). In the absence of endothelium, incubation with l-arginine alone resulted in contractions significantly larger than those observed in rings treated with l-NAME; no further potentiation was observed after combined incubation with l-arginine plus l-NAME (fig. 3; table 1).

l-arginine is the common substrate for nitric oxide synthases and arginases. In the urea cycle, l-arginine is hydrolyzed by arginase to produce urea and ornithine.²²  The arginase products, urea (100 μM) and ornithine (100 μM) significantly augmented the concentration-dependent contractions to dexmedetomidine in the presence of l-NAME (fig. 4; table 1). The augmentation was similar to that obtained with l-arginine. The potentiating effects of urea or ornithine alone were not significantly different (fig. 4; table 1).

Arginase consists of two isoforms: arginase I and II.²³  Fluorescence staining revealed that in rat aortae, two isoforms of the enzyme are localized mainly in the endothelium, and are also present in the vascular smooth muscle cells (fig. 5).

To determine whether or not the potentiation caused by l-arginine on the contraction to dexmedetomidine is due to the synthesis of downstream products catalyzed by arginases, urea levels were determined in the solution bathing aortic rings exposed to dexmedetomidine (10 μM), alone or in combination with l-NAME and/or l-arginine for three hours; this duration of treatment is approximately equal to the time needed for the functional studies reported above. The release of urea increased significantly in aortic rings treated with l-NAME or l-arginine (fig. 6).

After measuring the urea production, the preparations were collected for Western blotting to determine the presence of arginase I and II. The two isoforms were present in the rat aortae. In the presence of dexmedetomidine (10 μM), the presence of arginase I was significantly enhanced in rings treated with l-NAME and l-arginine (fig. 7), while the levels of arginase II were reduced in rings incubated with l-NAME (fig. 7).

The present experiments were performed on isolated rings of rat aortae and studied under isometric conditions. This is a standard procedure to pharmacologically evaluate endothelium-dependent and independent responses;^24–27  it permits the parallel study of several preparations from the same blood vessel, while exposing each preparation to only one concentration-response curve of dexmedetomidine, thus reducing biases due to interanimal variability or to sequential administration of the compound.

Dexmedetomidine is used clinically for sedation before and during surgical procedures, as well as in mechanically ventilated patients in the intensive care unit.²⁸  The peak plasma concentration of dexmedetomidine ranges between 0.71 and 1.71 ng/ml (approximately 3 to 7 nM) in patients requiring postoperative sedation in the intensive care unit, and receiving a loading infusion of 2.5 μg/kg over 10 min followed by a maintenance infusion of 0.7 μg · kg^–1 · h^–1.²⁹  With a dose of dexmedetomidine adjusted to maintain critically ill patients in the predefined target sedation range, the plasma concentrations of the drug vary greatly, from an undetectable level to more than 30 ng/ml (approximately 126 nM).^30,31  In this study, dexmedetomidine, at the clinically relevant concentration range (between 10 nM and 10 μM), induced a concentration-dependent contraction in quiescent isolated rat aortic rings, after endothelium-removal or l-NAME-treatment. These findings, in line with observations in either human or rat blood vessels,^8–11,32  show that absence (endothelium-removal) or impaired function of endothelial nitric oxide synthase (incubation with l-NAME) enhances the vasoconstrictor effect of dexmedetomidine. In the presence of prazosin, but not rauwolscine, the contractions to dexmedetomidine were abolished. These results differ from previous studies indicating that both α₁-adrenergic and α₂-adrenergic receptors contribute to dexmedetomidine-induced contraction in the mesenteric artery of the same species.¹¹  This can be explained by the differential distribution of the two adrenoceptor subtypes in individual vascular beds. Indeed, α₁-adrenergic receptors are dominant in the aorta³³  while the α₂-adrenergic subtype is more prominent in smaller arteries.³⁴ 

An unexpected observation in present experiments was that the nitric oxide synthase inhibitor l-NAME³⁵  also augmented the contraction to dexmedetomidine in arteries without endothelium. These results could indicate that the removal of the endothelium had not been complete, which is made unlikely by the disappearance of the relaxation induced by acetylcholine.²¹  An alternative explanation is the presence of neuronal nitric oxide synthase in rat aortae,³⁶  the activity of which would also be inhibited by l-NAME. Moreover, endothelium removal potentiated dexmedetomidine-induced contractions more than nitric oxide synthase inhibition in rings with endothelium, confirming the endothelial nitric oxide synthase-independent production of nitric oxide by the endothelial cells of the rat aorta.³⁷  The present experiments were performed in a high Po₂ environment (95% O₂); if anything, this may underestimate the involvement of nitric oxide.^38,39 

l-arginine is the substrate for nitric oxide production by nitric oxide synthase.¹²  Under most experimental conditions, the acute administration of l-arginine improves endothelial function, thus facilitating vasodilatation.^14–16,22  Exogenous l-arginine at 1 mM (the same concentration used for incubation in the present experiments), causes endothelium-independent relaxations of the rat aorta.²²  However, the present results demonstrate that l-arginine augments, rather than inhibits, dexmedetomidine-induced contraction in aortic rings incubated with l-NAME, under conditions where nitric oxide synthase is not likely to be operative.

l-arginine is a common substrate for nitric oxide synthase and arginase; the latter catalyzes the hydrolysis of l-arginine into urea and ornithine.⁴⁰  Both arginase I and II are present in rat aortae,⁴¹  as confirmed by the present Western blotting data. In this study, the increase in dexmedetomidine-induced contraction caused by l-arginine in aortae treated with l-NAME was abolished by the arginase inhibitors BEC and l-NOHA.⁴²  Although the use of l-NOHA as an inhibitor of arginase is complicated by the fact that it is also a precursor for nitric oxide synthesis,⁴⁰  the similarity of the inhibition observed with BEC permits the conclusion that arginase is involved in the potentiation of the response to dexmedetomidine by l-arginine. Not only did BEC and l-NOHA reverse the l-arginine potentiation, but they also reduced the contraction to dexmedetomidine in the presence of l-NAME. BEC and l-NOHA are competitive inhibitors that do not inhibit nitric oxide synthase at concentrations that inhibit arginases.⁴⁰  However, they reverse tolerance to acetylcholine, indicating an augmentation in nitric oxide bioavailability.⁴²  Moreover, the increase in dexmedetomidine-induced contraction by l-arginine was observed only in preparations with endothelium. This concurs with the present immunofluorescent staining results that showed the presence of the two isoforms, arginase I and II, mainly in the endothelial cells of the rat aorta, although they have been reported by others to also be present in vascular smooth muscle cells.^17,40  The specific subtype of arginase responsible for enhancing the contraction to dexmedetomidine could not be determined as isoform-specific arginase inhibitors are not available. However, the changes in protein presence of the two isoforms observed in the present experiments suggests a major role of arginase I. Although the protein level of arginase II was more prominent than that of arginase I, in rings treated with l-NAME, the level of arginase I was augmented, while that of arginase II was diminished. When incubated with l-NAME and l-arginine the presence of arginase I increased while that of arginase II remained unchanged. Taken into consideration together, these results suggest that the potentiation of contraction to dexmedetomidine by l-arginine in the presence of l-NAME is likely due to the up-regulation of arginase I.

As mentioned in the previous paragraph, urea and ornithine are the major products generated by arginase.⁴⁰  They do not alter nitric oxide production⁴³  and urea does not reverse the inhibition of endothelium-dependent relaxation caused by nitric oxide synthase inhibition.⁴⁴  The metabolism of l-arginine by arginases produces equal amounts of urea and ornithine.⁴⁵  Hence, the same concentration (100 μM) of the two arginase products was used in the present study.^13,44  Higher concentrations were not considered in order to prevent a potential inhibitory effect on arginases.⁴⁶  The two arginase products augmented dexmedetomidine-induced contraction to a comparable extent and shared similar potentiating effects with l-arginine . At least in regards to urea, the endothelial cells of the rat aorta abundantly contain the urea transporter-B permitting its intracellular uptake.⁴⁷  The current findings do not permit further speculation concerning the mechanism underlying the augmentation of the contractions to dexmedetomidine caused by the arginase products. However, the abundancy of arginases in endothelial cells and the strict endothelium-dependency of the response to l-arginine suggest the release of endothelium-derived contracting factors (e.g., vasoconstrictor prostanoids or endothelin-1, the release of which is potentiated when endothelial nitric oxide synthase is dysfunctional²⁰ ) diffusing to the vascular smooth muscle cells and reinforcing the α₁-adrenergic activation caused by the compound. Further investigations are warranted to verify this interpretation. In any case, the present results suggest that the increase in exogenous l-arginine supply, combined with the blockade of nitric oxide synthase by l-NAME which favors arginases, stimulates the activity of the latter, and therefore increases the endogenous production of urea (and ornithine). This interpretation is supported by the present measurements using a quantitative colorimetric urea assay, under conditions mimicking sustained infusions of dexmedetomidine in patients.^29,30  In the presence of the compound, the level of urea was augmented in aortae incubated with l-arginine and l-NAME, compared with those treated with l-NAME only. Ornithine formation was not measured in the present experiments. Nevertheless, increased generation of urea and ornithine, the downstream products of the transformation of l-arginine by arginase, can explain the potentiating effects of l-arginine on dexmedetomidine-induced contraction when endothelial nitric oxide synthase is dysfunctional due to the presence of l-NAME (fig. 8).

In terms of clinical relevance, the contractions observed in quiescent arteries that lack endothelium or have impaired nitric oxide synthase activity suggest that the effect of dexmedetomidine on systemic vascular resistance and blood pressure will depend on the functional state of the endothelium. When using dexmedetomidine, the main challenge is to avoid profound hypotension. Thus, the present study may provide an explanation for the clinical observation that dexmedetomidine seems well-tolerated in patients in whom endothelial function is impaired since the present results indicate that such dysfunction favors the vasoconstrictor effects of the compound.⁴⁸ 

Patients with endothelial dysfunction may receive acute l-arginine supplementation to improve their endothelial function.^14,15  However, the results of this study demonstrate that l-arginine may amplify the vasoconstrictor potential of dexmedetomidine. This amplification by l-arginine depends on the activity of arginase, with the resulting production of urea and ornithine. Therefore, caution may be required when dexmedetomidine is administered to patients with endothelial dysfunction and impaired vascular nitric oxide homeostasis. This could be of particular importance in arteries lined with regenerated endothelial cells, which selectively lose the beneficial activation of endothelial nitric oxide synthase due to stimulation of G_i-coupled cell membrane receptors, among which the α₂-adrenergic receptors activated by dexmedetomidine.^16,20,49  Vascular response data obtained with dexmedetomidine in healthy volunteers⁵⁰  may predict poorly the actual response in such patients. Further in vivo studies and clinical trials are required to demonstrate how coexisting conditions (e.g., intake of medications) that may affect the activities of endothelial nitric oxide synthase and/or arginases can affect responses to dexmedetomidine in patients.

Supported by General Research Fund (773509M) from the Research Grant Council of the Hong Kong Special Administrative Region, China, and by Seed Funding for Basic Research (201111159042) from the University Research Committee of the University of Hong Kong, Hong Kong Special Administrative Region, China.

The authors declare no competing interests.