L-Arginine-dependent relaxation and cGMP
vascular smooth muscle formation
MICHELE E. GOLD, KEITH S. WOOD, RUSSELL E. BYRNS, GEORGETTE M. BUGA, AND LOUIS J. IGNARRO Department of Pharmacology, University of California School of Medicine,
GOLD, MICHELE E., KEITH GEORGETTE M. BUGA, AND
S. WOOD,
E. BYRNS, L-Argininedependent vascular smooth muscle relaxation and cGMP formation. Am. J. Physiol. 259 (Heart Circ. Physiol. 28): H1813Hl821, 1990.-The objective of this study was to ascertain whether endothelium-dependent relaxation and guanosine 3’,5’-cyclic monophosphate (cGMP) formation in bovine pulmonary artery are dependent on L-arginine. Arterial rings responded to acetylcholine and A23187 with increased cGMP accumulation and relaxation and showed resting L-arginine levels of -300 PM. Addition of L-arginine failed to cause relaxation or cGMP accumulation. Arterial rings incubated under tension at 37°C for 24 h showed a three- to fourfold decline in L-arginine levels, and this decline was accompanied by a similar decline in resting cGMP levels as well as complete refractoriness to endothelium-dependent relaxation and cGMP formation in response to acetylcholine and A23187, without alteration of responsiveness to nitric oxide, s-nitrosothiols, or nitroglycerin. The endothelium in 24-h incubated arterial rings was normal morphologically, as assessed by scanning electron microscopy. L-Arginine caused endothelial-dependent relaxation and cGMP formation in L-arginine-depleted rings, which was antagonized by oxyhemoglobin and methylene blue. Bovine aortic endothelial cells grown in L-arginine-deficient medium supplemented with D-arginine during the final 24 h of growth failed to generate endothelium-derived nitric oxide, as assessed by bioassay cascade. L-Canavanine, but not L-lysine or L-ornithine, protected against the decline in L-arginine and cGMP levels and loss of endothelium-dependent relaxation that was characteristic of 24-h incubated arterial rings. The pharmacological properties of L-arginine were shared by L-arginine ethyl ester, L-arginine methyl ester, and L-homoarginine but not N-cu-benzoyl-L-arginine ethyl ester or L-canavanine. These observations indicate that L-arginine or a structural analogue may be obligatory for endothelium-dependent relaxation and cGMP formation.
LOUIS J.
RUSSELL IGNARRO.
bovine pulmonary artery; endothelium-dependent relaxation; L-arginine; nitric oxide; guanosine 3’,5’-cyclic monophosphate
IS substantial experimental evidence that the vascular smooth muscle relaxant activity of the endothelium-derived relaxing factor (EDRF) first described by Furchgott and Zawadzki (3) is largely due to either nitric oxide (NO) or to a chemically unstable nitroso precursor to NO (8). The pharmacological, biochemical, and chemical properties of EDRF and authentic NO or labile s-nitrosothiols are indistinguishable (9). These observations led to studies aimed at elucidating the biosynthetic pathway of endothelium-derived NO. Studies conducted 0363-6135/90
$1.50 Copyright
90024
to date with vascular endothelial cells suggest that L-arginine or a structural analogue of L-arginine serves as an endogenous substrate for a novel enzymatic pathway that is involved with the synthesis of NO. The chemical evidence supporting the view that L-arginine is the substrate is that NG-@‘N]arginine is converted to “‘NO by intact vascular endothelial cells (20, 26). Although the proposed enzymatic pathway for the biosynthesis at NO from endogenous L-arginine involves the simultaneous formation of NO and L-citrulline, the conversion of L-arginine only to L-citrulline but not to NO has been demonstrated in broken endothelial cell preparations (21). Supporting pharmacological evidence includes the observations that certain structural analogues of L-arginine, such as p-methyl-L-arginine, inhibit endothelium-dependent relaxation in vitro (7-10) and elevate the systemic blood pressure in anesthetized rabbits (23) and guinea pigs (1). Preliminary observations from this laboratory indicated that depletion of endogenous levels of L-arginine in isolated arterial rings causes tolerance or refractoriness to endothelium-dependent vasodilators (4). The addition of exogenous L-arginine to such vascular preparations restored endothelium-dependent relaxation. The basis of this in vitro model of tolerance to endotheliumdependent relaxation is that incubation of endotheliumintact rings of bovine pulmonary artery under tension at 37°C in oxygenated Krebs bicarbonate solution for 24 h results in a marked decrease in tissue L-arginine concentrations. The objective of this study was to characterize this model system further and to determine whether endothelium-dependent relaxation and guanosine 3’,5’cyclic monophosphate (cGMP) formation are dependent on endothelium-derived L-arginine or close structural analogues of L-arginine. MATERIALS
THERE
Los Angeles, California
AND
METHODS
Chemicals and solutions. Phenylephrine hydrochloride, acetylcholine chloride, bradykinin triacetate, A23187, indomethacin, hemoglobin (human), methylene blue, L-arginine, D-arginine, L-arginine ethyl ester dihydrochloride, L-arginine methyl ester dihydrochloride, L-homoarginine hydrochloride, N-a-benzoyl-L-arginine ethyl ester hydrochloride, L-canavanine, L-lysine, L-ornithine hydrochloride, and Dowex-50W-H+ (200-400 mesh, 8% cross-linkage) were purchased from Sigma Chemical. Nitroglycerin (10% wt/wt triturate in lactose) was a
0 1990 the American
Physiological
Society
H1813
Downloaded from www.physiology.org/journal/ajpheart by ${individualUser.givenNames} ${individualUser.surname} (139.184.014.150) on July 30, 2018. Copyright © 1990 American Physiological Society. All rights reserved.
Hl814
L-ARGININE-DEPENDENT
RELAXATION
generous gift from ICI Americas. U 46619, provided by Upjohn, was dissolved and stored in absolute ethanol (10 mg/ml). Solutions of hygroscopic acetylcholine chloride were prepared in distilled water, divided into aliquots, and stored at -20°C. All amino acids were soluble and prepared in distilled water fresh daily. Oxyhemoglobin was prepared from hemoglobin as described previously (13). s-Nitroso-N-acetylpenicillamine (SNAP) was prepared as described previously (l7), and fresh aqueous solutions (1 mM) were prepared in ice-cold distilled water and discarded after 4 h because of chemical instability. NO gas was obtained and aqueous solutions were prepared as described previously (11). Caution was exercised in the handling of NO gas and solutions so that 0, was excluded until aliquots of the NO solutions were delivered into the bath chambers. Under such conditions the mean effective concentration (EC& values for NO in relaxing precontracted rings of bovine pulmonary artery were approximately 10 nM. Krebs bicarbonate solution consisted of (in mM) 118 NaCl, 4.7 KCl, 1.5 CaCIZ, 25 NaHC03, 1.2 MgSO,, 1.2 KH2P04, 11 glucose, and 0.023 Naz EDTA. Preparation of arterial rings. Bovine lungs were obtained from an abattoir located nearby and transported to the laboratory. The second intrapulmonary arterial branch (4-5 mm OD) extending into the larger lobe was rapidly isolated, cleaned of parenchyma, fat, and adhering connective tissue, and placed in cold preoxygenated Krebs bicarbonate solution. Arterial segments were sliced into rings (4 mm wide) as described previously (12, 14). Rings prepared in this manner possessedan intact endothelium as assessedpharmacologically by 80-100% relaxation in response to 0.1-l ,uM acetylcholine and as assessedhistologically after silver staining. In some arterial rings the endothelial layer was completely removed without appreciable damage to the underlying smooth muscle by gently inverting the rings, rubbing the intimal surface lightly with moistened filter paper for 3 min, and returning the rings to their original everted configuration. Endothelium-denuded arterial rings contracted in response to acetylcholine, failed to respond or contracted slightly to A23187, and had three- to fourfold lower basal levels of cGMP (8-14 pmol/g) than did unrubbed rings (35-50 pmol/g). Recording of muscle tension. Arterial rings were mounted by means of fine Nichrome wires in jacketed 25-ml drop-away chambers (Metro Scientific) containing Krebs bicarbonate solution (37°C) gassedwith 95% O,-5% CO, (12). The upper Nichrome wire of each ring was attached to a force-displacement transducer (Grass FT03C), and changes in isometric force were recorded on a Grass polygraph (model 79D). Length-tension relationships were determined initially for unrubbed and rubbed arterial rings. Tension was adjusted to the optimal length for maximal isometric contractions to potassium by progressively stretching the rings and repeatedly obtaining contractile responses to 80 mM KC1 with 15 min of equilibration in between each contractile response (12). Optimal tensions did not vary significantly as a function of intimal rubbing. The optimal tensions determined in these initial experiments were used in all subsequent experiments. Optimal tensions and maximal
AND
cGMP
FORMATION
contractile tensions developed in response to KC1 at optimal lengths were 6 and 20-24 g, respectively, for arterial rings. Rings were routinely depolarized by addition of 120 mM KC1 following 2 h of equilibration at optimal tension and then washed and allowed to equilibrate for 45 min before initiating any given protocol (12, 14). This procedure increases and stabilizes any subsequent submaximal precontractile responses to phenylephrine, presumably by loading the smooth muscle cells with calcium. This procedure has been employed routinely for bovine pulmonary vessels in this laboratory. Submaximal precontractile responses to 1 PM-10 PM phenylephrine ranged from 60 to 80% of maximal contractions produced by 120 mM KCl. Determination of cGA4P Leuels.cGMP determinations were made in arterial rings that had been equilibrated under tension and depolarized with KCl, as was done routinely for all ring experiments. Tone was monitored until the time of quick freezing. The use of rapid dropaway bath chambers, quick freezing, tissue extraction for cGMP determinations, and radioimmunoassay procedures was described previously (12). None of the test agents added to bath chambers interfered with the radioimmunoassay procedure. Recoveries of standard quantities of added cGMP were determined periodically and were 92-104%. Preparation and use of bovine aortic endothelial cells grown on microcarrier beads. Endothelial cells were harvested from two or three bovine thoracic aortas by collagenase-incubation techniques (13). After the cells were washed and resuspended in Dulbecco’s modified essential medium containing 10% fetal calf serum and antibiotics, the cells were incubated in 75-cm2 culture flasks for 24 h to allow cell adherence. Nonadherent cells were removed by washing and the cultures were continued for an additional 48 h. The adherent cells were removed from the flasks with 0.1% trypsin-EDTA in phosphatebuffered saline at 37”C, washed, and incubated with 0.5 g of Cytodex 3 microcarrier beads coated with a fine film of collagen (Pharmacia). The beads, cells, and medium were transferred into a stirrer flask and stirred (20 rpm) for 1 min and left undisturbed for 29 min; this cycle was repeated for 4 h, after which time the mixture was stirred while incubating until confluency was attained (2-3 days, as assessedby phase-contrast microscopy). Multiple passaged cells (between 4 and 8 passes) were used in these studies. One milliliter of microcarrier beads containing lo-30 x lo6 cells was packed in specially designed disposable plastic columns (0.4 cm diam) maintained at 37°C. The column was continuously perfused with oxygenated Krebs bicarbonate solution at 37°C at a rate of 5 ml/min. Bioassay cascadeusing endothelial cells. A modification (10) of the technique developed by Vane in 1964 was used (28). Briefly, a small column packed with aortic endothelial cells grown on microcarrier beads was perfused with oxygenated Krebs bicarbonate solution as described above. The perfusate was allowed to superfuse three isolated, helically cut, precontracted strips (2.5 cm long, 4 mm wide) of endothelium-denuded artery arranged in a cascade in which each strip was separated in flow time by 3 s. A second superfusion line was also
Downloaded from www.physiology.org/journal/ajpheart by ${individualUser.givenNames} ${individualUser.surname} (139.184.014.150) on July 30, 2018. Copyright © 1990 American Physiological Society. All rights reserved.
L-ARGININE-DEPENDENT
RELAXATION
positioned over the cascade (5 ml/min total flow at 37°C). Indomethacin (10 PM) was present in both perfusion and superfusion media to prevent formation of prostaglandins that could affect smooth muscle tone. Arterial strips were equilibrated by superfusion with Krebs bicarbonate solution for 2 h before initiating any protocol, at which time the strips were precontracted with a mixture of phenylephrine (50 PM) and U 46619 (50 nM). The purpose for using this mixture was that contractile responses were well maintai .ned and reproducible, whereas phenylephrine alone could not maintain tone and U 46619 alone caused frequent oscillatory responses. Responses were recorded as changes in length of the smooth muscle preparations, measured with auxotonic levers attached to transducers (model 386, Harvard Apparatus), and expressed as centimeters in Fig. 5 (10). Nitroglycerin, a stable end .othelium-independent relaxant, was superfused over the strips to standardize the preparations (6, Tissue extraction and determination of arginine levels. Arterial rings (3 rings totaling 30-50 mg/sample) were homogenized in 1 ml of 6% trichloroacetic acid, allowed to settle for 20-30 min, and centrifuged at 10,000 g for 15 min. The supernatant was extracted with water-saturated diethyl ether, and the ether layer was discarded. The ether remaining in the aqueous samples was driven off by heating at 70°C in a water bath for 5 min. Samples were lyophilized and reconstituted in 4 ml of 0.1 M citrate buffer, pH 5.3, and then chromatographed on 0.7 X 6 cm columns of Dowex-Na+ (prepared from Dowex-50W-H+, 200-400 mesh, 8% cross-linkage). Columns were washed with 20 ml of citrate buffer to remove ci .trulline and ornithine (which could be further separated and assayed if desired) and then eluted with 4 ml of 0.2 N NaOH, which eluted the arginine quantitatively. Recoveries of Acetylcholine
l
-
0 A23187
AND
cGMP
Hl815
FORMATION
standard quantities of added L-arginine were determined in each experiment and were 95-100%. Thus values were not corrected for arginine recovery. Samples were assayed for arginine by a modification of the Sakaguchi reaction (4), which is a selective and sensitive spectrophotometric procedure for detecting and quantifying arginine. In each experiment a separate standard curve for authentic L-arginine was conducted using a range of arginine concentrations that bracketed those found in the tissue samples. Arginase assay. Arginase activity was determined by measuring the disappearance of arginine from enzyme reaction mixtures where aliquots were removed and assayed for arginine by a modification of the Sakaguchi reaction procedure described above. Arginase (4 pug protein) was preincubated for 10 min at 37°C in 1 ml of 50 mM glycine-NaC1 buffer, pH 9.5, containing 0.1 mM MnCIZ. Enzymatic reactions were initiated by addition of L-arginine (0.3 mM final concentration) and samples were further incubated for 8 min. Chemical agents tested for inhibitory activity against arginase were added to preincubation reaction mixtures. Arginase activity was terminated by addition of 1.9 ml of ice-cold 0.2 N NaOH to reaction mixtures, and samples were kept on ice. Aliquots of 0.1 ml were removed and assayed for arginine as described above. Scanning electron microscopy. Isolated arterial rings were prepared for scanning electron microscopy using standard techniques as described previously (2, 16). Briefly, after myographic protocols, arterial rings were immediately fixed in 4% glutaraldehyde. Arterial rings were subsequently postfixed in osmium tetroxide and l
Day 1
-
NO
100
Day 1 --- Day 2
100
-z u ez
25 0
25
-25 -50
0 I
I
-8
I
-7 log Molar
I
I
J
-6
Concentration
FIG. 1. Responses of arterial rings to acetylcholine and A23187 before and after 24 h incubation. Rings were submaximally precontracted with phenylephrine (l-10 PM). Indomethacin (10 pM) was present in the bathing medium. At peak contractile responses cumulative additions of acetylcholine or A23187 were made as indicated. Concentrations represent final bath concentrations. Day 1 and day 2 signify, respectively, freshly isolated rings and rings incubated under tension at 37°C for 24 h. Negative numbers on y-axis signify percent contraction. Values are means of: SE using 24 rings (for each test condition) from 4 animals (6 rings per animal). Values for rings on day 2 are significantly different (P < 0.05) from values for rings on day 1.
I
-9
I1
I
I
I
-7
-8
log Molar
I
-6
I
IJ
-5
Concentration
FIG. 2. Responses of arterial rings to NO and s-nitroso-N-acetylpenicillamine (SNAP) before and after 24 h incubation. Endotheliumdenuded rings were submaximally precontracted with phenylephrine (l-10 ,uM). Indomethacin (10 ,uM) was present in bathing medium. At peak contractile responses cumulative additions of NO or SNAP were made as indicated. Concentrations represent final bath concentrations. Day I and day 2 signify, respectively, freshly isolated rings and rings incubated under tension at 37°C for 24 h. Values are means k SE using 24 rings (for each test condition) from 4 animals (6 rings per animal). Values for rings on day 2 are not significantly different (P > 0.1) from values for rings on day 1.
Downloaded from www.physiology.org/journal/ajpheart by ${individualUser.givenNames} ${individualUser.surname} (139.184.014.150) on July 30, 2018. Copyright © 1990 American Physiological Society. All rights reserved.
H1816
L-ARGININE-DEPENDENT
RELAXATION
AND
cGMP
FORMATION
then coated with a gold-palladium film. Samples were examined and photographed in an ETEC Auto-Scan electron microscope. Calculations and statistical analysis. Relaxation or contraction of arterial rings was measured as the decrease or increase, respectively, in tension below or above the elevated tension elicited by submaximal precontraction with phenylephrine. Where shown, values in the figures are expressed as means + SE and represent unpaired data. Comparisons were made using either Duncan’s multiple range test (29), where comparisons with a common control were made (Figs. 4 and 8), or the Student’s t test for unpaired values for the other comparisons. The level of statistically significant difference was P < 0.05. RESULTS
Abolition of endothelium-dependent arterial relaxation by 24 h incubation. Incubation of arterial rings mounted
under tension at 37°C in oxygenated Krebs bicarbonate solution for 24 h resulted in the development of complete tolerance or refractoriness to the relaxant effects of acetylcholine and A23187 (Fig. 1). In contrast, refractoriness did not develop, under similar conditions, to the relaxant effects of either NO or the labile s-nitrosothiol SNAP (Fig. 2). Scanning electron microscopy revealed that no apparent structural or morphological damage to the endothelium was inflicted by incubating arterial rings for 24 h under the defined conditions (Fig. 3). Decline in arterial L-arginine levels after 24 h incubation. Basal or resting levels of L-arginine in freshly
prepared rings of bovine pulmonary artery were -300 PM. Incubation of arterial rings under tension at 37°C in oxygenated Krebs bicarbonate solution for 24 h resulted in a marked decline in L-arginine levels to
vascular smooth muscle formation
MICHELE E. GOLD, KEITH S. WOOD, RUSSELL E. BYRNS, GEORGETTE M. BUGA, AND LOUIS J. IGNARRO Department of Pharmacology, University of California School of Medicine,
GOLD, MICHELE E., KEITH GEORGETTE M. BUGA, AND
S. WOOD,
E. BYRNS, L-Argininedependent vascular smooth muscle relaxation and cGMP formation. Am. J. Physiol. 259 (Heart Circ. Physiol. 28): H1813Hl821, 1990.-The objective of this study was to ascertain whether endothelium-dependent relaxation and guanosine 3’,5’-cyclic monophosphate (cGMP) formation in bovine pulmonary artery are dependent on L-arginine. Arterial rings responded to acetylcholine and A23187 with increased cGMP accumulation and relaxation and showed resting L-arginine levels of -300 PM. Addition of L-arginine failed to cause relaxation or cGMP accumulation. Arterial rings incubated under tension at 37°C for 24 h showed a three- to fourfold decline in L-arginine levels, and this decline was accompanied by a similar decline in resting cGMP levels as well as complete refractoriness to endothelium-dependent relaxation and cGMP formation in response to acetylcholine and A23187, without alteration of responsiveness to nitric oxide, s-nitrosothiols, or nitroglycerin. The endothelium in 24-h incubated arterial rings was normal morphologically, as assessed by scanning electron microscopy. L-Arginine caused endothelial-dependent relaxation and cGMP formation in L-arginine-depleted rings, which was antagonized by oxyhemoglobin and methylene blue. Bovine aortic endothelial cells grown in L-arginine-deficient medium supplemented with D-arginine during the final 24 h of growth failed to generate endothelium-derived nitric oxide, as assessed by bioassay cascade. L-Canavanine, but not L-lysine or L-ornithine, protected against the decline in L-arginine and cGMP levels and loss of endothelium-dependent relaxation that was characteristic of 24-h incubated arterial rings. The pharmacological properties of L-arginine were shared by L-arginine ethyl ester, L-arginine methyl ester, and L-homoarginine but not N-cu-benzoyl-L-arginine ethyl ester or L-canavanine. These observations indicate that L-arginine or a structural analogue may be obligatory for endothelium-dependent relaxation and cGMP formation.
LOUIS J.
RUSSELL IGNARRO.
bovine pulmonary artery; endothelium-dependent relaxation; L-arginine; nitric oxide; guanosine 3’,5’-cyclic monophosphate
IS substantial experimental evidence that the vascular smooth muscle relaxant activity of the endothelium-derived relaxing factor (EDRF) first described by Furchgott and Zawadzki (3) is largely due to either nitric oxide (NO) or to a chemically unstable nitroso precursor to NO (8). The pharmacological, biochemical, and chemical properties of EDRF and authentic NO or labile s-nitrosothiols are indistinguishable (9). These observations led to studies aimed at elucidating the biosynthetic pathway of endothelium-derived NO. Studies conducted 0363-6135/90
$1.50 Copyright
90024
to date with vascular endothelial cells suggest that L-arginine or a structural analogue of L-arginine serves as an endogenous substrate for a novel enzymatic pathway that is involved with the synthesis of NO. The chemical evidence supporting the view that L-arginine is the substrate is that NG-@‘N]arginine is converted to “‘NO by intact vascular endothelial cells (20, 26). Although the proposed enzymatic pathway for the biosynthesis at NO from endogenous L-arginine involves the simultaneous formation of NO and L-citrulline, the conversion of L-arginine only to L-citrulline but not to NO has been demonstrated in broken endothelial cell preparations (21). Supporting pharmacological evidence includes the observations that certain structural analogues of L-arginine, such as p-methyl-L-arginine, inhibit endothelium-dependent relaxation in vitro (7-10) and elevate the systemic blood pressure in anesthetized rabbits (23) and guinea pigs (1). Preliminary observations from this laboratory indicated that depletion of endogenous levels of L-arginine in isolated arterial rings causes tolerance or refractoriness to endothelium-dependent vasodilators (4). The addition of exogenous L-arginine to such vascular preparations restored endothelium-dependent relaxation. The basis of this in vitro model of tolerance to endotheliumdependent relaxation is that incubation of endotheliumintact rings of bovine pulmonary artery under tension at 37°C in oxygenated Krebs bicarbonate solution for 24 h results in a marked decrease in tissue L-arginine concentrations. The objective of this study was to characterize this model system further and to determine whether endothelium-dependent relaxation and guanosine 3’,5’cyclic monophosphate (cGMP) formation are dependent on endothelium-derived L-arginine or close structural analogues of L-arginine. MATERIALS
THERE
Los Angeles, California
AND
METHODS
Chemicals and solutions. Phenylephrine hydrochloride, acetylcholine chloride, bradykinin triacetate, A23187, indomethacin, hemoglobin (human), methylene blue, L-arginine, D-arginine, L-arginine ethyl ester dihydrochloride, L-arginine methyl ester dihydrochloride, L-homoarginine hydrochloride, N-a-benzoyl-L-arginine ethyl ester hydrochloride, L-canavanine, L-lysine, L-ornithine hydrochloride, and Dowex-50W-H+ (200-400 mesh, 8% cross-linkage) were purchased from Sigma Chemical. Nitroglycerin (10% wt/wt triturate in lactose) was a
0 1990 the American
Physiological
Society
H1813
Downloaded from www.physiology.org/journal/ajpheart by ${individualUser.givenNames} ${individualUser.surname} (139.184.014.150) on July 30, 2018. Copyright © 1990 American Physiological Society. All rights reserved.
Hl814
L-ARGININE-DEPENDENT
RELAXATION
generous gift from ICI Americas. U 46619, provided by Upjohn, was dissolved and stored in absolute ethanol (10 mg/ml). Solutions of hygroscopic acetylcholine chloride were prepared in distilled water, divided into aliquots, and stored at -20°C. All amino acids were soluble and prepared in distilled water fresh daily. Oxyhemoglobin was prepared from hemoglobin as described previously (13). s-Nitroso-N-acetylpenicillamine (SNAP) was prepared as described previously (l7), and fresh aqueous solutions (1 mM) were prepared in ice-cold distilled water and discarded after 4 h because of chemical instability. NO gas was obtained and aqueous solutions were prepared as described previously (11). Caution was exercised in the handling of NO gas and solutions so that 0, was excluded until aliquots of the NO solutions were delivered into the bath chambers. Under such conditions the mean effective concentration (EC& values for NO in relaxing precontracted rings of bovine pulmonary artery were approximately 10 nM. Krebs bicarbonate solution consisted of (in mM) 118 NaCl, 4.7 KCl, 1.5 CaCIZ, 25 NaHC03, 1.2 MgSO,, 1.2 KH2P04, 11 glucose, and 0.023 Naz EDTA. Preparation of arterial rings. Bovine lungs were obtained from an abattoir located nearby and transported to the laboratory. The second intrapulmonary arterial branch (4-5 mm OD) extending into the larger lobe was rapidly isolated, cleaned of parenchyma, fat, and adhering connective tissue, and placed in cold preoxygenated Krebs bicarbonate solution. Arterial segments were sliced into rings (4 mm wide) as described previously (12, 14). Rings prepared in this manner possessedan intact endothelium as assessedpharmacologically by 80-100% relaxation in response to 0.1-l ,uM acetylcholine and as assessedhistologically after silver staining. In some arterial rings the endothelial layer was completely removed without appreciable damage to the underlying smooth muscle by gently inverting the rings, rubbing the intimal surface lightly with moistened filter paper for 3 min, and returning the rings to their original everted configuration. Endothelium-denuded arterial rings contracted in response to acetylcholine, failed to respond or contracted slightly to A23187, and had three- to fourfold lower basal levels of cGMP (8-14 pmol/g) than did unrubbed rings (35-50 pmol/g). Recording of muscle tension. Arterial rings were mounted by means of fine Nichrome wires in jacketed 25-ml drop-away chambers (Metro Scientific) containing Krebs bicarbonate solution (37°C) gassedwith 95% O,-5% CO, (12). The upper Nichrome wire of each ring was attached to a force-displacement transducer (Grass FT03C), and changes in isometric force were recorded on a Grass polygraph (model 79D). Length-tension relationships were determined initially for unrubbed and rubbed arterial rings. Tension was adjusted to the optimal length for maximal isometric contractions to potassium by progressively stretching the rings and repeatedly obtaining contractile responses to 80 mM KC1 with 15 min of equilibration in between each contractile response (12). Optimal tensions did not vary significantly as a function of intimal rubbing. The optimal tensions determined in these initial experiments were used in all subsequent experiments. Optimal tensions and maximal
AND
cGMP
FORMATION
contractile tensions developed in response to KC1 at optimal lengths were 6 and 20-24 g, respectively, for arterial rings. Rings were routinely depolarized by addition of 120 mM KC1 following 2 h of equilibration at optimal tension and then washed and allowed to equilibrate for 45 min before initiating any given protocol (12, 14). This procedure increases and stabilizes any subsequent submaximal precontractile responses to phenylephrine, presumably by loading the smooth muscle cells with calcium. This procedure has been employed routinely for bovine pulmonary vessels in this laboratory. Submaximal precontractile responses to 1 PM-10 PM phenylephrine ranged from 60 to 80% of maximal contractions produced by 120 mM KCl. Determination of cGA4P Leuels.cGMP determinations were made in arterial rings that had been equilibrated under tension and depolarized with KCl, as was done routinely for all ring experiments. Tone was monitored until the time of quick freezing. The use of rapid dropaway bath chambers, quick freezing, tissue extraction for cGMP determinations, and radioimmunoassay procedures was described previously (12). None of the test agents added to bath chambers interfered with the radioimmunoassay procedure. Recoveries of standard quantities of added cGMP were determined periodically and were 92-104%. Preparation and use of bovine aortic endothelial cells grown on microcarrier beads. Endothelial cells were harvested from two or three bovine thoracic aortas by collagenase-incubation techniques (13). After the cells were washed and resuspended in Dulbecco’s modified essential medium containing 10% fetal calf serum and antibiotics, the cells were incubated in 75-cm2 culture flasks for 24 h to allow cell adherence. Nonadherent cells were removed by washing and the cultures were continued for an additional 48 h. The adherent cells were removed from the flasks with 0.1% trypsin-EDTA in phosphatebuffered saline at 37”C, washed, and incubated with 0.5 g of Cytodex 3 microcarrier beads coated with a fine film of collagen (Pharmacia). The beads, cells, and medium were transferred into a stirrer flask and stirred (20 rpm) for 1 min and left undisturbed for 29 min; this cycle was repeated for 4 h, after which time the mixture was stirred while incubating until confluency was attained (2-3 days, as assessedby phase-contrast microscopy). Multiple passaged cells (between 4 and 8 passes) were used in these studies. One milliliter of microcarrier beads containing lo-30 x lo6 cells was packed in specially designed disposable plastic columns (0.4 cm diam) maintained at 37°C. The column was continuously perfused with oxygenated Krebs bicarbonate solution at 37°C at a rate of 5 ml/min. Bioassay cascadeusing endothelial cells. A modification (10) of the technique developed by Vane in 1964 was used (28). Briefly, a small column packed with aortic endothelial cells grown on microcarrier beads was perfused with oxygenated Krebs bicarbonate solution as described above. The perfusate was allowed to superfuse three isolated, helically cut, precontracted strips (2.5 cm long, 4 mm wide) of endothelium-denuded artery arranged in a cascade in which each strip was separated in flow time by 3 s. A second superfusion line was also
Downloaded from www.physiology.org/journal/ajpheart by ${individualUser.givenNames} ${individualUser.surname} (139.184.014.150) on July 30, 2018. Copyright © 1990 American Physiological Society. All rights reserved.
L-ARGININE-DEPENDENT
RELAXATION
positioned over the cascade (5 ml/min total flow at 37°C). Indomethacin (10 PM) was present in both perfusion and superfusion media to prevent formation of prostaglandins that could affect smooth muscle tone. Arterial strips were equilibrated by superfusion with Krebs bicarbonate solution for 2 h before initiating any protocol, at which time the strips were precontracted with a mixture of phenylephrine (50 PM) and U 46619 (50 nM). The purpose for using this mixture was that contractile responses were well maintai .ned and reproducible, whereas phenylephrine alone could not maintain tone and U 46619 alone caused frequent oscillatory responses. Responses were recorded as changes in length of the smooth muscle preparations, measured with auxotonic levers attached to transducers (model 386, Harvard Apparatus), and expressed as centimeters in Fig. 5 (10). Nitroglycerin, a stable end .othelium-independent relaxant, was superfused over the strips to standardize the preparations (6, Tissue extraction and determination of arginine levels. Arterial rings (3 rings totaling 30-50 mg/sample) were homogenized in 1 ml of 6% trichloroacetic acid, allowed to settle for 20-30 min, and centrifuged at 10,000 g for 15 min. The supernatant was extracted with water-saturated diethyl ether, and the ether layer was discarded. The ether remaining in the aqueous samples was driven off by heating at 70°C in a water bath for 5 min. Samples were lyophilized and reconstituted in 4 ml of 0.1 M citrate buffer, pH 5.3, and then chromatographed on 0.7 X 6 cm columns of Dowex-Na+ (prepared from Dowex-50W-H+, 200-400 mesh, 8% cross-linkage). Columns were washed with 20 ml of citrate buffer to remove ci .trulline and ornithine (which could be further separated and assayed if desired) and then eluted with 4 ml of 0.2 N NaOH, which eluted the arginine quantitatively. Recoveries of Acetylcholine
l
-
0 A23187
AND
cGMP
Hl815
FORMATION
standard quantities of added L-arginine were determined in each experiment and were 95-100%. Thus values were not corrected for arginine recovery. Samples were assayed for arginine by a modification of the Sakaguchi reaction (4), which is a selective and sensitive spectrophotometric procedure for detecting and quantifying arginine. In each experiment a separate standard curve for authentic L-arginine was conducted using a range of arginine concentrations that bracketed those found in the tissue samples. Arginase assay. Arginase activity was determined by measuring the disappearance of arginine from enzyme reaction mixtures where aliquots were removed and assayed for arginine by a modification of the Sakaguchi reaction procedure described above. Arginase (4 pug protein) was preincubated for 10 min at 37°C in 1 ml of 50 mM glycine-NaC1 buffer, pH 9.5, containing 0.1 mM MnCIZ. Enzymatic reactions were initiated by addition of L-arginine (0.3 mM final concentration) and samples were further incubated for 8 min. Chemical agents tested for inhibitory activity against arginase were added to preincubation reaction mixtures. Arginase activity was terminated by addition of 1.9 ml of ice-cold 0.2 N NaOH to reaction mixtures, and samples were kept on ice. Aliquots of 0.1 ml were removed and assayed for arginine as described above. Scanning electron microscopy. Isolated arterial rings were prepared for scanning electron microscopy using standard techniques as described previously (2, 16). Briefly, after myographic protocols, arterial rings were immediately fixed in 4% glutaraldehyde. Arterial rings were subsequently postfixed in osmium tetroxide and l
Day 1
-
NO
100
Day 1 --- Day 2
100
-z u ez
25 0
25
-25 -50
0 I
I
-8
I
-7 log Molar
I
I
J
-6
Concentration
FIG. 1. Responses of arterial rings to acetylcholine and A23187 before and after 24 h incubation. Rings were submaximally precontracted with phenylephrine (l-10 PM). Indomethacin (10 pM) was present in the bathing medium. At peak contractile responses cumulative additions of acetylcholine or A23187 were made as indicated. Concentrations represent final bath concentrations. Day 1 and day 2 signify, respectively, freshly isolated rings and rings incubated under tension at 37°C for 24 h. Negative numbers on y-axis signify percent contraction. Values are means of: SE using 24 rings (for each test condition) from 4 animals (6 rings per animal). Values for rings on day 2 are significantly different (P < 0.05) from values for rings on day 1.
I
-9
I1
I
I
I
-7
-8
log Molar
I
-6
I
IJ
-5
Concentration
FIG. 2. Responses of arterial rings to NO and s-nitroso-N-acetylpenicillamine (SNAP) before and after 24 h incubation. Endotheliumdenuded rings were submaximally precontracted with phenylephrine (l-10 ,uM). Indomethacin (10 ,uM) was present in bathing medium. At peak contractile responses cumulative additions of NO or SNAP were made as indicated. Concentrations represent final bath concentrations. Day I and day 2 signify, respectively, freshly isolated rings and rings incubated under tension at 37°C for 24 h. Values are means k SE using 24 rings (for each test condition) from 4 animals (6 rings per animal). Values for rings on day 2 are not significantly different (P > 0.1) from values for rings on day 1.
Downloaded from www.physiology.org/journal/ajpheart by ${individualUser.givenNames} ${individualUser.surname} (139.184.014.150) on July 30, 2018. Copyright © 1990 American Physiological Society. All rights reserved.
H1816
L-ARGININE-DEPENDENT
RELAXATION
AND
cGMP
FORMATION
then coated with a gold-palladium film. Samples were examined and photographed in an ETEC Auto-Scan electron microscope. Calculations and statistical analysis. Relaxation or contraction of arterial rings was measured as the decrease or increase, respectively, in tension below or above the elevated tension elicited by submaximal precontraction with phenylephrine. Where shown, values in the figures are expressed as means + SE and represent unpaired data. Comparisons were made using either Duncan’s multiple range test (29), where comparisons with a common control were made (Figs. 4 and 8), or the Student’s t test for unpaired values for the other comparisons. The level of statistically significant difference was P < 0.05. RESULTS
Abolition of endothelium-dependent arterial relaxation by 24 h incubation. Incubation of arterial rings mounted
under tension at 37°C in oxygenated Krebs bicarbonate solution for 24 h resulted in the development of complete tolerance or refractoriness to the relaxant effects of acetylcholine and A23187 (Fig. 1). In contrast, refractoriness did not develop, under similar conditions, to the relaxant effects of either NO or the labile s-nitrosothiol SNAP (Fig. 2). Scanning electron microscopy revealed that no apparent structural or morphological damage to the endothelium was inflicted by incubating arterial rings for 24 h under the defined conditions (Fig. 3). Decline in arterial L-arginine levels after 24 h incubation. Basal or resting levels of L-arginine in freshly
prepared rings of bovine pulmonary artery were -300 PM. Incubation of arterial rings under tension at 37°C in oxygenated Krebs bicarbonate solution for 24 h resulted in a marked decline in L-arginine levels to
