Interleukin 1 Activates Soluble Guanylate Cyclase in Human Vascular Smooth Muscle Cells Through a Novel Nitric Oxide-Independent

Interleukin 1 Activates Soluble Guanylate Cyclase in Human Vascular Smooth Muscle Cells through a Novel Nitric Oxide-independent Pathway By Debbie Beasley and Mary McGuiggin

From the Division of Nephrology, Department of Medicine, New England Medical Center Hospitals, Tufts University School of Medicine, Boston, Massachusetts 02111

Summary Recent demonstration of cytokine-inducible production of nitric oxide (NO) in vascular smooth muscle cells (VSMC) from rat aorta has implicated VSMC-derived NO as a key mediator of Downloaded from http://rupress.org/jem/article-pdf/179/1/71/1395607/71.pdf by guest on 29 September 2021 hypotension in septic shock. Our studies to determine whether an inducible NO pathway exists in human VSMC have revealed a novel cytokine-inducible, NO-independent pathway of guanylate cyclase activation in VSMC from human saphenous vein (HSVSMC). Interleukin 1 (IL-1), tumor necrosis factor (TNF), interferon 3, (IFN-'y) and Escherichia coli lipopolysaccharide (LPS) increased cGMP at 24 h, whereas IL-2 and IL-6 were ineffective. The effect of IL-1 on cyclic guanosine Y,5'-monophosphate (cGMP) was delayed, occurring after 6 h of exposure, and was maximal after 10 h. Methylene blue and LY83583 reversed the IL-l-induced increase in cGMP, suggesting that it was mediated by activation of soluble guanylate cyclase. However, IL-l-induced cGMP in HSVSMC was not inhibited by extracdlular hemoglobin. Also, the effect of IL-1 on cGMP was not reversed by nitro- or methyl-substituted g-arginine analogs, aminoguanidine, or diphenyleneiodonium, all of which inhibit IL-l-induced NO synthase in rat aortic VSMC (RAVSMC). IL-l-induced cGMP in HSVSMC was also independent of tetrahydrobiopterin and extracellular L-arginine, as it was not affected by 2,4-diamino-6-hydroxyprytimidine, an inhibitor of tetrahydrobiopterin biosynthesis, and was similar in C-arginine-free and L-arginine-containing media. Analysis of NO synthase mRNA with the use of polymerase chain reaction indicates that levels of mRNA for inducible NO synthase are several orders of magnitude lower in IL- l-treated human HSVSMC than in IL-l-treated KAVSMC. IL-l-induced cGMP was also NO independent in human umbilical artery VSMC, and NO dependent in rat vena cava VSMC. Together these results indicate that IL-1 activates a novel NO-independent pathway of soluble guanylate cyclase activation in human VSMC.

cent evidence has implicated the free radical, nitric oxide NO synthase as a mediator of the pathophysiological vasodi- R~ (NO) 1, as a key physiological and pathophysiological lation associated with septic shock. IL-1, TNF, and bacterial regulator of vascular tone. NO is produced enzymatically from LPS induce NO synthase activity in rat aortic vascular smooth r-arginine by a family of enzymes termed NO synthases (1). muscle cells (RAVSMC; 7-9), which is demonstrable as in- Activation of a calcium-calmodulin-dependent, constitutive creased production of cyclic guanosine Y,5'-monophosphate NO synthase within the endothelium accounts for the ac- (cGMP) and nitrite. NO-induced cGMP production is respon- tion of endothelium-dependent vasodilators in animal (2, 3) sible for diminished vascular contraction observed in intact and human blood vessels (4, 5), and contributes to vasodila- rat aorta treated with IL-1 in vitro (10, 11). To establish tation in normal animals and humans (4, 6). Studies with whether VSMC-derived NO may contribute to cytokine and rodent cells and tissues have implicated a cytokine-inducible LPS-induced hypotension in humans, we sought to determine whether a similar cytokine-inducible NO pathway exists in human VSMC. 1 Abbreviations used in thispaper: cGMP, cyclicguanosine 3',5'-monophos- In this study, We demonstrate that IL-1, TNF, IFN-y, and phate; DAHP, 2,4-diamino-6-hydroxypyrimidine; DPI, diphenyleneido- LPS induce a novel pathway of guanylate cyclase activation donium; HSVMC, vascular smooth muscle cells from human saphenous in VSMC from human saphenous vein (HSVSMC). This vein; IBMX, isobutylmethylxanthine; L-NAME, NC-nitro-t-arginine methyl ester; L-NMMA, NC-monomethyl-t-arginine acetate; t-NoArg, pathway is distinct from the NO synthase pathway activated N~-nitro-L-arginine; NO, nitric oxide; RAVSMC, rat aortic vascular in R,AVSMC in that it is not inhibited by NO synthase in- smooth muscle cells; VSMC, vascular smooth muscle cells. hibitors including nitro- and methyl-substituted t-arginine

71 J. Exp. Med. The Rockefeller University Press 0022-1007/94/01/0071/10 $2.00 Volume 179 January 1994 71-80 analogs, aminoguanidine, and diphenyleneiodonium (DPI), cell extracts was determined by RIA using a commercial kit (Ad- is independent of tetrahydrobiopterin and extracellular c-ar- vanced Magnetics, Inc., Cambridge, MA). ginine, and does not generate measurable nitrite. We suggest Nitrite Assay. Nitrite was measured by a standard method (14) that IL-1 induces synthesis of a novel NO-independent in which aliquots of cell supernatant were mixed with an equal volume of Greiss reagent (0.05% naphthylethylenediamine di- pathway of soluble guanylate cyclase activation in HSVSMC. hydrochloride, 0.5% sulfanilamide, and 2.5% phosphoric acid) and incubated at room temperature for 10 min. The absorbance at 550 nm was measured, and nitrite concentration determined using so- Materials and Methods dium nitrite as a standard and distilled water as a blank. The back- Culture of VSMC. Human VSMC were grown by explant tech- ground nitrite values of each media was determined and subtracted nique from unused portions of saphenous veins harvested for coro- from experimental values. nary artery bypass surgery (New England Medical Center) as pre- RTPCR. Total RNA was isolated by the acid guanidinium viously described (12, 13), or were obtained from Dr. Peter Libby thiocyanate-phenol-chloroform method (15). 2 /xg RNA was (Brigham and Women's Hospital, Boston, MA). VSMC were also reversed transcribed for 1 h at 37~ with 200 U of Superscript grown from human umbilical arteries. Umbilical cords were ob- Moloney murine leukemia virus reverse transcriptase (GIBCO tained at normal Cesarean births (New England Medical Center), BILL), using an oligo(dT) primer and the buffer supplied by the the umbilical arteries were dissected free from the cord, cleaned manufacturer, in a total volume of 20/tl. The reaction was termi- of adventitia and endothelium, and VSMC grown from explants nated by heating to 95~ for 5 rain, and 2.5/zl of this first strand of the media. Cells were passaged every 7-30 d by harvesting with cDNA was added to each PCR reaction. PCR mixes contained Downloaded from http://rupress.org/jem/article-pdf/179/1/71/1395607/71.pdf by guest on 29 September 2021 trypsin-EDTA and seeding at a 1:3 ratio. For experiments, cells 50 mM KC1, 10 mM Tris, pH 8.3, 1.0 mM MgC12, 0.01 mg/ml were grown in DMEM, supplemented with 10--20% FCS, gluta- gelatin, 200 #M of each deoxynucleotide triphosphate, 250 nM mine, penicillin, and streptomycin, and used between passages two of each primer, and 0.5 U Taq DNA polymerase (Pharmacia, Pis- and six after reaching confluence at 7-30 d. cataway, NJ) in a total, volume of 25/A, overlaid with 75 #1 of VSMC were isolated from rat aorta and rat vena cava by enzy- mineral oil. The inducible NO synthase PCR primers (TAGAG- matic dissociation using standard methods as described previously GAACATCTGGCCAGG and TGGCCGACCTGATGTTGCCA) (7). VSMC were grown in DMEM containing 10% FCS, gluta- were homologous to inducible human hepatocyte NO synthase (16) mine, and antibiotics, passaged once a week by harvesting with and to rat vascular smooth muscle NO synthase (17). Amplification trypsin-EDTA, and plated at a 1:5 ratio. VSMC were used between was performed as follows: 94~ for 2 min (initial melt); 26-40 passages 3 and 12 after reaching confluence at 4-7 d. cycles of 94~ for 1 min, 57~ for 2 min, 72~ for 2 min; and VSMC Incubations. HSVSMC were washed three times with then 72~ for 5 rain (final extension). Products were size sepa- HBSS and incubated with or without cytokines or LPS in DMEM rated by electrophoresis in either a 2% agarose or 5% polyacryl- containing 5% FCS (DMEM-5) for 6 or 24 h. The cytokines used amide gel and visualized by ethidium bromide staining. In some in this study included human rlL-1 (Cistron Biotechnology, Inc., cases, PCR bands were quantitated by adding 0.25 #Ci (33 nM) Pine Brook, NJ), human rTNF-c~ (Genzyme Corp., Boston, MA), [32p]dCTP to the PCR reaction, cutting the PCR band from the human IL-2 (Endogen, Inc., Boston, MA), human rlL-6 (Genzyme gel and counting in a liquid scintillation counter. Corp.), human rlFN-3' (Genentech, South San Francisco, CA), and Drugs. Human rlL-1 was a gift of Dr. Richard Dondero (Cis- E. coli LPS (serotype 055:B5; Sigma Chemical Co., St. Louis, MO). tron Biotechnology, Inc.). DPI was a gift of Dr. Owen T. G. Jones RAVSMC were incubated with or without IL-1 under identical (University of Bristol, Bristol, UK). Human hemoglobin was a conditions for 24 h. The phosphodiesterase inhibitor, isobutyl- gift of Dr. Robert Valeri (Naval Blood Research Laboratory, Boston, methylxanthine (IBMX; 0.1 or 1 raM) was added to the cells 15-30 MA), and contained 97.2% oxyhemoglobin and 2.8% methemo- min before obtaining cell extracts for cGMP determination. Inhib- globin. IBMX, N~-nitro-L-arginine methyl ester (L-NAME), N c- itors of NO synthesis/action were added either at the beginning nitro-L-arginine (L-NoArg), aminoguanidine bicarbonate, and of the incubation with IL-1 or towards the end, before the addi- 2,4-diamino-6-hydroxypyrimidine were obtained from Sigma tion of IBMX. In most experiments, the L-arginine concentration of the tissue culture media (DMEM) was 400/zm. For L-arginine depletion 300- studies, cells were incubated in MEM with or without t-arginine (600 #M; Select-Amine-kit; GIBCO BRL, Gaithersburg, MD) for 0 200" 48 h before activation with IL-1. HSVSMC were then activated E with IL-1 in fresh media with or without c-arginine. The effect of NO synthase inhibitors was tested in media containing either ts no (0/~M) or low (50 #M) r-arginine. HSVSMC were incubated 100" 0 for 6 h with or without IL-1 in MEM containing 50 #M t-arginine, 0 and inhibitors were added directly to this media at the end of the

incubation period. Alternatively, VSMC or HSVSMC were in- 0 t t i ! t t cubated for 24 h with or without IL-1 in DMEM-5 containing 0 4 8 12 16 20 24 400/xM L-arginine. 40 min before the end of this incubation period, the cells were washed and the media replaced with Earle's BSS (r- Time (hr) arginine free) with or without NO synthase inhibitors. Figure 1. IL-1 induces a delayed increase in cGMP accumulation in Cyclic GMP Measurements. cGMP was extracted from the cells HSVSMC. HSVSMC were incubated with IL-1 (200 U/ml) in DMEM-5 by rapid aspiration of the medium and addition of ice-cold 0.1 N and cGMP determined after 1-24 h. IBMX (0.1 raM) was added 30 min HC1 to each well. Cells were incubated I h on ice and the extracts before cGMP determination, n = 4 replicates. Error bars which are not frozen at -20~ before assay for cGMP. The cGMP content of visible are less than the width of the symbol

72 Cytokine-induced cGMP in Human Vascular Smooth Muscle Cells 300 10

A t- [] CON O 8 0 E 200" [] ILl x 6"

a. 4- 100 ' o o 2

CON ILl TNF IFN IL2 IL6 LPS -IBMX +IBMX Figure 3. Effect of cytokines on zGMP accumulation in HSVSMC. Figure 2. IL-1 increase cGMP in the presence and absence of a phos- HSVSMC were incubated with IL-1 (1 nM or 170 U/ml), TNF (1 riM), phodiesterase inhibitor, IBMX. HSVMSC were incubated with IL-1 (200 E. coil LPS (100/~g/ml), IFN-y (10 riM), IL-2 (10 nM), or IL-6 (10 nM) U/ml) in DMEM-5 and cGMP determined after 6 h. IBMX (1 mM) was for 24 h. IBMX (0.1 mM) was added 30 min before cGMP determination. added to some of the wells 20 min before cGMP determination, n = 4

n = 4 replicates. (*)p < 0.02, significantlydifferent from control HSVSMC. Downloaded from http://rupress.org/jem/article-pdf/179/1/71/1395607/71.pdf by guest on 29 September 2021 replicates.

Chemical Co. NG-monomethyl-r-arginine (L-NMMA) acetate and in cGMP reflect increased production rather than decreased L-NMMA F-hydroxyazobenzene-p'-sulfonate salt were obtained degradation by phosphodiesterase. AU other experiments were from Calbiochem-Novabiochem Corp. (La JoUa, CA). (6R)-5,6,7,8- conducted with IBMX. tetrahydrobiopterin dihydrochloride was from ICN Biomedicals (Ir- HSVSMC incubated 24 h with TNF (1 nM), IFN-'y (10 vine, CA). riM), or E. coli LPS (100 #g/ml), showed an increase in cGMP Data Analysis. Values are expressed as mean + SE. Significant similar to that of IL-1 (1 nM). In contrast, IL-2 (10 nM) differences were determined by unpaired Student's t test, and p <0.05 was considered statistically significant. and IL-6 (10 nM) did not affect cGMP production (Fig. 3). Methylene Blue and LY83,583 Reverse IL-l-induced cGMP in HSVSMC. IL-l-induced cGMP accumulation was re- Results versed by two soluble guanylate cyclase inhibitors, methy- IL-I, TNF, LPS, and IFN-T Increase cGMP in HSVSMC. lene blue and LY 83,583 (Fig. 4 A). Both methylene blue IL-1 induced a delayed increase in cGMP in HSVSMC. cGMP (10/.tM) and LY 83,583 (10 #M) lowered cGMP content in was significantly increased after 6 h of exposure to IL-1 (p nontreated HSVSMC, and decreased the IL-l-induced incre- <0.0001), continued to increase reaching maximal levels after ment in cGMP by 90 and 91% (2 <0.002). In contrast, ex- 10 h of exposure, and then decreased towards control levels tracellular hemoglobin (30 #M) did not significantly affect (Fig. 1). The magnitude of the IL-1 effect was similar whether cGMP levels in control or IL-treated HSVSMC (Fig. 4 B). cells were studied in the presence or absence of a phos- r-Arginine Analogs Do Not Reverse IL-1-induced cGMP in phodiesterase inhibitor, IBMX (Fig. 2). Whereas IBMX in- HSVSMC. IL-l-induced cGMP production was not attenu- creased basal levels of cGMP fivefold, a 6-h exposure to IL-1 ated by t-arginine analogs containing methyl or nitro sub- increased cGMP fourfold in the absence of IBMX, and 4.5- stitutions of the guanidino nitrogen. Neither L-NMMA, fold in the presence of IBMX. Thus, IL-l-induced changes t-NAME, r-NoArg, nor aminoguanidine inhibited IL-l-in-

A B

500 200

[] CON 400 " [] ILl 150

0 300 " Figure 4. IL-l-induced increases in cGMP in E HSVSMC are attenuated by methylene blue 100 (MB) and LY83583 (LY), but not by hemoglobin (Hb). D,. 200 HSVSMC were incubated with I1.-1 (200 U/ml) (9 for 6 h. MB or LY (each 10 #M; A) or Hb (30 o * * 50 #M; B) were added 40 min before, and IBMX 20 100 min before cGMP determination, n = 4 replicates. (+) p < 0.01, significantly different from control HSVSMC; (*) p < 0.002, significantly different CON MB LY CON Hb from IL-l-treated HSVSMC.

73 Beasleyand McGuiggin duced cGMP in HSVSMC, when tested at concentrations of 1 mM in the presence of 50/zM t-arginine (Fig. 5). Like- A wise, neither r-NoArg nor t-NMMA (0.03-1 raM) reversed 125. the increased cGMP accumulation in IL-l-treated human VSMC, when tested in balanced salt solution that contained E 100" no r-arginine (Fig. 6). Under similar experimental conditions, i aminoguanidine (0.01-3 raM) was equally ineffective at -1 75' reversing IL-l-induced cGMP in HSVSMC (Fig. 7), as were S0' \ ~ L-NoArg/human VSMC r-NAME and t-NMMA p-hydroxyazobenzene?'-sulfonate r* I \ --.--o--,.- L-NMMA/human VSMC salt (data not shown). In contrast, r-NMMA (Fig. 6) and \ L-NMMAJrat VSMC aminoguanidine (1-1,000 #M; Fig. 7) were potent inhibitors o 2s m of IL-l-induced cGMP in RAVSMC under identical experimental conditions. 0 ,~o ,ooo DPI Does Not Reverse IL-I-induced cGMP in HSVSMC. The flavoprotein inhibitor, DPI (0.3-3/zM) attenuated the Inhibitor (~tM) IL-1-induced increase in cGMP in RAVSMC in a concentra- Figure 6. Effect of t-NoArg and L-NMMA on IL-l-induced increases in cGMP in HSVSMC and RAVSMC, in the absence of extracellular

tion-dependent fashion (Fig. 8 A). These concentrations of Downloaded from http://rupress.org/jem/article-pdf/179/1/71/1395607/71.pdf by guest on 29 September 2021 DPI did not affect the increase in cGMP caused by sodium t-arginine. HSVSMC and RAVSMC were incubated with or without 1I.-1 (200 U/ml) in DMEM-5 for 24 h. 40 rain before cGMP determination, nitroprusside in these cells (data not shown). In contrast, DPI the cells were washed and the media replaced with Earle's BSS (0 r-arginine) (1-3 #M) did not significantly affect IL-l-induced cGMP in containing either no additions, t-NoArg, or t-NMMA (0.03-I raM). IBMX HSVSMC (Fig. 8 B). (0.1 mM) was added 20 min before cGMP determination, n = 4 replicates. IL-l-induced cGMP in HSVSMC Is Independent of Extracel- lular ~-Arginine. IL-l-increased cGMP was not attenuated in HSVSMC that had been cultured for 48 h in t-arginine-free synthase activity in several cell types by depleting intracel- MEM, then stimulated for 6 h with IL-1 in fresh t-arginine- lular levels of this cofactor. Addition of DAHP (2.5-10 mM) free MEM (Fig. 9). Also, repletion of r-arginine in some of at the beginning of the 24-h incubation with IL-1 block IL- the t-arginine-depleted VSMC I h before cGMP determina- l-induced cGMP in RAVSMC in a concentration-dependent tion did not affect the response to IL-1. manner (Fig. 10 A). The effect of DAHP (10 mM) was reversed IL-l-induced cGMP in HSVSMC Is Independent of Tetrahydrobi- by the addition of 100 #M tetrahydrobiopterin to the media opterin. 2,4-diamino-6-hydroxypyrimidine (DAHP) is an in- 80 rain before cGMP determination, supporting the hypoth- hibitor of tetrahydrobiopterin synthesis, which inhibits NO esis that the inhibitory effect of DAHP was due to depletion of tetrahydrobiopterin. In contrast, DAHP and tetrahydrobiopterin had no effect on IL-l-induced cGMP in HSVSMC, 250 when added under the same conditions (Fig. 10 B).

200 150

m 'O 125 ' o E 1so

100 ' a. \ 100 r3 '~ 75' r VSMC 50' HSVSMC 50

2s :

Z .~ ~ uJ ,..cn r o" 0 _ ~ ~ < < 1 1 0 100 1000 10000 o ~ "r o + z z z 3 -~ ', Aminoguanidine (~tM) + + + Figure 7. Effect of aminoguanidine on IL-l-induced increase in cGMP Figure 5. Effect of r-NMMA, t-NAME, t-NoArg, and aminoguani- in HSVSMC and RAVSMC, in the absence of extracellular r-arginine. dine on IL-l-induced increases in cGMP in HSVSMC, in media containing HSVSMC and RAVSMC were incubated with or without IL-1 (200 U/ml) 50/xM t-arginine. VSMC were incubated with or without 1I.-1 (200 U/ml) in DMEM-S for 6 and 24 h, respectively. 80 min before cGMP determina- in MEM containing 50/xM r-arginine for 6 h. NO synthase inhibitors tion, the cells were washed and the media replaced with Earle's BSS (0 (1 mM each) were added 40 min before, and IBMX 20 min before cGMP t-arginine) containing aminoguanidine (0-2 mM). IBMX (0.1 mM) was determination, n = 4 replicates. added 20 rain before cGMP determination, n = 4 replicates.

74 Cytokine-induced cGMP in Human Vascular Smooth Muscle Cells A B

600 250"

I~M DPI 50O 200- Iio ~tM DPI [] 0.3 Figure 8. Effect of the flavo- Am 400 [] 1 Iio protein inhibitor DPI on IL-l-in- 0 ~3 1SO. [] 1 duced cGMP in RAVSMC (A) and Q~ 300 HSVSMC (B). HSVSMC and 100 " RAVSMC were incubated with or without IL-1 (200 U/ml) in 0 200 J DMEM-5 for 6 and 24 h, respec- 50- tively. DPI (0.3 - 1 ~M) was added 80 min before, and IBMX 20 min before cGMP determination, n = CON ILl CON I L 1 4 replicates.

Extracellular Nitrite.Accumulation. Whereas nitrite was de- VSMC showed an exponential increase in PCR product be- Downloaded from http://rupress.org/jem/article-pdf/179/1/71/1395607/71.pdf by guest on 29 September 2021 tectable in supernatants of HSVSMC, treatment with IL-1 tween 20 and 26 cycles; no further increase in PCR product did not significantly affect nitrite accumulation. The nitrite was seen with 28-30 cycles of amplification (data not shown). concentration of supernatants from control and IL-l-treated Thus, 25 or 26 amplification cycles were used for further HSVSMC were 2.0 _+ 0.2 #M and 2.5 +_ 0.3/~M, respec- studies. Exposure of RAVSMC to IL-1 produced a time- tively, at 24 h, and 4.5 + 1.1/~M and 3.5 _+ 0.3/~M at 48 h. dependent increase in NO synthase mRNA, that was max- Detection of NO Synthase mRNA by RT-PCR. eDNA pre- imal at 4 and 24 h (Fig. 11 A). When cDNA from IL-l-treated pared from rat and human VSMC was amplified using a single RAVSMC was amplified for 25 cycles in the presence of set of primers that had complete homology to both rat and [32p]dCTP, an intense band was visible by ethidium bromide human inducible NO synthase (16, 17). Amplification of staining of products (Fig. 11 B), whereas control VSMC gave eDNA prepared from RAVSMC, which had been incubated only a faint band. Quantitation of the 32p content of the with IL-1 for 24 h, revealed a PCR product of the expected bands shown in Fig. 11 B by liquid scintillation counting size (255 bp) that was detected after as few as 20 cycles (data revealed that IL-1 treatment of rat VSMC produced an 18- not shown). Amplification of eDNA from IL-l-treated rat fold increase in NO synthase mRNA levels (control VSMC, 2,702 cpm; 24-h IL-l-treated VSMC, 49,455 cpm). In contrast, eDNA prepared from HSVSMC treated with IL-1 for 0-24 h showed no visible PCR products by either ethidium

120 bromide staining (Fig. 11 B), or by liquid scintillation counting (less than the background level of 228 cpm). When eDNA from HSVSMC was amplified further, for 30-32 cycles, faint 100 bands of the expected size were visible in IL-l-treated samples. Cytokine-induced cGMP in Human Umbilical Artery and Rat 8O Vena Cava VSMC. To determine whether the difference in O mechanism of IL-1 action in rat aorta and human saphenous E VSMC reflects differences between arteries and veins, or be- v 60 tween species, we studied IL-l-induced cGMP in VSMC from Q. human umbilical artery and rat vena cava. In human umbil- 40 ical artery VSMC, cGMP was not significantly increased by IL-1 alone (20 ng/ml), whereas treatment with IL-1 (20

20 ng/ml) and TNF (50 ng/ml) for 6 h resulted in a 44-130% increase in cGMP (data not shown), cGMP was no longer increased at 24 h, suggesting a more rapid time course, similar 0 to the time course in HSVSMC. Nitrite was not detectable +L-arg, 54 h - L-arg, 54 h - L-arg, 53 h in 24-h supernatants obtained from either control or IL-1- +L-arg, 1 h and TNF-treated human umbilical artery VSMC. Finally, the Figure 9. L-arginine depletion does not affect IL-l-induced cGMP in L-arginine analog L-NMMA did not affect cGMP in human HSVSMC. HSVSMC were incubated in MEM-5 with or without t-arginine umbilical artery VSMC stimulated with IL-1 and TNF (Fig. for 48 h. The media was then replaced with fresh MEM with or without 12 A). t-arginine, and with or without IL-1 (200 U/ml) and the cells incubated In rat vena cava VSMC, cGMP was significantly increased for an additional 6 h. L-arginine (600/~M) was added back to some of the t-arginine-free wells 60 rain before cGMP determination. IBMX was after 3-h incubation with IL-1 (20 ng/ml), and continued added 30 rain before cGMP determination, n = 4 replicates. (oFen bar) to increase up to 22-fold at 24 h (data not shown). Nitrite CON; (shaded bar) IL-1. was detectable in 24-h supernatants from control cells, and

75 Beasley and McGuiggin A B 200O 100

1500 80

O E ~., 60 1000 D,. 4o o o 500 20

0 0 0 0 2.5 5 10 10 mM DAHP 0 0 2.5 5 10 10 mM DAHP 100 I~M BH4 100 p.M BH4 CON I L 1 CON I L 1 Figure 10. Effectof inhibition of tetrahydrobiopterinbiosynthesis on ILl-induced cGMP in rat aortic (A) and human saphenous VSMC (B). RAVSMC

or HSVSMC were incubated in DMEM-5 with or without IL-1 (200 U/ml), and with or without DAHP for 24 h and cGMP determined. Tetrahydrobi- Downloaded from http://rupress.org/jem/article-pdf/179/1/71/1395607/71.pdf by guest on 29 September 2021 opterin (100/zM) was repleted in some of the wells by the addition to the media 80 min before cGMP determination. IBMX was added 20 rain before cGMP determination, n = 4 replicate wells.

was increased in IL-l-treated cells (0.7 + 0.4/zM and 6.4 Discussion + 0.4/xM, respectively). The methyl and nitro-substituted A rapidly expanding literature has implicated a role for L-arginine analogs, L-NMMA, L-NoArg, and L-NAME, cytokine-inducible NO synthesis in the pathogenesis of septic reversed the effect of IL-1 on cGMP in rat vena cava VSMC shock. Cytokines induce NO synthase activity in VSMC, (Fig. 12 B), as did aminoguanidine. suggesting that this intracellular source of NO may play a major role in vascular relaxation during sepsis. In RAVSMC, NO synthase activity is detectable by prolonged, t-arginine-dependent increases in cGMP production and accumulation of nitrite in the media (7), and is associated with an increase in inducible NO synthase mRNA (18). In this study, we report that IL-1 induces a similar delayed and prolonged increase in cGMP production in VSMC cultured from human saphenous vein. As in RAVSMC, TNF and E. coli LPS also increased cGMP in human VSMC. IFN-3', which is ineffective at inducing NO synthase in R.AVSMC (18, 19), also increased cGMP in HSVSMC. Thus, inflammatory cytokines increase the vasodilatory mediator cGMP in both R.AVSMC and HSVSMC. However, the mechanism of IL-1 action was distinct in human VSMC. First, the time course of IL-1 action differed, cGMP increases after a lag period of >3 h in both cell types, however the maximal response occurs earlier in HSVSMC than in RAVSMC (10 versus 36 h; 7 and Fig. 1). Second, studies with inhibitors indicated that the mechanism of IL-1 action in HSVSMC involves activation of soluble guanylate cyclase but does not involve NO. t-arginine analogs that contain methyl or nitro substitu- tions on the guanidino nitrogen are competitive inhibitors Figure 11. PCR analysisof inducibleNO synthasemRNA in RAVSMC of NO synthase (2, 20, 21), which consistently inhibit all and HSVSMC. cDNA prepared by reverse transcription of 250 ng RNA constitutive and inducible NO synthases described to date was amplified by PCR, and products electrophoresed on a 2% agarose gel (A) or 5% polyacrylamidegel (B), and stained with ethidium bro- in various cell types. Neither monomethyl- nor nitro-substi- mide. (A) LaneM, molecular size markers (310-, 271-281-, 234-, and 194- tuted L-arginine analogs attenuated IL-l-induced increases bp fragments of Haelll-digested q~X-174; lanes I-5: RAVSMC treated in cGMP in human VSMC when tested at concentrations with ILl for 0, 1, 2, 4, and 24 h, respectively.(B) LanesM, HaellI-digested ~<1 mM either in the absence of L-arginine or in the presence qSX-174 (310, 271, 281, 234, 194, and 118 bp); lanes I-6: HSVSMC treated with IL-1 for 0, 1, 2, 4, 8, and 24 h, respectively:tLAVSMC treated without of 50/~M t-arginine. In fact, there was a tendency for t-arginine (lane 7) or with (lane 8) IL-1 for 24 h; lane 9: control PCR reaction (no analogs to increase IL-l-induced cGMP in HSVSMC, sug- RNA in reverse transcription reaction). gesting that NO produced by the cells may attenuate the IL-

76 Cytokine-induced cGMP in Human Vascular Smooth Muscle Cells Figure 12. Effectoft-arginine analogs on cytokine-inducible cGMP in human umbilicalartery (A) and rat vena A B cava (B) VSMC. (A) Human umbilical 200 8000 artery VSMC were incubated 6 h with [] CON 1- or without IL-1 (200 U/ml) and TNF [] L-NMMA (50 ng/ml) in MEM containing 50/~M [] CON 150" z-arginine, t-NMMA (1 mM) was T 6000 [] L-NMMA added 40 rain beforeand IBMX (1 mM) o [] L-NoArg 20 min beforecGMP determination. (B) [] L-NAME Rat vena cava VSMC were incubated 100" 4000 BAG e* 24 h with or without IL-1 (200 U/ml) in DMEM-5.80 min beforecGMP de- 0 termination, VSMC were washed and 60- 2000 the media replaced with MEM containing 50 ~M L-arginine with or without L-arginine analogs (each 1 i. mM). IBMX (0.1 mM) was added 20 0 0 CON ILl +TN F CON ILl rain before cGMP determination. Downloaded from http://rupress.org/jem/article-pdf/179/1/71/1395607/71.pdf by guest on 29 September 2021

1-induced response. In contrast, all L-arginine analogs tested the IL-l-inducible guanylate cyclase activating pathway inhibit cytokine-induced NO production in RAVSMC, under in HSVSMC is independent of tetrahydrobiopterin. similar experimental conditions. Aminoguanidine is a nucleo- Cytokine-activated macrophages, endothelial ceils, and phillic hydrazine compound that contains two equivalent VSMC from rodents produce substantial amounts of NO guanidino nitrogens, and has been shown to inhibit cytokine- which is oxidized to nitrite and nitrate, and these end-products inducible NO synthase in pancreatic/3 cells (22), presum- accumulate extracellularly to micromolar levels. Whereas ni- ably by acting as a competitive inhibitor of NO synthase. trite was present at '~,4/~M in 48-h supernatants from human Our results indicate that aminoguanidine is also a potent in- VSMC, IL-1 did not affect nitrite production. Interestingly, hibitor of cytokine-inducible NO synthase in RAVSMC, and VSMC cultured from human mesenteric microarterioles have is effective at concentrations as low as 1/zM. In contrast, been reported to release NO in the basal state and this release IL-induced cGMP in HSVSMC was not affected by amino- is inhibited by L-NMMA (36). It is not clear whether NO guanidine. produced by human VSMC in our study or in the latter study Purified constitutive and inducible NO synthases are (36) is a product of an inducible or constitutive NO synthase. NADPH dependent (23) and contain flavin adenine dinucleo- Detection of inducible NO synthase mRNA by PCR in- tide and flavin mononucleotide (24, 25), which act as cofactors dicates that IL-l-treated HSVSMC express very low levels for NO synthase (26). Both macrophage and endothelial NO of NO synthase mRNA compared with IL-treated RAVSMC synthases are inhibited by DPI (27), an inhibitor of NADPH- (at least three orders of magnitude lower based on the number dependent ftavoproteins (28, 29). Our results indicate that of cycles required to obtain visible product by ethidium bro- cytokine-inducible NO synthase in RAVSMC is also in- mide staining, and by quantitation of incorporated radioac- hibitable by DPI, whereas IL-l-induced cGMP in HSVSMC tivity into the relevant PCR product). IL-l-treated HSVSMC was not affected. These results provide additional evidence even contained lower levels of inducible NO synthase mRNA that the increase in cGMP in HSVSMC does not involve NO than nontreated RAVSMC. The finding of low levels of NO synthase or a related NADPH-dependent flavoprotein. synthase mRNA suggests that an inducible NO synthase gene Purified constitutive and inducible NO synthases also con- may be present in HSVSMC, although it was not effectively tain tetrahydrobiopterin (30, 31) which is an essential cofactor induced by ILl under the experimental conditions employed. for NO synthase oxidation of L-arginine to citrulline and NO It remains to be determined whether the observed NO syn- (30). Cytokine-inducible NO synthesis in endothelial cells thase related PCR product in HSVSMC represents an induc- (32, 33), VSMC (19), macrophages (34), and fibroblasts (35) ible NO synthase having significance in HSVSMC in vivo. is dependent on tetrahydrobiopterin synthesis. Our studies Whereas our results rule out a role of NO synthase in IL- confirm that cytokine-inducible NO synthase activity in rat l-induced cGMP in HSVSMC, activation of cGMP was in- aortic VSMC is dependent on tetrahydrobiopterin synthesis. hibited by methylene blue and LY83,583, two inhibitors of Treatment of RAVSMC with DAHP, an inhibitor of GTP endothelium-dependent relaxation that inhibit soluble guany- cyclohydrolase, the rate-limiting enzyme in tetrahydrobiopterin late cyclase activation independent of NO production. The biosynthesis, blocked the effect of IL-1, whereas addition of mechanisms by which methylene blue and LY83,583 inhibit extraceUular tetrahydrobiopterin, restored IL-l-induced cGMP guanylate cyclase are not completely understood. Methylene in DAHP-treated VSMC, confirming that the action of DAHP blue may inhibit guanylate cyclase activation by oxidation was due to depletion of tetrahydrobiopterin. In contrast, of the regulatory hemoprotein moiety of the enzyme (37). DAHP and extracellular tetrahydrobiopterin had no effect on Recent in vivo evidence indicates that methylene blue and IL-l-induced cGMP in saphenous VSMC, indicating that LY83583 inactivate soluble guanylate cyclase by generating

77 Beasleyand McGuiggin hydroxyl radical (38). Our results indicate that methylene only a few human cell types in vitro, in contrast to the abun- blue and LY83,583 inhibit NO-independent as well as NO- dance of nonhuman cell types shown to have cytokine- dependent activation of soluble guanylate cyclase. The results inducible NO synthase activity. Human hepatocytes (16) and also indicate that IL-l-induced cGMP in human VSMC in- human chondrocytes (46) express NO synthase activity after volves NO-independent activation of soluble rather than par- treatment with cytokines or LPS. However, cytokine treat- ticulate guanylate cyclase. IL-l-induced cGMP in human ment does not affect nitrite or nitrate accumulation in 24--48-h VSMC was not inhibited by hemoglobin which inhibits supernatants of human monocytes or macrophages (47-51). guanylate cyclase activation by endothelium-dependent vasodi- In one study, TNF stimulated nitrite accumulation in 7-d lators by binding NO with high affinity (39). These results supernatants of human blood mononuclear cells (52), sug- may indicate that IL-l-treated human VSMC do not pro- gesting that nitrite may be produced but at very low levels. duce an activator of soluble guanylate cyclase which is bound Our studies may indicate that cytokine-inducible NO syn- by hemoglobin. However, we cannot rule out the possibility thase activity is not ubiquitous to all human VSMC. Alter- that hemoglobin is ineffective because it remains extracellular, natively, the experimental conditions which we used may not and thus does not block the effect of a mediator which acts mimic those required for effective induction of the NO syn- exclusively within the cell and not on adjacent cells. thase gene in human cells. Regardless, our studies establish Clear evidence has been presented for the regulation of vas- that IL-1 can induce a NO-independent pathway of guany- cular tone by constitutive endothelial NO synthase in humans. late cyclase activation in human VSMC, and thereby raise the Downloaded from http://rupress.org/jem/article-pdf/179/1/71/1395607/71.pdf by guest on 29 September 2021 r-NMMA blocks acetylcholine-induced increases in forearm question of the relative importance of NO-dependent and blood flow (5), and human endothelial NO synthase has been -independent pathways of guanylate cyclase activation in the recently cloned from human umbilical vein endothelial cells vascular actions of cytokines in humans. (40). However, evidence concerning the existence and nature In conclusion, IL-1 increases cGMP in VSMC from human of a cytokine-inducible NO synthase in human vascular cells saphenous vein by a mechanism that is distinct from NO. is limited. One report demonstrates IL-l-induced nitrite In contrast to the IL-l-inducible NO pathway in R_AVSMC, production in VSMC grown from a single human aorta (41), IL-l-inducible cGMP production in HSVSMC is not inhibited and inducible NO synthase transcripts have been reported by NO synthase inhibitors including nitro- and methyl- in human aortic VSMC treated with a mixture of cytokines substituted t-arginine analogs, aminoguanidine, and flavo- (16). Although in vivo studies also support the hypothesis protein inhibitors, and is independent of extracellular t-arginine that NO contributes to cytokine and sepsis-induced hypoten- and tetrahydrobiopterin. IL-1 fails to enhance nitrite produc- sion in humans, they are not definitive. L-arginine analogs tion by HSVSMC because of its inability to induce significant were effective at improving blood pressure in early clinical levels of NO synthase mRNA. This NO-independent pathway trials with septic shock patients (42), however, blood pres- of guanylate cyclase activation appears to be specific to human sure may improve in part because of inhibition of constitu- VSMC, as it is also present in human umbilical artery VSMC, tive NO synthase. There is also evidence that cytokines in- whereas VSMC from rat vena cava and rat aorta display NO- duce NO synthase activity in humans in vivo: arginine-derived dependent increases in cGMP. It will be important to estab- nitrate is increased in the plasma and urine of cancer patients lish whether this distinct guanylate cyclase activating pathway receiving high-dose IL-2 therapy (43, 44), and is increased is present in VSMC from other human arteries or veins, or in the plasma of septic patients (45). However, the cellular in other human cell types. NO-independent activation of source of the nitrate is not discernible from in vivo studies. soluble guanylate cyclase in VSMC may contribute to sys- Cytokine-inducible NO synthase has been demonstrated in temic vasodilation in septic patients.

The authors gratefully acknowledge Dr. Jeffrey B. Tatro (Department of Medicine, New England Medical Center) for helpful comments and suggestions. This work was supported by grants from Baxter Healthcare Corporation, Renal Division, Extramural Grant Program and the National Institutes of Health (HL-47569). Address correspondence to Dr. Debbie Beasley, New England Medical Center, Box 172, 750 Washington St., Boston, MA 02111.

Received for publication 14 April 1993 and in revisedform 2I September 1993.

References 1. Nathan, C. 1992. Nitric oxide as a secretory product of mam- derived relaxing factor. Proa Natl. Acad. Sci. USA. 85:8664. malian cells. FASEB (Fed. Am. Soc. Exp. Biol.) J. 6:3051. 3. Forstermann, U., J.S. Pollock, H.H.H.W. Schmidt, M. Heller, 2. Sakuma, I., D.J. Stuehr, S.S. Gross, C. Nathan, and R. Levi. and F. Murad. 1991. Calmodulin-dependent endothelium- 1988. Identification of arginine as a precursor of endothelium- derived relaxing factor/nitric oxide synthase activity is present

78 Cytokine-inducedcGMP in Human Vascular Smooth Muscle Cells in the particulate and cytosolic fractions of bovine aortic en- oxide synthesis: cell type-selective inhibition by NG-amino- dothelial cells. Pro~ Natl. Acad. Sci. USA. 88:1788. arginine, NG-nitroarginine and NG-methylarginine. Biochem. 4. Vallance, P., J. Collier, and S. Moncada. 1989. Effects of Biophys. Res. Commun. 170:96. endothelium-derived nitric oxide on peripheral arteriolar tone 22. Corbett, J.A., K.G. Triton, K. Chang, K.S. Hasan, Y. Ido, in man. Lancet. 2:997. J.L. Wang, M.A. Sweetland, J.R. Lancaster, Jr., and J.R. Wil- 5. Vallance, P., J. Collier, and S. Moncada. 1989. Nitric oxide liamson. 1992. Aminoguanidine, a novel inhibitor of nitric oxide synthesized from r-arginine mediates endothelium dependent formation, prevents diabetic vascular dysfunction. Diabetes. dilatation in human veins in vivo. Cardiovasc. Res. 23:1053. 41:552. 6. Rees, D.D., R.M.J. Palmer, and S. Moncada. 1989. Role of 23. Marietta, M.A., P.S. Yoon, R. Iyengar, C.D. Leaf, and J.S. endothelium-derived nitric oxide in the regulation of blood Wishnok. 1988. Macrophage oxidation of r-arginine to nitrite pressure. Proc. Natl. Acad. Sci. USA. 86:3375. and nitrate: nitric oxide is an intermediate. Biochemistry. 27:8706. 7. Beasley, D., J.H. Schwartz, and B.M. Brenner. 1991. Inter- 24. Hevel, J.M., K.A. White, and M.A. Marietta. 1991. Puri- leukin 1 induces prolonged L-arginine-dependent cyclic guano- fication of the inducible routine macrophage nitric oxide syn- sine monophosphate and nitrite production in rat vascular ~hase: Identification as a flavoprotein.J. Biol. Chem. 266:22789. smooth muscle cells. J. Clin. Invest. 87:602. 25. Bredt, D.S., C.D. Ferris, and S.H. Snyder. 1992. Nitric oxide 8. Busse, R., and A. Mulsch. 1990. Induction of nitric oxide syn- synthase regulatory sites: phosphorylation by cyclic AMP- thase by cytokines in vascular smooth muscle cells. FEBS (Fed. dependent protein kinase, protein kinase C, and calcium- Eur. Biochem. Soc.) Lett. 275:87. calrnodulin protein kinase; identification of flavin and calmo-

9. Rees, D.D., S. Cellek, R.M.J. Palmer, and S. Monada. 1990. dulin binding sites. J. Biol. Chem. 267:10976. Downloaded from http://rupress.org/jem/article-pdf/179/1/71/1395607/71.pdf by guest on 29 September 2021 Dexamethasone prevents the induction by endotoxin of a ni- 26. Stuehr, D.J., N.S. Kwon, and C.F. Nathan. 1990. FAD and tric oxide synthase and the associated effects on vascular tone: GSH participate in macrophage synthesis of nitric oxide. Bio- an insight into the endotoxin shock. Biochem. Biophys. Res. chem. Biophys. Res. Commun. 168:558. Commun. 173:541. 27. Stuehr, D.J., O.A. Fasehun, N. Soo Kwon, S.S. Gross, J.A. 10. Beasley, D. 1990. Interleukin 1 and endotoxin activate soluble Gonzalez, R. Levi, and C.F. Nathan. 1991. Inhibition ofmac- guanylate cyclase in vascular smooth muscle. Am. J. Physiol. rophage and endothelial cell nitric oxide synthase by diphenyl- 259:R38. eneiodonium and its analogs. FASEB (Fed. Am. Soc. Exp. Biol.) 11. French, J.F., L.E. Lambert, and R.C. Dage. 1991. Nitric oxide J. 5:98. synthase inhibitors inhibit interleukin-13-induced depression 28. Cross, A.R., and O.T.G. Jones. 1986. The effect of the inhib- of vascular smooth muscle. J. Pharmacol. Exp. Ther. 259:260. itor diphenylene iodonium on the superoxide-generating system 12. Libby, P., and K.V. O'Brien. 1983. Culture of quiescent ar- of neutrophils. Specific labelling of a component polypeptide terial smooth muscle cells in a defined serum-flee medium. of the oxidase. Biochem. J. 237:111. J. Cell Physiol. 115:217. 29. Hancock, J.T., and O.T.G. Jones. 1987. The inhibition by 13. Ross, R., and B. Kariya. 1980. Morphogenesis of vascular diphenyleneiodonium and its analogues of superoxide genera- smooth muscle in atherosclerosis and cell structure. In Hand- tion by macrophages. Biochem. J. 242:103. book of Physiology. The Cardiovascular System, Section 2. D.F. 30. Hevel, J.M., and M.A. Marietta. 1992. Maerophage nitric oxide Bohr, A.P. Somlyo, and H.Y. Sparks, editors. American Physi- synthase: relationship between enzyme-bound tetrahydrobi- ological Society, Bethesda, MD. 69. opterin and synthase activity. Biochemistry. 31:7160. 14. Green, L.C., D.A. Wagner, J. Glogowski, P.L. Skipper, J.S. 31. Klatt, P., B. Heinzel, B. Mayer, E. Ambach, G. Werner- Wishnok, and S.K. Tannebaum. 1982. Analysis of nitrate, ni- Felmayer, H. Wachter, and E.K. Werner. 1992. Stimulation trite and [15N] nitrate in biological fluids. Anal. Biochem. of human nitric oxide synthase by tetrahydrobiopterin and selec- 126:131. tive binding of the cofactor. FEBS (Fed. Eur. Biochem. Soa) Lett. 15. Chomczynski, P., and N. Sacchi. 1987. Single-step methods 305:160. of RNA isolation by acid guanidinium thiocyanate-phenol- 32. Gross, S.S., E.A. Jaffe, R. Levi, and R.G. Kilbourn. 1991. chloroform extraction. Anal. Biochem. 162:156. Cytokine-activated endothelial cells express an isotype of ni- 16. GeUer, D.A., C.J. Lowensten, R.A. Shapiro, A.K. Nussler, tric oxide synthase which is tetrahydrobiopterin-dependent, M. Di Silvio, S.C. Wang, D.K. Nakayama, R.L. Simmons, calmodulin-independent and inhibited by arginine analogs with S. Snyder, and T.R. Billiar. 1993. Molecular cloning and ex- a rank-order of potency characteristic of activated macrophages. pression of inducible nitric oxide synthase from human hepa- Biochem. Biophys. Res. Commun. 178:823. tocytes. Proc. Natl. Acad. Sci. USA. 90:3491. 33. Schmidt, K., E.R. Werner, B. Mayer, H. Wachter, and W.R. 17. Nunokawa, Y., N. Ishida, and S. Tanaka. 1993. Cloning of Kukovetz. 1992. Tetrahydrobiopterin-dependent formation of inducible nitric oxide synthase in rat vascular smooth muscle endothelium-derived relaxing factor (nitric oxide) in aortic en- cells. Biochem. Biophys. Res. Commun. 191:89. dothelial cells. Biochem. J. 281:297. 18. Beasley, D., and M. Eldridge. Interleukin-1/~ and tumor necrosis 34. Sakai, N, S. Kaufman, and S. Milstein. 1993. Tetrahydrobi- factor synergistically induce nitric oxide synthase in rat vas- opterin is required for cytokine-induced nitric oxide produc- cular smooth muscle cells. Am. J. Physiol. In press. tion in a murine macrophage cell line (RAW 264). Mol. Phar- 19. Gross, S.S., and R. Levi. 1992. Tetrahydrobiopterin synthesis. macol. 43:6. An absolute requirement for cytokine-induced nitric oxide 35. Werner-Felmayer, B.G., E.K. Werner, D. Fuchs, A. Hausen, generation by vascular smooth muscle.J. Biol. Chem. 267:25722. G. Reibnegger, and H. Wachter. 1990. Tetrahydrobiopterin- 20. Hibbs, J.B.J., R.R. Taintor, and Z. Vavrin. 1987. Macrophage dependent formation of nitrite and nitrate in murine fibroblasts. cytotoxicity: role for r-arginine deiminase and imino nitrogen J. Exp. Med. 172:1599. oxidation to nitrite. Science (Wash. DC). 235:473. 36. Bernhardt, J., M.K. Tschudi, Y. Dohi, I. Gut, B. Urwyler, 21. Gross, S.S., D.J. Stuehr, K. Aisaka, E.A. Jaffe, R. Levi, and F.R. Buhler, and T.F. Luscher. 1991. Release of nitric oxide O.W. Griffith. 1990. Macrophage and endothelial cell nitric from human vascular smooth muscle cells. Biochem. Biophys.

79 Beasleyand McGuiggin Res. Comm. 180:907. 45. Ochoa, J.B., A.O. Udekwu, T.R. Billiar, R.D. Curran, F.B. 37. Martin, W., G.M. Villani, D. Jothianandan, and R.E Furch- Cerra, R.L. Simmons, and A.B. Peitzman. 1991. Nitrogen oxide gott. 1985. Selective blockade of endothelium-dependent and levels in patients after trauma and during sepsis. Ann. Surg. glyceryl trinitrate-induced relaxation by hemoglobin and by 214:621. methylene blue in the rabbit aorta. J. Pharmacol. Exp. Ther. 46. Palmer, R.M.J., M.S. Hickery, I.G. Charles, S. Moncada, and 232:708. M.T. Bayliss. 1993. Induction of nitric oxide synthase in human 38. Kontos, H.A., and E.P. Wei. 1993. Hydroxyl radical-dependent chondrocytes. Biochem. Biophys. Res. Commun. 193:398. inactivation of guanylate cyclasein cerebral arterioles by methy- 47. Cameron, M.L., D.L. Granger, J.B. Weinberg, W.J. Kozumbo, lene blue and by LY83583. Stroke. 24:427. and H.S. Koren. 1990. Human alveolar and peritoneal macro- 39. Gibson, Q.H., and F.J.W. P,oughton. 1957. The kinetics and phages mediate fungistasis independently of L-arginine oxida- equilibria of the reactions of nitric oxide with sheep hae- tion to nitrite or nitrate. Am. Rev. Respir. Dis. 142:1313. moglobin, j, Physiol. (Lend.). 136:507. 48. Murray, H.W., and R.F. Teitelbaum. 1992. L-arginine- 40. Janssens, S.P., A. Shimouchi, T. Quertermous, D.B. Bloch, dependent reactivenitrogen intermediates and the antimicrobial and K.D. Bloch. 1992. Cloning and expression of a cDNA effect of activated human mononuclear phagocytes. J. Infect. encoding human endothelium-derived relaxing factor/nitric Dis. 165:513. oxide synthase, j. Biol. Chem. 267:14519. 49. Padgett, E.L., and S.B. Pruett. 1992. Evaluation of nitrite 41. Junquero, D.C., T. Scott-Burden, V.B. Schini, and P.M. Van- production by human monocyte-derivedmacrophages. Biochem. houtte. 1992. Inhibition of cytokine-induced nitric oxide Biophys. Res. Commun. 186:775.

production by transforming growth factor-j81in human smooth 50. Schneemann, M., G. Schoedon, S. Hofer, N. Blau, L. Guer- Downloaded from http://rupress.org/jem/article-pdf/179/1/71/1395607/71.pdf by guest on 29 September 2021 muscle cells. J. Physiol. 454:451. rero, and A. Schaffner. 1993. Nitric oxide synthase is not a 42. Petros, A., D. Bennett, and P. Vallance. 1991. Effect of nitric constituent of the antimicrobial armature of human mono- oxide synthase inhibitors on hypotension in patients with septic nuclear phagocytes. J. Infect. Dis. 167:1358. shock. Lancet. 338:1557. 51. Sakai, N., and S. Milstien. 1993. Availability of tetrahydrobi- 43. Hibbs, J.B., C. Westenfelder, P,. Taintor, Z. Vavrin, C. Kab- opterin is not a factor in the inability to detect nitric oxide litz, R. Baranowski,J.H. Ward, R.L. Menlove, M.P. McMurry, production by human macrophages. Biochem. Biophys. Res. J. Kushner, and W.E. Samlowski. 1992. Evidence for cytokine- Commun. 193:378. inducible nitric oxide synthesis from r-arginine in patients 52. Denis, M. 1991. Tumor necrosis factor and granulocyte receiving interleukin-2 therapy. J. Clin. Invest. 89:867. macrophage-colony stimulating factor stimulate human mac- 44. Ochoa, J.B., B. Curti, A.B. Peitzman, R.L. Simmons, T.R. rophages to restrict growth of virulent Mycobacterium avium Billiar, R. Hoffman, R. Rault, D.L. Longo, W.J. Urba, and and to kill avirulent M. avium: killing effector mechanism de- A.C. Ochoa. 1992. Increased circulating nitrogen oxide after pends on the generation of reactive nitrogen intermediates, j. human tumor immunotherapy: correlation with toxic hemo- Leukocyte Biol. 49:380. dynamic changes. J. Natl. Cancer. Inst. (Bethesda). 84:864.

80 Cytokine-inducedcGMP in Human Vascular Smooth Muscle Cells