INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

NITRIC OXIDE (NO) IS A SMALL inter- and intracellular signaling molecule generated by a family of enzymes termed NO synthases (NOS) through the oxidation of L-arginine to citrulline (25). There are three well-characterized isoforms of NOS that are named after the tissues from which they were originally isolated and include endothelial (eNOS), neuronal, and inducible (iNOS) (12). Both eNOS and neuronal NOS are constitutively expressed and require increased Ca²⁺ levels for maximal activity, whereas iNOS is induced in response to proinflammatory molecule exposure (27). Calmodulin is bound to iNOS even at resting Ca²⁺ concentrations, leading to the production of greater amounts of NO (8). Evidence from genetic models, as well as experiments with pharmacological inhibitors, has shown that low amounts of NO generated by eNOS are important for normal vascular function, regulating vascular tone (20), remodeling (38), angiogenesis (34), and inhibiting platelet aggregation and leukocyte adhesion to the vessel wall (33). The higher amounts of NO produced through the iNOS play an important role in host-defense mechanisms but also contribute to the pathophysiology of several inflammatory conditions.

Endothelial cells (ECs) are not only a source of NO in the vessel wall but also respond to NO through activation of soluble guanylyl cyclase (sGC) (1). sGC is a heterodimeric molecule comprising two subunits ( and ) with approximate molecular masses of 82 and 70 kDa, respectively (1). Activation of sGC by NO results from the binding of NO to the heme moiety of sGC, which induces a conformational change that increases its activity, leading to the conversion of GTP to cGMP. cGMP is intimately involved in many signal transduction pathways, targeting cGMP-dependent protein kinase, cGMP-gated cation channels, and cGMP-regulated phosphodiesterases (1).

Until now, most studies have focused on and characterized the expression and function of NOS in ECs from large conduit vessels of different origins, with little information existing in cells isolated from resistance vessels. However, it is known that NO generation differs, depending on the size of the vessel (21). Differences in the expression of eNOS have been documented in transgenic mice containing 1.6 kb of the 5' flanking region of the human eNOS promoter (17). Transgene expression was documented in micro- and macrovascular ECs of the heart, brain, and skeletal muscle but not in the vasculature of the liver, kidney, and spleen. In the pulmonary vascular bed, the role of NO in endothelial-dependent relaxation is enhanced in conduit compared with resistance pulmonary arterial rings (2). In the present study, we investigated regulation and expression of components of the NOS-NO-cGMP signaling pathway in mesenteric ECs (MEC). We report here that MEC do not express eNOS mRNA or protein but generate large amounts of NO through the iNOS after treatment with cytokines. Moreover, we provide evidence for differential regulation of iNOS expression within the vessel wall, as the presence of the microtubule depolymerizing agent colchicine attenuates iNOS protein levels in smooth muscle cells treated with cytokines but has no effect on nitrite accumulation and iNOS protein expression in cytokine-treated MEC.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Chemicals and reagents. Sprague-Dawley rats were purchased from Harlan (Indianapolis, IN). Tissue culture plasticware was obtained from Corning Glass (Corning, NY), whereas medium 199 growth medium was purchased from GIBCO Laboratories (Grand Island, NY). Fetal calf serum was obtained from Hyclone Laboratories (Logan, UT), and ¹²⁵I was obtained from DuPont NEN (Boston, MA). RNAzol was purchased from Biotecx Laboratories (Houston, TX), and the GeneAmp RNA PCR kit was purchased from Perkin Elmer (Norwalk, CT). Enhanced chemiluminescence detection system was obtained from Amersham International (Buckinghamshire, UK). Protein binding dye, polyvinylidene difluoride membrane, dry milk, Tween 20, and other immunoblotting reagents were all purchased from Bio-Rad (Richmond, CA). X-ray film was obtained from Kodak, and interleukin (IL)-1 was purchased from Boehringer Mannheim (Mannheim, Germany). All other chemical reagents, including penicillin, streptomycin, succinyl tyrosine cGMP methyl ester, IBMX, sodium nitroprusside (SNP), atriopeptin II (APII), bradykinin, ACh, BSA, Nonidet P-40, phenylmethylsulfonyl fluoride, aprotinin, and EDTA, were obtained from Sigma Chemical (St. Louis, MO).

Isolation and culture of rat vascular cells. Animal handling and euthanasia were in accordance with guidelines from the Institutional Committee on Animal Use for Research and Education. Rat aortic smooth muscle cells (RASMC) were isolated and grown as previously described (36). Briefly, aortas were dissected from four male rats, fatty tissue was removed, and the vessels were incubated for 25 min in Hanks' balanced salt solution containing (per ml) 150 units Worthington collagenase II, 1 mg Worthington soybean trypsin inhibitor, and 0.75 unit Worthington elastase. The adventitia was then removed, and the lumen of the vessels was gently flushed to remove the endothelium. The vessels were rinsed in Hanks' balanced salt solution and then incubated with fresh enzyme solution for 1 h in a shaking 37°C water bath. The cell suspension was centrifuged, the supernatant discarded, and the pellet resuspended in DMEM-F12 supplemented with 1.4 mM L-glutamine and 10% fetal bovine serum. RASMC thus obtained were characterized by their appearance and expression of -actin. For isolation of MEC, the mesenteric vascular bed along with its surrounding adipose tissue was excised from three to four animals and incubated with an enzyme solution (containing collagenase, deoxyribonuclease, papain, dithiothreitol, and BSA) for 45 min at 37°C in a shaking water bath, as previously described (41). After digestion for 45 min, cells were then centrifuged and plated in medium 199 supplemented with 10% fetal bovine serum, 0.2 g/l L-glutamine, 100 U/ml penicillin, and 0.1 mg/ml streptomycin. MEC thus obtained were characterized as ECs by their cobblestone monolayer appearance, angiotensin-converting enzyme activity, uptake of acetylated low-density lipoprotein, and expression of von Willebrand factor. Smooth muscle cells were used at passages 3-6, and MEC were used at passages 1 or 2.

Determination of intracellular cGMP levels. ECs were exposed to SNP (0-100 µM) or rat APII (0-0.1 µM) for 15 min at 37°C in the presence of the phosphodiesterase inhibitor IBMX (1 mM). Medium was then aspirated, and 500 µl of 0.1 N HCl were added to each well to stop enzymatic reactions and to extract cGMP. Thirty minutes later, the HCl extract was collected, and cell remnants were removed from the wells by adding hot 1.0 N NaOH and scraping the well with a rubber policeman. The HCl extract was analyzed for cGMP by RIA, and the NaOH-solubilized samples were used to determine protein concentration. Because MEC express low levels of sGC, we used a bioassay system to determine the release of biologically active NO from these cells. For these experiments, RASMC were grown in 24 multiwell plates and served as reporter cells to detect NO release from the MEC, which were grown on glass coverslips. NO-stimulated cGMP accumulation in IBMX (1 mM, 10 min)-pretreated reporter cells was determined in short-term cocultures, established by transferring the coverslips with the ECs into wells containing the smooth muscle. After 15 min, the IBMX containing solution was aspirated, the coverslips with MEC were removed from the wells, and intracellular cGMP content was quantified as described above. To test for NO release from the endothelial isoform of NOS (eNOS), MEC were exposed to either vehicle for basal NO release or bradykinin (1 µM) for stimulated NO release during the 15 min cocultivation period. To test for the expression of the inducible isoform of NOS (iNOS), MEC were pretreated with IL-1 (135 U/ml) and tumor necrosis factor (TNF)- (150 units) for 24 h before the 15 min coculture was established with the reporter cells.

RIA for cGMP. The radioligand ¹²⁵I-labeled succinyl cGMP-tyrosine methyl ester was prepared in our laboratory. Stock solutions of the succinyl tyrosine methyl ester of cGMP were diluted in 50 mM sodium acetate buffer (pH 4.75) and iodinated using carrier-free ¹²⁵I. The iodination reaction products were separated by reverse-phase high-performance liquid chromatography. With the use of a monoclonal antibody for cGMP, radioimmunoassay was performed in the gamma flow automated RIA system (36). Standard stock solutions of cGMP (20 µM) were prepared in 0.1 N HCl, and the absorbance of the solution was routinely monitored spectrophotometrically (Shimadzu, UV 160U). Standard dilutions (0.63-80 nM) were made from the stock solution. The HCl extract containing cGMP was used for RIA directly.

Protein determination. The NaOH solubilized samples were centrifuged (2,000 rpm for 5 min at room temperature), supernatants were collected, and protein concentration was subsequently determined by the Bradford method (6). Sample aliquots were combined with the protein binding dye, and optical density was subsequently determined at 630 nm using a multiwell plate reader (Dynetech Laboratories). BSA fraction V was used as the protein standard.

Nitrite determination. Nitrite, a stable breakdown product of NO, was measured in the supernatants of ECs or smooth muscle cells. Cells were pretreated with the agent of interest for 30 min and then incubated with a cytokine mixture of IL-1 (135 U/ml) and TNF- (150 U/ml). After 24 h, aliquots of the cell culture supernatants (150 µl of media) were removed and centrifuged at 1,000 rpm to pellet cells. Nitrite concentration was then determined by the Griess reaction. Briefly, medium was combined with an equal volume of the Griess reagent (1% sulfanilamide and 0.1% napthylethylenediamide in 5% phosphoric acid), and the optical density was measured at 550 nm using a Molecular Dynamics microplate reader. Nitrite concentration was calculated by comparison with 550-nm optical density of standard solutions of sodium nitrite prepared in cultured medium.

RT-PCR. With the use of published sequences, primers for the eNOS and iNOS, and the ₁- and ₁-subunits of sGC, RNA was reverse transcribed and amplified using a commercially available kit (GeneAmp RNA PCR kit) in a DNA Thermal Cycler 480 (Perkin Elmer). cDNA was amplified as previously described (36). After amplification, 10 µl of the PCR reaction mixture were electrophoresed on 0.9% agarose gels, stained with ethidium bromide, visualized on a ultraviolet transilluminator, and photographed. A molecular weight standard consisting of 100-bp increments between 100 and 2,600 bp (Pharmacia, LKB Biotechnology) was used to confirm the predicted PCR product size.

Immunoblotting. Cells were lysed in lysis buffer (1% Nonidet P-40, 150 mM NaCl, 20 mM HEPES, pH 7.0, 1 mM EDTA, 1 µM aprotinin, and 1 mM phenylmethylsulfonyl fluoride), lysates were centrifuged at 20,000 rpm, the supernatant fraction was subsequently collected, and protein concentration was determined by the Bradford method. Proteins were then resolved on 7.5% SDS-polyacrylamide gels and transferred to polyvinylidene difluoride membrane at 60 V for 1.5 h at 4°C in buffer containing 25 mM Tris and 700 mM glycine. Membranes were subsequently incubated overnight at 4°C with 5% dry milk in buffer containing 0.1% (vol/vol) Tween 20 in TBS (TTBS) to block nonspecific binding. The next day, membranes were incubated with isoform-specific antibodies in 5% milk in TTBS for 1 h at room temperature, washed three times with TTBS for 20 min each time, and then blocked additionally for 1 h with 5% milk in TTBS. Membranes were then incubated for 1 h with a horseradish peroxidase-conjugated secondary antibody. Immunoreactive protein bands were subsequently visualized using the enhanced chemiluminescence system after exposure to X-ray film. To check for equality in loading and transfer, membranes were subsequently incubated with a monoclonal antibody against tubulin.

Measurement of small mesenteric artery vascular relaxation. Male Wistar rats (300-400 g) were anesthetized with pentobarbital sodium (50 mg/kg ip), and heparin (100 units) was administered into the left ventricle. A portion of the mesentery was removed and placed in modified Krebs-Ringer bicarbonate solution [composition (in mM): 118.3 NaCl, 4.7 KCl, 2.5 CaCl₂, 1.2 MgCl₂, 1.2 KH₂PO₄, 25 NaHCO₃, 11.1 dextrose], which had been chilled and oxygenated (20% O₂ and 5% CO₂). A section of the mesenteric artery (200-300 µm diameter and 1-2 mm long) was isolated from the surrounding tissues microscopically. Vessels were transferred to the tissue bath and mounted between two glass micropipettes (100-µm-diameter tips) with 10-0 ophthalmic suture. The tissue bath was then transferred to the stage of an Olympus inverted light microscope coupled to a monitor and video dimension analyzer (Living Systems Instrumentation, Burlington, VT). Mesenteric artery intraluminal diameter was continuously observed on a video monitor and recorded on a Grass polygraph.

9

5

4

Data and statistical analysis. Data are presented as means ± SE of the indicated number of individual observations. Statistical comparisons between groups were performed using one-way ANOVA followed by Dunnett's or Newman-Keuls post hoc test or Student's t-test as appropriate. For vascular reactivity measurements, statistical comparisons were performed using repeated-measures analysis of variance. Significance levels were determined by Student's modified t-test with Bonferroni correction for multiple comparisons utilizing the error mean square term from the ANOVA. Differences among means were considered significant at P < 0.05.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

To establish optimal growth conditions, MEC were seeded in complete growth medium containing FCS and EC growth factor (ECGF), a preparation that mainly contains fibroblast growth factor from bovine cerebellum. The next day, medium was changed, and cells were maintained either in complete growth medium, in medium with FCS alone, or in medium containing no FCS at all. As shown in Fig. 1, cells grown in medium supplemented with FCS and ECGF proliferated rapidly, reaching confluence after 3 days. In the presence of medium containing FCS alone, MEC growth rate was retarded, and MEC number declined in cultures maintained in the absence of FCS and ECGF.

View larger version (13K): [in this window] [in a new window]   Fig. 1.   Growth characteristics of mesenteric endothelial cells (MEC). MEC were seeded in complete medium [FCS + endothelial cell growth factor (ECGF)]. The next day, medium was changed to complete medium (), FCS alone (), or no FCS (), and cell no. was assessed for 5 days. Data are expressed as means ± SE; n = 4 wells/group.

To examine the expression of eNOS and iNOS in MEC, we measured mRNA and protein levels of the two NOS isoforms. As shown in Fig. 2A, no detectable levels of eNOS mRNA were observed in MEC, whereas high levels were found in whole rat heart mRNA. MEC expressed iNOS mRNA only when exposed to a cytokine mixture, whereas no signal was observed in control cells (Fig. 2B). Furthermore, MEC express mRNA for both the ₁- and ₁-subunits of sGC, albeit in low amounts (Fig. 2C). Immunoblot analysis of NOS isoforms revealed that neither MEC nor RASMC expresses eNOS (Fig. 3A). On the other hand, iNOS expression is induced when both MEC and RASMC are treated with a cytokine mixture, an effect that is inhibited in both cell types by the tyrosine kinase inhibitor genistein (Fig. 3B). iNOS expression level was inhibited by colchicine in RASMC but not in MEC, suggesting that iNOS expression is differentially regulated in these two cell types (Fig. 3C).

View larger version (31K): [in this window] [in a new window]   Fig. 2.   Nitric oxide synthase (NOS) and soluble guanylyl cyclase (sGC) expression in MEC. A: RT-PCR analysis using endothelial NOS (eNOS)-specific primers. Total rat heart RNA was used as a positive control. No product can be detected in MEC after 35 cycles of amplification. B: MEC express inducible NOS (iNOS) when exposed to a cytokine mixture [CMIX; 130 U/ml interleukin (IL)-1 and 150 U/ml tumor necrosis factor (TNF)-] for 24 h. C: MEC express low levels for both the 1- and 1-subunits of sGC. For all RT-PCR reactions, 500 ng were reverse transcribed and amplified for 35 cycles. A marker consisting of DNA fragments in increments of 100 bp (lower band 100 bp) was used to estimate the size of the PCR products. + and  indicate either the presence or absence of RT in the reverse transcription reaction, respectively. CTL, control.

View larger version (20K): [in this window] [in a new window]   Fig. 3.   Immunoblotting analysis and comparison of NOS isoforms in MEC and rat aortic smooth muscle cells (RASMC). A: eNOS is not expressed in RASMC [aortic smooth muscle (ASM)] and MEC in culture. Rat heart homogenates were used as positive controls. Protein (50 µg) for each treatment was subjected to SDS-PAGE, transferred to polyvinylidene difluoride membrane, and then detected with eNOS-specific antibody. B: iNOS expression is inhibited in both MEC and RASMC by the tyrosine kinase inhibitor genistein (GEN). Cells were treated with a CMIX (130 U/ml IL-1 and 150 U/ml TNF-) for 24 h in the presence or absence of 10 µM GEN. C: iNOS protein expression in RASMC, but not MEC, is inhibited by the microtubule-depolymerizing agent colchicine (COL). Cells were treated with a CMIX for 24 h in the presence or absence of 10 µM COL.

To study further the functionality of iNOS expression in MEC, we treated these cells with cytokines and compared their responses with those in RASMC. Nitrite production was increased in both cell types after exposure to IL-1, whereas TNF- alone had no effect on NO formation (Fig. 4). Unlike RASMC, cotreatment with TNF- and IL-1 did not further enhance NO production in MEC. Pretreatment with N^G-nitro-L-arginine methyl ester (L-NAME) abolished nitrite accumulation in response to cytokine treatment in both cell types. Confirming the protein data, nitrite levels in cytokine-treated RASMC, but not MEC, were decreased by the microtubule depolymerizing agent colchicine. Moreover, taxol reversed the effect of colchicine on nitrite levels in RASMC.

View larger version (17K): [in this window] [in a new window]   Fig. 4.   Regulation of nitrite production from MEC cultures exposed to cytokines. Cells were treated with IL-1 (130 U/ml), or TNF- (150 U/ml), or both for 24 h in the presence of NG-nitro-L-arginine methyl ester (L-NAME; 1 mM), COL (10 µM), or COL + taxol (TAX) (10 µM each). Nitrite concentration in the culture medium was then determined using the Griess reaction. Data are expressed as means ± SE; n = 3 wells/group. * P < 0.05 vs. control. # P < 0.05 vs. IL-1 + TNF- control group.

To determine whether MEC express guanylate cyclases, we measured GMP accumulation in response to the sGC and particulate GC (pGC) activators SNP and APII. As shown in Fig. 5A, minimal cGMP accumulation in MEC was only noted at higher SNP concentrations (3-fold at 100 µM), in contrast to RASMC. On the other hand, APII stimulated higher cGMP production in MEC compared with RASMC (Fig. 5B). Because of the low sGC activity in MEC, we assessed NO production from these cells by measuring accumulation of its surrogate marker cGMP in reporter RASMC. Under basal conditions or in response to bradykinin, MEC do not release any NO (Fig. 6). However, there was significant NO production from MEC treated with a cytokine mix (150 U/ml TNF- and 135 U/ml IL-1), supporting the finding that MEC can express iNOS if exposed to proinflammatory stimuli. In NOS activity assays, L-NAME-inhibitable arginine-to-citrulline conversion was only noted in MEC treated with the cytokine mixture (data not shown). The evidence supporting the presence of different NOS isoforms in MEC is summarized in Table 1.

View larger version (15K): [in this window] [in a new window]   Fig. 5.   MEC express low sGC and high particulate guanylyl cyclase activity. Cells were exposed to the indicated concentration of sodium nitroprusside (SNP; A) or atriopeptin II (APII; B) for 15 min in the presence of 1 mM IBMX. cGMP was extracted in HCl and measured by RIA. C, control. Data are expressed as means ± SE; n = 4 wells/group.

View larger version (15K): [in this window] [in a new window]   Fig. 6.   Release of nitric oxide from MEC. Untreated MEC or MEC treated with TNF- (150 U/ml) and IL-1 (130 U/ml) for 24 h were washed with Earle's balanced salt solution and incubated for an additional 30 min in the presence or absence of L-NAME (100 µM). Release of biologically active nitric oxide from MEC under basal (vehicle) or stimulated conditions (bradykinin; 10 µM) or from MEC treated with the cytokine mixture (TNF- + IL-1) was determined in a bioassay system using smooth muscle cells as reporters, as described in MATERIALS AND METHODS. Data are expressed as means ± SE; n = 4 wells/group. * P < 0.05 vs. vehicle.

To test whether mesenteric vessels constitutively express biologically active NO, small arteries were preconstricted with PE, and ACh-induced relaxation was determined. As shown in Fig. 7, vessels dilated in the presence of ACh in a concentration-dependent manner. Pretreatment of the vessels with L-NNA attenuated the dilatory response to ACh, suggesting that these vessels release NO ex vivo. Moreover, high concentrations of extracellular K⁺ blocked relaxation to ACh, confirming reports that dilation in response to ACh in this vascular bed is mainly dependent on endothelium-derived hyperpolarizing factor (9).

View larger version (14K): [in this window] [in a new window]   Fig. 7.   ACh-induced relaxation of small mesenteric arteries. Vessels were pretreated with vehicle, N-nitro-L-arginine (L-NNA; 300 µM), or KCl (40-50 mM) and preconstricted with phenylephrine (PE). After pretreatments, dose-response curves to ACh were determined, and %relaxation was determined. Data are presented as means ± SE; n = 7 wells/group. , PE; , L-NNA + PE; , KCl. * P < 0.05 vs. PE.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Although ECs isolated from different vascular beds express common markers, such as platelet-endothelium adhesion molecule-1, von Willebrand factor, and angiotensin-converting enzyme, their biological properties differ greatly, depending on the vascular bed of origin (4, 16, 18, 43). Most of the published studies on NOS regulation have been performed on ECs isolated either from conduit vessels (aorta, pulmonary artery, or umbilical vein) or microvascular beds of the lung, brain, or heart of different species. In the present study, we investigated the expression of different components of the NO-cGMP pathway in cells isolated from the mesenteric bed of the rat.

cGMP accumulation is the most reliable way to determine release of biologically active NO in cells. The "receptor" for NO, sGC, is a ubiquitous enzyme, expressed at high levels in the lung, liver, and brain, that is also expressed in cultured vascular ECs (1). Stimulation of MEC with a NO donor revealed that these cells do respond to NO by increasing their cGMP content; however, sGC levels are low compared with smooth muscle cells. On the contrary, MEC express higher levels of pGC activity compared with smooth muscle cells. The high pGC and low sGC activity exhibited by MEC is in agreement with what has been reported for pulmonary arterial and aortic ECs from different species (29). The reduced ability of MEC to produce cGMP in response to NO made it necessary to use reporter smooth muscle cells to determine the release of NO from MEC.

ECs have the capacity to express both a low-output constitutive NOS and an iNOS (12). Under resting conditions or in response to the eNOS activator bradykinin, MEC did not produce measurable NO as evidenced by the lack of L-NAME-inhibitable cGMP accumulation in reporter smooth muscle cells. Optimal NOS enzymatic activity requires the presence of cofactors, such as tetrahydrobiopterin, heme, flavin mononucleotide, and NADPH (25). NOS activity is also negatively and positively regulated through phosphorylation and protein-protein interactions (14, 15, 23). The inability of MEC to produce NO is not due to a deficiency in cofactors, but rather to the lack of eNOS expression, as suggested by the absence of eNOS mRNA and protein in these cells. Our results are in line with the findings of Balligand et al. (3), who demonstrated that cardiac microvascular ECs express no constitutive NOS activity but have a robust increase in iNOS activity in response to inflammatory cytokines. In addition, mice expressing a 5.2-kb eNOS promoter-reporter transgene demonstrated high -galactosidase expression in large- and medium-sized vessels but not small arterioles, capillaries, and venules (42). Similarly, in humans, robust endogenous expression of eNOS mRNA and protein appears to be predominantly restricted to large vessels (45). Although MEC cannot be used to study eNOS function and regulation, this unexpected phenotype of MEC can be exploited in transfection experiments. Biochemical and molecular studies on eNOS regulation are routinely performed in non-ECs that do not express eNOS constitutively, such as COS-7 and HEK cells (26). To study eNOS in an eNOS-deficient EC, cells from eNOS/ mice would have to be obtained, with the obvious limitations of cost and difficulty in isolating murine EC. Because of the ease in the isolation procedure, MEC could offer an alternative to study eNOS function in the context of a primary or low-passage nontransformed EC.

One of the ways ECs respond to proinflammatory stimuli is by increasing NO generation through the iNOS (12). In rat cells, IL-1 has been demonstrated to be the most potent inducer of iNOS expression in cells isolated from the aorta, as well as heart and brain microvessels (3, 5, 24). Exposure of MEC to IL-1, but not TNF-, increased nitrite accumulation in an L-NAME-inhibitable manner. Moreover, the cytokine mixture (IL-1 and TNF-) increased iNOS gene expression, release of biologically active NO, and NOS activity, suggesting that MEC, although they do not show eNOS expression, can produce NO through an iNOS pathway.

Exposure of cells to IL-1 promotes phosphorylation on serine and threonine residues of a number of intracellular substrates (35). In addition, IL-1 and TNF- have been shown to trigger tyrosine phosphorylation in several cell types, including ECs (31). Some of these tyrosine phosphorylation cascades couple to increased gene expression, and tyrosine kinase inhibitors have been demonstrated to inhibit several cytokine-inducible molecules in ECs (19, 32). To further study iNOS expression in MEC, ECs were pretreated with a broad-spectrum tyrosine kinase inhibitor, genistein, before stimulation with cytokines. Genistein abolished the increase in iNOS protein levels in MEC as well as in RASMC. This observation is in line with the finding that genistein and other nonspecific tyrosine kinase inhibitors abolish iNOS expression in vascular cells, including smooth muscle, neutrophils, and macrophages (11, 28, 37).

A variety of observations suggest that changes in the cytoskeleton are linked to altered gene expression. Ribosomes, RNA, and signaling molecules (GTPases, tyrosine and serine/threonine kinases, as well as transcription factors) are known to localize to the cytoskeleton (22). Both reduced and increased gene expression have been documented in response to microtubule network disruption (27, 44). In ECs, microtubule depolymerization decreases tissue plasminogen activator expression (39). Moreover, molecules important for iNOS expression (members of the mitogen-activated protein kinase family and nuclear factor-B) bind to microtubules in a manner that affects their activity (22, 40). Our laboratory has previously characterized the role of microtubules in mediating the activation of iNOS in smooth muscle cells. Pretreatment of the cultured rat vascular smooth muscle cells with the microtubule depolymerizing agent colchicine inhibits lipopolysaccharide-induced iNOS expression (27). Taxol, which stabilizes microtubule polymerization, reverses the actions of colchicine (27). Contrary to what has been observed in smooth muscle cells, pretreatment of macrophages with colchicine only minimally reduced nitrite production, whereas it stimulated iNOS expression in catecholaminergic neurons (10, 44). In the present study, colchicine prevented the IL-1 + TNF--induced accumulation of nitrite and abolished iNOS protein levels in smooth muscle cells. In contrast, however, modulation of microtubule polymerization by colchicine did not affect iNOS expression or nitrite formation in MEC after cytokine exposure. Differences in the regulation of iNOS expression between endothelial and smooth muscle cells have been previously reported, as interferon- induces iNOS expression in cardiac myocytes but not in cardiac microvascular ECs (40). Moreover, dexamethasone minimally inhibits iNOS mRNA and protein levels in cardiac microvascular cells but abolishes iNOS expression in smooth muscle cells (3). Our findings suggest that microtubules do not play a prominent role in the induction of iNOS in MEC and support the notion of differential regulation of iNOS in EC.

In vivo, endothelium-dependent regulation of vascular tone and relaxation occurs through the production of endothelium-derived relaxing factors, including NO, vasodilator prostaglandins, and endothelium-derived hyperpolarizing factor (30). The contribution of each of the above endothelium-derived relaxing factors to vasorelaxation depends on the vasodilator used, the size of the vessel, the vascular bed, and the species examined. For example, ACh-induced relaxation in the mouse aorta is abolished in eNOS/ animals, whereas ACh-induced relaxation in hamster skeletal microvessels is NO independent and occurs through endothelium-derived hyperpolarizing factor generation (9, 20). To test whether rat mesenteric microvessels have the ability to generate NO, we evaluated ACh-induced relaxations in the presence and absence of a NOS inhibitor. Similar to what has been reported for the rat and hamster mesenteric beds, preincubation of the vessels with L-NNA attenuated the vasodilatory action of ACh (9, 13). Collectively, our data suggest that the mesenteric endothelium releases biologically active NO ex vivo but stops expressing eNOS when placed in culture.

In summary, we have characterized the NO-cGMP signaling pathway in MEC and have demonstrated that these cells do not express eNOS but are capable of generating large amounts of NO if stimulated with cytokines. Moreover, iNOS expression in MEC, vs. that to smooth muscle cells, is differentially regulated because microtubule depolymerization does not affect MEC iNOS induction. Because of the unique expression profile of NOS, MEC might be useful in studying the role of iNOS in EC biology (growth, migration, angiogenesis) without having to account for the low-level constitutive production of NO through eNOS. In addition, these cells could be useful in studies investigating the biochemical and cell biological properties of eNOS in vivo in the context of an EC in transfection experiments.