INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

ENDOTHELIAL CELL (EC) activation and subsequent expression of leukocyte adhesion molecules are initial events in multiple physiological and pathological processes. An initial step in the development of an inflammatory response involves the adherence of blood-borne leukocytes to activated EC (reviewed in Ref. 34). Leukocyte adhesion to activated endothelium is mediated through the expression of leukocyte adhesion proteins on the EC surface. Two specific proteins that medimonocyte adhesion to vascular endothelium are E-selectin and vascular cell adhesion molecule 1 (VCAM-1) (7). E-selectin (formerly known as endothelial-leukocyte adhesion molecule 1) is a surface glycoprotein in the selectin family that is expressed only in EC after treatment with various EC-activating agents (6). VCAM-1 belongs to the immunoglobulin superfamily of adhesion molecules and was first reported (29) to mediate the binding of lymphocytes and monocytes to activated EC.

Recently, viral double-stranded RNA (dsRNA) has been shown (36) to induce leukocyte adhesion to EC. dsRNA is an intermediate within the replication cycle for RNA viruses, as well as some DNA viruses, and has been demonstrated in many viral-infected cells (8). A viral cause for atherosclerosis has been investigated for many years (14). Etingin et al. (14) have reported that virus-induced vascular injury initiates events leading to inflammation, thrombosis, and atherosclerosis. Intracellular dsRNA from virus infection of a cell induces gene expression that can be mimicked by exogenous dsRNA (15). Polyinosinic-polycytidylic acid [poly(I-C)] is a synthetic dsRNA that is often used to simulate the viral-infected state when added exogenously to cells. Poly(I-C) has been recently reported (13, 25, 40) to induce leukocyte adhesion to EC through the expression of the EC adhesion proteins E-selectin, VCAM-1, and intercellular adhesion molecule 1 (ICAM-1). The induction of adhesion molecules in EC by poly(I-C) is thought to occur independently of its ability to induce type I interferons in that interferon- was shown not to induce the expression of these genes (40).

Nitric oxide (NO), the product of the activity of the constitutively expressed enzyme endothelial nitric oxide synthase (NOS), has been shown (10, 21) to decrease cytokine-induced EC activation in confluent cultures. Because NO is a readily available response mediator in EC and because viral infections elicit changes in NO levels, we investigated whether NO was involved in the poly(I-C)-induced expression of leukocyte adhesion molecules on vascular EC.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Materials. Purified human monocyte interleukin-1 (IL-1) was a gift from Otsuka Pharmaceutical (Rockville, MD). The monoclonal antibodies 7A9 and 2G7 against E-selectin and VCAM-1, respectively, and the cDNA probes for E-selectin and VCAM-1 were generous gifts from Dr. Walter Newman (Otsuka Pharmaceutical). Heparin, lipopolysaccharide (LPS), and all media components, as well as other reagents not specifically mentioned, were purchased from Sigma Chemical (St. Louis, MO). Poly(I-C) was purchased from Pharmacia Biotech (Piscataway, NJ) and freshly prepared for each experiment according to the recommendations of the manufacturer. N^G-monomethyl-L-arginine monoacetate (L-NMMA) and L-N⁵-(1-iminoethyl)ornithine (L-NIO) were obtained from Alexis (San Diego, CA). ¹²⁵I-labeled streptavidin was purchased from Amersham (Arlington Heights, IL). [³²P]dCTP and sodium [⁵¹Cr]chromate were the products of Du Pont NEN (Boston, MA). All tissue culture plastic ware was purchased from Costar (Cambridge, MA). Biotin-conjugated, affinity-purified F(ab')₂-fragment goat anti-mouse immunoglobulin (Ig) G + IgM [heavy + light chain (H+L)] was obtained from Jackson ImmunoResearch Labs (West Grove, PA). Nytran membranes used for Northern analysis were purchased from Schleicher and Schuell (Keene, NH). All reagents used in cell culture were tested for endotoxin contamination using the limulus endotoxin detection assay kit from BioWhitaker (Walkersville, MD). Only endotoxin-free reagents were used for these studies.

Cell culture. Human umbilical vein EC (11) and human aortic EC (37) were isolated from human umbilical vein and human infrarenal arteries, respectively, by collagenase treatment as previously described. Cells were cultured in MCDB 107 medium supplemented with heparin (90 µg/ml), 15% fetal bovine serum (FBS; vol/vol), and EC growth supplement (150 µg/ml) derived from bovine hypothalamus. Culture plates were coated with fibronectin (1 µg/cm²). Primary cultured human umbilical vein EC or passaged human aortic EC between passages 1 and 4 were used where specified. U-937 cells originally derived from a human histocytic lymphoma were procured from the American Type Culture Collection (Rockville, MD), grown in suspension culture in RPMI-1640 medium containing 5% FBS, and routinely subcultured at a 1:5 ratio three times per week. All cell types were grown under 5% CO₂ at 37°C.

Monocytic cell adhesion to EC. U-937 cell adhesion was measured as previously described (11). Briefly, EC were grown in 24-well plates in appropriate media. U-937 cells were labeled for 90 min at 37°C with ⁵¹Cr (100 µCi, sodium chromate) in culture medium. The labeled cells were washed by centrifugation, and 10⁶ radiolabeled viable cells were added to each well of EC. After binding occurred for 1 h at 4 °C, the wells were washed and the cells lysed with 1% Triton X-100, and an aliquot was removed for gamma radiation counting. The number of U-937 cells bound per well was calculated from the initial specific activity [counts per minute (cpm) per cell] of the monocytic cell preparation. Spontaneous release of Cr from the monocytic cells during the assay incubation was <5% of the total count. We report the data using U-937 cells because the majority of the replicate experiments were done using this monocytic cell line; however, poly(I-C) induction of adhesion to EC was also confirmed using freshly isolated human monocytes.

EC surface expression of leukocyte adhesion proteins. EC were plated in 48-well cluster plates. EC were washed twice with medium containing 1% bovine serum albumin (BSA) and incubated for 1 h at 4°C with antibody (1 µg/ml). The wells were washed three times, biotin-conjugated, affinity-purified F(ab')₂-fragment goat anti-mouse IgG + IgM (H+L) (1:1,000 dilution) was added, and the plate was incubated at 4°C for 30 min. After the plates were washed three times, ¹²⁵I-streptavidin (0.25 µCi/well) was added to each well and the cells were incubated at 4°C for 15 min. Subsequently, the wells were washed four times, the cells lysed with 1% Triton X-100, and the radioactivity quantitated.

RNA isolation and Northern analysis. EC were washed with phosphate-buffered saline (PBS), and total RNA was extracted using the Trizol reagent from GIBCO BRL (Gaithersburg, MD) per manufacturer's instructions. Total RNA (20 µg) was electrophoresed through 1% formaldehyde agarose gels in the presence of ethidium bromide and capillary-blotted to Nytran membranes. High-specific-activity probes were prepared by labeling cDNA corresponding to VCAM-1, E-selectin, and GAPDH (glyceraldehyde-3-phosphate dehydrogenase) with the use of a random primer labeling kit (Pharmacia Biotech) using [-³²P]dCTP (3,000 Ci/mmol). The incorporated counts were separated from free label by gel filtration using a spin column (Sepharose CL-4B), and total cpm were determined from Cerenkov counts. The Nytran membranes were prehybridized and hybridized according to the manufacturer's instructions for at least 4 h at 42°C. The filters were washed serially in 2× sodium chloride-sodium citrate buffer (SSC)-0.5% sodium dodecyl sulfate (SDS) and 0.2× SSC-1% SDS at 42°C for a total of 1 h. Loading efficiency was determined by both ethidium bromide staining and intensity of GAPDH message for each sample. The relative intensity of message expression for either E-selectin or VCAM-1 was normalized using the expression of the "housekeeping" gene.

Nuclear extract preparation and electrophoretic mobility shift analysis. Nuclear proteins from EC were prepared using a modification of a previously described method (12). Briefly, cells were washed with iced PBS and then scraped and spun down. The cell pellets were then resuspended in 500 µl of buffer A [10 mM N-2-hydroxyethylpiperazine-N'-2-ethanesulfonic acid (HEPES) pH 7.9, 1.5 mM MgCl₂, 10 mM KCl, 1 mM dithiothreitol (DTT), and 1 mM phenylmethylsulfonyl fluoride (PMSF)] and spun down. The pellet was then resuspended in buffer A (80 µl/10⁷ cells) to which 0.1% Triton X-100 was added. After 10 min of incubation on ice, the samples were spun at 3,500 revolutions/min for 4 min at 4°C. The supernatant was removed, and the nuclear pellet was washed in buffer A (500 µl). The nuclear pellet was then resuspended in 50 µl of buffer C (20 mM HEPES, 25% glycerol, 0.42 M NaCl, 1.5 mM MgCl₂, 0.2 mM EDTA, 1 mM DTT, and 1 mM PMSF). After 30 min of incubation at 4°C, the samples were spun at 20,000 g for 10 min. The supernatant containing nuclear proteins was stored at 70°C until used for electrophoretic mobility shift analysis (EMSA). Protein concentration of the nuclear extracts was determined using the bicinchoninic acid (BCA) assay method (Pierce, Rockford, IL). The sequence of the 28-residue oligonucleotide probe was taken from the nuclear factor-B (NF-B) binding domain of the E-selectin promoter (5'-AGGCCATTTCTTTACTGG-3') (27). The oligonucleotide probes were annealed and labeled by a "filling-in" reaction using [-³²P]dCTP; unlabeled dTTP, dATP, and dGTP; and the Klenow fragment of DNA polymerase 1. Nuclear extracts (5 µg protein) were incubated for 20 min at room temperature in 20 µl total reaction mixture containing 5 × 10⁴ cpm ³²P-labeled probe (~1 ng), 225 µg/ml BSA, 0.1 mg/ml polydeoxyinosinic-polydeoxycytidylic acid, and binding buffer [12 mM HEPES pH 7.9, 4 mM tris(hydroxymethyl)aminomethane (Tris), 60 mM KCl, 1 mM EDTA, 12% glycerol, 1 mM DTT, and 1 mM PMSF]. Competition analyses were performed to verify the specificity of the shifted bands using 100- to 200-fold excess of unlabeled oligonucleotide that was coincubated with the nuclear extracts in the binding reaction for 15 min at room temperature before the radiolabeled oligonucleotide probe was added. Samples were loaded on nondenaturing 4% polyacrylamide gels in 1× Tris-glycine buffer, pH 8.5. Electrophoresis was performed at 15 V/cm. The gels were then dried and analyzed by autoradiography.

Transient transfection and luciferase assay. Each well of primary cultured human umbilical vein EC at ~60% confluence was transfected for 6 h in OptiMEM (modified Eagle's minimum essential medium) with 10 µg/ml Lipofectin reagent (GIBCO BRL) and 1 µg E-selectin promoter-driven pGL3 luciferase reporter vector. The cells were then washed twice and allowed to recover overnight in MCDB 107 complete media + 15% FBS. After this incubation period, the cells were induced with poly(I-C) in the presence or absence of N^G-methyl-L-arginine (L-NMA; 500 µM) pretreatment. Cells were then washed twice in PBS and harvested by scraping in 200 µl of lysis buffer (Promega, Madison, WI). Cell debris was pelleted by centrifugation, and the supernatant was assayed for luciferase activity and protein content. Luciferase activity was determined by combining cell extracts (75 µl) with luciferin reagent (150 µl; Promega) and measuring luminescence using a microplate luminometer. Transfections were carried out in triplicate on two separate occasions, and the data were corrected for protein content. Protein content of the cell extracts was measured using the BCA assay method (Pierce).

Measurement of NO production by chemiluminescence. The NO production by primary and passaged EC was measured using an NO analyzer (Sievers 280 NOA, Boulder, CO). Nitrate and nitrite present in overlying tissue culture media (4 µl) were converted to NO by a saturated solution of VCl₃ in 0.8 M HCl, and the NO was detected based on a gas-phase chemiluminescent reaction between NO and ozone (2). Nitrite and nitrate standards were also tested. Nitrite and nitrate standards displayed linearity between 0.05 and 30 µM (r²  0.95 for all experiments) and were detected with equal efficiency. NO levels were determined by interpolation from the known standard curves.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

It has been shown (3, 33) that primary cultured EC have higher levels of NO and NOS activity than cells passaged in culture. We wanted to confirm these results and measured NO in the conditioned media of primary cultured EC and passaged EC. We found that primary EC had nearly five times the basal levels of NO (20 µM) as EC carried out to passage 4 (4 µM) (Fig. 1) and that the NOS inhibitor L-NMA (500 µM) decreased the production of NO (Fig. 1). We also demonstrated that histamine-stimulated NO production by primary EC was significantly inhibited by L-NMA at concentrations as low as 100 µM (data not shown). Because primary cultured EC exhibited a significantly higher basal level of NO, we performed our studies using these cells.

View larger version (33K): [in this window] [in a new window]   Fig. 1.   Measurement of basal nitric oxide levels in human umbilical vein endothelial cells (HUVEC). Primary cultured EC and passage 4 (p.4) HUVEC were incubated in MCDB 107 medium + 2% fetal bovine serum (FBS) for 72 h in presence (hatched bars) or absence (solid bars) of NG-methyl-L-arginine (L-NMA; 500 µM). After incubation period, nitric oxide within a sample of conditioned medium was measured as described in MATERIALS AND METHODS. Data are reported as means ± SE (n = 3).

NO is readily available within EC and has been shown (20) to elicit antiviral effects. It is for these reasons that we investigated whether NO was involved in the signaling pathway of dsRNA within EC. We tested whether NOS inhibitors affected poly(I-C)-induced U-937 cell adhesion to primary cultured human umbilical vein EC. The NOS inhibitor L-NMA decreased the dsRNA-induced adhesion by ~70% with a 50% inhibitory concentration of ~200 µM (Fig. 2A). In repeat experiments, maximal inhibition occurred at 500 µM NOS inhibitor; therefore, we used this concentration in subsequent assays. These concentrations are consistent with those reported (19) to inhibit NOS activity in EC. L-NMA (1 mM) had no effect on cell viability as measured by the rate of protein synthesis. The same concentrations of L-NMA had no effect on IL-1-induced U-937 cell adhesion (Fig. 2B), indicating that the inhibitory effect of L-NMA was specific for the poly(I-C)-induced signaling pathway. Poly(I-C)-induced U-937 cell adhesion to subconfluent human aortic EC was also inhibited by L-NMA (45%) but did not affect IL-1 or LPS-induced adhesion to these cells (data not shown). Hence, the inhibition of NOS activity specifically affected poly(I-C)-induced adhesion in two types of human vascular EC. Three other endothelial NOS inhibitors, N^G-nitro-L-arginine methyl ester (L-NAME; data not shown), L-NMMA, and L-NIO, also inhibited poly(I-C)-induced U-937 cell adhesion to human umbilical vein EC (40-70%) without altering basal adhesion (Fig. 3).

View larger version (14K): [in this window] [in a new window]   Fig. 2.   Effect of L-NMA on polyinosinic-polycytidylic acid [poly(I-C)]- and interleukin-1 (IL-1)-induced U-937 cell adhesion to EC. EC were incubated with either MCDB 107 medium + 5% FBS alone or with 20-min pretreatment of L-NMA (100-300 µM) in absence () or presence () of poly(I-C) (100 µg/ml; A) or absence () or presence () of IL-1 (10 ng/ml; B) for 6 h. U-937 cell adhesion was then measured as described in MATERIALS AND METHODS. Data are representative of 4 experiments and are reported as means ± SE (n = 3). * P  0.05 vs. no L-NMA.

View larger version (16K): [in this window] [in a new window]   Fig. 3.   Effect of nitric oxide synthase (NOS) inhibitors on poly(I-C)-induced U-937 cell adhesion to primary HUVEC. EC were incubated with either MCDB 107 medium + 5% FBS alone or with poly(I-C) (100 µg/ml) in presence or absence of 20-min pretreatment of NOS inhibitors NG-monomethyl-L-arginine (L-NMMA; 500 µM), L-NMA (500 µM), or L-N5-(1-iminoethyl)ornithine (L-NIO; 250 µM) for 6 h. U-937 cell adhesion was then measured as described in MATERIALS AND METHODS. Data are representative of 3 experiments and are reported as means ± SE (n = 3).

We investigated whether L-NMA inhibited poly(I-C)-induced adhesion by reducing the surface expression of leukocyte adhesion molecules. We found that L-NMA reduced poly(I-C)-induced E-selectin surface expression by 25% and VCAM-1 expression by 44% (Fig. 4). Our previous observations (14a) have indicated that a critical mass of adhesion proteins on the cell surface is necessary to mediate monocytic cell adhesion; hence, the collective partial reduction in adhesion protein levels on the cell surface may account for the observed 70% inhibition of adhesion seen in Fig. 2. L-NMA did not significantly affect IL-1-induced E-selectin and VCAM-1 surface expression on the same cells (data not shown). The binding of an irrelevant antibody, mouse IgG1, (MOPC-21), indicated that the nonspecific binding in these experiments was minimal (Fig. 4).

View larger version (15K): [in this window] [in a new window]   Fig. 4.   Effect of L-NMA on poly(I-C)-induced surface expression of leukocyte adhesion molecules. EC were treated with poly(I-C) (100 µg/ml) in absence (open bars) or presence (solid bars) of L-NMA (500 µM) and assayed for cell surface expression of E-selectin and vascular cell adhesion molecule 1 (VCAM-1) measured by monoclonal antibody binding and subsequent 125I-streptavidin binding to the secondary biotinylated antibody (MOPC) as described in MATERIALS AND METHODS. Data represent 1 of 3 similar experiments and are reported as means ± SE (n = 3).

L-NMA significantly inhibited poly(I-C)- but not IL-1-induced mRNA levels of adhesion molecules. L-NMA (500 µM) inhibited poly(I-C)-induced E-selectin and VCAM-1 mRNA levels in primary cultured human umbilical vein EC by 62 and 74%, respectively (Fig. 5). The data were quantified using phosphorimaging analysis and normalized to GAPDH mRNA expression (Fig. 5C). L-NMA augmented IL-1-induced levels of E-selectin, whereas it inhibited VCAM-1 mRNA levels slightly (Fig. 5, A and B). This marginal decrease in IL-1-induced VCAM-1 mRNA was not significant when normalized for GAPDH levels.

View larger version (46K): [in this window] [in a new window]   View larger version (44K): [in this window] [in a new window]   Fig. 5.   Effect of NOS inhibition on steady-state levels of leukocyte adhesion molecule mRNA in EC. EC were incubated with MCDB 107 + 5% FBS alone or with either poly(I-C) (100 µg/ml) or IL-1 (10 ng/ml) in absence or presence of 30-min pretreatment with L-NMA (500 µM) for 2 h at 37°C. Total cellular RNA was then isolated, electrophoretically separated, and probed with [32P]dCTP-labeled cDNA for E-selectin (A) or VCAM-1 (B) as described in MATERIALS AND METHODS. C: relative intensity of E-selectin (solid bars) and VCAM-1 (hatched bars) mRNA was quantified via phosphorimaging analysis and normalized for glyceraldehyde-3-phosphate dehydrogenase mRNA levels in the same experiment. Data represent 1 of 3 similar experiments.

We tested whether the L-NMA effect on E-selectin expression was at the transcriptional level by transiently transfecting primary cultured human umbilical vein EC with a plasmid containing the luciferase reporter gene driven by an E-selectin promoter sequence (540 base pairs 5' to the transcription initiation start site). Figure 6 shows that in the presence of L-NMA, poly(I-C)-induced E-selectin promoter-driven luciferase expression was reduced to basal levels. These results were confirmed using bovine aortic EC in which the data were normalized for transfection efficiency as measured by -galactosidase activity (data not shown). These data indicated that the NOS inhibitor inhibited E-selectin transcription in response to poly(I-C). To determine whether L-NMA reduced E-selectin transcription via the availability within the nucleus of the necessary transcription factor NF-B, we performed EMSA. We incubated nuclear extracts [prepared from primary EC induced with poly(I-C) in the presence or absence of L-NMA] with radiolabeled oligonucleotides corresponding to the consensus NF-B binding domain of the E-selectin promoter. We determined the specificity of the band shifts using excess specific and nonspecific unlabeled oligonucleotides in competition studies (data not shown). We found that L-NMA reduced the amount of activated NF-B in the EC nucleus in response to poly(I-C) by ~70% (Fig. 7). Collectively, these data indicate that L-NMA attenuated poly(I-C)-induced adhesion protein expression at the transcriptional level by inhibiting poly(I-C)-induced NF-B activation.

View larger version (16K): [in this window] [in a new window]   Fig. 6.   Effect of NOS inhibition on E-selectin promoter activity in primary cultured HUVEC. EC were transiently transfected with plasmids containing an E-selectin promoter-driven luciferase reporter gene as described in MATERIALS AND METHODS. Cell extracts were collected after 6 h of induction with poly(I-C) (100 µg/ml) in presence or absence of L-NMA (500 µM). Luciferase activity was quantified and corrected for protein content of each sample as described in MATERIALS AND METHODS. Data represent 1 of 2 similar experiments and are reported as means ± SE (n = 3).

View larger version (56K): [in this window] [in a new window]   View larger version (15K): [in this window] [in a new window]   Fig. 7.   Effect of L-NMA on poly(I-C)-induced nuclear factor-B (NF-B) activation. A: EC were treated with poly(I-C) (100 µg/ml) in presence and absence of L-NMA (500 µM) for 2 h. Nuclear extracts (5 µg protein/sample) were then prepared as described in MATERIALS AND METHODS and incubated with radiolabeled oligonucleotides corresponding to consensus NF-B binding domain of E-selectin promoter. Samples were subjected to electrophoresis through a 4% polyacrylamide gel, autoradiography, and phosphorimaging analysis ( B). Data represent 1 of 3 similar experiments.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

The chemical properties of NO make it an attractive signaling molecule. It has an unpaired electron that allows it to react readily with molecular oxygen, heme, and iron-sulfur- and thiol-containing proteins (reviewed in Ref. 24). NO is membrane diffusible and has a long biological half-life and diffusion distance, considering its reactivity. The reactions that NO can undergo depend on the rate of NO synthesis and local environmental factors (32). Once NO reacts with molecular oxygen, peroxynitrite (ONOO) is formed. This NO derivative is a strong oxidant itself, can readily catalyze membrane lipid peroxidation, and has been shown to be involved in the oxidant stress in porcine myocardial ischemic-reperfusion injury (reviewed in Ref. 16).

The role of NO in heart disease is controversial. The fact that the activity of NOS has been shown to be increased during atherosclerosis (32) suggests that NO or NO derivatives (5) are associated with this disease process. However, the role of NO in leukocyte adhesion to vascular EC is unclear. Although most reports (4, 10, 21, 22) describe a protective effect of NO against leukocyte adhesion, our studies presented here and those of others (39) indicate that, under certain circumstances and depending on the specific agonist under study, NO may be involved in augmenting adhesion to EC.

In this study we have tested the hypothesis that NO may play a signaling role in dsRNA-induced EC activation. We chose to employ primary versus passaged EC due to the reports (3, 33) that endothelial NOS activity and subsequent NO levels were substantially greater in primary versus passaged EC. Our finding that NO augmented poly(I-C)-induced adhesion to EC is consistent with that of Vidal et al. (39) in a different system, tumor cell adhesion to EC.

Recent reports (10, 21, 31) have shown that NO inhibits cytokine-induced adhesion molecule expression through an NF-B-sensitive pathway in confluent monolayers of passaged EC. Although our studies focused on the role of NO in the signaling pathway of poly(I-C), we have also observed, consistent with the results of others, that NOS inhibition by L-NMA slightly augmented IL-1-induced monocytic cell adhesion to passaged EC. However, we report here that, in primary EC, L-NMA inhibited poly(I-C)-induced EC activation with little to no effect on the IL-1 response. The discrepancy with our results may be attributed primarily to the fact that others have focused on cytokine induction of EC, whereas we studied the effects of NO in the signaling cascade induced by dsRNA. Also, we used primary cultured EC, which differ in basal levels of NO from passaged EC. Because the role that NO plays is dependent on the concentration and the rate of its synthesis, it may be that in cells with higher basal levels of NO and NOS activity, the endogenous signaling events differ.

Elevated NO levels have been shown to occur concomitantly with induced endothelial-leukocyte adhesion protein expression in several cases, including malaria infection (35, 38), septic shock (28), histamine treatment (26, 33), and lysophosphatidylcholine induction (23, 41). NO levels and adhesion protein expression may be linked in the case of vascular injury as well. The role of leukocyte adhesion proteins in EC during ischemia and reperfusion injury is well defined; however, the role of NO during this process of injury remains unclear. Although there are many reports describing a protective role for NO in the prevention of ischemia and reperfusion injury, there are also a large number of reports contending that NO promotes ischemia and reperfusion injury (reviewed in Ref. 18). The fact that an increase in leukocyte adhesion proteins occurs during some injury processes in which there is a concurrent increase in NO levels again provides support for the hypothesis that NO may be involved in the expression of these genes.

Because poly(I-C) (acting as an interferon inducer), NO, and LPS have been shown (9, 17) to regulate xanthine oxidase activity in animal models, we considered the possible role for this enzyme in our system. Xanthine oxidase is present in vascular EC and is thought to be involved in free radical generation during reperfusion injury; however, its activity is not detectable in human EC under normoxic conditions (30). We do not believe that xanthine oxidase is playing a significant role in our observation for several reasons. First, our experiments were performed under normoxic conditions using short incubation periods with poly(I-C). These conditions are inconsistent with the known induction of xanthine oxidase activity. Second, the expression of E-selectin and VCAM-1 is not regulated by type I interferons; hence, poly(I-C) acts directly to induce the expression of these genes. Finally, our observations are specific for poly(I-C) in that the expression of these proteins in response to LPS was not affected. Because LPS activates xanthine oxidase in a manner similar to poly(I-C) (17), both signaling pathways would be expected to be affected if this enzyme played a significant role in our observations.

In conclusion, our results suggest the possibility that NO is a response mediator for dsRNA in EC. Hence, infecting viruses may mediate their signals through the endogenous NO available within EC. This may prove to be a target for inhibition when considering therapy for organ transplant recipients who have viral infections during immunosuppression, especially because increases in NOS expression have been identified in models of transplant arteriosclerosis (1). Patients with viral infections experience a higher rate of posttransplant atherosclerosis (36); therefore, blocking a response mediator of viral signaling may be beneficial. In addition, the augmentation by NO of poly(I-C)-induced expression of leukocyte adhesion proteins in EC and subsequent recruitment of immune cells may prove to be an additional aspect of the anti-viral effects of NO already identified (20).