Inhibition of Inducible Nitric Oxide Synthase Expression by Interferons α and β in Bovine Retinal Pigmented Epithelial Cells*

Bovine retinal pigmented epithelial (RPE) cells express an inducible nitric oxide synthase (NOS-2) after activation with interferon (IFN)-γ and lipopolysaccharide (LPS). Experiments were performed to investigate the effects of IFN-α and IFN-β on NOS-2 activity. These types of interferons did not aid LPS in the production of nitrite, but markedly inhibited in a concentration-dependent manner the nitrite release due to LPS/IFN-γ. Analysis by Western and Northern blots showed that RPE cells co-stimulated with IFN-α or IFN-β with LPS/IFN-γ accumulated lower levels of NOS-2 protein and mRNA than in the presence of LPS/IFN-γ alone. The presence of IFN-α or IFN-β did not accelerate mRNA degradation, implying that these interferons did not affect NOS-2 mRNA stability, but more probably NOS-2 gene expression. Furthermore, IFN-γ binding studies demonstrated that the inhibitory effect of IFN-α and IFN-β is not caused by a blocking of IFN-γ receptors. Analysis of NF-κB activation by electrophoretic mobility shift assay demonstrated that LPS/IFN-γ-induced NF-κB binding was not changed by the presence of IFN-α. However, similar experiments revealed that the activation of interferon regulatory factor-1 (IRF-1) by LPS/IFN-γ was decreased by IFN-α. This phenomenon could be due to the decline of IRF-1 mRNA and the up-regulation of IRF-2 mRNA, an IRF-1 repressor, by IFN-α. These results suggest that the inhibitory effect of IFN-α and -β on NOS-2 induction could be partially explained by their effect on the induction of the IRFs, which were involved in NOS-2 gene transcription.

Nitric oxide (NO) 1 is an important signaling molecule that mediates a variety of essential physiological processes including neurotransmission, vasodilatation, and host cell defense (1,2). NO is synthesized from L-arginine by NO synthase (NOS), a family of enzymes with distinct functional, biochemical, and regulatory properties (1,2). The constitutive NOS isoforms, originally described in endothelial cells and in neurons, produce small amounts of NO in response to an intracellular calcium increase (1,3). Cytokine-inducible NO synthase, whose expression requires protein synthesis, has been demonstrated in a wide variety of mammalian cells and tissues (2,4), and was cloned in different cell types such as murine macrophages (5)(6)(7), human hepatocytes (8), and human chondrocytes (9). The role of sustained NO production by the inducible NOS (NOS-2) is well described in murine macrophages, where NO is responsible for their cytostatic and cytolytic activities toward invading organisms (2,10). On the other hand, NO generated by NOS-2 is also involved in some pathophysiological states, generally related to local and systemic inflammation (2,4).
In the retina, we have demonstrated that Mü ller glial cells can express the macrophage-type NOS, NOS-2, after endotoxin and cytokine stimulation (11). Retinal pigmented epithelial (RPE) cells from bovine (12), human (13), and murine (14,15) species also contain an inducible isoform of NOS. In rat and bovine RPE cells, NOS-2 mRNA activity is induced only by cooperation between IFN-␥ and LPS and can be potentiated in the rat by the addition of TNF-␣ (15). RPE cells resemble other cell types with respect to the combination of mediators capable of inducing NO generation (4). Several lines of evidence indicate that in murine macrophages all three species of interferon, IFN-␣, IFN-␤, and IFN-␥, are able to aid LPS in the release of nitrite (16,17), a stable end product of NO, and for the expression of NOS-2 mRNA (18). However, only IFN-␥ induced NO release alone, and the combination of LPS with IFN-␥ was the more potent for NOS-2 induction (16 -18).
The purpose of the present study was to investigate whether IFN-␣ and IFN-␤, which have been previously described to be able to interact with RPE cells (19,20), could regulate the production of NO in bovine RPE cells. Our results indicate that, instead of enhancing LPS-induced NOS-2, IFN-␣ and IFN-␤ markedly inhibit NO production stimulated by the LPS/IFN-␥ combination in bovine RPE cells. We have found that the inhibitory effect of IFNs on NOS-2 activity could be correlated with a decrease of NOS-2 protein and mRNA accumulation. This phenomenon is not due to a competition of IFN-␣ and IFN-␤ for IFN-␥ receptors. In addition, there was no inhibition of the activation of the transcription factor NF-B. Analysis of the effects of IFN-␣ on the induction of the two transcription factors, IRF-1 and IRF-2, revealed that IFN-␣ inhibits the LPS/IFN-␥-induced IRF-1 mRNA accumulation and increased IRF-2 mRNA accumulation in RPE cells.

MATERIALS AND METHODS
Cell Cultures-Bovine RPE cells were prepared, subcultured, and characterized as previously reported (21). Cells of passages 1-5 were used for the experiments.
Assay of NOS Activity in RPE Cell Extracts-Confluent cells were incubated for 72 h with different combinations of LPS, IFN-␥, IFN-␣, and IFN-␤. Cells were harvested by trypsinization, washed once in phosphate-buffered saline (PBS), and resuspended in 50 mM Tris/HCl, pH 7.8, 10 mM dithiothreitol. After sonication, the homogenate was centrifuged for 20 min at 4°C at 100,000 ϫ g. The supernatant was passed through a Dowex AG 50W-X8 column to remove endogenous arginine, and NOS activity was evaluated by the conversion of L-[ 3 H]arginine to L-[ 3 H]citrulline, as described previously (23,24), after the determination of protein content (Bio-Rad assay kit).
Western Blot Analysis-After treatment with LPS and IFN-␥ with or without IFN-␣ and IFN-␤ for different periods, cells were washed with PBS and scraped into lysis buffer containing protease inhibitors. Samples were centrifuged, and after one freeze/thaw cycle, 100 g of supernatant proteins were subjected to SDS-polyacrylamide gel electrophoresis. Proteins were then transferred to an Immobilon membrane (Millipore, Saint Quentin en Yvelines, France) by electroblotting. Western blot analysis using a polyclonal antibody specific for liver-inducible NOS (22) was performed as described previously (25). The intensity of the bands was quantified using densitometric measurements with One Descan densitometric software (Scanalytics, Billerica, MA).
RNA Isolation and Northern Blot Analysis-Total RNA was extracted from cultured cells treated by cells lysis in guanidinium isothiocyanate followed by phenol acid extraction. The RNA was denatured, electrophoresed (25 g/lane) in 1% formaldehyde-agarose gel, and then transferred to a nylon membrane. Blots were hybridized with a randomly primed 32 P-labeled NOS-2 cDNA probe (SmaI and EcoRI digestion of pGEM plasmid containing the cloned murine macrophage NOS cDNA) as described previously (11,25). The hybridized blots were then washed and autoradiographed. To correct for differences in RNA loading, membranes were stripped and rehybridized with a full-length glyceraldehyde-3-phosphate dehydrogenase (GAPDH) cDNA probe as a control. A high resolution camera coupled to an image processor (Ultra-Lum Inc., Carson, CA) driven by One Descan densitometric software (Scanalytics) was used to perform densitometric measurements. Results are expressed in arbitrary units as the ratio of NOS-2/GAPDH expression. For mRNA stability analysis, RPE cells were stimulated with LPS (1 g/ml) and IFN-␥ (100 units/ml) in the presence or absence of either IFN-␣ or IFN-␤ for 18 h. Actinomycin D (5 g/ml) was added, and total RNA was prepared at the time indicated and further processed for Northern hybridization as described above.
125 I-IFN-␥ Binding Studies-IFN-␥ was iodinated according to the chloramine T method as described previously (26). The specific activity of the labeled IFN-␥ obtained was usually 25,000 cpm/ng. RPE cells were grown to confluence in 24-well dishes in the presence of Dulbecco's modified Eagle's medium. Cultures were transferred at 4°C, washed with PBS, and then incubated 30 min at 4°C in binding buffer (serumfree Eagle's modified medium, 25 mM Hepes, pH 7.4, and 0.1% bovine serum albumin). Cells were then washed with ice-cold PBS and incubated at 4°C with 125 I-IFN-␥ at 20 units/ml in the absence or in the presence of increasing amounts of IFN-␥, IFN-␣, or IFN-␤. After 3 h, the cells were washed with ice-cold PBS and then solubilized for 15 min in 0.1 N NaOH (26). Cell-bound radioactivity was analyzed using a ␥ scintillation counter.
RT-PCR Analysis of IRF-1 and IRF-2 mRNA-One g of RNA was reverse-transcribed for 90 min at 42°C with 200 units of superscript Moloney murine leukemia virus reverse transcriptase (Life Technologies SARL, Eragny, France), using random hexamers, and 2 l of cDNA were added to each PCR reaction, as described previously (27). Amplification was performed as follows: 94°C for 2 min; 24 cycles for IRF-1 and GAPDH and 30 cycles for IRF-2 (number of cycles that were below saturating conditions) at 94°C for 30 s, 55°C for 30 s, 72°C for 45 s; and then 72°C for 2 min. The amplified fragments were separated on a 1.2% agarose gel and transferred onto Nylon membrane (Amersham, Les Ulis, France). Specificity of the amplification process was verified by hybridization of blots with 32 P-labeled specific internal oligonucleotide probe, washed three times in 1 ϫ SSC, 0.1% SDS at 50°C, and exposed to x-ray films. The intensity of the bands was quantified using densitometric measurements as describe above and expressed in arbitrary units as the ratio of IRF-1/GAPDH or IRF-2/GAPDH expression.
Statistical Analysis-Results were expressed as mean Ϯ S.E. They were analyzed statistically by Mann Whitney U test. p values less than 0.05 were considered as significant.
We therefore determined how long RPE cells had to be exposed to IFN-␣ and IFN-␤ to inhibit LPS/IFN-␥-stimulated nitrite formation. The inhibitory effect of hIFN-␣ and hIFN-␤ could also be observed (Fig. 1B) when cells were pretreated for 12 h with these IFNs before the addition of the inducers (LPS and IFN-␥), but not as well as that observed when IFN-␣ or IFN-␤ were continuously present with LPS/IFN-␥. The addition of hIFN-␣ and hIFN-␤ 24 h after the treatment of RPE cells with LPS and IFN-␥ decreased the inhibitory effect of these interferons, and when hIFN-␣ or hIFN-␤ were added 36 h after LPS/IFN-␥, no inhibitory effect was observed (Fig. 1B). These results suggested that the inhibitory response of hIFN-␣ and hIFN-␤ on the release of NO caused by LPS and IFN-␥ required the presence of the inhibitors (hIFN-␣ and hIFN-␤) during the early period of the induction.
Absence of NOS Activity in Extracts from Bovine RPE Cells Stimulated with IFN-␣ and IFN-␤-The effect of IFNs on NOS activity in RPE cells was further characterized using cytosolic extracts isolated from cells cultured for 72 h with LPS and IFN-␥ with or without hIFN-␣ and hIFN-␤. As  (Table I), which was inhibited by addition of L-NMMA in the cytosol extract. More interestingly, the results in Table I demonstrate that enzyme activity was largely reduced in cells previously co-incubated with hIFN-␣ or hIFN-␤, at concentrations that largely prevented nitrite and citrulline release in the culture medium (see above).
Decrease of NOS-2 Protein by IFN-␣ and IFN-␤-To determine whether the inhibitory effect of IFN-␣ and IFN-␤ on LPS/IFN-␥-induced NOS activity is due to the direct inhibition of NOS expression or to an indirect effect involving the synthesis of a coenzyme necessary for enzymatic activity, the expres-sion of inducible RPE NOS protein was investigated by Western blot analysis after 72 h of treatment with LPS and IFN-␥ with or without hIFN-␣ and hIFN-␤ (Fig. 2). In LPS/IFN-␥treated cells, but not in untreated cells, a band at 130 kDa, corresponding to the size described for NOS-2 protein (2, 22, 25) was observed. Furthermore, densitometric analysis revealed that a simultaneous treatment with hIFN-␣ or with hIFN-␤ greatly decreased the 130-kDa protein signal by 92 and 71.5%, respectively, compared with LPS/IFN-␥ (Fig. 2), demonstrating that IFN-␣ and -␤ markedly prevent NOS protein expression.
Decrease of NOS-2 mRNA Accumulation by IFN-␣ and IFN-␤-The expression of NOS-2 messenger was investigated by Northern blot to determine if the decrease of inducible RPE NOS protein could result from a decrease in NOS-2 mRNA accumulation. Total RNA was extracted from RPE cells after 24 h of treatment, corresponding to the maximal expression of NOS-2 mRNA (25). Only one detectable mRNA signal at 4.4 kilobase pairs was detected, while in the unstimulated RPE cells, NOS-2 mRNA was not detectable (Fig. 3), as we recently reported (25). Inclusion of hIFN-␣ or -␤ in the culture medium inhibited to a large extent the expression of mRNA of inducible NOS in RPE cells stimulated with LPS/IFN-␥ in concordance with NOS activity evaluated by nitrite release (Fig. 3). A similar inhibition was obtained with bIFN-␣ (data not shown). Hybridization with a probe for GAPDH revealed a single message at about 1.4 kilobase pairs, with equal intensity in all lanes, demonstrating a similar RNA loading throughout.
The effect of IFNs on NOS-2 mRNA stability was assessed by experiments using actinomycin D. Total RNA was isolated at various times after the addition of actinomycin D and examined by Northern blot successively for NOS-2 and GAPDH, to correct the loading differences. Fig. 4 shows that the addition of hIFN-␣ did not decrease the stability of NOS-2 mRNA induced by LPS and IFN-␥. In experiments not reported here, similar results were obtained with hIFN-␤ and bIFN-␣.    hIFN-␤, and bIFN-␤ were unable to compete with IFN-␥ for binding in their receptors.

activity in cytosolic extracts of RPE-stimulated cells by IFN-␣ and IFN-␤ Confluent cells were incubated for 72 h in fresh Dulbecco's modified Eagle's medium DMEM alone (control) or in the presence of LPS (1 g/ml) plus IFN-␥ (100 units/ml) with or without hIFN-␣ (100 units/ml) or hIFN-␤ (10 3 units/ml). After this time cells were harvested by trypsinization, and formation of L-[ 3 H]citrulline was determined in
Absence of Effect of hIFN-␣ on NF-B Activation-Because the NOS-2 gene promoter region has been shown to contain NF-B consensus sequences (28) and because NF-B was activated in RPE cells after LPS/IFN-␥ treatment, 2 we examined the effect of IFN-␣ on NF-B activation by electrophoretic mobility gel-shift assays. Fig. 6 shows a rapid activation of NF-B after LPS and LPS/IFN-␥ stimulation. The addition of excess unlabeled consensus oligonucleotide completely prevented the complex formation, demonstrating the specificity of the DNA/protein interaction. Coincubation with hIFN-␣ did not change the LPS/IFN-␥-induced NF-B activation (Fig. 6,  lane 3).
Effect of hIFN-␣ on IRF-1 Activation-We tested if stimulation of RPE cells with IFN-␣ modify the activation of IRF-1, a transcriptional factor involved in NOS-2 induction (29, 30). EMSA analysis (Fig. 7) revealed the presence of an induced 2 V. Faure and O. Goureau, manuscript in preparation.

FIG. 2. Effect of IFN-␣ and IFN-␤ on NOS-2 protein.
Cells were incubated with or without LPS (1 g/ml) and IFN-␥ (100 units/ml) in combination with either hIFN-␣ (100 units/ml) or hIFN-␤ (10 3 units/ml) for 72 h. Cell lysates (100 g) were subjected to SDS-PAGE, followed by immunoblotting with anti NOS-2 antibody as described under "Materials and Methods." Molecular size markers (in kDa) are shown at the right. The experiment shown represents one of three independent blots that gave identical results.

FIG. 3. Down-regulation of NOS-2 mRNA by IFN-␣ and IFN-␤.
RPE cells were grown to confluence, then exposed to LPS (1 g/ml) and IFN-␥ (100 units/ml) in the absence or in the presence of either 100 units/ml hIFN-␣ or 10 3 units/ml hIFN-␤. After 24 h the nitrite accumulated in the culture medium was measured by the Griess reaction after which total RNA was isolated, and the levels of RPE NOS-2 and GAPDH mRNAs were assessed by Northern analysis as described under "Materials and Methods." Densitometric analysis of the NOS-2 Northern blot corrected for GAPDH expression is shown in histogram. The experiment shown represents one of three independent trials that gave similar results.  5 and 7, compared,  respectively, to lanes 3 and 6), indicating that this interferon induces a decrease of IRF-1 binding to its specific DNA target sequence.
Differential Regulation of IRF-1 and IRF-2 mRNA Accumulation by hIFN-␣-Total RNA was extracted from RPE cells after 3 h of treatment, corresponding to the maximal accumulation of IRF-1 mRNA. 2 RT-PCR analysis showed that IRF-1 mRNA was induced in RPE cells by IFN-␥ alone and by LPS/ IFN-␥, while hIFN-␣ alone had no significant effect (Fig. 8). IRF-1 mRNA accumulation induced by IFN-␥ or by the combination LPS/IFN-␥ was decreased by hIFN-␣ treatment at concentrations that completely abrogated NOS-2 expression (Fig. 8).
Since IRF-2 has been demonstrated to function as a repressor of interferon-stimulated genes (31), we have investigated the effect of hIFN-␣ on its mRNA level. RT-PCR analysis, depicted in Fig. 9, demonstrated that IRF-2 mRNA expression was constitutive in RPE cells and that hIFN-␣ treatment increased its accumulation, while IFN-␥ largely decreased the level of IRF-2 mRNA. The combination LPS/IFN-␥, responsible for NOS-2 induction, also decreased the IRF-2 mRNA level. Furthermore, hIFN-␣, which inhibited NOS-2 induction, counteracted the decrease of IRF-2 mRNA induced by IFN-␥ or by LPS/IFN-␥ (Fig. 9).

DISCUSSION
In this work, we demonstrate that type I IFN (IFN-␣ and IFN-␤) can suppress the LPS/IFN-␥-dependent expression of NOS activity in bovine RPE cells as assessed by nitrite accumulation in the culture medium. Measurements of L-citrulline synthesis by cytoplasmic extracts further support the conclu- sion that levels of NOS activity are largely decreased. This loss of activity reflected the absence of NOS protein in the IFN-␣and IFN-␤-treated RPE cells, as revealed by Western blot. Finally, the marked inhibition by IFN-␣ and IFN-␤ of the LPS/IFN-␥-induced increase in NOS mRNA suggests that IFN-␣ and IFN-␤ affect the regulation of the inducible form of NOS at the mRNA level. The detected levels correlated with the levels of NOS protein activity. Experiments with actinomycin D revealed that hIFN-␣ and hIFN-␤ do not decrease the half-life of NOS-2 mRNA in RPE cells stimulated with LPS/ IFN-␥, demonstrating that the decreased expression of NOS-2 observed after hIFN-␣ or hIFN-␤ treatment is not attributable to decreased mRNA stability. Therefore, inhibition of the transcription rate of the NOS-2 gene seems to be the most likely mechanism involved. The negative regulation of NO synthase induction was previously reported (2) for different cytokines and growth factors, such as interleukin-4 and -10, fibroblast growth factor, and transforming growth factor ␤ (2,24,25,32).
Our results contrast with those of previous studies (16 -18, 33, 34), in which IFN-␣ and IFN-␤ were generally considered to be activators of NOS induction. Indeed, exogenous addition of IFN-␣ or IFN-␤ on murine peritoneal macrophages could induce nitrite release (16,17), and blockade of endogenous IFN-␣ or IFN-␤ production could reduce LPS-or IFN-␥-induced NOS-2 mRNA accumulation in macrophages (33,34). In preliminary experiments not reported herein, we have tested the ability of IFN-␣ to regulate the nitrite release in peritoneal macrophages from thioglycollate-treated mice. We have found that IFN-␣ is unable to potentiate nitrite release due to LPS or IFN-␥ and had no effect on LPS/IFN-␥-induced nitrite release, suggesting that the inhibitory effect of IFN-␣ on NOS-2 induction in RPE cells is specific from the species and/or from the cell type. However, this phenomenon is not exclusive to RPE cells since recent reports demonstrated the ability of IFN-␣ and IFN-␤ to largely reduce the IFN-␥-mediated nitrite release in rat peritoneal macrophages (35) and in human thyrocytes (36). In rat macrophages this decrease of nitrite production was due to a down-regulation of NOS-2 mRNA (37), as in bovine RPE cells.
Concerning the molecular mechanism which operates for the suppressive action of NOS-2 by IFN-␣ and -␤, we have postulated that they may interrupt some steps in the signaling pathway by which LPS/IFN-␥ induces NOS-2 gene expression. First, on the basis of our binding experiments the inhibition of NOS induction by type I IFNs was not due to a competition between IFN-␣ or IFN-␤ with IFN-␥ at the receptor level. We have also looked for the activation of two transcription factors, NF-B and IRF-1, involved in NOS-2 induction (28 -30). We demonstrated, by EMSA, that IFN-␣ did not inhibit LPS/IFN-␥-induced NF-B activation, indicating that the inhibitory effect of this compound on LPS/IFN-␥-induced NOS-2 appeared to be independent of NF-B activation. The down-regulation of NF-B activity by other compounds could be an inhibitory regulatory pathway for NOS induction in RPE cells, as it has been recently reported for the inhibition of NOS-2 induction by the glucocorticoids in human alveolar epithelium A549/8 cells (38).
NF-B is not the only regulatory factor of NOS-2 gene expression, but could function in concert with several other transcription factors, such as IRF-1 (29,30). We demonstrated in this study that treatment of RPE cells with IFN-␥ alone or with IFN-␥ plus LPS resulted in an increase of IRF-1 mRNA and in an appearance of a prominent IRF-1-specific DNA-protein complex, which were clearly inhibited in the presence of IFN-␣. Interestingly, a similar result has been reported by Politis et al. (39) with another member of interferon-induced transcription factor, the interferon consensus sequence binding protein. They demonstrated that IFN-␣ could suppress the IFN-␥-induced interferon consensus sequence binding protein mRNA accumulation in peritoneal macrophages. Besides these interferoninduced transcription factors, there is another transcription factor, termed IRF-2, that is structurally similar to IRF-1, but that represses the effect of IRF-1 (31,40). We demonstrated that IRF-2 mRNA is differentially regulated by the IFNs, since IFN-␥ largely decreased its accumulation, while IFN-␣ significantly increased it. Furthermore, IFN-␣ is able to prevent the IFN-␥-and the LPS/IFN-␥-induced IRF-2 mRNA decrease. These results with IRFs demonstrated that inducers of NOS-2 (LPS/IFN-␥) increased IRF-1 and decreased IRF-2 mRNA, while the inhibitor of NOS-2 induction (IFN-␣) decreased IRF-1 and increased IRF-2 mRNA. A very similar regulation of IRF-1 has been recently reported in murine macrophages by NOS inducers and inhibitors (41,42). It would be consistent to propose that IFN-␣, by increasing IRF-2 mRNA and decreasing IRF-1 mRNA accumulation, could favor the neosynthesis of the repressor IRF-2, which is more stable than IRF-1 (43). As previously suggested, IRF-1 and IRF-2 could compete for the specific ISRE sequence (31,40). Then, the accumulation of IRF-2 and diminution of IRF-1 could favor the interaction of IRF-2 with the NOS-2 gene promoter in place of IRF-1 and could suppress the activation of NOS-2 gene by IRF-1, after IFN-␣ treatment. It has been recently reported that closed tyrosine kinases and transcription factors in the JAK/STAT pathway could be involved in the transduction of the IFNs signals, leading to the transcription of early response genes (reviewed in Darnell et al. (44)). There are at least two distinct sets of genes that can be stimulated in response to IFNs: those containing GAS elements, which bind STAT-1 homodimeres, FIG. 9. Differential regulation of IRF-2 mRNA accumulation by IFN-␥ and hIFN-␣ in RPE cells. Confluent RPE cells were exposed to different combination of LPS (1 g/ml), IFN-␥ (100 units/ml), and hIFN-␣ (100 units/ml) as noted. After 3 h, total RNA was isolated, and the levels of IRF-2 and GAPDH mRNAs were assessed by RT-PCR analysis as described under "Materials and Methods." Densitometric analysis of the IRF-2 band corrected for GAPDH expression is shown below. The experiment shown represents one of three independent trials that gave similar results. and those containing ISRE sequences, which bind STAT1 homodimeres or STAT1/STAT2 heterodimeres coupled to p48 (ISGF3␥), respectively, named STAT1-p48 complex and ISGF3 (45). IFN-␣ activates both STAT1 and STAT2 by phosphorylation on tyrosine residues, while IFN-␥ phosphorylated tyrosines only on STAT1, leading to a difference in complex assembly that results in a functional difference in gene expression (45). In this context Li et al. (46) recently reported that activation of IRF-1 transcription in response to IFN could be modulated by the level of p48. It is conceivable that IFN-␣ and -␤ might counteract the initial phase of the IFN-␥ signaling cascade in bovine RPE cells, possibly by a phosphorylationdephosphorylation process and/or by a competition between the different STAT complexes (44 -46).
Although NO might act as an antimicrobial and antiviral effector molecule in the retina (47), the sustained NO release by NOS-2 may cause retinal cell damage as we previously observed in the case of retinal light damage (48) or during endotoxin-induced uveitis (27). Type I IFNs are produced by various cell types in the early phase of viral or bacterial infection. Thus, the production of IFN-␣ and -␤ during viral infection could decrease the potential antiviral NO activity (49) and might weaken the antiviral defense in the retina. This could represent a mechanism by which the virus protects itself against a large release of NO. This hypothesis is in agreement with recent results in macrophages (50), in which the viral infection alters the induction of NOS-2, by increasing the synthesis of endogenous IFN-␣ and -␤. However, we cannot exclude that actions of IFN-␣ and -␤ on RPE cells might also prevent NO-mediated tissue damage induced by LPS and IFN-␥ in addition to an important role in the regulation of immune response in the retina.