2'-5' Oligoadenylate synthetase plays a critical role in interferon-gamma inhibition of respiratory syncytial virus infection of human epithelial cells.

Respiratory syncytial virus (RSV), associated with bronchiolitis and asthma, is resistant to the antiviral effects of type-I interferons (IFN), but not IFN-gamma. However, the antiviral mechanism of IFN-gamma action against RSV infection is unknown. The molecular mechanism of IFN-gamma-induced antiviral activity was examined in this study using human epithelial cell lines HEp-2 and A549. Exposure of these cells to 100-1000 units/ml of IFN-gamma, either before or after RSV infection, results in a significant decrease in RSV infection. After 1 h of exposure, IFN-gamma induces protein expression of IFN regulatory factor-1 (IRF-1) but not IRF-2, double-stranded RNA-activated protein kinase, and inducible nitric-oxide synthase in these cells. The mRNA for IRF-1, p40, and p69 isoforms of 2'-5' oligoadenylate synthetase (2-5 AS) are detectable, respectively, at 1 and 4 h of IFN-gamma exposure. Studies using cycloheximide and antisense oligonucleotides to IRF-1 indicate a direct role of IRF-1 in activating 2-5 AS. Cells transfected with 2-5 AS antisense oligonucleotides inhibit the antiviral effect of IFN-gamma. A stable cell line of HEp-2 overexpressing RNase L inhibitor, RLI-14, which exhibits an IFN-gamma-induced gene expression pattern similar to that of the parent cell line, shows a significant reduction in RNase L activity and IFN-gamma-mediated antiviral effect, compared with HEp-2 cells. These results provide direct evidence of the involvement of 2-5 AS in IFN-gamma-mediated antiviral activity in these cells.

The respiratory syncytial virus (RSV) 1 is the most important cause of lower respiratory tract infection in infants and young children worldwide and is a risk factor for the development of asthma. In the United States alone, RSV causes ϳ4 million cases of respiratory tract infection annually, which results in 95,000 hospitalizations and 4,500 deaths (1,2). An effective prophylaxis or treatment against RSV is not available. Intranasal administration of a plasmid expressing IFN-␥ cDNA or a recombinant RSV expressing IFN-␥ attenuates virus replication in mice without compromising immunogenicity (3,4). The potential and mechanism of IFN-␥ as an antiviral agent against RSV-induced lung disease remain to be elucidated.
IFN-␥, a type II interferon, is a pleiotropic cytokine that plays an important role in modulating nearly all phases of immune and inflammatory responses. IFNs bind to specific receptors on cells and activate a JAK-STAT (Janus kinasesignal transducer and activator of transcription) signaling cascade that culminates in the transcriptional induction of IFNstimulated genes (ISGs) (5). The Jak1 and Jak2 phosphorylate STAT-1 following the binding of IFN-␥ to its receptor (5)(6)(7). Once phosphorylated, STAT molecules dimerize and translocate to the nucleus and bind to ␥-activated sequence elements present in the regulatory regions of various ISGs. The antiviral mechanism of IFN-␥ may involve one or more of a number of ISG-encoded products, including interferon regulatory factor-1 (IRF-1) (8), double-stranded RNA-activated protein kinase (PKR) (9,10), the Mx family of proteins (11), a family of 2Ј-5Ј oligoadenylate synthetases (2-5 AS) (12,13), and RNase L (14).
The RNase L is constitutively expressed in most of the mammalian cells and is found in an inactive form being bound to RNase L inhibitor, RLI, a 68-kDa protein not regulated by IFN-␥ (15). The 2-5 AS produces a series of 5Ј-phosphorylated and 2Ј-and 5Ј-linked oligoadenylates (2-5A) from ATP when activated by double-stranded RNA (12,16). It has been suggested that the binding of 2-5A with RNase L results in the release of RLI, dimerization, and the activation of RNase L (17), and subsequently activated RNase L mediates the cleavage of single-stranded RNA. However, the mechanism of the induction and activation of each of these genes vary in different cells and for the type of viruses. The mechanism of the IFN-␥mediated antiviral activity remains to be elucidated for many clinically important viruses.
Previously, we have shown that in BALB/c mice, RSV infection produces IFN-␥, which coincides with the resolution of infection (18), and prophylactic intranasal IFN-␥ gene transfer decreases RSV replication and infection (3). Whereas alveolar pneumocytes are the primary target of RSV infection in mice (19), RSV targets epithelial cells in human airways and lungs (20). In this study, the mechanism of IFN-␥-mediated anti-RSV activity was investigated in the HEp-2 cells, which are normally used for RSV propagation, and in the human lung alveolar epithelial cell line A549. Our results indicate that the IFN-␥-mediated inhibition of RSV infection in human epithelial cells involves the 2-5 AS/RNase L pathway.
Immunoblot Analysis-Cells were harvested in a 2-pack volume of cell lysis buffer, and equal amounts of proteins were separated on 10% SDS-PAGE and transferred to nitrocellulose membrane. For the detection of iNOS, the lysate of the IFN-␥ and lipopolysaccharide-stimulated murine macrophage (RAW 264.7) was used as positive control. Immunoblotting was performed as described earlier (21) with monoclonal antibodies to IRF-1, IRF-2, PKR, cytokeratin-18 (Santa Cruz Biotechnology, Inc., Santa Cruz, CA), iNOS (Transduction Laboratories, Lexington, KY), and phospho-eIF-2␣ (Cell Signaling, Beverly, MA).
Nitrite Assay-Nitrite, a stable breakdown product of NO in physiological systems, was assayed using the Griess reaction from cell culture supernatant (23).
Antisense Treatment-Phosphorothioate antisense oligonucleotides (ODNs) were used in all experiments. Antisense sequences against p40 and p69 subunits of 2-5 AS and IRF-1 are as follows: p40 subunit, 5Ј-TTT CTG AGA TCC ATC ATT GA-3Ј; p69 subunit, 5Ј-TCC CCA TTT CCC ATT GC-3Ј; IRF-1, 5Ј-CGA GTG ATG GGC ATG TTG GC-3Ј. The control ODN sequences 5Ј-GTC TAT GAA TAC TTT CCT AG-3Ј, 5Ј-CAC CTC TAT CTC TCT CG-3Ј, and 5Ј-CGA GTG GTA GAC GTA TTG GC-3Ј are a scramble of the antisense sequence to p40, p69, and IRF-1, respectively, i.e. identical in base composition. For 2-5 AS blocking, HEp-2 cells were treated with 1000 units/ml of IFN-␥ 20 h pre-infection. Cells were transfected with equimolar concentrations of p40 and p69 2-5 AS antisense ODNs 3 h before RSV infection. IFN-␥ was added back to the cells following RSV infection. Following 72 h of RSV infection, cells were washed three times with cold phosphate-buffered saline, pH 7.4, and harvested, and the clear cell homogenate was used for plaque assay. For IRF-1 blocking, cells were transfected with IRF-1 antisense for 3 h and then treated with 1000 units/ml of IFN-␥. Cells were harvested 24 h after IFN-␥ addition, and total RNA was isolated.
2-5 AS Assay-2-5 AS assay was done following the method described earlier (25). Briefly, 20 l of reaction mixture containing 10 g of the cell homogenate, 20 mM Tris-HCl, pH 7.5, 20 mM magnesium acetate, 2.5 mM dithiothreitol, 5 mM ATP, 5 Ci of [␣-32 P] ATP, and 50 g/ml poly(I)-poly(C) was incubated for 3 h at 30°C. The reaction was stopped by boiling for 3 min and centrifuged and was incubated for 3 h at 37°C with 3 l of 1 unit/l calf intestine alkaline phosphatase to convert the unreacted [␣-32 P]ATP to inorganic phosphate. 2 l of the sample were then spotted on a polyethyleneimine-cellulose thin layer chromatography plate (Sigma) and resolved in 750 mM KH 2 PO 4 , pH 3.5. The 2-5A formed was then quantified using Advanced Quantifier software and expressed as arbitrary units.
Generation of a Stable Cell Line Overexpressing RNase L Inhibitor-Human RLI cDNA was amplified as KpnI-BamHI cassette and cloned in mammalian expression vector, pcDNA3.1 (Invitrogen) by standard procedure (24). HEp-2 cells were transfected with 5 g of pcDNA3.1-RLI using Lipofectin (Invitrogen). The empty pcDNA3.1 vector was used as a control. Stable transfectants were selected with G418 (Invitrogen). Individual clones were isolated and analyzed for the expression of RLI mRNA. The clone that expressed RLI at the highest level and had normal morphology and growth pattern was selected and named RLI-14.
Statistical Analysis-Experiments were repeated two to four times for each experiment as indicated. Statistical significance was analyzed using paired two-tailed student's t tests. Differences were considered statistically significant when the p value was less than 0.05.

IFN-␥ Attenuates RSV Infection in Human Epithelial
Cells-To examine the effect of IFN-␥ on RSV infection, HEp-2 cells were pre-incubated for 4 -20 h with different concentrations of IFN-␥ and subsequently infected with RSV. Respective concentrations of IFN-␥ were added back to the cells in complete medium after the removal of viral inoculum. Cells were harvested at 72 h post-infection, and viral titer was determined by plaque assay. A 97% inhibition of viral replication was observed in HEp-2 and A549 cells, at 1000 units/ml of IFN-␥ added from 20 h pre-infection ( Fig. 1, A and B). Cells treated with 1000 units/ml of IFN-␥ from 4 h pre-infection also showed significant reduction (p Ͻ 0.01) in RSV titer (50% reduction). A significant decrease (p Ͻ 0.01) in RSV titer (39% reduction) was observed in A549 cells, which were not treated with IFN-␥ before infection but were only treated at 1 h post-infection (Fig.  1C). To rule out the possibility that the reduction in RSV titers could be because of cytotoxicity of IFN-␥, an MTT cytotoxicity assay was performed. The results indicate that the cells were metabolically as viable as the untreated control cells when treated with the highest concentrations of IFN-␥ (1000 units/ ml; see Fig. 1D). Thus, IFN-␥ did not exhibit any cytotoxic or growth inhibitory effect on these cells. These results suggest that the treatment of cells with soluble IFN-␥ results in a significant decrease in RSV infection of epithelial cells.
IFN-␥ Induces IRF-1 Protein Expression-ISGs implicated in the antiviral activity of IFNs include IRFs, PKR, and iNOS. To identify the ISGs in these cells potentially involved in protection against RSV infection, proteins were analyzed from cells at various time points following IFN-␥ treatment (1000 units/ml). Western blot analysis was performed using specific antibodies to IRF-1, IRF-2, and PKR ( Fig. 2A). There was increased expression of IRF-1 but no change in the expression of IRF-2 following IFN-␥ addition. Expression of IRF-1 increased after 30 min of IFN-␥ addition. The expression of PKR decreased gradually over time ( Fig. 2A), and no change in the expression of phospho-eIF-2␣ was observed following IFN-␥ addition (data not shown). Cytokeratin-18 was used as an internal control, the expression of which did not change with the addition of IFN-␥. To examine whether IFN-␥-induced iNOS plays a role in antiviral action, iNOS expression was examined by Western blotting (Fig. 2B). The expression of iNOS protein could not be detected before and after IFN-␥ addition. Murine macrophage cell lysate containing iNOS was used as a positive control that did not bind to the cytokeratin-18 antibody used as internal control. To completely rule out the involvement of iNOS in the antiviral effect of IFN-␥, levels of soluble NO were examined in the culture supernatant of both HEp-2 and A549 cells before and after the addition of IFN-␥ at various time points. No detectable level of NO (lowest concentration of standard was 0.25 M) was observed in both cell lines at any time point. A similar expression pattern was observed for IRF-1, IRF-2, PKR, and iNOS in A549 cells (data not shown). These results indicate that IFN-␥ up-regulates IRF-1 in these cells, and neither PKR nor iNOS plays any role in the antiviral activity of IFN-␥ against RSV infection in human epithelial cell lines.
IFN-␥ Up-regulates IRF-1 and 2-5 AS mRNA-To investigate the role of IRF-1 in IFN-␥-induced antiviral activity, the mRNA expression of IRF-1 and 2-5 AS was examined. Of the four isoforms of 2-5 AS (p40, p46, p69, and p100) detected in human cells to date, the p40 and p46 isoforms are dependent on double-stranded RNA for activation, are derived from the same gene by differential splicing between the fifth and an additional sixth exon of this gene (13,27,28), and are thus identical for the first 346 residues, except for their C-terminal ends. Of the two high molecular weight isoforms, p69, but not p100, requires double-stranded RNA for activation (29). Therefore, IFN-␥ induced expression of IRF-1 and 2-5 AS mRNA was analyzed in Northern analyses using gene-specific probes for IRF-1 and the p40 and p69 isoforms of 2-5 AS (Fig. 3A). The IRF-1 mRNA was induced at 30 min after addition of IFN-␥ and continued to increase gradually thereafter until 48 h post-IFN-␥ exposure. The induction of the p40 and p69 isoforms of 2-5 AS was observed starting at 4 h and peaked at 24 h post-IFN-␥ exposure. The p40 probe hybridized to two transcripts of 1.8 and 1.6 kilobase pairs. Similarly, the p69 probe hybridized to two major transcripts of 5.7 and 4.5 kilobase pairs, of which 5.7 kilobase pairs was the major transcript.
IRF-1 induces 2-5 AS genes, which contain consensus bind-ing sites for IRF-1 within their promoter (30 -33). To examine whether IRF-1 is involved directly in the induction of 2-5 AS transcripts, HEp-2 cells were transfected with antisense IRF-1 ODN and subsequently treated with IFN-␥ (1000 units/ml) for 24 h. Pretreatment of cells with IRF-1 antisense ODN decreased 2-5 AS (p40 and p69) mRNA expression, whereas treatment with scrambled mismatch control ODN had no effect on 2-5 AS mRNA expression (Fig. 3B). To determine whether other intermediates are involved in IRF-1 activation of 2-5 AS mRNA expression, HEp-2 cells were treated with 10 g/ml of cycloheximide (CHX) to block protein synthesis after a 2-h IFN-␥ induction. CHX at this concentration was shown previously to inhibit protein synthesis in HEp-2 cells (34). Cells were harvested 22 h after CHX treatment, and p40 and p69 isoforms of 2-5 AS mRNA were examined by Northern analysis (Fig.  3C). Results showed that addition of CHX to the cells following 2 h of IFN-␥ treatment did not inhibit induction of 2-5 AS mRNAs. In contrast, a moderate increase in 2-5 AS transcripts was observed in CHX-treated cells. Taken together, these results confirm that IFN-␥ induces IRF-1, which, in turn, directly up-regulates 2-5 AS, suggesting that the latter may be involved in the anti-RSV mechanism of IFN-␥. and p69 isoforms of 2-5 AS. Scrambled mismatch of the antisense ODN sequences at the same concentration were used as control. RSV infection was barely detectable in cells treated either with IFN-␥ alone or with IFN-␥ and control ODNs but not in those treated with IFN-␥ and antisense ODNs (Fig. 4A). Addition of antisense ODN significantly reverted (p Ͻ 0.01) the antiviral effect of IFN-␥ against RSV infection, and this reversal was dose-dependent, which increased with increasing concentrations of antisense ODNs. 2-5 AS activity was also reduced in a dose-dependent manner (Fig. 4, B and C) in the cells treated with antisense ODNs to 2-5 AS but not control ODNs. These results indicate that the addition of 2-5 AS antisense ODNs to IFN-␥-treated cells reduced 2-5 AS activity in these cells and, in turn, the antiviral effect of IFN-␥.

2-5 AS Antisense Oligonucleotides Abrogate the Antiviral Effect of IFN-␥ in
Overexpression of RLI Does Not Alter the IFN-␥ Responses in HEp-2 Cells-In addition to RNase L, RLI has been implicated in the antiviral effect of IFN-␥. To determine the role of 2-5A/ RNase L-mediated antiviral mechanism, a stable cell line expressing RLI, RLI-14, was established. Northern analysis of RNAs from RLI-14 and HEp-2 using a gene-specific probe for RLI showed a major 3.5-kb transcript and a minor 2.8-kb transcript (Fig. 5A). A 7-fold increase in the major RLI transcript expression was observed in RLI-14 cells when compared with HEp-2 cells. The analysis of IFN-␥-induced proteins in the RLI-14 cell line by Western blotting showed that IFN-␥ induced expression of IRF-1, but not IRF-2, at 30 min post-induction, and IRF-1 expression continued to increase thereafter until 48 h (Fig. 5B) as in HEp-2 cells (Fig. 2A). Also, a time-specific decrease in PKR protein concentration was observed after IFN-␥ addition in the RLI-14 cell line. The expression of cytokeratin-18, used as an internal control, remained unchanged with IFN-␥ addition. The level of mRNA expression of IRF-1, p40, and p69 isoforms of 2-5 AS was observed by Northern analysis, and the expression level showed a gradual increase over time following IFN-␥ stimulation (Fig. 5C) as in HEp-2 cells (Fig. 3). These results suggest that overexpression of RLI does not change the expression pattern of the IFN-␥-induced genes involved in antiviral activity of these cells.
RLI Overexpression Decreases the Antiviral Activity of IFN-␥-To examine the antiviral effect by the overexpression of RLI, both HEp-2 and RLI-14 cells were treated with IFN-␥ at 100 and 1000 units/ml at 20 h pre-infection and were subsequently infected with RSV. IFN-␥ was added back to the cells at respective concentrations following RSV infection. HEp-2 cells treated with 100 and 1000 units/ml of IFN-␥ showed significant inhibition (p Ͻ 0.001) of RSV infection (72 and 97% reduction, respectively) when compared with untreated cells. In marked contrast, RLI-14 cells showed significantly lower inhibition of infection (only 12 and 22% reduction, respectively) compared with HEp-2 cells at respective concentrations of IFN-␥ (Fig.  6A). In the absence of IFN-␥ treatment, both cell lines exhibited identical RSV titers upon infection. However, the viral titer decreased significantly (p Ͻ 0.01) when the concentration of IFN-␥ was increased from 100 to 1000 units/ml in RLI-14 and HEp-2 cells (Fig. 6A). To examine whether this increase in antiviral activity was because of an increased expression of 2-5 AS with increasing concentrations of IFN-␥, the expression of 2-5 AS mRNA following treatment with increasing concentrations of IFN-␥ was analyzed. Results showed that the expression of 2-5 AS mRNA in cells was IFN-␥ dose-dependent, i.e. 1000 units/ml of IFN-␥ induced a higher level of 2-5 AS mRNA than 100 units/ml of IFN-␥ (Fig. 6B). 2-5 AS, in turn, produces 2-5A, which subsequently binds to RNase L and increases the level of active RNase L by releasing RNase L from RLI (17). Reduction in virus replication was also lower in RLI-14 cells (18%) when compared with HEp-2 cells (86%) (Fig. 6C). To examine whether the reduction in inhibition of RSV infection in RLI-14 cells was because of reduced RNase L activity in these cells, RNase L assay was done using rRNA cleavage assay. This reaction uses cell lysate as a source of both substrate and enzyme, thus giving a comparison of the ribonuclease activity of RNase L in different cell types. The results confirm that ribonuclease activity of RNase L is indeed reduced in RLI-14 cells when compared with HEp-2 cells as evident from the rRNA cleavage products (Fig. 6D). Together, these results confirm the involvement of 2-5 AS/RNase L in the antiviral effect of IFN-␥ against RSV infection.

DISCUSSION
This report focuses on the elucidation of the mechanism underlying IFN-␥-mediated resistance to RSV infection in human epithelial cells. The salient findings from these studies are as follows. (i) IFN-␥ given before or after RSV infection significantly attenuates RSV infection of human epithelial cells. (ii) Exposure of these cells to IFN-␥ induces IRF-1, which, in turn, induces a key antiviral protein, 2-5 AS. (iii) The latter produces 2-5A, which activates RNase L, responsible for degradation of viral RNAs.
RSV is resistant to the antiviral effects of type-I interferons and human MxA (22). Therefore, the finding that treatment of HEp-2 and A549 cells from 20 h pre-infection with as low as 100 units/ml of IFN-␥ inhibits RSV infection and replication when compared with untreated cells has significant therapeutic implications. HEp-2 and A549 cells treated with 1000 units/ml of IFN-␥ from 20 h pre-infection to 72 h post-infection exhibited a 97% (30 -31-fold in log 10 plaque-forming units/ml) reduction in RSV titer. Furthermore, treatment of these cells with IFN-␥ 1 h post-RSV infection without any prior treatment with IFN-␥ significantly decreased (by 39%; 1.7-fold reduction in log 10 plaque-forming units/ml) RSV titer. The anti-RSV effect of IFN-␥ in the most permissive cell lines, such as HEp-2 and A549, implicates its potential as an important therapeutic modality against RSV infection.
The mechanism of antiviral action of IFN-␥ is complex and may be unique for individual cell lines and viruses. Because HEp-2 and A549 represent the most RSV-permissive epithelial cell lines, they were selected to investigate the expression profile of ISGs including IRF-1, IRF-2, PKR, and iNOS, which are potentially relevant to IFN-␥-induced antiviral activity in these cells. The results indicate that IFN-␥ induces both the mRNA and protein for IRF-1 but not IRF-2. These findings in HEp-2 cells are consistent with those found for human macrophages, where IFN-␥ treatment does not enhance IRF-2 gene expression, despite strong up-regulation of IRF-1 mRNA expression (35). Previous studies have shown that in non-induced cells the IRF-2 protein functions as a repressor of ISGs (36). IFN-␥ induction temporarily removes this repression and activates ISGs including IRF-1. IRF-1 and IRF-2 compete for the same cis-acting recognition sequences but with opposite effects (36 -38). PKR has been implicated in antiviral activity. IFN-␥-activated PKR phosphorylates and inactivates eIF-2␣ and leads to restriction of cellular, as well as viral, protein synthesis (39,40). The results that IFN-␥ treatment induces a gradual de- GAPDH was used as an internal control. The relative intensity of the major RNase L inhibitor band (3.5 kb) normalized to GAPDH is plotted (lower panel). B, RLI-14 cells were treated with 1000 units/ml of IFN-␥ and harvested at various time points after IFN-␥ treatment. 30 g of total proteins were separated on SDS-PAGE and immunoblotted using specific antibodies. Cytokeratin-18 was used as an internal control. C, total RNA from IFN-␥-treated RLI-14 cells was isolated, and Northern analysis was performed for IRF-1, 2-5 AS, p40, and p69 isoforms. GAPDH was used as an internal control. The mRNA sizes are indicated at the right of the band by an arrow. Each of these experiments was repeated twice, and the result of a representative experiment is shown. cline in PKR protein expression and no change in phosphorylation of eIF-2␣ (data not shown) indicate that PKR is not involved in IFN-␥-mediated antiviral activity in these cells. Furthermore, iNOS is known to mediate antiviral property of IFN-␥ (41). The lack of detectable levels of iNOS protein or NO in IFN-␥-stimulated HEp-2 and A549 cells indicates that iNOS does not play a role in IFN-␥-mediated inhibition of RSV infection in these cells.
IFN-␥ is known to induce IRF-1, which, in turn, activates 2-5 AS expression (30 -33). The results of this study show that the expression of 2-5 AS p40 and p69 are induced by IFN-␥ in these cells at 4 h and peaks at 24 h post-IFN-␥ addition (Fig. 3A). This 2-5 AS mRNA expression is consistent with the antiviral effect of IFN-␥ observed in these cells when the cells are treated with IFN-␥ from 4 h pre-infection and is highest when treated from 20 h pre-infection (Fig. 1) as the level of 2-5 AS is maximum at that time. The result that expression of 2-5 AS mRNA is decreased in cells treated with IFN-␥ and antisense IRF-1 ODN, but not with scrambled mismatch ODN, demonstrates that the IRF-1 is involved directly in 2-5 AS induction. These data are also consistent with the reports that IRF-1 induces 2-5 AS genes, which contain consensus binding sites for IRF-1 within their promoter (30 -33). 2-5 AS transcripts are induced at 4 h following IFN-␥ addition, although highest expression is observed only at 24 h. This large time gap between IRF-1 expression and 2-5 AS induction could suggest involvement of other intermediates in induction of 2-5 AS. To discern whether IRF-1 directly activates 2-5 AS mRNA expression, 2-5 AS expression was examined in HEp-2 cells after the highest expression of IRF-1 (2 h of IFN-␥ induction) and then by blocking protein synthesis with CHX treatment. A moderate increase in expression of 2-5 AS in CHX-treated cells compared with controls precludes involvement of other intermediates in the induction of 2-5 AS following IRF-1 up-regulation. The reason for the increase in 2-5 AS mRNA levels in CHX-treated cells is unknown. It is likely that other protein(s) synthesized following IRF-1 expression repress(es) IRF-1 induction of 2-5 AS.
A significant finding of this report is the demonstration that 2-5 AS plays a central role in the antiviral activity in these cells. The abrogation of IFN-␥-mediated anti-RSV activity by treatment of cells with an equimolar mixture of antisense ODNs to p40 and p69 but not by the scrambled mismatch ODNs in a dose-dependent fashion and decrease in synthetase activity following ODN treatment provide definitive evidence supporting the role of 2-5 AS in the antiviral mechanism of IFN-␥. The 2-5 AS is pivotal to anti-RSV activity of IFN-␥, as 2-5 AS-induced 2-5A (13,17) binds to and activates RNase L, which cleaves double-stranded RNA 3Ј of UpN residues (42,43).
In contrast to 2-5 AS, the biological activity of RNase L is controlled at the level of enzymatic activation rather than through regulation of its transcription and translation. Increasing endogenous levels of 2-5A leads to enhanced RNase L activity (44), which suggests that intracellular levels of 2-5A are rate-limiting in the activation of RNase L, whereas cellular levels of RNase L are sufficient for maximal biological activity. RNase L remains in an inactive form in the cells being bound to an inhibitor, RLI, which antagonizes the binding of 2-5A to RNase L, thus inhibiting the latter's dimerization, activation, and nuclease activity (45). To dissect the role of RNase L in 2-5 AS-induced anti-RSV mechanism, RLI-14, a stable cell line overexpressing RLI, was established from HEp-2 cells and characterized. The finding that RLI-14 was almost identical to the parent HEp-2 cells in its response to IFN-␥ shows that RLI overexpression does not alter the induction of ISGs in these cells (Fig. 5). This result is consistent with the report that RLI expression is not regulated by IFN-␥ (15).
The demonstration that RLI-14 cells with reduced RNase L activity exhibit a slight reduction in viral titers in the presence of IFN-␥ (Fig. 6) is a significant finding as it confirms that the activation of RNase L is indeed critical to the antiviral effect of IFN-␥ and that the antiviral state is controlled by the elevated expression of 2-5 AS in these cells following IFN-␥ treatment. The role of RLI found in this study is consistent with that reported for human immunodeficiency virus, where RLI is induced during human immunodeficiency virus, type 1 infection and down-regulates the 2-5A/RNase L pathway in human T cells (46). On the other hand, the reduction in antiviral effect of IFN-␥ in these cells is dependent on the dose of IFN-␥, indicating that the level of 2-5 AS and, in turn, 2-5A, is also crucial to the antiviral effect of IFN-␥. A significant reduction (p Ͻ 0.01) in RSV infection is observed when HEp-2 cells are treated with 100 units/ml of IFN-␥ from 20 h pre-infection and transfected with 1 M 2-5A 2 h pre-infection when compared with the cells treated with 100 units/ml of IFN-␥ alone. 2 Thus, both 2-5 AS-induced 2-5A and RLI are major determinants of antiviral activity of cells.
In summary, these results demonstrate that IFN-␥ inhibits RSV infection of HEp-2 and A549 human epithelial cells and the mechanism underlying this inhibition. IFN-␥ induces IRF-1, which, in turn, up-regulates 2-5 AS that generates 2-5A, the activator of RNase L. RNase L is normally found in the cytoplasm in inactive state bound to RLI. Thus, RNase L-mediated cleavage of viral RNA is governed by the yin-yang mechanism involving 2-5A and RLI. In a 2-5A-dominant state cells are protected from RSV infection because of the activation of RNase L. In contrast, an RLI-dominant condition attenuates the antiviral effect by inactivation of RNase L. Because 2-5A and RLI are, respectively, governed by IFN-␥-dependent and -independent mechanisms, treatment with IFN-␥ or overexpression of 2-5 AS should provide an efficient means to redirect the 2-5A:RLI ratio toward a shift in favor of 2-5A and achieve a profound antiviral effect.