|Home | About | Journals | Submit | Contact Us | Français|
Respiratory syncytial virus (RSV) infection of airway epithelial cells results in persistent NF-κB activation and NF-κB-mediated interleukin-8 production. Previous studies in airway epithelial cells demonstrated that tumor necrosis factor alpha (TNF-α)-induced NF-κB activation is transient due to regulation by IκBα. However, during RSV infection, IκBα has only a partial inhibitory effect on NF-κB activation. Studies presented here demonstrate that neither increased IκBα production which occurs as a result of RSV-induced NF-κB activation nor inhibition of proteasome-mediated IκBα degradation results in a reversal of RSV-induced NF-κB activation. Thus, while manipulation of IκBα results in reversal of TNF-α-induced NF-κB activation, manipulation of IκBα does not result in a reversal of RSV-induced NF-κB activation.
Respiratory syncytial virus (RSV) infection in the airway is a major cause of morbidity and mortality in children (7). In vitro and in vivo studies have demonstrated that the pathophysiology of RSV infection involves airway inflammation (17). A key component to the inflammatory response is the production of proinflammatory mediators, such as interleukin-1 (IL-1), IL-6, IL-8, tumor necrosis factor alpha (TNF-α), and granulocye-macrophage colony-stimulating factor, by airway epithelial cells (18). All of these proinflammatory mediators are regulated at the level of gene transcription by the nuclear factor NF-κB (3). Studies in our and other labs have demonstrated that RSV replication in airway epithelial cells is associated with NF-κB activation (5, 8, 12). Therefore, amelioration of NF-κB activation offers a potential means of reversing RSV-induced inflammation.
In unstimulated cells, NF-κB is sequestered in the cytoplasm by inhibitors in the IκB family (3). Studies using A549 cells have demonstrated that with TNF-α stimulation, IκBα is targeted for degradation within 5 to 10 min (10). This is associated with NF-κB activation (10, 14). However, when the cells are treated with the proteasome inhibitor MG-132, IκBα degradation and NF-κB activation are reversed. The studies presented here were designed to determine whether augmentation of IκBα protein levels was associated with a reversal of RSV-induced NF-κB activation.
Studies in our lab and others have demonstrated that A549 cells respond to RSV infection similarly to primary airway cells in culture (2, 11). For all experiments, the A549 cells were between passages 80 and 95. The cells were maintained in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with lipopolysaccharide-free 8% fetal calf serum and 2 mM l-glutamine. No antimicrobial agents were used at any time, and the cells were free of mycoplasma infection. The cells were infected with RSV by adding the virus at a multiplicity of infection of 1.0 in DMEM for 2 h. The RSV-containing DMEM was then removed, and the cells were washed several times with DMEM.
Previous studies have demonstrated that RSV infection of airway epithelial cells results in persistent NF-κB activation. Therefore, our focus in this study was on whether newly synthesized IκBα could inhibit RSV-induced NF-κB activation and whether inhibition of IκBα degradation would limit RSV-induced NF-κB activation.
Following RSV infection, NF-κB activation was observed for up to 72 h by electrophoretic mobility shift assays (EMSA) (9) of nuclear extracts (Fig. (Fig.1B).1B). The NF-κB probe consisted of a 32P-labeled double-stranded DNA corresponding to the NF-κB binding site present in the IL-8 gene. The sequence of the probe is CAGCTACGCAGCGTGGAATTTCCT, which corresponds to a mutated NF-IL-6 site that does not bind NF-IL-6 (data not shown) and an intact NF-κB site from the IL-8 gene (15, 16). As illustrated in Fig. Fig.1B1B (which is representative of five experiments), control cells had minimal NF-κB activation. In contrast, in RSV-infected cells, NF-κB activation was apparent 24 h after infection and remained elevated at 48 and 72 h after infection.
Western blot analysis was used to determine the effects of RSV infection on IκBα protein levels. A549 cells were infected with RSV, and cytoplasmic extracts were obtained 24, 48, and 72 h after infection. As illustrated in Fig. Fig.1A1A (which is representative of five experiments), 24 h after infection, IκBα levels are lower in RSV-infected cells than in control cells. However, IκBα is still detectable in the RSV-infected cells. This is in contrast to TNF-α stimulation, which results in complete loss of IκBα, as previously reported (10, 14). Furthermore, in RSV-infected cells, IκBα levels approach those of control cells at 48 and 72 h. Thus, RSV infection is associated with an initial decline in IκBα levels, though near-control levels are detected later in the infection. However, NF-κB activation persists despite the near-control levels of IκBα, suggesting that IκBα cannot inhibit NF-κB nuclear translocation and DNA binding in the RSV-infected cells.
In other cell types, activation of NF-κB has been associated with increased IκBα gene expression (13). Three NF-κB binding sites are present in the 5′ flanking region of the IκBα gene (13). By increasing IκBα gene expression and protein production, NF-κB essentially down-regulates its own activation. To determine whether RSV-induced NF-κB activation was associated with IκBα gene expression, Northern blot analysis was performed. As illustrated in Fig. Fig.2A,2A, at 24, 48, and 72 h after infection IκBα mRNA expression is higher in RSV-infected cells than in controls. The increase in IκBα mRNA expression occurs during viral replication, as evidenced by expression of the RSV F gene (8). Thus, this result indicates that increased IκBα mRNA expression is associated with viral replication during RSV infection.
The next experiment was designed to determine whether the increase in IκBα mRNA levels observed with RSV infection was due to prolongation of the IκBα mRNA half-life. Control and RSV-infected cells were treated with actinomycin D for 1, 2, or 4 h, and Northern blot analysis performed on cellular RNA. As illustrated in Fig. Fig.2B,2B, treatment with 10 μg of actinomycin D per ml results in decreased IκBα mRNA expression in both control and RSV-infected cells. Furthermore, as illustrated in Fig. Fig.2C,2C, the slope of the degradation curve for the IκBα mRNA signal in RSV-infected cells was identical to the slope of the curve in control cells. The curve for the RSV-infected cells was shifted upward due to a higher initial expression level of IκBα in RSV-infected cells. This result indicates that the increase in IκBα mRNA expression in RSV-infected cells is not due to stabilization of the message but likely is due to increased gene transcription.
The next experiment was designed to determine whether RSV infection resulted in the synthesis of IκBα protein. Control and RSV-infected cells were placed in methionine- and cysteine-free medium for 4 h, and then 35S-labeled methionine and cysteine were added. Cells lysates were obtained and subjected to protein precipitation using an antibody to IκBα. The resulting precipitate was subjected to electrophoresis on a 10% Tris-glycine gel. As illustrated in Fig. Fig.3,3, minimal IκBα is detected in control cells (lane 1). However, in RSV-infected cells (lane 2), two bands are detected at the appropriate size for IκBα and phosphorylated IκBα. This result indicates that RSV infection is associated with increased IκBα protein synthesis in A549 cells.
It was intriguing that IκBα protein levels were increasing after 24 h, but there was no apparent inhibition of NF-κB activation. This result indicates that NF-κB-regulated IκBα production did not result in effective inhibition of NF-κB activation as was previously reported for TNF-α stimulation (22). To determine whether IκBα production had any effect on RSV-induced NF-κB activation, the next set of experiments were performed to determine the effects of limiting IκBα gene expression and the effects of limiting IκBα protein production during RSV infection.
To determine whether limiting RSV-induced IκBα mRNA expression altered NF-κB activation, control and RSV-infected cells were treated with actinomycin D for 4 h. As illustrated in Fig. Fig.2B,2B, this results in decreased IκB mRNA expression. Treatment with actinomycin D also resulted in decreased IκBα protein detection (data not shown). EMSA analysis of nuclear extracts for NF-κB activation is illustrated in Fig. Fig.4A.4A. Treatment of control cells with actinomycin D resulted in greater NF-κB activation than in untreated control cells (Fig. (Fig.4A;4A; compare lanes 1 and 2). Furthermore, treatment of RSV infected cells with actinomycin D for 4 h resulted in a greater NF-κB response than in untreated RSV-infected cells (Fig. (Fig.4A;4A; compare lanes 3 and 4). This result is consistent with the notion that inhibition of IκBα mRNA expression results in an exaggerated NF-κB response during RSV infection.
The next experiment focused on determining whether inhibition of IκBα protein synthesis would alter RSV-induced NF-κB activation. Control and infected cells were treated with 10 μg of cycloheximide per ml for 4 h. As illustrated in the Western blot in Fig. Fig.4B,4B, a small difference in IκBα protein detection is observed between control (lane 1) and RSV-infected (lane 2) cells. Treatment of control cells with cycloheximide had little effect on IκBα protein detection in control cells (lane 3) but had a marked effect on IκBα protein detection in RSV-infected cells (lane 4).
Analysis of nuclear extracts for NF-κB activation by EMSA analysis is illustrated in Fig. Fig.4C.4C. Treatment of control cells with cycloheximide resulted in no apparent difference in NF-κB activation (compare lanes 1 and 3). However, treatment of RSV-infected cells with cycloheximide resulted in an increase in NF-κB activation (compare lanes 2 and 4). Together, the data from Figs. Figs.4B4B and C indicate that inhibition of IκBα protein synthesis during RSV infection is associated with an increase in NF-κB activation.
The combined results from the experiments using actinomycin and those using cycloheximide indicate that inhibition of IκBα mRNA expression and IκBα protein synthesis results in a exaggerated response to RSV infection with respect to NF-κB activation. The results of these experiments are consistent with (but not proof of) the notion that RSV-induced IκBα mRNA expression and protein synthesis are important in at least partially limiting NF-κB activation.
As illustrated in Fig. Fig.1A,1A, on days 1, 2, and 3, RSV-infected cells contained less IκBα than control cells, which indicates that IκBα degradation was occurring during RSV infection. The next set of experiments was designed to determine whether inhibition of RSV-induced IκBα degradation would result in altered NF-κB activation in response to RSV. For these studies, control cells or cells infected with RSV were treated with various concentrations of the proteasome inhibitor MG-132 for 2 h. Previous studies in our lab demonstrated that treatment of A549 cells with MG-132 reversed TNF-α-induced NF-κB activation by limiting IκBα degradation (10).
The results of the experiments using the proteasome inhibitor MG-132 are illustrated in Fig. Fig.5.5. Western blot analysis (Fig. (Fig.5A)5A) demonstrates that treatment of control and RSV-infected cells with increasing concentrations of MG-132 results in increased IκBα protein detection. The IκBα bands detected in the MG-132 treated cells migrated slightly more slowly than those not treated with MG-132. Previous studies in our lab using MG-132 indicated that the bands represented phosphorylated IκBα (10). Western blot analysis using an antibody to phosphorylated IκBα confirmed that the slower-migrating bands were phosphorylated IκBα (data not shown). Thus, treatment of RSV-infected cells with the proteasome inhibitor MG-132 resulted in reversal of IκBα protein loss but not IκBα phosphorylation.
To determine whether inhibition of proteasome-mediated IκBα degradation resulted in a reversal of RSV-induced NF-κB activation, EMSA analysis of nuclear extracts was performed. As illustrated in Fig. Fig.5B,5B, treatment with increasing concentrations of MG-132 had a partial effect on reversing RSV-induced NF-κB activation. Thus, despite the observation that treatment with MG-132 resulted in complete reversal of RSV-induced IκBα degradation, only a partial reversal of RSV-induced NF-κB activation was observed.
An interesting finding from the EMSA analysis in Fig. Fig.5B5B was that treatment with MG-132 resulted in a slower-migrating NF-κB–DNA complex. The next experiment was designed to characterize this complex. As illustrated in Fig. Fig.6,6, comparison of lane 1 (control cells) to lane 2 (RSV-infected cells) confirms that NF-κB activation is present during infection. Treatment of infected cells with 100 μM MG-132 results in a decrease in the NF-κB signal which is slower migrating (lane 3). When EMSA analysis of the same extracts from lane 3 is performed in the presence of a 100-fold excess of unlabeled NF-κB DNA probe, the signal is completely abolished (lane 4), indicating that the slower-migrating signal represents NF-κB–DNA binding.
Lanes 5 to 10 in Fig. Fig.66 represent attempts to supershift the NF-κB–DNA complex with polyclonal antibodies to individual potential components of NF-κB. The nuclear extracts used are from RSV-infected cells treated with 100 μM MG-132 for 2 h. The addition of antibodies to p65 (lane 5) and p50 (lane 7) resulted in supershifting of the NF-κB DNA complex. However, addition of antibodies specific for c-Rel (Lane 6) and p52 (Lane 8) did not result in any alterations in binding of NF-κB to the DNA probe, nor did they affect migration of the complex. Our results confirms an earlier finding by Bitko et al. (5). Furthermore, this result indicates that the slower-migrating complex observed in lane 3 is composed at least in part of p65 and p50 and does not represent heterodimers involving c-Rel or p52.
Since MG-132 inhibits the degradation of IκBα, we attempted to determine whether IκBα could be complexed to NFκB while bound to the DNA probe. Therefore, two different commercially available polyclonal antibodies to IκBα (C-15 [Fig. 6, lane 9] and C-21 [lane 10]; Santa Cruz Biotechnology) were added to the EMSA reactions. Neither of these antibodies resulted in supershifting of the NF-κB–DNA signal or competitive inhibition of binding of NF-κB to DNA probe. To further investigate whether IκBα remained bound to NF-κB during treatment with MG-132, whole-cell extracts were subjected to protein precipitation using a polyclonal antibody to p65. The precipitated protein was subjected to Western blot analysis using an antibody to IκBα (C-21). As illustrated in Fig. Fig.7,7, the IκBα protein is easily detected in control cells (lane 1). However, in RSV-infected cells, less IκBα coprecipitated with the p65 antibody (lane 2). Furthermore, no increase in IκBα coprecipitation with the p65 antibody was observed with RSV-infected cells treated with 100 μM MG-132 (lane 3). Thus, treatment with MG-132 does not result in increased p65-IκBα binding. Combined, the results of the EMSA analysis and coprecipitation studies do not support the notion that the slower-migrating NF-κB–DNA signal observed in MG-132-treated cells infected with RSV is due to persistent IκBα binding. Furthermore, the coprecipitation study indicates that despite the observation that MG-132 results in decreased IκBα degradation, it does not increase binding of IκBα to p65.
The EMSA analysis in Fig. Fig.5B5B indicates that treatment of RSV-infected cells with 100 μM MG-132 resulted in a small decrease in NF-κB activation. To determine whether this was of biological significance, the effects of RSV infection on IL-8 gene transcription and IL-8 protein production were determined. IL-8 gene transcription was assessed in a luciferase reporter assay previously described (9). In this assay, NF-κB-mediated IL-8 gene transcription is responsible for luciferase activity measured in cell lysates. Luciferase activity was measured in cells transfected with the reporter plasmid (controls) and cells transfected with the reporter plasmid and infected with RSV. For the control cells (n = 12), each of the individual values is divided by the mean in order to obtain a standard deviation. For the RSV-infected cells, each of the individual values (n = 12) is divided by the mean of the control cell values. Therefore, the values are reported as an increase in luciferase activity as a function of RSV infection. As illustrated in Fig. Fig.8A,8A, RSV infection of A549 cells results in a nearly sevenfold increase in luciferase activity (and therefore IL-8 gene transcription). Treatment with 100 μM MG-132 for 4 h resulted in a small but statistically significant (P < 0.05 by analysis of variance [ANOVA], n = 12) decrease in luciferase activity (Fig. (Fig.8;8; compare RSV to RSV+MG-132). However, similar to the observed changes in NF-κB activation, the luciferase activity in RSV-infected cells treated with 100 μM MG-132 was markedly greater than that in control cells.
The next set of experiments focused on determining whether the small but statistically significant decrease in IL-8 gene transcription correlated with IL-8 protein production and release. For these experiments, the medium from control cells and RSV-infected cells was collected 4 h after the medium was changed. Cells treated with MG-132 received the agent at a concentration of 100 μM for 4 h. Cell supernatants, representative of the release of IL-8 over 4 h, were subjected to a previously described (9) enzyme-linked immunosorbent assay (ELISA) for IL-8. Cell lysates were also obtained after 4 h and subjected ELISA. As illustrated in Fig. Fig.8B,8B, RSV infection is associated with a marked increase in the amount of IL-8 released from the cells (error bars represent standard deviation for the supernatant, P < 0.001 by ANOVA, n = 12) and the amount of cell-associated IL-8 (lysates, P < 0.001 by ANOVA, n = 12). However, treatment of RSV-infected cells with 100 μM MG-132 for 4 h did not result in any statistically significant difference in IL-8 release into the supernatant or the cell-associated IL-8 (lysates) compared to untreated RSV-infected cells. Thus, although inhibition of proteasome-mediated IκBα degradation resulted in a small reversal of RSV-induced NF-κB activation and NF-κB-mediated IL-8 gene transcription, no significant change in IL-8 release or IL-8 production was observed.
The purpose of the study reported here was to determine whether augmentation of endogenous IκBα protein levels would reverse RSV-induced NF-κB activation and NF-κB-mediated inflammation. Our findings indicate that RSV-induced IκBα protein production partially limits virus-induced NF-κB activation in A549 cells. Data presented here indicate that RSV infection is associated with increased IκBα mRNA expression and protein production. Limiting either IκBα mRNA expression or protein production results in exaggerated NF-κB activation in response to RSV infection. Additionally, data presented here indicate that inhibition of endogenous IκBα degradation did not reverse RSV-induced NF-κB activation. Combined, these results indicate that endogenous IκBα plays a role in limiting RSV-induced NF-κB activation but cannot be recruited to completely reverse RSV-induced NF-κB activation.
Recently Thomas et al. reported that adenovirus-mediated expression of a nondegradable form of IκBα resulted in inhibition of RSV-induced NF-κB activation and NF-κB-mediated RANTES production in airway epithelial cells (20). The studies reported by Thomas et al. differ from those presented here in two major respects. First, the cells were transfected with the adenovirus-containing IκBα construct simultaneously with RSV infection. This enabled the cells to produce nondegradable IκBα during the entire course of RSV infection. In contrast, we did not treat the infected cells until RSV infection had progressed for 2 days. Thus, one reason for the differences in results observed maybe accounted for by experimental design. Our studies were focused on reversing NF-κB activation after it was initiated, whereas the studies of Thomas et al. were designed to inhibit NF-κB activation from the initial point of infection.
A second difference between the studies involves their use of a mutant, exogenous IκBα as opposed to our attempts to augment endogenous IκBα protein levels. The recombinant IκBα was rendered resistant to phosphorylation and subsequent degradation by mutating serines at positions 32 and 36. Thus, any recombinant protein made by the cells theoretically would not be degraded. In contrast, our studies focused on maintaining the endogenous IκBα protein at control levels, as this approach was successful with TNF-α stimulation (10). Inhibition of proteasome degradation does not alter IκBα phosphorylation. It is possible that nonphosphorylated IκBα maintains interactions with NF-κB to a greater degree than phosphorylated IκBα. Determining whether phosphorylation of IκBα alters its interactions with p65 was beyond the scope of this study.
Recently, Jamaluddin et al. reported that the major component of IκBα proteolysis occurs independent of the proteasome pathway in RSV-infected cells (14). In this study, the authors found that treatment with 25 μM Mg-132 for 4.5 h did not reverse IκBα degradation. The results of our study, in comparison, indicate that treatment with MG-132 in dose range of 1 to 100 μg/ml for 2 h (all nontoxic as assayed by neutral red uptake) resulted in a significant reversal of IκBα degradation. The 2-h treatment course was chosen based on preliminary studies comparing treatment courses with MG-132 of 20 min to 4 h. These studies demonstrated that peak inhibition of IκBα degradation occurred at 2 h (data not shown). The differences in our reported findings may be explained by the fact that MG-132 is a reversible inhibitor of the proteasome, and the optimal time course was chosen for our study.
Bitko and Barik recently reported on the persistent activation of the RelA component of NF-κB by RSV (4); they demonstrated that MG-132 (100 μM) reversed IκBα degradation and NF-κB activation at certain time points. The effects of MG-132 were reported as modest. The authors also presented data which suggest that IκBα may act as a chaperone for NF-κB, protecting it from IκBα inhibition. The results presented here are consistent with the finding of Bitko and Barik. The unsuccessful attempts to inhibit RSV-induced NF-κB activation by increasing IκBα protein levels may be due to binding of IκBβ to NF-κB. This would suggest that a pool of NF-κB (which can no longer be regulated by IκBα) is present in the RSV-infected cells. An analysis of the role of IκBβ was beyond the scope of our study.
In summary, the study presented here focused on testing the hypothesis that manipulation of IκBα levels would alter RSV-induced NF-κB activation and NF-κB-mediated IL-8 production; the results do not support this hypothesis. While RSV-induced IκBα synthesis does limit NF-κB activation, complete reversal does not occur. Furthermore, augmentation of IκBα levels by inhibition of its degration does not reverse RSV-induced NF-κB activation.
While our results are not in complete agreement with those of Thomas et al. (20) and Jamaluddin et al. (14), a recurrent finding is that RSV induces persistent NF-κB activation. The results presented here, along with those of Jamaluddin et al. and Bitko and Barik, may begin to explain why some children with RSV infection have persistent respiratory symptoms (1, 6, 21). It has been well documented that inflammation plays a key role in RSV-induced morbidity and mortality (6, 21). Persistent NF-κB-mediated inflammatory gene production in airway epithelial cells likely contributes to the morbidity associated with RSV infection. Furthermore, this persistent inflammatory response may help explain the morbidity associated with RSV infection in children with chronic inflammatory airway disorders such as cystic fibrosis (1). It will be critical in future studies to define the mechanisms by which persistent NF-κB activation occurs, in order to begin to reverse the inflammatory effects associated with its activation.
This study was supported by grants from the NIH (HL03451 to M.A.F.) and the Cystic Fibrosis Foundation (FIEDLEGO97).