Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Human rhinovirus (HRV) infections are the predominant cause of the common cold and have been associated with exacerbations of asthma (1, 2), chronic bronchitis (3), and sinusitis (4). Although upper respiratory viral infections can induce airway inflammation and changes in airway physiology, the mechanisms underlying these effects are not well understood.

It has been shown that rhinovirus infection of asthmatic and atopic subjects can enhance lower airway inflammation. Infection increased recruitment of eosinophils into bronchoalveolar lavage fluid (BALF) 48 h after allergen challenge of allergic but not nonallergic subjects, a finding that persisted for a month after infection (5). Rhinovirus-induced exacerbation of lower airway inflammation has also been observed, without allergen provocation, in asthmatic but not normal subjects. Rhinovirus infection increased the numbers both of submucosal lymphocytes and of eosinophils in the epithelial layer. Again, the eosinophilia persisted into convalescence (6). Given that eosinophil numbers correlate with the degree of airway hyperresponsiveness and with disease severity in asthma (7, 8), it is reasonable to hypothesize that rhinovirus infections induce asthma exacerbations by increasing lower airways eosinophilia.

The airway epithelial cell is the primary site of rhinovirus infection (9, 10), and there is growing evidence that virally infected respiratory epithelial cells produce several cytokines that could play an important role in the recruitment and activation of inflammatory cells in the lung (11- 14). Among the cytokines produced by epithelial cells in response to rhinovirus infection is granulocyte macrophage colony-stimulating factor (GM-CSF), a hematopoietic cytokine that could contribute significantly to airway eosinophilia induced by rhinovirus infection (11). GM-CSF not only promotes differentiation, chemotaxis, and survival of eosinophils, but also can enhance their activation state, induce adhesion molecule expression, and act as a cofactor for superoxide production and degranulation (15, 16). GM-CSF is known to be important in eosinophilic inflammation in the lung. A substantial amount of GM-CSF is produced in the lungs after antigen challenge in asthmatic subjects, and GM-CSF appears to be the major cytokine in the BALF of symptomatic asthmatics responsible for eosinophil survival (16).

It is clear, however, that upper respiratory viral infections do not lead to exacerbations of airway inflammation and of asthma symptoms in all asthmatics. Susceptibility to disease exacerbation may depend upon many factors, including viral load and the host antiviral and immune responses. Recent studies have demonstrated that the free radical nitric oxide (NO) may play an important role in host defense against viruses (1, 17, 18). Replication of a wide range of DNA and RNA viruses is inhibited in vitro by the addition of chemical donors of NO or by the induction of type II nitric oxide synthase (NOS), one of three isoforms of the enzyme that produces NO. Studies in animal models have shown that inhibitors of NOS can increase viral load and decrease survival, and mice deficient in NOS II have been reported to be more susceptible to viral infection (19). In previous in vitro studies of rhinovirus-infected human bronchial epithelial cells, we have shown that NO can inhibit virus replication as well as virus-induced production of the proinflammatory cytokines interleukin (IL)-8 and IL-6 (18).

Because of the potentially important role of GM-CSF in inducing and maintaining eosinophilic inflammation during viral exacerbations of asthma, the present studies were undertaken to further delineate the kinetics of viral production of GM-CSF from human respiratory epithelial cells, and to define the effects of NO on virally induced production of GM-CSF. Studies were conducted in primary human bronchial and adenoid epithelial cells and in the BEAS-2B bronchial epithelial cell line, on the basis of the previous demonstration that this cell line responds to rhinovirus infection by producing a profile of cytokines similar to that of primary human cells. We demonstrate that rhinovirus infection leads to a rapid and sustained increase in production of GM-CSF from human respiratory epithelial cells and that NO significantly inhibits rhinovirus-induced production of GM-CSF. In contrast to its effects on production of IL-8 and IL-6, which appear to occur primarily at the post-transcriptional level, NO inhibits epithelial production of GM-CSF upon infection with HRV by repressing virally induced transcription of GM-CSF in these cells.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Materials

The following reagents were purchased from the indicated suppliers: Dulbecco's modified Eagle's medium (DMEM), Eagle's minimal essential medium, Ham's F-12 medium, Hanks' balanced salt solution (HBSS), L-glutamine, penicillin-streptomycin-amphotericin B (Fungizone), trace elements, and retinoic acid (Biofluids, Rockville, MD); hydrocortisone, epithelial cell growth factor, and endothelial cell growth supplement (Collaborative Research, Bedford, MA); fetal bovine serum (FBS) (Gemini Bio Products, Inc., Calabasas, CA); transferrin, acrylamide, N,N'-methylene- bis-acrylamide, ammonium persulfate, N,N,N',N'-tetramethyl-ethylenediamine liquid, and insulin (GIBCO BRL, Grand Island, NY); 3- (2-hydroxy-2-nitroso-1-propyl-hydrazino)-1-propanamine (NONOate) (Cayman Chemical Company, Ann Arbor, MI); RNAzol B (Tel-Test, Inc., Friendswood, TX); agarose (FMC Bioproducts, Rockland, ME); 3-(N-Morpholino)propanesulfonic acid (MOPS), DNA Polymerase I (Klenow fragment), and phenylmethylsulfonyl fluoride (PMSF) (Boehringer Mannheim, Indianapolis, IN); anti-p65 antibody, anti-p50 antibody, and anti-c-Rel antibody (Santa Cruz Biotechnologies, Santa Cruz, CA); and [³²P]deoxycytidine triphosphate (dCTP), poly dI-dC, oligolabeling kit (Amersham-Pharmacia, Piscataway, NJ). All other chemicals were purchased from Sigma Chemical Company (St. Louis, MO).

The following stock buffers were used: 10× MOPS (0.2 M MOPS, 0.05 M sodium acetate, and 0.01 M EDTA), 50× Denhardt's solution (1% Ficoll, 1% polyvinylpyrrolidone, 1% bovine serum albumin), and 20× saline sodium phosphate ethylenediaminetetraacetic acid (SSPE) (175.3 g NaCl, 27.6 g NaH₂PO₄ H₂O, and 7.4 g EDTA in 1 liter H₂O [pH 7.4]).

Cell Lines and Viruses

The BEAS-2B cell line was a gift from Curtis Harris (National Cancer Institute, Bethesda, MD). HRV type 16 and WI-38 cells were purchased from the American Type Culture Collection (Rockville, MD). HRV-16 viral stocks used for experiments were generated by passage in WI-38 cells and were purified by centrifugation through sucrose gradients as previously described (18). Purified virus was used for all experiments. To assure that cytokine induction by purified HRV-16 was specifically due to viral infection and not due to residual factors from WI-38 cells that may have copurified with the virus, three modifications of the viral preparation were tested. First, inactivation of virus by exposure to ultraviolet (UV) light was performed as previously described (18). Second, a 2-ml aliquot of a purified HRV-16 virus preparation was subjected to ultracentrifugation through a Centricon membrane with 30 kD molecular weight cut-off (Amicon, Beverly, MA). Centrifugation of the preparation was interrupted when approximately 1 ml of the solution had passed through the filter. At this time, the approximately 2-fold concentrated retentate and the filtrate were removed and used to infect epithelial cells. The filtrate and the concentrated retentate were tested for their ability to generate GM-CSF and compared with the original virus preparation. Third, epithelial cell cultures were preincubated with 20 µg/ml of mouse blocking monoclonal antibody (mAb) to human intercellular adhesion molecule (ICAM)-1 (84H10; Coulter-Immunotech, Miami, FL) as previously described (11). Cells were exposed to varying doses of HRV-16 and cytokine production was measured in supernatants recovered at 24 h after infection.

Epithelial Cell Culture

Primary human bronchial epithelial cells and primary human adenoid epithelial cells were obtained by protease digestion of human tissue as previously described (20). Primary bronchial epithelial cells were grown on culture plates coated with rat-tail collagen (type VII; Sigma) in culture medium consisting of DMEM/ Ham's F12 with 5% heat-inactivated FBS, penicillin (100 U/ml), streptomycin (100 U/ml), amphotericin B (250 ng/ml), and L-glutamine (2 mM). Primary adenoid epithelial cells were grown in serum-free bronchial epithelial growth medium (BEGM) (Clonetics, San Diego, CA). BEAS-2B cells (between passages 34 and 45) were grown in culture medium consisting of Ham's F12 nutrient medium with penicillin (100 U/ml), streptomycin (100 U/ml), amphotericin B (250 ng/ml), L-glutamine (2 mM), phosphoethanolamine-ethanolamine (0.5 mM), transferrin (10 µg/ml), endothelial cell growth supplement (3.75 µg/ml), epidermal growth factor (12.5 ng/ml), insulin (5 µg/ml), hydrocortisone (10⁷ M), cholera toxin (10 ng/ml), 3,3',5-triiodothyronine (3 × 10⁹ M), retinoic acid (0.1 ng/ml), and trace elements. This medium is hereafter referred to as F12/10×. Additional Ham's F-12 was prepared with all of the above additives except hydrocortisone and is hereafter referred to as F12/9×. The epithelial cell cultures were incubated at 37°C in 95% air and 5% CO₂. For the experiments, cells were plated on six-well plates (Costar, Cambridge, MA) at a density of 1.5 × 10⁵ cells per well or in 75-cm² tissue culture (T-75 cm²) flasks (Costar) at a density of 1 × 10⁶ cells per flask. For experiments with primary human epithelial cells, each experiment was performed with cells obtained from a different donor tissue.

Viral Infection of Epithelial Cells

We have shown previously that the production of GM-CSF by primary human epithelial cells is markedly inhibited by glucocorticoids (21), and we have confirmed this observation for HRV-16- induced production of GM-CSF by BEAS-2B cells (unpublished observations). Because BEAS-2B cells and primary human bronchial epithelial cells are usually grown in medium containing low levels of hydrocortisone, the monolayers of BEAS-2B cells and primary epithelial cells (75 to 80% confluent) were placed in medium without hydrocortisone (F12/9×) for 24 h before infecting the cells with HRV-16. Using Northern blot analysis of the messenger RNA (mRNA) for GM-CSF, we verified that 24 h is sufficient time to deplete hydrocortisone and its inhibitory effects on message expression. After 24 h in F12/9×, the monolayers of cells were washed three times with HBSS and then HRV-16 was added to the cells at a concentration of 10⁴ 50% tissue culture infective dose (TCID₅₀) units/ml HBSS (18). This equates to an infectious dose of 0.01 TCID₅₀ units per cell. This number, however, represents a multiplicity of infection relative to the WI38 cells in which the TCID₅₀ was determined and may not be directly equivalent for the infection of epithelial cells inasmuch as we and others have previously determined that infectivity varies depending upon the type of host cell (18). The cells were incubated with the virus for 1 h at 34°C and washed three times with fresh HBSS, then placed in fresh F12/9× medium. For kinetic studies, the time at which fresh F12/9× medium was added was defined as time 0. After incubation at 34°C for various times, supernatants were removed from the cells and stored at 80°C for later analysis of cytokine protein production and viral content. For some experiments, total cellular RNA was extracted from the cells at various times after infection and stored at 80°C for later analysis.

Primary human adenoid epithelial cells were infected using a similar protocol, with two exceptions: the cells were grown in BEGM, and the cells were placed in fresh BEGM after the incubation with virus.

Titration of Viruses

Supernatants from infected epithelial cells were collected at various times and assessed for viral content by cytotoxicity assays in WI-38 cells as previously described (18).

Quantification of GM-CSF

Levels of GM-CSF in cell supernatants were determined by a commercial enzyme-linked immunosorbent assay sensitive to 7.8 pg of GM-CSF/ml (R&D Systems, Minneapolis, MN).

Probes for Northern Blotting

The full-length complementary DNA (cDNA) for GM-CSF was kindly provided by Steven Gillis (Immunex, Seattle, WA). The full-length cDNA for glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was purchased from Clontech (Palo Alto, CA). Probes for GM-CSF and GAPDH were labeled to a high specific activity by the random-primer method with [³²P]dCTP and a random-primer DNA labeling kit (Boehringer Mannheim) as previously described (18). The radiolabeled cDNA was isolated from Sephadex G-50 minicolumns (Pharmacia) by centrifugation at 800 × g for 2 min.

RNA Extraction and Northern Blotting

Total cellular RNA was extracted from epithelial cells using RNAzol B as previously described (18). The integrity of each RNA sample was assessed by electrophoresis of an aliquot (0.5 µg) on a 1% agarose gel with 0.5 µg ethidium bromide/ml buffer. RNA was stored at 80°C.

Northern analysis was performed using equal amounts (20 µg) of RNA from each experimental condition as previously described (18). Films were routinely developed for varying times to ensure that band intensities assessed by densitometry were within the linear range for the film. Densitometry was performed with a scanning densitometer (Kodak, Rochester, NY) and densitometric analysis was performed with National Institutes of Health Image software or Kodak software.

Real-Time Reverse Transcriptase/Polymerase Chain Reaction

Using the Applied Biosystems 7700 Sequence Detector, RNA (100 ng) was reverse transcribed to cDNA, followed by polymerase chain reaction (PCR) amplification in the presence of specific primers and a fluorescently labeled probe for GM-CSF that were selected using Primer Express software. GM-CSF primers were: forward, 5'-AGGGCCCCTTGACCATGA-3'; reverse, 5'-CAAAGGTGATAGTCTGGGTTGCA-3'; and GM-CSF probe, FAM-CAGCACTGCCCTCCAACCCCG-TAMRA.

The GM-CSF primers and probes were synthesized by PE Biosystems (Foster City, CA). The primers and probes for the housekeeping gene GAPDH were obtained from the same company. Fold induction in mRNA was calculated using the comparative method recommended by the Applied Biosystems technical bulletin.

Effect of NONOate on Cytokine Production and Viral Replication

NONOate was prepared in alkaline solution (0.01 M NaOH) as a 100-mM stock solution immediately before each experiment and kept at 4°C until added to the cells as previously described (18). Aliquots of the stock solution were added to the epithelial cell culture medium (pH 7.4) in a final concentration range of 100 µM to 1 mM, and were present both during and after the virus exposure. The half-life of the NONOate used in these studies is 76 min at physiologic pH and room temperature (Cayman Chemical Co.). Using a CHI 832 electrochemical detector with a WPI NO electrode and a Faraday Cage (CH Instruments, Inc., Cordova, IN), the precise concentration of free NO in the culture medium was measured at 34°C. The following peak concentrations of free NO were reached within 10 min of adding NONOate to the culture medium: 1 mM NONOate generated 11 µM NO; 500 µM NONOate, 8 µM NO; 300 µM NONOate, 6 µM NO; and 100 µM NONOate, 3 µM NO. The concentration of free NO declined in the solutions at a rate of about 0.02 µM/min. The range of concentrations generated by the NONOate used in these experiments is consistent with that produced by macrophages expressing type II NOS (22).

Supernatants from the epithelial cells incubated with or without NONOate were removed at various times after viral infection and stored at 80°C for later analysis of GM-CSF protein and viral content. In some experiments, RNA was extracted from the cells at various times after infection and stored at 80°C for later Northern or real time reverse transcriptase (RT)/PCR analysis.

RNA Stability

BEAS-2B cells in T-75 cm² flasks were infected with HRV-16 as described earlier. At 1 h after infection, actinomycin D (10 µg/ml) was added. NONOate (1 mM) or vehicle control was also added at this point. Total RNA was extracted immediately after the addition (baseline) and at 1, 3, and 4 h later. The RNA was stored at 80°C for later Northern analysis.

Extraction of Nuclear Proteins

Two T-75 cm² flasks containing approximately 1 × 10⁷ BEAS-2B cells, or one six-well plate containing approximately 5 × 10⁶ primary human adenoid epithelial cells, were used for each treatment condition: uninfected control, HRV-16 infected with and without NONOate (500 µM). The monolayers of cells were mechanically detached in ice-cold phosphate-buffered saline (10 ml). The cell suspension was transferred to a 15-ml conical tube and centrifuged at 1,200 × g and 4°C for 8 min. The cell pellets were combined and lysed by resuspension in 100 µl of lysis buffer (10 mM N-2-hydroxyethylpiperazine-N'-ethane sulfonic acid [pH 7.9], 60 mM KCl, 1 mM EDTA, 1 mM dithiothreitol [DTT], 1 mM PMSF, 0.5% Nonidet P-40 [NP-40]). An aliquot (10 µl) was removed, mixed with an equal volume of trypan blue, and examined under ×40 microscopy to confirm the presence of round, intact nuclei. The remainder of the lysed cell suspension was centrifuged at 1,200 × g and 4°C for 5 min. The resultant nuclear pellet was washed with lysis buffer without NP-40 and centrifuged at 1,200 × g and 4°C for 5 min. The pellet was resuspended in 100 µl of nuclear resuspension buffer (25 mM Tris-HCl [pH 8.0], 400 mM KCl, 1 mM DTT, 1 mM PMSF, and 20% wt/vol glycerol), rapidly frozen and thawed three times, and then centrifuged at 7,000 × g and 4°C for 12 min. The supernatant containing the nuclear proteins was removed and transferred to an Eppendorf tube. An aliquot (10 µl) was used for protein determinations that were performed using the Bio-Rad Protein Assay adapted for microtiter plates (Bio-Rad Chemical Division, Richmond, VA). The remaining nuclear protein extract was frozen at 80°C for later analysis by electrophoretic mobility shift assay (EMSA).

EMSA

The oligonucleotide 5'-GTTCAGGTAGTTCCCCCGCCTC-3' was synthesized by Genosys Biotechnologies (The Woodlands, TX) and contained the reported nuclear factor (NF)-B binding site from the promoter of the GM-CSF gene (23). The oligonucleotide (100 ng) was radiolabeled by random priming (24) using 6 units of DNA Polymerase I (Klenow fragment), [-³²P]dCTP (50 µCi), and 10 µl of Pharmacia reaction mix (deoxyadenosine triphosphate, deoxythymidine triphosphate, and deoxyguanidine triphosphate) in 50 µl final volume of 10 mM Tris-EDTA (pH 8.0) at 37°C for 60 min. The radiolabeled probe was isolated from a Sephadex G-50 minicolumn (Pharmacia) by centrifugation at 800 × g for 2 min.

Extracts of nuclear proteins (200 ng) were incubated with 2 ng of [-³²P]-labeled oligonucleotide (50,000 to 100,000 counts per min per µl) and 1 µg of poly(dI-dC) in 10 µl of binding buffer (10 mM Tris-HCl [pH 7.4], 10% wt/vol glycerol, and 65 mM KCl) for 30 min at 25°C. For supershift experiments, antisera were added and the reaction mixture was incubated for 10 additional minutes at 25°C. Electrophoresis of the entire reaction solution was carried out in vertical 5% polyacrylamide gels using 0.045 M Tris-borate/0.001 M EDTA (pH 8.0) buffer. The gel was fixed in 10% acetic acid/10% ethanol for 10 min and dried before assessment by autoradiography (Biomax film; Kodak). Films were routinely developed for various times to ensure that band intensities assessed by densitometry were within the linear range for the film.

Statistical Analysis

Data are expressed as means ± standard error of the mean (SEM). Comparisons of the kinetics of the RNA expression were performed by one-way analysis of variance. For post hoc analysis, pairwise comparisons of the means were performed with the least significant difference multiple-range test (25). The effects of NONOate on cytokine production, RNA expression, and viral titers were compared using Student's t test for paired samples. Differences were considered significant for values of P < 0.05.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Kinetics of GM-CSF mRNA Expression and Protein Secretion

To examine the time course of virally induced GM-CSF mRNA and protein production, total cellular RNA and supernatants from BEAS-2B cells were harvested at various times after HRV-16 infection. Figure 1 shows that there was a time-dependent increase in GM-CSF mRNA production after HRV-16 infection. Maximal mRNA expression occurred by 1 h after infection, but mRNA levels were still higher than in noninfected controls at 24 h after infection. Induction of mRNA was followed by significant increases in GM-CSF protein in the supernatants. Increased cytokine production occurred by 2 h after infection, reached maximal concentrations by 24 h, and remained significantly elevated at 48 h. The most rapid rate of GM-CSF protein production occurred from 2 to 7 h.

View larger version (32K): [in this window] [in a new window]   Figure 1.   Time course of induction of steady-state mRNA levels and protein for GM-CSF from HRV-16 infected BEAS-2B cells. The upper panel shows a representative Northern blot for GM-CSF and the housekeeping gene GAPDH. The center panel shows the means plus SEM of the densitometric ratios of the two signals for four experiments. The lower panel shows the means plus SEM of protein levels produced for four experiments at each of the time points noted. Asterisks indicate significant increases in each parameter relative to the control (P < 0.05).

The induction of GM-CSF was a specific effect of HRV-16. Not only was this response induced using a purified virus preparation, but also the induction of GM-CSF was markedly abrogated by prior brief UV treatment of the purified virus. In each of two experiments, specific viral induction of GM-CSF (expressed as increase above the medium control) was reduced from 180 to 70 pg/ml (61% inhibition) and from 80 to 28 pg/ml (65% inhibition), respectively. Together, the time course of enhanced GM-CSF expression and the inability of UV inactivation to completely inhibit rhinovirus-induced production of this cytokine would be consistent with an early induction of GM-CSF occurring independently of viral replication, followed by a replication-dependent, more chronically maintained, phase of increased GM-CSF expression.

Because the HRV-16 viral stocks are propagated in WI-38 cells, it is possible that some unidentified factors derived from these cells may copurify with the virus and lead to virus-independent induction of cytokines. To confirm that GM-CSF induction was due to viral infection and not to residual factors from the WI-38 cells, two additional modifications of the virus infection protocol were used. First, the purified HRV-16 viral stock was subjected to ultracentrifugation through a Centricon membrane with 30 kD molecular weight cut-off. The filtrate, which should contain the same concentration of most of the potential cytokine contaminants as the original preparation, and the concentrated retentate, which should contain the virus, were tested for their ability to generate GM-CSF and compared with the original virus preparation. The original viral preparation released 95 pg/ml of GM-CSF, whereas the concentrated retentate released 210 pg/ml. The filtrate did not release any detectable GM-CSF. Similar results were obtained for virally induced IL-8 production (26). These results strongly suggest that the effects of the original stock virus preparation are specific to the virus. We also examined the effect of preincubation of BEAS-2B cells (n = 1) and primary human adenoid epithelial cells (n = 2) with a blocking mAb to ICAM-1, the cell-surface receptor for HRV-16. There was a dose-dependent inhibition of cytokine generation relative to the amount of virus used for infection. At an infectious dose of 10⁴ TCID₅₀ GM-CSF production was inhibited by 100%, and at an infectious dose of 3 × 10⁴ TCID₅₀ GM-CSF production was inhibited by an average of 58% (range 35 to 100%). These data demonstrate that blocking the cell-surface receptor for HRV-16 results in inhibition of the production of GM-CSF, and suggest that the induction of cytokine generation is related to the viral load. The inability to fully inhibit cytokine production at the higher infectious dose is consistent with our earlier observation that very little virus is necessary to begin a productive infection (11). Overall, these experiments demonstrate that the induction of GM-CSF by purified HRV-16 is virus-specific.

Effect of a NO Donor on HRV-16-Induced GM-CSF Production

To examine the effects of the NO donor NONOate on GM-CSF production, comparisons were made in BEAS-2B cells and in primary human bronchial epithelial cells infected with HRV-16 in the presence or absence of NONOate. As shown in Figure 2, NONOate (1 mM) significantly inhibited GM-CSF production by BEAS-2B cells at both 4 and 24 h after infection. A range of NONOate concentrations (100, 300, and 500 µM, and 1 mM) were tested (not shown). Complete inhibition of GM-CSF production was observed at 4 h with concentrations as low as 300 µM, whereas approximately 50% inhibition was observed at 24 h with 500 µM NONOate. The levels of cytokine generated were inhibited more at 4 h than at 24 h presumably due to the waning levels of NO at 24 h. As shown in Figure 3, NONOate also inhibited virally induced GM-CSF production by primary human bronchial epithelial cells. The effects of NONOate on absolute levels of GM-CSF produced by primary cells did not reach statistical significance as assessed by Student's t test due to the variability in GM-CSF levels produced by each cell preparation. In each of four experiments, however, the production of GM-CSF from NONOate-treated primary cells was lower than that from control infected cells (P < 0.05 for paired comparison of normalized data).

View larger version (26K): [in this window] [in a new window]   Figure 2.   Effects of NONOate (1 mM) on GM-CSF production from HRV-16-infected BEAS-2B cells at 4 and 24 h. Data represent the means plus SEM of values from four experiments. The asterisks indicate significant inhibition compared with levels produced at the same time after infection in the absence of NONOate (P < 0.05 in each case).

View larger version (27K): [in this window] [in a new window]   Figure 3.   Effects of NONOate (1 mM) on GM-CSF production from HRV-16-infected primary human bronchial epithelial cells at 4 and 24 h. Data represent the means plus SEM of values from four experiments. The asterisks indicate significant inhibition compared with levels produced at the same time after infection in the absence of NONOate (P < 0.05 in each case).

As shown in Figures 2 and 3, the absolute levels of GM-CSF produced by primary human bronchial epithelial cells were about 8- to 10-fold higher than the levels of GM-CSF produced by BEAS-2B cells. To further investigate the effects of virus and NONOate on primary human respiratory epithelial cells, we isolated epithelial cells from human adenoid tissue. These cells not only provided another source of primary human epithelial cells but also were derived from the upper airway epithelium, the predominant site of HRV infection. In each of three experiments, the production of GM-CSF at 24 h was significantly increased in HRV-16-infected cells (control = 120 ± 39 pg/ml; HRV-16 = 158 ± 30 pg/ml; P < 0.05). The addition of NONOate (500 µM) inhibited the production of GM-CSF in all three experiments (115 ± 11 pg/ml). NONOate (500 µM) had no significant effects on basal cytokine production in BEAS-2B cells (n = 4) or primary human adenoid epithelial cells (n = 2).

Effect of a NO Donor on GM-CSF mRNA

Comparisons of GM-CSF mRNA expression from HRV-16-infected BEAS-2B cells in the presence or absence of the NO donor NONOate (1 mM) were made at 1 and 4 h. This concentration of NONOate was selected due to its maximal effect on inhibition of GM-CSF protein production. Our previous studies have shown that this concentration has no effect on cell viability at these time points (18). As shown in Figure 4, NONOate significantly inhibited GM-CSF mRNA expression in HRV-16-infected BEAS-2B cells at both times examined.

View larger version (53K): [in this window] [in a new window]   Figure 4.   Effects of NONOate (1 mM) on steady-state mRNA levels for GM-CSF at 1 and 4 h after infection with HRV-16. The upper panel shows a representative Northern blot for GM-CSF and the housekeeping gene GAPDH. The lower panel shows the means plus SEM of the densitometric ratios of the two signals for three experiments at each time point. The asterisks indicate significant inhibition by NONOate compared with levels produced by virus alone (P < 0.05).

Comparisons of GM-CSF mRNA expression from HRV-16 infected primary human adenoid epithelial cells were made at 1 h after infection using real-time RT-PCR. In the absence of NONOate there was an average 6.5-fold induction (range, 1.8 to 14.3) in GM-CSF mRNA. In the presence of NONOate (500 µM) there was a 75 ± 10% (range, 58 to 94%) inhibition of the virus-induced GM-CSF mRNA.

Stability of GM-CSF mRNA

To determine whether the inhibitory effects of NONOate were due to decreased mRNA stability, experiments were performed to determine the decay rate of GM-CSF mRNA in the presence and absence of NONOate. In the absence of NONOate the half-life of GM-CSF mRNA was less than 1 h, whereas in the presence of NONOate the half-life was increased to approximately 1.5 to 2 h, suggesting that the inhibition of GM-CSF mRNA by NONOate was not due to a decreased stability of the mRNA (Figure 5).

View larger version (42K): [in this window] [in a new window]   Figure 5.   Stability of GM-CSF mRNA in HRV-16-infected BEAS-2B cells in the presence and absence of NONOate. Infected cells were incubated for 1 h after virus infection and actinomycin D (10 µg/ml) was added to the infected cells. NONOate (1 mM) or vehicle control was also added at this point and RNA was extracted at the time points indicated. The figure is a representative Northern blot for GM-CSF and the housekeeping gene GAPDH for two experiments.

Effect of a NO Donor on HRV-16 Induction of NF-B

Previous studies have shown that the transcription factor NF-B plays a role in the induction of GM-CSF gene expression by proinflammatory cytokines (27). Given that NO has been reported to inhibit tumor necrosis factor (TNF)--induced NF-B activation in endothelial cells via induction and stabilization of IB (28), we determined whether NONOate may inhibit transcription of GM-CSF in HRV-16-infected primary human epithelial cells and BEAS-2B cells by inhibiting viral activation of NF-B. Viral induction of NF-B was maximal by 30 min after infection and declined to baseline within 3 h (26). As shown in Figure 6, however, although HRV-16 infection induced activation of NF-B within 30 min after infection, the formation of specific NF-B complexes was not altered in the presence of NONOate (500 µM) in primary or in BEAS-2B cells.

View larger version (69K): [in this window] [in a new window]   Figure 6.   Effect of the NO donor NONOate (500 µM) on NF-B activation in BEAS-2B epithelial cells (upper right panel) and primary human adenoid epithelial cells (upper left panel) infected with HRV-16. Nuclear protein extracts were obtained from BEAS-2B cells 30 min after rhinovirus infection or sham infection, and EMSA was performed using the 32P-radiolabeled oligonucleotide sequence corresponding to the NF-B binding sites from the promoters of the GM-CSF gene. Lower panel: Nuclear protein extracts were prepared from HRV-16-infected BEAS-2B cells and EMSAs were performed in the presence and absence of specific antibodies (1 µg) to p65, p50, and c-Rel Rel protein family members.

To determine the subunit composition of NF-B complexes involved in the GM-CSF promoter sequences in HRV-16-infected cells, nuclear extracts from BEAS-2B cells were incubated with antibodies to specific Rel proteins. Addition of anti-p65 or anti-p50 antibodies led to decreased intensities of the NF-B complex that were accompanied by the formation of supershift bands of higher molecular weights (Figure 6). By contrast, antibodies to c-Rel had no effect on the NF-B complex.

Viral Titers after HRV-16 Infection

Supernatants were collected at 24 h after infection and assessed for viral titers in the WI-38 cell cytotoxicity assay for HRV-16. The supernatants removed from HRV-16-infected cells in the absence of NONOate contained 10³ TCID₅₀ /ml virus. Viral content in the supernatant of infected BEAS-2B cells was significantly inhibited by 1 mM NONOate (10² TCID₅₀/ml, n = 3; P < 0.05). Viral content in the supernatant of primary bronchial epithelial cells was also inhibited by 1 mM NONOate (10³ TCID₅₀ /ml in the absence of NONOate, 10^2.1 TCID₅₀/ml in the presence of NONOate; P = 0.15).

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

In contrast to other respiratory viruses, such as influenza or respiratory syncytial virus, cytotoxic damage of infected epithelial cells does not appear to play a role in the pathogenesis of rhinovirus infections, inasmuch as cytotoxicity is not observed either in infected human epithelial cell cultures (11) or in the nasal mucosa of infected individuals (29). In light of this, emphasis has focused on the concept that induction or exacerbation of airway symptoms may result from the actions of proinflammatory mediators that are generated as a consequence of rhinovirus infection. Clearly, the generation of proinflammatory cytokines by infected airway epithelial cells could play an important role in the pathogenesis of upper respiratory viral infections, and of viral-induced exacerbations of diseases such as asthma or rhinosinusitis. One of the cytokines that can be generated from respiratory epithelial cells upon rhinovirus infection is GM-CSF (11). Although the role of GM-CSF in upper respiratory viral infections has not been fully defined, the ability of this glycosylated polypeptide to enhance the differentiation, survival, and activation status of several type of inflammatory cells suggests that it could contribute to enhanced airways inflammation during asthma exacerbations. Further support for this concept is provided by the demonstration that increased levels of GM-CSF are produced in asthmatic airways (30) and that GM-CSF represents the major cytokine enhancing eosinophil survival in asthmatic airways (16). Moreover, increased production of GM-CSF is associated with changes in airways function (31). In the present study we have further examined the molecular mechanisms regulating the production of GM-CSF in primary human epithelial cells and the BEAS-2B bronchial epithelial cell line upon rhinovirus infection.

Infection with HRV-16 led to a rapid accumulation of mRNA for GM-CSF with observed levels being maximal at 1 h after infection. Steady-state mRNA levels remained elevated, however, through 24 h. Consistent with the time course for mRNA production, GM-CSF protein was detected in the supernatants by 2 h and was significantly elevated by 7 h after infection. Protein production remained elevated through 48 h, reaching maximal levels by 24 h. The absolute quantity of GM-CSF in cell supernatants varied among the epithelial cells examined. Primary human bronchial epithelial cells infected with HRV-16 produced about 8- to 10-fold higher levels of GM-CSF in comparison with primary human adenoid epithelial cells or BEAS-2B cells. These quantitative differences may be due to a variety of factors, including variability among the tissue donors, the culture medium in which the cells were grown, host cell antiviral responses, or, in the case of the BEAS-2B cell line, passage number. Overall, however, the qualitative aspects of the virally induced cytokine generation were similar in all three cell types.

The profile for HRV-16 induction of GM-CSF is similar to that which we have previously reported for viral induction of IL-8 and IL-6 in terms of its rapid onset, but differs in its prolonged nature compared with the transient induction of IL-8 and IL-6 (18). Prolonged production of GM-CSF from epithelial cells has also been observed in response to other stimuli, including cytokines (21). The specificity of GM-CSF induction as a viral effect was evidenced by the fact that a purified virus preparation induced this response and that preincubation of the epithelial cells with a mAb to ICAM-1, the cell-surface receptor for HRV-16, inhibited GM-CSF production from infected cells. Cytokine generation was also markedly abrogated, although not completely inhibited, by prior UV treatment of the virus. These results are consistent with the concept that there may be two components to the signal transduction pathways leading to rhinovirus-induced GM-CSF expression in epithelial cells. We suggest that the rapid expression of the GM-CSF gene that peaks at 1 h after rhinovirus infection occurs via a pathway that is independent of viral replication. Precedent for this is provided from our earlier studies demonstrating that the rapid, transient production of IL-8 and IL-6 seen in infected cells is also a specific viral-dependent but replication-independent effect. In contrast to the profile seen for IL-8 and IL-6, however, expression of GM-CSF in response to viral infection also shows a more sustained and chronic phase that is likely to be dependent upon viral replication. The partial inhibition of GM-CSF production in response to UV-treated (replication-deficient) rhinovirus could be explained by the inhibition of this chronic phase of production.

We also examined the effects of NO on GM-CSF protein generation. Recent evidence has demonstrated that among its many pharmacologic effects, NO has potent antiviral properties. Because epithelial cells lose the expression of NOS when placed in culture, the NO donor NONOate was used for the studies. This donor releases NO in a time-dependent manner at physiologic pH. Our previous studies showed, for the first time, that NO inhibited both rhinovirus replication and rhinovirus-induced production of the cytokines IL-8 and IL-6 in human respiratory epithelial cells (18). Our new data extend these observations and show that NONOate also significantly inhibits HRV-16- induced GM-CSF generation from epithelial cells. The inhibitory effects were dose-dependent and were observed in both primary human epithelial cells and BEAS-2B cells infected with HRV-16. As with our previous findings for IL-8 and IL-6, the degree of GM-CSF protein inhibition was more pronounced at 4 h than at 24 h after infection, consistent with the ability of NONOate to cause inhibition being dependent upon the release of sufficient amounts of NO.

To further examine the mechanisms of NONOate inhibition of GM-CSF production, the effects of NO on virus-induced expression of mRNA were studied. NONOate significantly inhibited steady-state cytokine mRNA expression at 1 and 4 h after infection. These results were in marked contrast to those observed in our earlier studies on HRV-16-induced production of IL-8 and IL-6, where NONOate failed to inhibit HRV-16-induced cytokine mRNA expression at any time point examined but significantly inhibited IL-8 and IL-6 protein production, implying that NO was acting via a post-transcriptional mechanism. Clearly, the capacity of NONOate to inhibit steady-state mRNA expression for GM-CSF in HRV-16-infected cells could occur due to inhibition of transcription and/or as a result of decreased stability of mRNA. It is well established that AU-rich sequences in the 3' untranslated region of the GM-CSF gene can tightly regulate mRNA stability (32, 33), and a variety of stimuli that increase mRNA levels of GM-CSF, including activators of protein kinase C and cytokines, such as IL-1, do so, at least in part, by enhancing mRNA stability (33). We examined, therefore, whether the capacity of NONOate to inhibit HRV-16-induced increases in levels of mRNA for GM-CSF could be due to a capacity to decrease mRNA stability. By contrast, our data clearly showed that NONOate actually increased the stability of GM-CSF mRNA, demonstrating that the inhibition of HRV-16 induction of steady-state mRNA by NONOate was not due to decreases in mRNA stability but was most likely due to effects on transcription.

Previous studies have shown that HRV infection of epithelial cells leads to a rapid and transient activation of the transcription factor NF-B (34). Binding sites for several transcription factors, including NF-B, have been identified in the proximal promoter region upstream of the transcription start-point in the human GM-CSF gene (27). Because NO has been reported to inhibit the activation of NF-B in endothelial cells by increasing the expression and nuclear translocation of IB (35), we tested the hypothesis that the inhibitory effect of the NO donor NONOate on GM-CSF mRNA production in epithelial cells was due to an inhibition of HRV-16-induced activation of NF-B. In contrast to the reported effects in endothelial cells, NO did not inhibit activation of NF-B in HRV-16-infected epithelial cells. These results imply that the effects of NO on NF-B may be cell- or stimulus-specific, and that NO may be inhibiting other transcription factors that are important in regulating GM-CSF mRNA synthesis (27).

In summary, our data confirm that there is a rapid and sustained induction of GM-CSF production from human bronchial and adenoid epithelial cells infected with HRV-16. In contrast to previous observations for virally induced epithelial cell production of IL-8 and IL-6, NO markedly inhibited virally induced GM-CSF expression by reducing mRNA levels in these cells. These effects were not due to decreased mRNA stability or the inhibition of the transcription factor NF-B. Although additional studies will be necessary to elucidate the mechanisms by which NO inhibits GM-CSF production, our data indicate that the inhibition involves suppression of transcription. Thus, in addition to inhibiting viral replication during rhinovirus infections, NO can reduce the generation of proinflammatory cytokines by exerting effects at both the transcriptional and post-transcriptional levels, depending upon the cytokine in question. In light of these properties, NO is likely to be an important component of the host response to rhinovirus infection.