Human rhinovirus (HRV) infections trigger exacerbations of asthma and chronic obstructive pulmonary disease (COPD) and are associated with lymphocytic infiltration of the airways. We demonstrate that infection of primary cultures of human airway epithelial cells, or of the BEAS-2B human bronchial epithelial cell line, with human rhinovirus type 16 (HRV-16) induces expression of CXCL10 [IFN-γ-inducible protein 10 (IP-10)], a ligand for the CXCR3 receptor found on activated type 1 T lymphocytes and natural killer cells. IP-10 mRNA reached maximal levels 24 h after HRV-16 infection then declined, whereas protein levels peaked 48 h after infection with no subsequent new synthesis. Cytosolic levels of AU-rich factor 1, a protein associated with mRNA destabilization, increased beginning 24 h after HRV-16 infection. Generation of IP-10 required virus capable of replication but was not dependent on prior induction of type 1 interferons. Transfection of synthetic double-stranded RNA into epithelial cells induced robust production of IP-10, whereas transfection of single-stranded RNA had no effect. Induction of IP-10 gene expression by HRV-16 depended upon activation of NF-κB, as well as other transcription factor recognition sequences further upstream in the IP-10 promoter. In vivo infection of human volunteers with HRV-16 strikingly increased IP-10 protein in nasal lavages during symptomatic colds. Levels of IP-10 correlated with symptom severity, viral titer, and numbers of lymphocytes in airway secretions. Thus IP-10 may play a role in the pathogenesis of HRV-induced colds and in HRV-induced exacerbations of COPD and asthma.
- interferon-γ-inducible protein 10
- asthma exacerbations
- innate immunity
- host defense
human rhinoviruses are not only the predominant cause of the common cold, the most prevalent acute respiratory illness in humans, but also have been implicated as the major viral pathogen associated with exacerbations of asthma and chronic obstructive pulmonary disease (COPD) (9, 33). The mechanisms by which human rhinovirus (HRV) induces exacerbations of these diseases, however, remain to be fully elucidated.
The respiratory epithelial cell is the principal site of HRV infections in the airway (4, 24), but such infections lead to no discernible changes in epithelial integrity or viability (34, 36), indicating that induction of symptoms is not due to direct cytotoxicity. Rather, it seems likely that host defense and inflammatory responses to HRV infection may contribute to symptoms. Experimental HRV infections lead to rapid increases in recruitment of neutrophils and lymphocytes to the airways of normal subjects (15), with increased eosinophilia also being seen in some asthmatic subjects, and it has been reported that the percentage of lymphocytes in recovered secretions correlates directly with airway symptoms (15). There is substantial evidence that HRV-infected epithelial cells can elaborate a range of proinflammatory and regulatory molecules that may contribute to the inflammatory and host defense responses to infection. These include pleiotropic cytokines, such as IL-1 and IL-6 that exert multiple actions in the airways (29, 34), and chemokines, such as IL-8 and regulated on activation, normal T cell expressed, and presumably secreted (RANTES) that could contribute to the recruitment and activation of selected inflammatory cell populations including neutrophils and eosinophils, respectively (29, 32, 34). Increased levels of these cytokines and chemokines also have been detected in airway secretions during in vivo infections (25, 35).
The current paradigm is that T lymphocytes recruited during viral infections are predominantly of the “type 1” phenotype, characterized by production of a pattern of cytokines that includes interferon (IFN)-γ and IL-2. (By contrast, type 2 T cells produce cytokines including IL-4, IL-5, and IL-13). Consistent with this paradigm, increased production of type 1 cytokines has been reported to be associated with more rapid viral clearance during HRV infections (11). It has been shown recently that type 1 and type 2 T cells express distinct subsets of chemokine receptors that may regulate selective recruitment of T cell subtypes to inflammatory sites. In particular, activated type 1 cells preferentially express the CXCR3 chemokine receptor (28). Although type 1 T lymphocytes play a central role in rhinovirus infections, we are aware of no studies to examine the production of selective chemoattractant stimuli that induce recruitment of this cell type during HRV infections.
Interferon-γ-inducible protein 10 (IP-10, CXCL10) is a member of the non-ELR CXC chemokine family that is produced by several cell types, particularly epithelial cells (31). This chemokine is a ligand for the CXCR3 receptor and serves as a selective chemoattractant for both activated type 1 T lymphocytes and natural killer cells. Increased expression of IP-10 not only has been observed in the airways of subjects with COPD (27), but also in patients with asthma (20), where levels correlate with lymphocyte numbers (16). Consistent with this latter observation, IP-10 has been shown to contribute both to airway inflammation and airway hyperresponsiveness in a murine model of asthma (18). The current studies were undertaken, therefore, to test the hypothesis that HRV infection of epithelial cells could induce expression of IP-10 in vitro and to begin to examine mechanisms that may be responsible for such an induction. Furthermore, we examined whether IP-10 levels are increased in airway secretions in vivo during experimental HRV infections and whether IP-10 levels relate to airway symptoms, viral titers, and numbers of infiltrating lymphocytes.
MATERIALS AND METHODS
The following reagents were purchased from the indicated suppliers: Eagle's minimal essential medium, Ham's F-12 medium, HBSS, l-glutamine, penicillin-streptomycin-amphotericin B, trace elements, and retinoic acid (Medicorp, Montreal, PQ, Canada); hydrocortisone, epithelial cell growth factor, and endothelial cell growth supplement (Becton Dickinson, Mississauga, Ontario, Canada); FBS, nonessential amino acids, sodium pyruvate, TRIzol reagent, transferrin, insulin, and dNTPs (Invitrogen, Burlington, Ontario, Canada); Bronchial epithelial cell growth medium (BEGM) (Bio Whittaker, Walkersville, MD); calpain inhibitor I (Calbiochem-Novabiochem, San Diego, CA); Taqman master mix, 20× GAPDH, RNase inhibitor, and reverse transcriptase (Applied Biosystems, Foster City, CA); Pfu polymerase (Stratagene, Foster City, CA); IP-10 antibodies, recombinant IP-10, recombinant IFN-β, antibody to the type 1 IFN receptor (CD118) and class-matched IgG2a antibody (R&D, Minneapolis, MN); the firefly luciferase reporter plasmid pGL3-basic, the renilla luciferase plasmid pRLV, genomic DNA, and the dual luciferase reporter assay system (Promega, Madison, WI). All other chemicals were purchased from Sigma Chemical (St. Louis, MO).
Viruses and cell lines.
The BEAS-2B cell line was a gift from Curtis Harris (National Cancer Institute, Bethesda, MD). Human rhinovirus type 16 (HRV-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 propagation in WI-38 cells and were purified by centrifugation through sucrose to remove ribosomes and soluble factors as previously described (12). Viral titers were determined using WI-38 cells as previously described, and the identity of HRV-16 was confirmed by antibody inhibition (29). Replication-deficient HRV-16 was produced by exposure of aliquots of the same purified stock used for live virus experiments for 5 min to a Spectroline model XX-15F high-intensity short-wavelength (254 nm) UV lamp at a distance of 5 cm. Inactivation was confirmed by the inability of stocks to replicate and cause cytopathic effects in WI-38 cells.
Epithelial cell culture.
Primary human airway epithelial cells were obtained by protease digestion of human adenoid tissue as previously described (30). Primary epithelial cells were grown on six-well culture plates (Costar, Cambridge, MA) in serum-free epithelial growth medium (BEGM). Primary cultures were established to be >98% epithelial cells using cytokeratin staining as previously described (6). Cells were cultured in BEGM from which hydrocortisone had been withdrawn for 18 h before the addition of HRV-16. Normal human bronchial epithelial cells were obtained from Clonetics (San Diego, CA) and were grown in BEGM. BEAS-2B cells were grown in serum-free culture medium consisting of Ham's F-12 nutrient medium with 10 additives (F-12/10×), as previously described (29). Additional Ham's F-12 was prepared with all of the above additives except hydrocortisone and is hereafter referred to as F-12/9×. For experimentation, cells between passages 45 and 65 were plated on six-well plates in F-12/10×. For 18 h before addition of virus, F-12/10× was replaced by F-12/9×. All epithelial cell cultures were incubated at 37°C in 5% CO2. All data presented are from primary airway epithelial cells, except data for luciferase transfections, which were performed in BEAS-2B cells.
Viral infection of epithelial cells.
HRV-16 was added to the cells at a concentration ranging from 103 to 3 × 104 50% tissue culture infective dose (TCID50) units/ml and incubated at 34°C in 5% CO2 for appropriate times. In some instances, cells were cultured in the presence of calpain inhibitor I (10 μM) for 4 h before and then during culture with HRV-16. At various time points, supernatants were harvested and RNA was extracted with TRIzol. Concentrations of RNA were calculated based on absorbance. Assay of lactate dehydrogenase activity (Promega) was performed to establish that drugs were not cytotoxic under these conditions. Experiments were also performed to test the effects of calpain inhibitor I on replication of HRV-16 in epithelial cells. In these experiments, cells were preincubated with calpain inhibitor I and exposed to HRV-16 for 1 h. Cells were washed to remove free virus, and calpain inhibitor I was again added to the cells. Supernatants were collected at appropriate times and viral titer was assayed using WI-38 cells. The ability of calpain inhibitor I to inhibit translocation of NF-κB to the nucleus was performed using an electrophoretic mobility shift assay (EMSA) as previously described (14).
Transfection of epithelial cells with double-stranded or single-stranded RNA.
The synthetic double-stranded RNA (dsRNA) polyinosinic-polycytodylic acid (poly I-C) or single-stranded (ss) RNA (poly U) (0.1 μg) was transfected into primary cells in basal culture medium containing no additives (BEBM, Bio Whittaker) using Fugene 6 (Roche, Laval, PQ, Canada) as previously described (10). Concurrently, 0.1 μg of dsRNA was added to cells in BEBM in the absence of Fugene 6. After 4 h, cells were washed three times with HBSS and then cultured in BEGM without hydrocortisone for 48 h. Supernatants were collected for ELISA analysis.
Type 1 interferon receptor (CD118) blockade.
Monolayers of primary airway epithelial cells were preincubated for 1 h with 5 μg/ml of anti-CD118 or class-matched control antibody and then exposed to 104 TCID50/ml of HRV-16 or 3 ng/ml of recombinant IFN-β. After 24 h, supernatant was removed and cellular RNA was extracted.
Culture supernatants were assayed for IP-10 protein as per manufacturer's protocol (R&D) at room temperature. Sensitivity of the assay was 30 pg/ml. A commercial ELISA that detects 12 of the known isoforms of IFN-α (R&D) with a sensitivity of 12.5 pg/ml was used to measure levels of this cytokine. IFN-β was assayed with an ELISA sensitive to 8.3 pg/ml (Biosource, Montreal, PQ, Canada).
IP-10 gene expression analysis was performed using the Applied Biosystems model 7900 Sequence Detector. Input RNA (400 ng) was reverse transcribed into cDNA, followed by PCR amplification in the presence of specific forward 5′GAAATTATTCC TGCAAGCCAATTT-3′ and reverse 5′-TCACCCTTCTTTTTCATTGTAGCA-3′ primers (University of Calgary DNA Services, Calgary, Alberta, Canada) and probe 5′-FAM-TCCACGTGTTGAGATCA-MGB-3′ (Applied Biosystems, Foster City, CA). Primers and probe for detection of IFN-α gene expression were designed to recognize a conserved sequence that would permit detection of 11 IFN-α isoforms. The forward primer sequence was 5′AGAATCACT CTCTATCTGAAAGAGAAGAAATA-3′; reverse primer 5′TCATGATTTCTGCTCTGACAACCT-3′ and the probe sequence was 5′-FAM-AGCCCTTGTGCCTGG-MGB-3′. Expression of IFN-β and IFN-γ was performed using primer and probe kits available from Applied Biosystems. Tristetraprolin (TTP) gene expression was assayed using specific forward 5′-GGGAATCCTGGTGCTCAAATT-3′ and reverse 5′-GGGTTTGGCAACGGCTTT-3′ primers and probe 5′-FAM-CCTCCAAAAGCAAGTGA-MGB-3′. Expression of the housekeeping gene GAPDH was also assayed using reagents obtained from Applied Biosystems. Efficiency curves were performed for each gene of interest, relative to the housekeeping gene and data were calculated as fold increase over control as described previously (30).
Preparation of promoter constructs.
A 972-bp IP-10 promoter construct, corresponding to the sequence from −875 to +97 (relative to the transcriptional start site) of the 5′-flanking region of the human IP-10 gene, was generated from human genomic DNA using forward 5′-GCGTAGGTACCTAGAACCCCATCGTAAATC-3′ and reverse 5′-GCGTAGCTAGCTAGCAGCAAATCAGAATGG-3′ primers incorporating KpnI and NheI restriction sites at the 5′- and 3′-ends, respectively. These restriction sites were used to insert the resulting amplicon upstream of a luciferase reporter in the pGL3-basic plasmid using the Rapid DNA Ligation Kit (Roche). This 972-bp construct has previously been reported to correspond to the full-length human IP-10 promoter (17). The sequence of the construct was determined by the University of Calgary DNA sequencing facility and was found to be identical to that reported for IP-10 as part of the sequence of human chromosome 4 (accession no. AC112719). Potential transcription factor binding sites were identified using the Genomatix MatInspector program. A 376-bp truncated IP-10 promoter construct (sequence from −279 to +97 of the 5′-flanking region of the human IP-10 gene) was similarly generated except the 972-bp IP-10 promoter construct was used as a template and the forward 5′-GCGTAGGTACCTAGAGAATGGATTGCAACC-3′ primer with a KpnI restriction site was employed. Point mutations in two putative NF-κB sites, NF-κB1 and NF-κB2, were generated in the full-length construct using standard site-directed mutagenesis procedures. Boldface, lowercase characters in the following sequences denote mutation sites. For NF-κB1, a mutant reverse primer 5′-GCAACATGtGACTTCaCCAGG-3′was annealed in combination with the previously described forward primer and subjected to PCR amplification. Concurrently, a mutant forward primer 5′-CCTGGtGAAGTCaCATGTTGC-3′ was annealed in combination with the previously described reverse primer and PCR amplification was, again, performed. The resulting two amplicons were gel purified, combined, and subjected to PCR amplification. The resulting amplicon was gel purified, digested with KpnI and NheI, and ligated into pGL3. A point mutant construct for NF-κB2 was generated in the same fashion utilizing a mutant forward primer 5′-GCAGAGtGAAATTaCGTAACTTGG-3′ and a mutant reverse primer 5′-CCAAGTTACGtAATTTCaCTC TGC-3′. Successful generation of all constructs was confirmed by sequencing.
Luciferase construct transfection.
Each promoter construct was cotransfected with the pRLV plasmid into subconfluent (50–60%) monolayers of BEAS-2B cells using Fugene 6 in Ham's F-12 basic medium as per manufacturer's protocol. After 5 h of transfection, cells were washed and allowed to recover overnight in fresh F-12/9× medium. Cells were then washed and infected with 3 × 104 TCID50 units of HRV-16 in F-12/9×. After a 24-h incubation, firefly luciferase activity was measured in each sample and normalized relative to renilla luciferase activity. Data were expressed as fold induction of HRV-treated cells over cells cultured with medium alone.
Western blotting for AU-rich factor 1 or TTP.
Primary epithelial cells were grown on p100 culture plates until ∼70% confluent. Cells were then exposed to HRV-16 or medium control. After 12, 24, and 48 h of culture, supernatant was removed and monolayers were scraped in phosphate-buffered saline and centrifuged. The cell pellet was then dissolved in 0.5% Nonidet P-40 for 5 min on ice then centrifuged to pellet intact nuclei and collect the soluble cytosolic fraction. Protein concentrations were quantified using a DC Protein Assay (Bio-Rad, Montreal, PQ, Canada). Equivalent amounts of cytosolic protein (7.5 μg per lane) were then separated by SDS-PAGE and then electrotransferred to a polyvinyl difluoride membrane. Membranes were blocked with skim milk and incubated with 1:750 dilution of a rabbit anti-AU-rich factor 1 (AUF1) antibody (Upstate Biotechnology, Lake Placid, NY) overnight at 4°C. For TTP, membranes were exposed either to a 1:200 dilution of a commercially available mouse antibody to TTP (Santa Cruz Biotechnology, Santa Cruz, CA) or to 1:200 dilution of a rabbit antibody to TTP that was kindly provided by Dr. Keith Blackwell (Harvard Medical School, Boston, MA). Membranes were washed and then incubated for 1 h with a 1:7,500 dilution of peroxidase-conjugated anti-rabbit (or anti-mouse) immunoglobulin. Proteins were visualized with ECL substrate reagent (Amersham, Piscataway, NJ). Equal loading was confirmed by Coomassie staining of blots and comparison of specific band intensities.
Experimental rhinovirus infection.
The protocol was approved by the Institutional Review Board of the Johns Hopkins Bayview Medical Center, and informed consent was obtained from all subjects. Six healthy adults (five male and one female) who had no detectable serum neutralizing antibody to HRV-16 or any history of cold symptoms within the previous 6 wk were recruited. Subjects were required to come to the laboratory on 5 consecutive days (days 0–4) and again on day 7. After baseline nasal lavage, viral challenges were done by intranasal administration of a previously described safety-tested inoculum of rhinovirus type 16 (30). The virus stock was diluted 1:20 in HBSS containing 0.5% gelatin. Volunteers were inoculated twice intranasally during a 1-h period with 0.250 ml/nostril delivered by pipette and three sprays per nostril (70 μl/spray) from a pump spray bottle giving a total dose of 1,000 TCID50 HRV-16.
Nasal lavages were done with 10 ml of lactated Ringer solution in the morning on days 0–4 and day 7. The volume of recovered lavage fluids was recorded, and aliquots were taken for cell enumeration using a hemocytometer and cell differential counts using Diff-Quik (Baxter, Mississauga, Ontario, Canada)-stained cytocentrifuge preparations (Cytospin; Shandon Southern Instruments, Sewickley, PA). Lymphocyte counts were performed by two different observers, both of whom were blinded to the samples being examined. A minimum of 200 cells were counted for each slide, and mean data were presented. Nasal lavage fluids were divided into aliquots of 1-ml volume and stored at −80°C for subsequent assay of IP-10 and of viral titers by bioassay in WI-38 cells. Lavages were made 0.1% BSA before analysis. Infection was established by demonstrating an increase in serum neutralizing antibody to HRV-16 of at least fourfold and/or by the presence of HRV-16 in recovered secretions. The identity of HRV-16 was confirmed by neutralization with specific antibody.
Subjects were given a symptom questionnaire to fill out each morning and evening of all study days regarding the following eight symptoms: sneezing, stuffy nose, runny nose, sore throat, cough, headache, chills/fever, and malaise. Symptoms were rated for severity as follows: 0, none; 1, mild; 2, moderate; and 3, severe. As previously reported, criteria for a cold required a total symptom score of ≥5 over each of 4 days after challenge and/or the belief of the subject that a cold had occurred (25, 30).
For normally distributed data, appropriate one-way or repeated-measures ANOVA were used to assess significant differences, with post hoc analysis using Fisher's least-significant-difference tests. Alternatively, paired t-tests were used. For data that were not normally distributed, analysis was performed using Kruskal-Wallis or Friedman ANOVA, followed by post hoc analysis using Wilcoxon matched-pairs signed-ranks test. Correlations were assessed using linear regression or Spearman rank analysis as appropriate. For all statistical tests, a P value of ≤ 0.05 was assumed to be significant.
HRV induces IP-10 mRNA and protein release in a time- and dose-dependent fashion.
Real-time PCR analysis demonstrated that HRV-16 infection of primary adenoid epithelial cells caused a significant (P < 0.01), time-dependent induction of mRNA expression for IP-10. Modest increases were detected at 6 h after exposure to 104 TCID50/ml of HRV-16 but reached a maximum at 24 h (40.4 ± 28-fold) and then declined to low levels at 48 h (Fig. 1A). Although it is not possible to achieve statistical significance by post hoc Wilcoxon analysis with n = 4, increases in mRNA relative to medium control were seen in all four experiments at each time point. Assessment of viral titers at these same points was also performed (Fig. 1B). Although virus was not detected at 6 h, significant elevations were detected at both 24 h and 48 h postinfection. Levels of shed virus detected at each time point were similar to those we have previously reported (30). ELISA analysis of cumulative levels of IP-10 protein in culture supernatants from HRV-16-infected cells also demonstrated a significant (P < 0.005) time-dependent increase compared with medium control. Consistent with the time course of mRNA production, virally induced protein release was not detected at 6 h postinfection but was markedly increased by 24 h. A maximal level of 606 ± 270 pg/ml of IP-10 protein (compared to control levels of 59 ± 17 pg/ml) was observed at 48 h postinfection. No further increase in secretion was seen at 72 h (Fig. 1C).
Viral induction of IP-10 protein was not restricted to upper airway epithelial cells. In four experiments using cells from different donors, bronchial epithelial cells produced 119 ± 34 pg/ml in response to medium alone and 15,616 ± 5,100 pg/ml in response to HRV-16 infection (P < 0.05, data not shown).
We also examined whether HRV-16-induced IP-10 protein release was dependent on the initial dose of HRV-16 to which epithelial cells were exposed. After 48 h of exposure, primary epithelial cells released 36 ± 4 pg/ml of IP-10 in response to medium, 111 ± 29 pg/ml in response to 103 TCID50/ml of HRV-16, 430 ± 229 pg/ml to 3 × 103 TCID50/ml, and 852 ± 310 pg/ml to 104 TCID50/ml (n = 4). Analysis of these data by one-way ANOVA confirmed a significant (P < 0.05) dependence on virus dose.
Induction of IP-10 responses requires HRV-16 capable of replication.
To determine whether induction of IP-10 was dependent on viral replication, we first compared responses to HRV-16 that had been rendered incapable of replication by brief exposure to UV light, with those of fully functional virus. Measurement of IP-10 protein 48 h after viral exposure of primary human epithelial cells revealed striking differences between the two viral preparations. As expected, intact HRV-16 induced a significant (P < 0.01) increase in IP-10 production compared with medium control. By contrast, IP-10 levels from cells exposed to UV-treated HRV-16 were significantly (P < 0.01) lower than those seen with intact virus and, indeed, were not statistically different than those seen with medium control (Fig. 2). Consistent with these data, IP-10 mRNA levels measured at 24 h after infection were increased by 32.5 ± 16.2-fold compared with medium control in the presence of replicating HRV-16 but by only 2.7 ± 0.9-fold in cells exposed to UV-treated virus (n = 4, data not shown).
dsRNA is not present in normal eukaryotic cells but is generated during the replication of HRV. It has been reported that dsRNA is an important trigger of host antiviral responses and can induce the production of several cytokines (13). To evaluate the potential role of intracellular dsRNA in HRV induced IP-10 production from epithelial cells, we examined IP-10 production from primary epithelial cells transfected with synthetic dsRNA. Transfection of ssRNA was an ineffective stimulus for IP-10 production (Fig. 3). In some experiments, IP-10 generation was observed when dsRNA was simply applied to the cell surface, but when data from six experiments were analyzed this did not achieve statistical significance (P = 0.15). By contrast, transfection of dsRNA led to strikingly increased IP-10 production, with mean levels exceeding 10,000 pg/ml. IP-10 production in cells transfected with dsRNA was significantly increased compared with each of the other treatment groups (P < 0.05 in each case).
HRV-16 effects on IP-10 promoter activity.
To begin to define transcriptional regulation factors that play a role in HRV-16-induced expression of IP-10 in epithelial cells, we performed studies using several IP-10 promoter-luciferase constructs (Fig. 4) transfected into the BEAS-2B human airway epithelial cell line. Preliminary studies confirmed that the BEAS-2B cell line was capable of producing IP-10 within 24 h after HRV infection [31 ± 0 pg/ml (limit of detection) in response to medium vs. 13,350 ± 2,930 pg/ ml in response to HRV-16 infection; n = 5, P < 0.02], indicating that this cell line was a reasonable model for primary cells. In the first series of experiments, we compared responses of a 972-bp promoter construct (IP-10GL3) with those of a truncated 376-bp construct that no longer contained two STAT, two interferon-stimulated responsive element (ISRE), and two CCAAT/enhancer-binding protein-β binding sites but retained two proximal sites for NF-κB (tIP-10GL3). HRV-16 induced an average 6.6-fold induction of activity in the full-length promoter, but there was a significant (P < 0.05) reduction to ∼50% of this response with the truncated mutant (Fig. 5A).
Because it has been shown previously that HRV infection of airway epithelial cells induces activation of NF-κB (14, 29), and this transcription factor has been implicated in the expression of IP-10 in response to other stimuli (17, 37), we also examined the effects of HRV-16 infection on activation of variants of the 972-bp construct in which each of the two proximal NF-κB recognition sites were mutated. Compared with the native promoter, mutation of either of the NF-κB sites led to a significant (P < 0.0001 in each case) reduction in promoter activity (Fig. 5B). Interestingly, mutation of the more proximal NF-κB binding site (κB1) led to a significantly lower response than mutation of the more distal κB2 site (P < 0.001).
Inhibitor of NF-κB activation blocks endogenous IP-10 mRNA expression.
To confirm the importance of NF-κB activation for endogenous gene expression, we investigated the effects of calpain inhibitor I, which blocks proteasomal degradation of IκB, on HRV-16-induced IP-10 mRNA expression. We confirmed that calpain inhibitor I blocked nuclear translocation of NF-κB in HRV-16-infected cells using EMSA (Fig. 6A). In each of four experiments, the inhibitor reduced (P < 0.02) HRV-16-induced expression of IP-10 mRNA (Fig. 6B). This was not due to an effect of the drug in viral replication, as there was no significant effect of calpain inhibitor I on viral titers in infected cells (Fig. 6C).
Induction of IP-10 by HRV-16 is not dependent on production of interferons.
HRV-16 induced expression of mRNA for IP-10 decreases markedly from 24 to 48 h postinfection (Fig. 1). Both type 1 and type 2 interferons are known to be able to induce IP-10 expression and both have AU-rich elements (AREs) in the 3′-untranslated regions (UTRs) of their mRNA that can mediate rapid decay. We examined, therefore, whether HRV-16-induced expression of IP-10 was dependent on interferon induction. Primary cultures of epithelial cells expressed no detectable mRNA for IFN-γ either under basal conditions or at any time point (3, 6, 12, 24 h) postinfection with HRV-16. In four experiments, there was no significant increase in gene expression for IFN-α as assessed by Kruskal-Wallis ANOVA. Maximum levels of IFN-α mRNA were seen at 12 h after HRV-16 infection (3 ± 0.6-fold) compared with control (Fig. 7). No IFN-α protein was detected in supernatants from HRV-16-infected cells. HRV-16 infection led to a significant (P = 0.025) time-dependent, but transient, induction of mRNA for IFN-β that was maximal at 6 h postinfection (Fig. 7). Three experiments adding exogenous IFN-β to the primary epithelial cells indicated that at least 300 pg/ml of IFN-β were needed to induce the production of comparable levels of IP-10 protein induced by HRV-16 infection in the same cell preparations (data not shown). Surprisingly, however, no IFN-β protein could be detected either in supernatants, or from whole cell lysates, of infected cells using an assay capable of detecting 8.3 pg/ml of IFN-β protein. To further evaluate the potential role of type 1 interferons, we compared the effects of an antibody to the type 1 interferon receptor (CD118) on IP-10 gene expression induced by HRV-16 infection or by exposure to IFN-β (Fig. 8). HRV-16 induced a 48 ± 28-fold induction of IP-10 mRNA compared with untreated cells, and this was not significantly affected in the presence of anti-CD118. By contrast, IP-10 mRNA induction by IFN-β was strikingly and significantly (P < 0.05) inhibited by anti-CD118. Class-matched control antibodies did not affect IP-10 gene expression by either stimulus. Together, these data indicate that HRV-16 induction of type 1 interferons plays, at best, a minimal role in IP-10 induction in virally infected primary epithelial cells.
Viral induction of transcriptional destabilizers.
The 3′-UTR of IP-10 mRNA contains a type 1 ARE. Destabilization of mRNAs that contain such elements occurs in the cytosol upon binding of specific proteins, such as TTP or AUF1 (5, 8). We therefore examined the effects of HRV-16 infection of primary airway epithelial cells on cytosolic expression of each of these destabilizers. Because TTP is itself partially transcriptionally regulated (26), we examined levels of TTP mRNA at varying time points after HRV-16 infection. Although basal expression of TTP mRNA was readily detectable in each of three experiments, HRV-16 infection induced no increase above these basal levels at any time point (data not shown). We also used antibodies to examine the cytosolic expression of TTP protein. Under the conditions used, we were not able to detect TTP in cytosolic extracts with either of the antibodies available to us. By contrast, under the same conditions, we were able to readily detect expression of AUF1. Moreover, whereas infection of primary epithelial cells with HRV-16 did not increase cytosolic expression of AUF1 at 12 h postinfection, when IP-10 mRNA was still approaching maximum levels, clear increases in cytosolic levels of AUF1 were seen beginning at 24 h and persisting 48 h after infection (Fig. 9). Thus AUF1 induction occurred in a time frame that paralleled that for loss of IP-10 mRNA.
IP-10 levels are increased in airway secretions during symptomatic HRV-16 infections in vivo.
All six subjects became infected, as evidenced both by recovery of shed virus from nasal secretions and by appropriate increases in serum neutralizing antibody, and developed symptomatic colds. Infection induced significant increases in symptom scores (P < 0.001 by Kruskal-Wallis). Compared with preinfection, symptom scores were significantly increased on each day postinfection. Viral titers in nasal secretions were significantly increased on days 1, 2, 3, and 4 postinfection (P < 0.05 in each case). Only one of the six subjects had detectable levels of IP-10 in baseline nasal lavages. Significant increases (P < 0.001) in IP-10 levels were seen, however, in all subjects after infection. Mean levels of IP-10 increased to 720 pg/ml on day 1 after infection before peaking at ∼60,000 pg/ml on day 2 (Fig. 10A). Compared with day 0, IP-10 levels were significantly increased on each of days 2-7 (P < 0.05 in each case). Moreover, levels of IP-10 in nasal lavages from all subjects at every time point correlated with symptom scores (rho = 0.545, P < 0.005). In agreement with previous reports, symptomatic colds were also associated with significantly increased numbers of lymphocytes in nasal secretions on days 2, 3, and 4 (P < 0.05 in each case) (Fig. 10B). In support of the hypothesis that IP-10 plays a role in lymphocyte recruitment there was a significant correlation between levels of IP-10 and lymphocyte numbers using all samples from this subject population (rho = 0.63, P = 0.0002). Finally, we examined the relationship in all samples from all subjects between IP-10 production and viral titer, each expressed in log units (Fig. 10C). Consistent with the observation that IP-10 generation in vitro was dependent on viral replication, there was a significant correlation between log IP-10 and log viral titers in vivo (r = 0.68, P < 0.0001).
Although the mechanisms underlying the manifestation of symptoms in response to HRV infections remain to be delineated, it is clear that symptomatic infections are associated with elevated levels of several mediators and proteins in airway secretions and with increased inflammatory cell recruitment. The epithelial cell is the primary site of HRV infection in the airway, and growing evidence suggests that viral modulation of epithelial function may play an important role in the host inflammatory and immune responses. The current studies provide further support for this concept by providing the first demonstration that HRV infection of both upper and lower airway epithelial cells induces production of IP-10, a chemokine that could be responsible for the selective recruitment of lymphocytes of the type 1 subtype to the airway mucosa.
Induction of IP-10 was clearly dependent upon the initial dose of HRV-16 to which cells were exposed, presumably reflecting increasing numbers of infected cells with increasing doses. In contrast to chemokines such as IL-8, which are induced rapidly after viral exposure (29), induction of IP-10 mRNA was not observed until at least 6 h after infection, with protein secretion not being detected until 24 h. Although mRNA induction showed significant dependence on time postinfection by Kruskal-Wallis ANOVA, achieving significance on suitable post hoc analysis is difficult due to substantial variation of the absolute magnitude of responses of primary cultures derived from individual donors, and the need for large n values for each experiment if nonparametric analysis must be used. Nonetheless, the pattern of response seen in each experiment was identical. Given that the replication cycle of Picornaviridae, such as HRV, is usually in the range of 8–10 h, it seemed likely that IP-10 gene induction may be dependent upon events occurring during the replication cycle. Consistent with this concept, increased expression of IP-10 was associated with increases in viral titers, and HRV-16 that had been rendered replication deficient by brief exposure to UV light did not induce IP-10 mRNA or protein expression. Although it has been shown that infected epithelial cells support ongoing replication of HRV for several days (34), IP-10 gene expression peaked at ∼24 h after infection with a sharp decline in mRNA levels by 48 h. These data differ somewhat from those seen in a murine glioblastoma cell line infected with measles virus, where mRNA levels for mouse IP-10 were sustained up to 32 h (22), but are similar to data observed in human epithelial cells stimulated with cytokines, where mRNA levels began to decline after 16 h (31).
We considered two potential mechanisms that may explain the transient elevations of IP-10 gene expression in HRV-16-infected epithelial cells. The first was that HRV-16-induced expression of IP-10 may be mediated by virally induced interferon production. Both type 1 and type 2 interferons are known to be inducers of IP-10, and both have mRNA with destabilizing AREs in their 3′-UTR (3). A role for IFN-γ could be ruled out as we could not detect any gene expression for IFN-γ in primary epithelial cells at any time after HRV-16 infection. In terms of type 1 interferons, a modest induction of IFN-α gene expression was detected at 12 h postinfection. There was also a more robust increase in IFN-β gene expression observed at 6 h postinfection. However, no IFN-α or IFN-β protein was detected at any time point in supernatants or lysates of HRV-16-infected epithelial cells. Given the lack of measurable IFN-α or IFN-β and the lack of a significant effect of antibody to the type 1 interferon receptor on HRV-16-induced IP-10 gene expression compared with the marked inhibition by the same antibody of IFN-β-induced IP-10 expression, our data do not support a substantial role of interferons in HRV-16-induced IP-10 expression. These observations are consistent with several reports in which a variety of viruses are also able to induce interferon-responsive genes, including IP-10, independently of interferon induction, by direct viral activation of steps in the interferon signal transduction pathway (2, 7, 21, 26).
As an alternative mechanism to account for transient HRV-16-induced IP-10 gene expression, we determined whether HRV-16 infection of epithelial cells led to increased cytosolic expression of proteins that could potentially destabilize IP-10 mRNA. The 3′-UTR of the human IP-10 gene contains an AUUUA motif in the context of an AU-rich sequence region. This combination has been defined as constituting a class I ARE, which is associated with mRNA deadenylation and decay (38). Among the proteins known to mediate such mRNA destabilization are TTP and AUF1 (5, 8). HRV-16 infection did not increase TTP mRNA expression, nor were we able to detect cytosolic expression of TTP protein by Western blot. This may simply reflect a lack of sensitivity, but, under the same conditions, we were readily able to detect cytosolic expression of the known isoforms of AUF1. HRV-16 infection did not increase cytosolic levels of AUF1 12 h after infection but led to marked increases at 24 and 48 h postinfection, with the higher-molecular-weight forms of AUF1 being most enhanced. This time frame of AUF1 induction is consistent with the loss of HRV-16-induced IP-10 gene expression. Moreover, AUF1 has been shown to bind to a single AUUUA motif, such as is present in the 3′-UTR of IP-10 mRNA, albeit with a weaker affinity than to 3′-UTR containing multiple AUUUA motifs (8). This difference in binding affinities correlates with the rate of mRNA decay induced by AUF1, such that molecules with a single AUUUA motif are degraded more slowly than those with multiple repeats (8). Together, therefore, these data are consistent with the possibility that AUF1 contributes to the relatively slow cytosolic destabilization of IP-10 mRNA, although additional studies will be needed to definitively establish such a role.
When other potential molecules generated during viral replication that may induce IP-10 generation are considered, an obvious candidate is dsRNA, which has been shown to be a potent stimulus not only for host cell antiviral responses, but also for cytokine induction (10, 13). Although several studies have attempted to mimic effects of intracellular dsRNA generated during viral replication by adding dsRNA to the extracellular milieu on the assumption that it is somehow internalized into cells, the recent demonstration that extracellular dsRNA can serve as a ligand for Toll-like receptor 3 (TLR3), and thereby induce activation of NF-κB, makes interpretation of data using this approach difficult (1). In the current studies, therefore, we performed pilot dosing experiments to establish doses of dsRNA that induced little IP-10 induction when applied extracellularly and then examined responses when cells were transfected with this dose. Our data demonstrate that extremely high levels of IP-10 are produced when epithelial cells are transfected with 0.1 μg of dsRNA. Based on this response it is likely that levels of dsRNA generated in response to viral infection are significantly >0.1 μg. In agreement with a recent report (10), we found that 10–30 μg of extracellular dsRNA were required to induce a response comparable to that seen with transfection of 0.1 μg (data not shown). This response was not an artifact of transfection because little or no response was observed when ssRNA was transfected. These data suggest that intracellular dsRNA, such as would be produced during viral infection, is a much more effective stimulus for IP-10 induction than stimulation via cell surface TLR3. Alternatively, the affinity of cell surface TLR3 for dsRNA may be relatively low. We cannot definitively rule out that transfected dsRNA may interact with intracellular TLR3 but are aware of no data supporting the existence of intracellular TLR3 in epithelial cells. Although our data are consistent with a role of viral dsRNA in IP-10 generation, additional studies will be needed to confirm this and to examine potential pathways involved. For example, although several studies have implicated activation of the dsRNA-activated protein kinase (PKR) in viral induction of cytokines, it has recently been shown that dsRNA-induced cytokine production from epithelial cells can also occur via an as yet to be delineated PKR-independent pathway (19). Further studies are needed to examine the molecular pathways, and potential intermediates, regulating IP-10 production in response to dsRNA and intact virus.
It has been shown previously that NF-κB response elements in the promoter region of the IP-10 gene are involved in transcriptional activation of IP-10 in response to a variety of stimuli (17, 22, 37), and HRV infection is known to activate NF-κB in epithelial cells (14, 29). Consistent with these observations, administration of calpain inhibitor I, a proteasomal inhibitor that blocks NF-κB activation, inhibited IP-10 gene induction from HRV-16-infected epithelial cells. Furthermore, mutation of either of the two NF-κB response elements in a 972-bp promoter-luciferase construct markedly reduced luciferase activity induced by HRV infection. It is of interest, however, that mutation of the more proximal NF-κB1 site, which almost completely abrogated luciferase responses to HRV-16 infection, had significantly greater effects compared with mutation of the NF-κB2 site (Fig. 5B). This is in marked contrast to earlier studies using the murine IP-10 promoter in which activation by dsRNA was completely unaffected by mutation of the NF-κB1 site but was profoundly inhibited by mutation of the NF-κB2 site (37). Similarly, promoter activation by measles virus, which also generates dsRNA, was also much more affected by mutation of the NF-κB2 site as opposed to the NF-κB1 site (22). Thus the reduction of promoter responses to HRV-16 by mutation at the NF-κB2 site may be consistent with inhibition of activation by viral dsRNA, but the striking effects of mutating the NF-κB1 site imply that additional signals are involved in upregulating transcription in response to HRV-16. This signal(s) may be provided by other intermediates of the viral replication pathway or may be due to effects of secondary stimuli induced in response to viral infection.
Although our initial studies have focused on the role of NF-κB in IP-10 gene activation by HRV-16, it is clear that additional regulatory sites are involved in transcriptional control. Previous studies have reported that maximal stimulation of IP-10 activity by either IFN-γ or by dsRNA is observed when ∼240 bp of promoter sequence containing the ISRE/interferon regulatory factor 2 site and the two NF-κB sites are used (23, 37). By contrast, promoter activity in response to HRV-16 was reduced by ∼50% when the region between −229 and −875 bp upstream of the transcriptional start site was deleted. Detailed studies will be performed to establish which putative transcription factor recognition sites play a role in the enhanced response of this longer promoter construct.
To evaluate in vivo IP-10 production, we experimentally infected healthy, sero-negative volunteers with HRV-16. Our data provide the first demonstration of a striking induction of IP-10 in airway secretions during symptomatic HRV infections. Although it is not possible to definitively establish epithelial cells as the source of this IP-10, the epithelium is the principal site of viral infection and replication within the airways, and our in vitro data demonstrate that epithelial induction of IP-10 is dependent on viral replication. The strong correlation between viral titer and IP-10 generation in vivo, therefore, is consistent with a major role for the epithelium as a contributor to IP-10 levels. Additional studies will be needed to determine whether a specific population of epithelial cells produces IP-10 during in vivo HRV infections. The functional relevance of IP-10 in vivo is supported both by the large levels of chemokine detected and by the correlation between IP-10 levels and lymphocyte numbers. Although we have not established the cytokine profiles generated by these lymphocytes, it is likely that they are, at least in part, of the type 1 phenotype, given that type 1 cytokines are detected during HRV infections (11) and that this subtype of lymphocytes selectively expresses the CXCR3 receptor (28). Finally, IP-10 also correlated with symptoms, suggesting that this chemokine may be one contributor to those aspects of the proinflammatory host response to viral infection that underlie the induction of symptoms. Together, these data provide strong support for a pathophysiological role of IP-10 in HRV infections.
In summary, our studies provide the first demonstration that HRV-infected epithelial cells produce IP-10 mRNA and protein in a time- and dose-dependent fashion. Induction of IP-10 requires HRV that is capable of replication, and the potency of intracellular dsRNA as a stimulus for IP-10 release suggests this intermediate may play a role in the induction of IP-10 by replicating virus. HRV-induced IP-10 gene expression involves transcriptional activation at NF-κB recognition sequences, but, in contrast to other known stimuli, the proximal NF-κB1 site appears to play a central role. Other, upstream transcription factors are also implicated in IP-10 gene induction by HRV-16. Rhinovirus infection of epithelial cells induced increased cytosolic expression of the mRNA destabilizer AUF1. Finally, we have established, for the first time, that generation of IP-10 also occurs in the airways during in vivo HRV-16 infection. The correlation of IP-10 with both lymphocyte numbers and symptoms suggest a role of this chemokine in the pathogenesis of HRV infections and potentially in HRV-induced airway exacerbations of asthma and COPD.
This work was supported by Canadian Institutes for Health Research Grant 43923.
The authors are grateful to Dr. Keith Blackwell for providing antibody to TTP. D. Proud is the recipient of a Canada Research Chair in Inflammatory Airway Diseases.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2005 the American Physiological Society