Rhinovirus Infection Induces Expression of Its Own Receptor Intercellular Adhesion Molecule 1 (ICAM-1) via Increased NF-κB-mediated Transcription*

Virus infections, the majority of which are rhinovirus infections, are the major cause of asthma exacerbations. Treatment is unsatisfactory, and the pathogenesis unclear. Lower airway lymphocyte and eosinophil recruitment and activation are strongly implicated, but the mechanisms regulating these processes are unknown. Intercellular adhesion molecule-1 (ICAM-1) has a central role in inflammatory cell recruitment to the airways in asthma and is the cellular receptor for 90% of rhinoviruses. We hypothesized that rhinovirus infection of lower airway epithelium might induce ICAM-1 expression, promoting both inflammatory cell infiltration and rhinovirus infection. We therefore investigated the effect of rhinovirus infection on respiratory epithelial cell ICAM-1 expression and regulation to identify new targets for treatment of virus-induced asthma exacerbations. We observed that rhinovirus infection of primary bronchial epithelial cells and the A549 respiratory epithelial cell line increased ICAM-1 cell surface expression over 12- and 3-fold, respectively. We then investigated the mechanisms of this induction in A549 cells and observed rhinovirus-induction of ICAM-1 promoter activity and ICAM-1 mRNA transcription. Rhinovirus induction of ICAM-1 promoter activity was critically dependent upon up-regulation of NF-κB proteins binding to the −187/−178 NF-κB binding site on the ICAM-1 promoter. The principal components of the rhinovirus-induced binding proteins were NF-κB p65 homo- or heterodimers. These studies identify ICAM-1 and NF-κB as new targets for the development of therapeutic interventions for virus-induced asthma exacerbations.

Asthma is increasingly common and now affects up to 30% of the population in westernized countries (1). Asthma exacerbations are the major cause of asthma morbidity and mortality. Respiratory viral infections have recently been associated with the majority of asthma exacerbations in both adults and children. In community-based studies, viral infections were identified in 80 -85% of exacerbations in children and 44% of exacerbations in adults (2,3). Viral infections have also been strongly implicated in more severe asthma exacerbations requiring hospitalization in both children and in adults (4). In all of these studies, rhinovirus infections caused around 65% of exacerbations in which a virus was identified (2)(3)(4). Rhinovirus-induced asthma exacerbations therefore cause enormous morbidity, especially among children, and represent a major health and economic problem.
To date, no safe effective therapy is available, since prophylactic inhaled steroids are ineffective (5), while intervention with high dose inhaled steroids is only partially effective (6). A better understanding of the mechanisms involved in rhinovirus-induced asthma exacerbations would greatly aid the development of new therapies for this common condition.
The mechanisms by which rhinoviruses trigger asthma exacerbations are poorly understood. Asthma is an inflammatory disease of the lower respiratory airways, and lower airway inflammatory changes have been described during experimental rhinovirus colds. A marked bronchial CD3 ϩ , CD4 ϩ , and CD8 ϩ T lymphocyte and eosinophil infiltration was observed in biopsies taken at the height of cold symptoms in both normal and in asthmatic subjects (7). However, the eosinophil infiltrate was more prolonged in the asthmatic subjects, still present 6 -8 weeks after infection, while the eosinophil counts in normal subjects had returned to base line (7). Rhinovirus experimental infections have also been reported to increase allergen-induced eosinophil numbers in bronchial lavage fluid in atopic rhinitic subjects, while no change in eosinophil numbers was observed in normal subjects (8), and to increase eosinophil products in sputum supernatants in asthmatic subjects (9). These data combined strongly suggest that rhinovirus-induced bronchial lymphocyte and eosinophil infiltration and activation are probably very important mechanisms in virus-induced asthma exacerbations.
The increased airway reactivity demonstrated in experimental rhinovirus infections in asthmatic subjects (9,10) and atopic subjects (11) and the induction by rhinovirus of late asthmatic responses to inhaled allergen (8,11) also provide indirect evidence of a link between lower respiratory inflammation during rhinovirus experimental infections and the mechanisms of virus-induced asthma exacerbations. Finally, evidence that the lymphocytic and eosinophilic inflammation observed during rhinovirus experimental infections (7)(8)(9) is probably also an important mechanism involved in virus-induced asthma exacerbations comes from the fact that asthma exacerbations have been induced by experimental rhinovirus infections (9,12).
Rhinovirus RNA has recently been detected in bronchial lavage cells taken during experimentally induced colds, suggesting that rhinovirus can promote local inflammation by direct infection of the lower airways (13). Indeed, rhinoviruses are capable of prolonged, noncytolytic infection of respiratory epithelial cells and induce production of proinflammatory cy-tokines such as IL-6, 1 IL-8, and granulocyte-macrophage colony-stimulating factor (14 -17). Taken together, these data suggest that lower airway rhinovirus-induced inflammatory cell recruitment is a critical event in rhinovirus-induced asthma exacerbations.
Intercellular adhesion molecule-1 (ICAM-1) is a cell surface glycoprotein belonging to the immunoglobulin supergene family that is involved in leukocyte trafficking and accumulation at sites of inflammation. In addition to modulating eosinophil infiltration by binding to its ligand CD18/CD11b, ICAM-1 is involved in lymphocyte infiltration by binding to CD18/CD11a expressed on both CD4 ϩ and CD8 ϩ T lymphocytes. There is now considerable evidence that ICAM-1 plays a central role in recruitment and activation of these inflammatory cells in the pathogenesis of asthma. Epithelial ICAM-1 expression is increased by allergen challenge in both allergic rhinitis and conjunctivitis (18 -20). High levels of epithelial ICAM-1, along with intraepithelial inflammatory cell infiltration, have been described in bronchial biopsies from subjects both with stable asthma and after allergen challenge (21,22). Direct evidence for the important role of ICAM-1 in eosinophil recruitment came from a primate study that demonstrated that allergen challenge causing a dual asthmatic response up-regulated ICAM-1 expression in airway epithelium and endothelium, that eosinophil infiltration into the airways correlated with the epithelial ICAM-1 levels, and that anti-ICAM-1 antibodies prevented both the eosinophil influx and airway hyperreactivity (23). The critical importance of ICAM-1 in asthma pathogenesis has been emphasized by two further recent studies in murine models of asthma, which demonstrated ICAM-1 regulation of lymphocyte and eosinophil recruitment to the lower airway (24,25).
Since epithelial infiltration of each of these cell types is specifically implicated in rhinovirus induced asthma (7)(8)(9), respiratory epithelial ICAM-1 expression is likely to play a very important role in inflammatory cell infiltration associated with rhinovirus-induced asthma exacerbations.
In addition to its important role in inflammatory cell recruitment and activation, ICAM-1 may play a critical role in rhinovirus-induced asthma exacerbations, since it is also the cellular receptor for the major group (90%) of rhinoviruses (26,27). Having observed previously that rhinoviruses are able to infect lower respiratory epithelial cells for several days without causing cytopathic effect and able to induce proinflammatory protein expression (16), we hypothesized that rhinovirus infection of lower respiratory epithelial cells would also induce increased expression of ICAM-1. Rhinovirus induction of ICAM-1 in lower respiratory epithelial cells could then not only promote intraepithelial inflammatory cell recruitment and activation but also increase the severity of the epithelial cell infection and therefore further exacerbate the airway inflammation. Modulation of ICAM-1 expression would then be expected to have profound effects on the epithelial inflammatory cell infiltration associated with rhinovirus-induced asthma exacerbations.
To investigate this hypothesis, studies were undertaken to determine whether rhinovirus infection has the ability to modulate respiratory epithelial ICAM-1 expression in an in vitro model. Having found marked rhinovirus-induced ICAM-1 upregulation in both primary bronchial epithelial cells and a human lower respiratory cell line, we investigated the intracellular mechanisms of rhinovirus induction of ICAM-1 expression to identify potential targets for modulation of rhinovirusinduced ICAM-1 in the therapy of rhinovirus-induced asthma exacerbations.

Cell Culture
A549 cells, a type II respiratory epithelial cell line, were obtained from the American Type Culture Collection (ATCC; Rockville, MD), and Ohio HeLa cells were obtained from the Medical Research Council Common Cold Unit (Salisbury, UK). Cells were split weekly and cultured at 37°C in 5% carbon dioxide in Eagle's minimal essential medium supplemented with 4 mM L-glutamine, 80 mg/ml of gentamycin, and 10% fetal bovine serum (Sigma, Poole, UK). Primary human bronchial epithelial cells were obtained by bronchial brushing from normal patients undergoing surgery. Cells were removed from the brush by vigorous shaking and were disaggregated in Clonetics (San Diego, CA) bronchial epithelial cell medium containing 3 mM DTT for 15 min. After washing, cells were plated onto collagen-coated 16-mm diameter culture wells and grown to confluence in bronchial epithelial cell medium. Cells were passaged in 100-mm Petri dishes and used in the assays at passages 3 and 4. For experiments, 70% confluent cells were detached using 0.05% trypsin with 0.02% EDTA and seeded at 2 ϫ 10 5 cells/well in 12-well culture plates. These cells are Ͼ95% cytokeratin 18-immunoreactive epithelial cells as assessed by immunofluorescence microscopy.

Viral Stocks
Rhinovirus types 16, 9 (major group), and 2 (minor group) were obtained from the Medical Research Council Common Cold Unit, and their identity was confirmed by neutralization with specific antiserum (ATCC). Viral stocks were generated by infecting monolayer cultures of HeLa cells until cytopathic effects were fully developed. Cells and supernatants were harvested, cells were disrupted by freezing and thawing, cell debris was pelleted by low speed centrifugation, and the resulting clarified supernatants were frozen at Ϫ70°C.
Rhinovirus titration was performed on the frozen aliquots by exposing confluent monolayers of HeLa cells in 96-well plates to serial 10-fold dilutions of viral stock. Plates were cultured for 5 days in 4% minimal essential medium at 37°C in 5% CO 2 . Cytopathic effect was assessed by visual assessment and by assessment of the continuity of the monolayer after fixation in methanol and staining with 0.1% crystal violet. Tissue culture infective dose 50% (TCID 50 )/ml values were determined (28), and virus at a multiplicity of infection (MOI) of 1 was used for all of the experiments except where indicated.

Rhinovirus Inactivation
For selected experiments, inactivation/filtration of the virus was performed by three different methods.

Prevention of Virus-Receptor Binding
Viruses were precoated with excess soluble receptor to saturate the receptor binding sites on the virus capsid. Virus stock solutions were preincubated with recombinant soluble ICAM-1 (sICAM; a gift of P. Esmon, Bayer Corp., Berkeley, CA) at a concentration of 1 mg/ml for 30 min at room temperature.

Prevention of Virus Replication
Viruses were inactivated by exposure to UV light at 1200 mJ/cm 2 for 30 min.

Filtration of Virus from Inoculum
Virus particles were removed from inocula by ultrafiltration through membranes (Amikon, London, UK) to remove all molecules greater than 30 kDa, performed according to the manufacturer's instructions.
For each method, confirmation of complete inactivation was carried out by microtiter plate assay for rhinovirus infectivity as described above.
effects of inactivated/filtered virus and the effects on primary bronchial epithelial cells were studied at 8 h. At the desired time points, cells were detached intact by incubation with 0.5 ml/well cell dissociation solution (Sigma) at 37°C for 10 min. More than 95% of cells were viable as determined by trypan blue dye exclusion. 10 5 cells were then washed and resuspended in PBA (phosphate-buffered saline, 1% bovine serum albumin, 0.1% sodium azide) and incubated with saturating amounts of fluorescein isothiocyanate-conjugated anti-human ICAM-1 (CD54) antibody or isotype-specific control antibody (Serotec, Oxford, UK) for 30 min at 4°C in the dark. After washing, 10 4 cells were analyzed for fluorescence by single color flow cytometry on a FACScan analyzer (Becton Dickinson, San Jose, CA). Mean fluorescence intensity was measured and normalized relative to noninfected control cells after subtraction of background staining.
ICAM-1 mRNA Analysis 5 ϫ 10 6 A549 cells were cultured in 100-mm plates until confluent, and medium alone or rhinovirus type 16 was added for various times between 1 and 24 h. Studies with inactivated/filtered virus were performed at 8 h. At the desired time points, cells were harvested, and ICAM-1 mRNA expression was evaluated by RT-PCR. Whole cell RNA was extracted using Trizol according to the manufacturer's instructions (Life Technologies, Inc., Paisley, UK). One g of total RNA was reverse transcribed by superscript reverse transcriptase (100 units; Promega, Southampton, UK) in a total volume of 10 l at 37°C for 1 h using P1 (24 ng/ml) as specific primer (Table I). The cDNA (2.5 l) was amplified by PCR in the presence of a master mix containing PCR buffer, MgCl 2 (1.5 mM), 1.25 units of Taq DNA polymerase (Promega), 0.2 mM dNTPs, and 0.6 mM specific primer pair (P1 and P2; Table I). Cycling conditions were 1 min at 94°C, 1 min at 55°C, and 2 min at 72°C for 25 cycles. First round products were diluted 1:100, and 2.5 l was thereafter used for a nested amplification under the same PCR conditions, with P3 and P4 as inner primers (Table I). Final PCR products (10 l) were electrophoresed through 1.5% agarose gels, stained in ethidium bromide, and photographed under UV light. In parallel, mRNA for adenine phosphoribosyltransferase (APRT), using primers indicated in Table I for 40 cycles at 56°C, was evaluated in each sample as housekeeping gene control. Densitometry was performed using a scanning densitometer, and densitometric analysis was performed using the Phoretix program (Biometra Ltd., Newcastle-upon-Tyne, UK) to express ICAM-1 mRNA relative to APRT mRNA.
To confirm the quantitative nature of the PCR, cell lysate was diluted 3-fold in triplicate and subjected to RNA extraction, RT-PCR, and densitometric analysis.

ICAM-1 Nuclear Transcription Analysis
Isolation of nuclei and in vitro nuclear transcription were performed using standard procedures (29). Confluent cell monolayers (2.5 ϫ 10 7 cells) were incubated with medium alone or rhinovirus 16 for 1 h. Cells were then washed twice with phosphate-buffered saline, harvested, and centrifuged at 500 ϫ g for 5 min. The cell pellet was resuspended in 1 ml of 10 mM Tris-HCl, pH 8.4, 1.5 mM MgCl 2 , 0.14 M NaCl and lysed by the addition of 5% Nonidet P-40. The progress of lysis was monitored by trypan blue exclusion. Nuclei were pelleted by centrifugation at 500 ϫ g for 1 min and were washed twice in 20 mM Tris-HCl, pH 8.3, 20% glycerol, 100 mM KCl, 4.5 mM MgCl 2 , 2 mM DTT.
In vitro nuclear transcription was carried out for 45 min at 30°C in 200 ml of this buffer supplemented with 1 mM each of ATP, GTP, CTP, and UTP.
For each sample, total RNA was extracted from 10 7 nuclei, before and after in vitro transcription, in the presence or absence of an RNA polymerase II inhibitor ␣-amanitin (1 mg/ml) (30). ICAM-1 RT-PCR was subsequently performed (see above) for each different condition to detect in vitro transcribed products.

Reporter Gene Constructs
The ICAM-1 promoter-chloramphenicol acetyltransferase (CAT) constructs were a generous gift of Dr. K. Degitz (Ludwig-Maximilians University, Munchen). They contained sequential deletions (Ϫ1160, Ϫ277, Ϫ182, Ϫ135, Ϫ88) of the ICAM-1 5Ј-flanking region linked to the coding region of the CAT reporter gene (31). A plasmid containing the longest promoter deletion with a mutated Ϫ187/Ϫ178 NF-B sequence (Ϫ1160m; mutated from TGGAAATTCC to TctAgATTag and confirmed by sequencing) and pBRAMScat2 vectors, composed of the CAT reporter gene and the herpes simplex virus minimal thymidine kinase promoter alone or linked to fragments of the ICAM-1 promoter (Ϫ199/Ϫ170 or Ϫ199/Ϫ182), were also kindly provided (31,32).

Cell Transfection and CAT Assay
A549 cells were transfected with reporter constructs (10 g) at 80% confluency by the calcium phosphate precipitation technique for 5 h, glycerol-shocked with 1ϫ HeBS (0.02 M Hepes, 0.135 M NaCl, 0.5 mM Na 2 HPO 4 , 5.5 mM D-glucose, pH 7.1), 15% glycerol for 30 s and washed. Transfected cells were cultured in 10% minimal essential medium for 24 h, and rhinovirus 16 or medium alone was added. At designated time points, cells were harvested, and cell extracts were prepared by three cycles of rapid freeze-thawing in 0.25 M Tris, pH 8.0. Each sample was incubated at 65°C for 15 min to inactivate endogenous transacetylases. Protein content was determined photometrically using the Bio-Rad protein assay (Bio-Rad). Protein-equivalent aliquots were assayed for CAT activity according to standard protocols (33). The assay was performed at 37°C for 60 min in a reaction mixture of 1 mM acetyl coenzyme A (Amersham Pharmacia Biotech) and 0.1 Ci of 14 C-chloramphenicol. Acetylated and unacetylated forms were resolved by thin layer chromatography, visualized by autoradiography, and measured on a scintillation counter. CAT activity was expressed as percentage of chloramphenicol converted to its acetylated derivatives.

Electrophoretic Mobility Shift Assay (EMSA)
Preparation of Nuclear Extracts-Uninfected and rhinovirus-infected A549 cells were prepared as described previously. At the desired time points (0, 30, 60, 90, and 120 min), the cells were mechanically detached, and nuclear extracts were obtained by a modification of the the method of Dignam et al. (34). Briefly, after washing with phosphatebuffered saline, cells were centrifuged at 4°C and resuspended in buffer A (10 mM Hepes, pH 7.9, 1.5 mM MgCl 2 , 10 mM KCl, 0.5 mM DTT) with freshly added protease inhibitors (10 mg/ml leupeptin, 5 mg/ml aprotinin, 1 mM phenylmethylsulfonyl fluoride). Membrane lysis was achieved by adding 0.5% Nonidet P-40 followed by vigorous agitation and incubation on ice for 5 min. The nuclei were then pelleted at 4°C and resuspended in buffer C (20 mM Hepes, pH 7.9, 25% glycerol, 0.42 M NaCl, 1.5 mM MgCl 2 , 10 mM KCl, 0.2 mM EDTA, 0.5 mM DTT, and freshly added protease inhibitors as above). This suspension was incubated on ice for 15 min and centrifuged. The protein concentration of the nuclear extracts was photometrically determined using the Bio-Rad protein assay.
Oligonucleotide Probes (Table II)-Double-stranded oligonucleotides containing wild-type and mutated sequences of ICAM-1 AP-1, NF-B, Sp1, and C/EBP recognition sequences were obtained commercially (Oswell DNA Service, Southampton, UK). Mutant sequences were identical to those used in the mutant reporter constructs. Probes containing NF-B, AP-1, or Sp1 consensus sequences were commercially obtained (Promega).
Supershift EMSA-Supershift assays were used to study which members of the NF-B family were involved in rhinovirus-induced formation of complexes with the ICAM-1 promoter sequence Ϫ199/ Ϫ170. One l of rabbit polyclonal antibodies against each of p65, p50, p52, c-Rel, and Rel-B (Santa Cruz Biotechnology, Inc., Santa Cruz, CA) were added to 2 g of nuclear extracts for 15 min at 4°C before the incubation with radiolabeled probe as described above. Rabbit preimmune serum was used as negative control.

Statistical Analysis
Data were expressed as mean Ϯ S.E., and comparison between groups was performed by analysis of variance for multiple comparisons and by paired Student's t tests for individual comparisons. All experiments were carried out at least three times.

Rhinovirus Induces ICAM-1 Cell Surface Protein Expression in A549 Cells and in Primary Bronchial Epithelial Cells-
Preliminary studies indicated that rhinovirus infection of A549 cells up-regulated ICAM-1 surface expression at 8 h postinoculation. Dose-response studies in A549 cells exposed to rhinovirus 16 were therefore carried out to determine if the induction of ICAM-1 occurred in a dose-response manner. Cell surface ICAM-1 expression was studied by flow cytometry 8 h after infection; enhanced expression of ICAM-1 relative to uninfected cells was observed at 0.1 TCID 50 /cell and peaked at 1 TCID 50 /cell, where there was 3.5-fold induction over uninfected cell levels of ICAM-1 expression (Fig. 1). Based on these doseresponse data, a MOI of 1 TCID 50 /cell was utilized in all subsequent studies.
To evaluate the temporal kinetics of ICAM-1 induction by rhinovirus, surface ICAM-1 expression was studied at 0, 1, 4, 8, 16, 24, 48, and 72 h post-rhinovirus 16 infection. Significant up-regulation was apparent within 4 h, was maximal at 8 h, and was still significantly increased at 48 and 72 h after infection (Fig. 2). The levels of rhinovirus-induced ICAM-1 were similar in magnitude to those observed with interferon-␥ (10 units/ml) treatment, which was used as a positive control (data not shown). In view of the time course results, an 8-h infection was chosen for comparative studies to investigate the receptor specificity and virus specificity of the up-regulation.
Having investigated rhinovirus regulation of ICAM-1 expression in the human lung carcinoma epithelial cell line A549, we wished to determine whether the same effects could be observed in primary human bronchial epithelial cells. The effect of rhinovirus infection on respiratory epithelial ICAM-1 surface expression was studied by flow cytometry in primary human bronchial epithelial cells. We observed that rhinovirus infection for 8 h increased ICAM-1 expression 12.7 times the values of control sham-infected cells (Fig. 3). These data confirmed that rhinovirus infection of both A549 cells and primary bronchial epithelial cells were associated with markedly increased ICAM-1 surface expression. We therefore investigated the mechanisms of this induction in A549 cells, since the numbers of cells required for subsequent experiments precluded the use of primary cells.

The Effects of Rhinovirus Replication and Receptor Binding on Rhinovirus-induced ICAM-1 Cell Surface Expression-Since
the virus inoculum was a crude preparation, we wished to confirm that the induction of ICAM-1 surface expression was a result of virus-specific effects rather than a result of stimulation by other soluble products such as cytokines present in the inoculum. We were also interested in investigating whether any rhinovirus-specific effect observed was a result of virus replication or of virus-receptor binding. We therefore elected to inactivate rhinovirus by two methods: UV inactivation to prevent replication but not receptor binding and precoating with soluble receptor (sICAM) to prevent receptor binding. Finally we filtered the inoculum through a molecular weight filter to remove all virus particles and RNA but not small molecules such as cytokines.
As can be observed in Fig. 4, incubation with UV-inactivated virus resulted in marked inhibition of rhinovirus-induced ICAM-1 surface protein expression (from 3.49 Ϯ 0.3-to 1.63 Ϯ 0.1-fold induction); however, there was still a small but significant induction with UV-inactivated rhinovirus over control cells (p Ͻ 0.05). In contrast, sICAM inactivated and filtered virus completely abrogated the induction observed with rhinovirus (Fig. 4). These results suggest that approximately onequarter of the ICAM-1 up-regulation observed with live rhinovirus occurs independently of viral replication, as a result of virus-receptor interaction, while the major part is dependent upon viral replication.
Rhinovirus Induction of ICAM-1 Is Not Virus Receptor/ Strain-specific-The major group (90%) of rhinoviruses use ICAM-1 as their cell surface receptor (26,27), while the remainder (minor group) use members of the low density lipopro-  tein receptor family (35). Having observed rhinovirus induction of ICAM-1 expression with a major group rhinovirus, rhinovirus 16, we wished to investigate whether rhinovirus induction of ICAM-1 up-regulation is strain-or group (receptor)-restricted.
We therefore compared the stimulatory effect of rhinovirus 16, rhinovirus 9 (both major group), and rhinovirus 2 (minor group), all at an MOI of 1, on A549 cell ICAM-1 surface expression at 8 h postinfection. As shown in Fig. 3, rhinovirus 16, rhinovirus 9, and rhinovirus 2 were equally effective at increas-ing ICAM-1 surface expression, demonstrating that rhinovirusinduced ICAM-1 up-regulation occurs with at least three of the many different rhinovirus serotypes and that the induction was not receptor-restricted. Furthermore, pretreatment of rhinovirus 2 with sICAM did not alter the ability of this minor group rhinovirus to induce ICAM-1. Having observed that sICAM pretreatment completely abolished (Fig. 4) the ICAM-1 expression induced by the major group rhinovirus, rhinovirus 16, these findings support our interpretations of the preceding data relating to the respective contributions of virus-receptor binding and virus replication to ICAM-1 induction by rhinoviruses.
Induction of ICAM-1 mRNA in A549 Cells by Live and Inactivated Rhinovirus-Having found rhinovirus-induced increases in ICAM-1 epithelial cell surface protein expression, we wished to test the effects of rhinovirus infection on epithelial cell ICAM-1 mRNA expression.
First, we wished to determine that the PCR analysis was quantitative over the range of input RNA used in the study. To investigate this, a cell lysate known to produce a strong band upon PCR was serially diluted 3-fold and subjected to extraction, RT-PCR, and densitometric analysis in triplicate. These studies clearly demonstrated that the ICAM-1 RT-PCR used in the subsequent studies was quantitative in a linear fashion (Fig. 5).
The time course of ICAM-1 mRNA induction in response to  (Fig. 6). In accordance with our findings on surface expression, a consistent response to rhinovirus infection was noted, with clear time-dependent increases in ICAM-1 mRNA expression being induced by rhinovirus infection (Table III). A representative experiment is depicted in Fig. 6, where an early increase in levels of ICAM-1 mRNA was detectable at 1 h, peaked at 8 h, and reduced toward but not as far as base line up to 24 h.
Rhinovirus Infection of A549 Cells Up-regulates ICAM-1 Gene Transcription-To determine whether the observed increases in ICAM-1 mRNA and protein expression in response to rhinovirus infection of A549 cells were mediated by increased ICAM-1 gene transcription, de novo synthesis of ICAM-1 mRNA (nuclear run-off) was studied in nuclei obtained from A549 cells after a 1-h rhinovirus infection and in control noninfected cells.
In accordance with the observed mRNA time course studies, ICAM-1 mRNA was undetectable in nuclei from control noninfected cells, either before (Fig. 8, lane 1) or after (Fig. 8, lane 2) in vitro transcription, while a weak band of ICAM-1 mRNA was detectable after a 1-h rhinovirus 16 infection but without in vitro transcription (Fig. 8, lane 3). The amount of ICAM-1 mRNA was markedly increased by 45-min in vitro transcription (Fig. 8, lane 4), indicating that rhinovirus infection of A549 cells resulted in increased de novo ICAM-1 mRNA transcription (Table V). This was confirmed by the fact that the rhinovirus-induced increase in ICAM-1 mRNA observed during in vitro transcription was abolished in the presence of ␣-amanitin, a DNA-dependent RNA polymerase II inhibitor (30), (Fig. 8, lane 5; Table V). From these results, we concluded that rhinovirus infection of A549 cells induces a rapid increase in ICAM-1 gene transcription.
Rhinovirus Infection of A549 Cells Increases ICAM-1 Promoter Activity-Having demonstrated rhinovirus induction of ICAM-1 gene transcription, we then carried out studies to determine whether rhinovirus infection of A549 cells increased ICAM-1 promoter activity. A549 cells were transiently transfected with constructs containing the CAT reporter gene, whose transcription was regulated by portions of the ICAM-1 promoter. Time course experiments with a construct containing the full-length promoter (Ϫ1160 bp) showed that rhinovirus 16-infected cells had significantly increased ICAM-1 promoter activity compared with control cells at all of the tested time points (24,48,72,96, 120 h, data not shown). At 24 h of infection, induction of ICAM-1 promoter activity was maximal, promoter activity being barely detected in control cells (acetylation 2 Ϯ 0.7%), while it was markedly increased in the rhinovirus-infected cells (acetylation 31.9 Ϯ 9%, p Ͻ 0.01). The 24-h time point was therefore utilized in subsequent experiments to investigate rhinovirus induction of shorter deletions of the ICAM-1 promoter, to identify the precise sites in the promoter induced by rhinovirus infection.

Rhinovirus Infection Induces Multiple Transcription Factors
Binding to the ICAM-1 Promoter-Sequence analysis of the proximal ICAM-1 promoter has revealed potential binding sites for several transcription factors including AP-1 (Ϫ284/Ϫ278), Sp1 (Ϫ206/Ϫ201), C/EBP (Ϫ199/Ϫ196), and NF-B (Ϫ187/ Ϫ178) and (Ϫ62/Ϫ53), which also overlaps with a further Sp1 binding motif (Ϫ59/Ϫ54) (36 -39). To further understand the mechanisms by which rhinovirus induces ICAM-1 promoter activity, studies were undertaken to investigate whether rhinovirus infection of A549 cells could increase the binding activity of relevant transcription factors in nuclei extracted from infected and uninfected lung epithelial cells, using labeled probes containing each of the potential binding sites in EMSAs.
Ϫ199 to Ϫ170 Probe Containing C/EBP (Ϫ199/Ϫ196) and NF-B (Ϫ187/Ϫ178) Sites-Two retarded complexes were observed using nuclear extracts from rhinovirus 16-infected A549 cells that were absent in nuclear extracts from uninfected cells. Time course experiments demonstrated that induction of these complexes was maximal 30 min after infection and decreased with longer incubations up to 2 h (Fig. 9A). Competition experiments were then carried out to confirm the specificity of the binding. The addition of excess unlabeled specific (Ϫ199/Ϫ170) oligonucleotide blocked binding of both of the protein complexes (Fig. 9B, lanes 1 and 2), confirming the specificity of the binding.
Further competition experiments were carried out to identify the transcription factors binding to the probe. The addition of unlabeled consensus NF-B probe completely blocked the binding of complexes to the specific probe, while the addition of an unlabeled consensus AP-1 probe had no effect (Fig. 9B, lanes 3  and 4), suggesting that both binding complexes were formed of     In order to confirm this, competition experiments with Ϫ190/ Ϫ170 probes containing mutated NF-B and C/EBP sites were then carried out. Binding of these complexes was not affected by competition with the probe containing an intact C/EBP site but a mutated ICAM-1 NF-B site (M 2 ), but binding was completely abrogated by competition with the probe containing an intact NF-B site but a mutated C/EBP site (M 1 ) (Fig. 9C, lanes  3-5). These results confirmed that rhinovirus infection of A549 cells induces transcription factors binding to the Ϫ187/Ϫ178 NF-B cis element but not the Ϫ199/Ϫ196 C/EBP element in the ICAM-1 promoter.
To determine which members of the NF-B/Rel protein family were responsible for the formation of the two inducible nuclear complexes, specific antisera directed against members of the NF-B/Rel family (p50, p52, p65, c-Rel, and Rel-B) were studied. Antiserum specific for p65 clearly supershifted DNAprotein complexes, while the antiserum directed against p50 and c-Rel significantly diminished the formation of the inducible complexes (Fig. 10). In contrast, antibodies to p52, Rel-B and preimmune serum had no significant effect on complex formation. These studies demonstrated that the most important members of the NF-B/Rel family mediating rhinovirusinduced NF-B element binding were p65, c-Rel, and p50, with p65 being the major component of the homo-or heterodimers formed.
Ϫ227 to Ϫ200 probe Containing an Sp1 Binding Site (Ϫ206/ Ϫ201)-The EMSA resulted in the retardation of two complexes, but no induction was observed in nuclear extracts from rhinovirus-infected cells at any time points up to 2 h (data not shown), indicating that proteins binding to this DNA segment containing an Sp1 binding site are not induced by rhinovirus infection of A549 cells.
Ϫ294 to Ϫ266 Probe Containing an AP-1 Binding Site (Ϫ284/Ϫ278)-A single protein-DNA complex was clearly induced in nuclear extracts from rhinovirus-infected A549 epithelial cells compared with noninfected cells, induction being maximal at 30 min and fading thereafter (data not shown). Competition experiments with specific and consensus AP-1 competitors completely abrogated the signal, while an irrelevant (Sp1) competitor did not, confirming the AP-1 specificity of the signal (data not shown). These data suggest that proteins binding to the AP-1 motif at Ϫ284/Ϫ278 in the ICAM-1 promoter are also induced in the nuclei of A549 cells by rhinovirus infection.
Ϫ74 to Ϫ43 Probe Containing an NF-B Binding Motif (Ϫ62/Ϫ53) Overlapping with an Sp1 Binding Motif (Ϫ59/ binding to the Ϫ199/Ϫ170 (containing a C/EBP and an NF-B binding site, Table II) portion of the ICAM-1 promoter was assessed by EMSA. Nuclear extracts were prepared from uninfected (time 0) and rhinovirus 16-infected A549 cells at various time points (30,60,90, and 120 min) after infection and incubated with radiolabeled Ϫ199/Ϫ170 ICAM-1 (Table II) probe. Resolution of binding complexes was accomplished on 5% polyacrylamide gels. A representative radiograph of one of at least three separate experiments is shown. Two retarded complexes binding to the Ϫ199/Ϫ170 portion of the ICAM-1 promoter were induced in nuclei from rhinovirus-infected cells, with peak induction of binding activity observed within 30 min of rhinovirus infection. Induction of binding activity by rhinovirus reduced gradually from 30 min but was still present up to 120 min after infection. B, the specificity of the binding activity induced by rhinovirus infection of A549 cells was examined by EMSA with competition studies. Nuclear extracts from A549 cells uninfected (lane 5) or rhinovirus 16 infected for 30 min (lanes 1-4) were incubated with radiolabeled Ϫ199/Ϫ170 ICAM-1 probe in the absence (lane 1) or in the presence of excess unlabeled specific Ϫ199/Ϫ170 probe (SC, lane 2) and excess unlabeled probes containing consensus NF-B and AP-1 binding sequences (lanes 3 and 4, respectively). Resolution of binding complexes was accomplished on 5% polyacrylamide gels. A representative radiograph of one of at least three separate experiments is shown. Specificity of the binding activity induced by rhinovirus infection of A549 cells to the Ϫ199/Ϫ170 portion of the ICAM-1 promoter was confirmed by complete inhibition of the induction of binding activity with excess unlabeled specific probe (lanes 1 and 2).  Ϫ54)-Two retarded complexes were induced in nuclear extracts from rhinovirus-infected A549 cells, with a maximal induction again observed at 30 min (data not shown). This binding activity could be competed away by excess unlabeled specific competitor identical oligonucleotide or by excess unlabeled oligonucleotide containing an NF-B consensus binding motif (data not shown). The binding was unaffected by a excess unlabeled double-stranded oligonucleotides containing an Sp1 consensus binding motif, suggesting that rhinovirus infection of A549 cells induces nuclear proteins binding to the Ϫ62/Ϫ53 NF-B site of the ICAM-1 promoter but not the Ϫ59/Ϫ54 Sp1 site.
Having observed induction of proteins capable of binding to different transcription factor binding sites (both NF-B and AP-1) within the ICAM-1 promoter, we then carried out reporter gene assays to determine which of the potential candidate transcription factor binding sites were functional in rhinovirus induction of ICAM-1 promoter activity.
Identification of the Ϫ187/Ϫ178 NF-B Site as the Rhinovirus Response Region in the ICAM-1 Promoter-CAT constructs containing serial deletions of the ICAM-1 promoter were studied to identify the ICAM-1 promoter regions involved in rhinovirus induction of ICAM-1 promoter activity. As seen in Fig. 11, CAT constructs under the control of the proximal Ϫ1160 and Ϫ277 bp of the ICAM-1 promoter were strongly and similarly induced by rhinovirus 16 infection of A549 cells. However, further deletions of the ICAM-1 promoter to Ϫ182 bp or shorter completely abolished the capacity of rhinovirus infection to induce ICAM-1 promoter activity. These studies indicated the presence of DNA sequences necessary for rhinovirus induction of ICAM-1 promoter activity between positions Ϫ277 and Ϫ182 relative to the transcription initiation site.
The Ϫ187/Ϫ178 NF-B binding motif, which is already known to play a role in ICAM-1 induction by cytokines (31,32), is located within this region of the ICAM-1 promoter, and its sequence is truncated by the Ϫ182 deletion. Furthermore, the EMSAs clearly demonstrated rhinovirus induction of nuclear proteins binding to this site. Therefore, for further investigations, constructs were used that specifically tested this site.
First, we examined whether DNA fragments containing either the complete binding motif or the Ϫ182 deletion of the motif could confer responsiveness to rhinovirus 16 in a heterologous promoter, the herpes simplex minimal thymidine kinase promoter contained in the plasmid pBRAMScat2 (31). Constructs containing the complete NF-B Ϫ187/Ϫ178 site (pBRAMScat2 Ϫ199/Ϫ170 ICAM-1), the truncated site (pBRAMScat2 Ϫ199/Ϫ182 ICAM-1), and the minimal promoter alone with no ICAM-1 promoter sequence (pBRAMScat2) were transfected in A549 cells. Fig. 12 shows that only when the entire ICAM-1 Ϫ187/Ϫ178 NF-B site is present (pBRAMScat2 Ϫ199/Ϫ170 ICAM-1), is the heterologous thymidine kinase promoter responsive to rhinovirus, confirming that this site is required intact for rhinovirus induction of ICAM-1 promoter activity and that it is also sufficient in the presence of a heterologous minimal promoter.
Finally, to investigate whether the Ϫ187/Ϫ178 NF-B site in the ICAM-1 promoter is essential for rhinovirus induction of ICAM-1 promoter activity to occur, constructs containing either the longest ICAM-1 promoter sequence (Fig. 13, Ϫ1160 ICAM-1) or the same construct with mutations from TG-GAAATTCC to TctAgATTag at the Ϫ187/Ϫ178 NF-B site (Ϫ1160 mICAM-1) were used to transfect A549 cells. As shown in Fig. 13, mutation of the Ϫ187/Ϫ178 NF-B site completely abrogated rhinovirus induction of ICAM-1 promoter activity, confirming that this NF-B binding site is necessary for rhinovirus induction of ICAM-1 gene transcription.

DISCUSSION
In these studies, we have investigated mechanisms involved in rhinovirus-induced asthma exacerbations by studying the effect of rhinovirus infection on airway epithelial cell ICAM-1 expression. These studies were performed, since ICAM-1 is the receptor for 90% of rhinoviruses and is an adhesion protein that has a central role in inflammatory cell recruitment to the lower airway following rhinovirus infection. ICAM-1 is therefore likely to play a very important role in the mechanisms of virus-induced asthma exacerbations.
We have demonstrated that rhinovirus infection of both primary bronchial epithelial cells and the type II respiratory epithelial cell line A549 markedly increases cell surface expression of ICAM-1. We then investigated the mechanisms of rhinovirus regulation of ICAM-1 expression in A549 cells and observed induction of ICAM-1 promoter activity and increased ICAM-1 mRNA transcription. The rhinovirus induction of ICAM-1 promoter activity was found to involve up-regulation of NF-B proteins binding to the Ϫ187/Ϫ178 NF-B binding site on the ICAM-1 promoter, and this site was required intact for rhinovirus up-regulation to occur. The principal component of the rhinovirus-induced NF-B-binding proteins were p65 subunits. These studies elucidate mechanisms probably involved in rhinovirus induction of asthma exacerbations and identify ICAM-1 and NF-B p65 as potential new targets for development of therapeutic intervention strategies for virus-induced asthma.
Our initial studies demonstrated that rhinovirus infection of A549 respiratory epithelial cells increased the cell surface expression of ICAM-1 protein in a dose-response manner, with peak induction occurring at an MOI of 1 (Fig. 1). The observed increase in ICAM-1 expression peaked at 8 h after virus inoc- ulation but remained elevated above noninfected cells for up to 72 h after inoculation (Fig. 2).
We then investigated the group and serotype specificity of the induction of ICAM-1 and demonstrated that both major group and minor group rhinoviruses and at least three different serotypes are equally able to up-regulate ICAM-1 surface protein expression (Fig. 4). Since major and minor group viruses bind to different cellular receptors (26,27,35), the observed induction of ICAM-1 by rhinoviruses is clearly not receptor-or serotype-restricted. These observations are in keeping with previous observations relating to rhinovirus induction of IL-8 (14,16,17) and IL-6 (15,17) and suggest that the mechanisms involved are likely to have broad applicability across all rhinovirus serotypes. Furthermore, since rhinoviruses are responsible for two-thirds of asthma exacerbations in which a virus is identified (2)(3)(4), the mechanisms involved are likely to be pertinent to the majority of asthma exacerbations.
The observation that rhinovirus infection of respiratory epithelial cells up-regulated ICAM-1 expression is of great interest, since this molecule is not only important in inflammatory cell recruitment and activation in asthma (23)(24)(25) but is also the receptor for the major group (90%) of rhinoviruses (26,27). Induction of increased expression of its own cell surface receptor is an unusual property for viruses, since previous observations demonstrate that virus infection down-regulates expression of virus cell surface receptors. For example, measles virus and human immunodeficiency virus both induce down-regulation of their receptors CD46 and CD4, respectively, a process thought to prevent superinfection (40,41). Our findings that rhinoviruses induce increased expression of their own receptor suggest that the converse occurs and that rhinovirus infection may render cells more, rather than less, susceptible to infection by other major group virus particles. A recent report goes some way toward confirming this hypothesis, in that rhinovirus infection of primary cultures of human tracheal epithelium was found to increase ICAM-1 mRNA expression and similar magnitude increases in ICAM-1 mRNA expression induced by IL-1␤ were found to increase susceptibility to rhinovirus infection (42). We were therefore interested in investigating whether our finding of rhinovirus-induced increased ICAM-1 expression in A549 cells could also be observed in primary bronchial epithelial cells. We observed a much greater induction of ICAM-1 protein (12-versus 3-fold) in the primary cells ( Figs. 1-4), suggesting that any observations made in A549 cells might actually be more rather than less marked in primary cells. Given that rhinovirus infection of respiratory epithelial cells is low grade and noncytopathic (16), induction of the viral receptor is likely to render the cells more susceptible to infection and, therefore, further ICAM-1 up-regulation, creating a "vicious cycle" of rhinovirus infection and ICAM-1 upregulation. The effect of rhinovirus infection on the expression of the minor group rhinovirus receptor (35) has not been studied, but similar mechanisms could in theory also operate for minor group rhinoviruses. This process, combined with increased ICAM-1 expression in asthma (20 -22) is likely to play an important role in increasing the susceptibility of asthmatic subjects to lower airway rhinovirus infections.
We have previously reported rhinovirus induction of IL-8 protein and mRNA from rhinovirus-infected A549 cells and used sICAM-coated and UV-inactivated virus to investigate the relative contributions of virus-receptor binding and virus replication to the observed up-regulation of IL-8 (16). We reported that UV inactivation reduced by about two-thirds, while precoating the virus with soluble receptor completely abrogated, the induction of IL-8. These data suggest that part of the observed up-regulation of IL-8 was related purely to virus- receptor binding, while the remainder required viral replication (16). In the present studies, we observed the same findings relating to sICAM and UV inactivation to be true for rhinovirus induction of ICAM-1 and have extended them to demonstrate that filtering virus particles but not substances with a molecular mass of Ͻ30 kDa (most cytokines) from the inoculum completely abrogated the observed ICAM-1 induction (Fig. 4). We also confirmed the receptor specificity of the sICAM inactivation by demonstrating that precoating a minor group (rhinovirus 2) virus with sICAM had no effect on ICAM-1 upregulation (Fig. 4). These additional data add further weight to our previous hypothesis that part of the signal to up-regulate ICAM-1 or IL-8 protein synthesis occurs consequent to virusreceptor binding, but the major part occurs through processes associated with viral replication.
The ability of rhinovirus infection to up-regulate epithelial cell surface expression of ICAM-1, an important molecule in asthma pathogenesis (23), may have particular importance in the mechanisms of virus-induced asthma exacerbations independent of the effects on rhinovirus replication. We have previously demonstrated that rhinovirus colds induce bronchial mucosal CD3 ϩ , CD4 ϩ , and CD8 ϩ lymphocyte and eosinophil infiltration, with a more persistent intraepithelial eosinophilia in asthmatic subjects (7). Epithelial expression of ICAM-1 is likely to play an important function in retaining both types of inflammatory leukocyte in the epithelium by binding to its ligands CD18/11a and CD18/11b on lymphocytes and granulocytes, respectively. In addition, binding of ICAM-1 to its integrin ligands on leukocytes may activate these cells and lead to secretion of proinflammatory cytokines and mediators (43)(44)(45)(46). Induction of ICAM-1 expression on respiratory epithelial cells is therefore likely to be an important mechanism regulating the bronchial mucosal CD3 ϩ , CD4 ϩ , and CD8 ϩ lymphocyte and eosinophil infiltration observed in rhinovirus infections. Given the important regulatory role of lymphocytes in promoting airway inflammation in asthma, the induction by rhinoviruses of ICAM-1 is a mechanism that represents an attractive target for development of therapeutic interventions aimed at reducing inflammatory cell recruitment and activation in virus-induced asthma exacerbations.
The observations on the time course and lack of receptor or serotype restriction of ICAM-1 induction by rhinoviruses are in keeping with previous reports of rhinovirus induction of proinflammatory cytokine expression in respiratory epithelial cells (14 -17). These similarities suggest that there may be common mechanisms involved in the induction of several proinflammatory proteins by rhinoviruses and that further investigation of the cellular/molecular mechanisms involved might lead to the identification of common pathways suitable for targeting of further future therapeutic intervention strategies. Therefore, having demonstrated that rhinovirus infection of respiratory epithelial cells increased ICAM-1 surface protein expression, we wished to investigate the effects of rhinovirus infection on A549 cell ICAM-1 mRNA expression to elucidate the cellular mechanisms involved in more detail. We observed rhinovirus induction of ICAM-1 mRNA occurring within 1 h of virus inoculation and peaking at 8 h (Fig. 6, Table III). Rhinovirus induction of ICAM-1 mRNA was observed up to 24 h post-virus inoculation (no studies were carried out beyond this time point). As we had observed with ICAM-1 protein expression, UV inactivation partially abrogated the rhinovirus-induced ICAM-1 mRNA expression while inactivating the virus by filtration or by precoating with soluble receptor completely abrogated the signal (Fig. 7, Table IV). These studies confirmed that, as with ICAM-1 protein expression, rhinovirus induction of ICAM-1 mRNA was also consequent partly upon virus-receptor binding and was partly related to viral replication.
Having observed rhinovirus-induction of both ICAM-1 protein and mRNA expression, we hypothesized that rhinovirus infection of A549 cells increased ICAM-1 expression by upregulating ICAM-1 gene transcription. To investigate this possibility, we analyzed in vitro transcription of ICAM-1 mRNA in rhinovirus-infected and noninfected cells. As seen in Fig. 8 and Table V, we observed clear induction of ICAM-1 gene transcription by rhinovirus infection and inhibition of this induction by an inhibitor of RNA polymerase II, ␣-amanitin. These data confirmed that rhinovirus infection of A549 respiratory epithelial cells rapidly increased de novo transcription of ICAM-1 mRNA.
Next, we wished to determine the molecular mechanisms involved in rhinovirus induction of ICAM-1 mRNA transcription, since these mechanisms might identify a target for development of new therapeutic intervention strategies. We focused on transcription factor-mediated activation of the ICAM-1 promoter, since previous observations indicated that both IL-6 induction by rhinoviruses (15) and IL-8 induction by respiratory syncytial virus (47) were dependent on proteins binding to NF-B sites in the relevant promoters, and our own observations had indicated that the presence of both an NF-B site and an AP-1 site was required for induction of the IL-8 promoter by rhinovirus infection (48).
The ICAM-1 promoter contains several potential transcription factor binding sites, several of which have been implicated in the induction of ICAM-1 gene transcription in response to various proinflammatory stimuli, such as PMA and cytokines (31,32). In the present study, we observed that rhinovirus infection of A549 epithelial cells induced proteins binding to FIG. 13. The ؊187/؊178 NF-B element of the ICAM-1 promoter is essential for rhinovirus induction of ICAM-1 gene transcription in A549 respiratory epithelial cells. A549 cells were transiently transfected with CAT constructs containing full-length ICAM-1 promoter (Ϫ1160 ICAM-1) or full-length ICAM-1 promoter mutated at the Ϫ187/Ϫ178 NF-B site (Ϫ1160 mICAM-1). Transfected cells were incubated with medium alone (C) or rhinovirus 16 at an MOI of 1 for 24 h (RV). Epithelial cells were harvested, and CAT activity in proteinequivalent cell lysates was assessed as described under "Materials and Methods." ICAM-1 promoter activation is expressed as -fold induction CAT activity in infected over control cells. Data are mean Ϯ S.E. of four separate experiments. Rhinovirus infection of A549 respiratory epithelial cells induced markedly increased ICAM-1 promoter activity of fulllength (Ϫ1160 ICAM-1) promoter. This rhinovirus responsiveness was completely abolished by mutation of the Ϫ187/Ϫ178 NF-B site (Ϫ1160 mICAM-1), confirming that this site is essential for rhinovirus induction of ICAM-1 promoter activity in A549 cells to occur. the both the Ϫ187/Ϫ178 and the Ϫ62/Ϫ53 ICAM-1 NF-B sites ( Fig. 9 and data not shown) and the Ϫ284/Ϫ278 AP-1 site (data not shown) but neither the Ϫ206/Ϫ201 nor Ϫ59/Ϫ54 Sp1 site. We therefore performed reporter gene assays to determine which of these sites was functional in rhinovirus induction of ICAM-1 promoter activity. We observed that sequential deletion of the ICAM-1 promoter up to Ϫ277 base pairs from the transcription initiation site had no effect on rhinovirus induction of ICAM-1 promoter activity (Fig. 11). These data suggest that despite the fact that proteins binding to this site are induced by rhinovirus infection of A549 cells, the Ϫ284/Ϫ278 AP-1 site is not required for rhinovirus induction of ICAM-1 promoter activity to occur. In contrast, deletion of the promoter to Ϫ182 bases resulted in complete loss of rhinovirus inducibility, suggesting that elements contained within the Ϫ182 to Ϫ277 region were necessary for, and that the Ϫ62/Ϫ53 NF-B site alone was insufficient for, rhinovirus-induced up-regulation of ICAM-1 promoter activity to occur (Fig. 11).
The Ϫ182 deletion interrupts the Ϫ187/Ϫ178 NF-B site. Given the previously observed functionality of this site in ICAM-1 induction by cytokines (32), we hypothesized that this site might also be important in rhinovirus induction of ICAM-1 promoter activity. Mutational analysis was therefore carried out, with reporter gene assays being performed with full-length ICAM-1 promoter and with full-length ICAM-1 promoter with the Ϫ187/Ϫ178 NF-B site mutated in four nucleotide positions. Despite the presence of the full-length promoter, mutation of this Ϫ187/Ϫ178 NF-B site completely abrogated rhinovirus induction of ICAM-1 promoter activity (Fig. 13), confirming that this site was required intact for rhinovirus induction of ICAM-1 promoter activity to occur. The importance of this Ϫ187/Ϫ178 NF-B site in rhinovirus induction of promoter activity was then investigated using plasmids containing the reporter gene linked to the thymidine kinase minimal promoter alone and with the truncated and intact versions of the Ϫ187/Ϫ178 NF-B site (Fig. 12). These data confirmed that this site was required intact for rhinovirus induction of a heterologous promoter to occur and that it was sufficient in the presence of a basic minimal promoter.
Previous studies have demonstrated that members of the NF-B family of transcription factors are important in induction of proinflammatory cytokines by both rhinovirus and respiratory syncytial virus (15,47,48). These data suggest that NF-B may play a very important role in the induction of proinflammatory cytokines by virus infections in general, and our data reported herein extend these observations to include proinflammatory adhesion molecules. The NF-B/Rel family of transcription factors contains several members, so far including p50, p52, p65, c-Rel, and Rel-B, which are capable of forming homo-or heterodimers. We performed supershift experiments to investigate which members of the NF-B/Rel family were induced by rhinovirus infection and demonstrated that the major component of rhinovirus-induced proteins binding to the Ϫ187/Ϫ178 NF-B site in the ICAM-1 promoter were p65 proteins, with smaller amounts of p50 and c-Rel (Fig. 10). These data support previous observations that the major rhinovirus-induced proteins binding to the IL-6 promoter in A549 cells are also p65 homodimers or heterodimers (15).
In addition to the early important study in primates by Wegner et al. (23), two further recent studies have reported important roles for ICAM-1 in promoting inflammation in asthma by demonstrating its important role in promoting lymphocyte and eosinophil infiltration in murine models (24,25). Given the marked bronchial lymphocyte and eosinophil infiltration observed in virus-induced asthma (7), its role as the rhinovirus receptor, and our demonstration that rhinovirus infection induces increased ICAM-1 expression in respiratory epithelial cells, we believe that ICAM-1 is likely to play a critical role in promoting lower airway inflammation in virusinduced asthma. This hypothesis is supported by the fact that we 2 and others (49) have observed increased sICAM levels in nasal secretions during rhinovirus infections. Recent support for the role of ICAM-1 up-regulation in rhinovirus-induced asthma also comes from the demonstration that experimental rhinovirus infection of asthmatic subjects up-regulates ICAM-1 expression on bronchial epithelial cells in vivo (50). These data make epithelial ICAM-1 a prime target for therapeutic intervention strategies in virus-induced asthma exacerbations. The fact that anti-ICAM-1 monoclonal antibodies were able to reduce allergen-induced airway hyperreactivity and eosinophil influx in a primate model of asthma (23) also strongly supports the potential of ICAM-1-targeted intervention to ameliorate virus-induced asthma exacerbations.
We have also investigated the mechanisms of rhinovirus induction of ICAM-1 and have observed an important role for NF-B-mediated transcriptional up-regulation, the major component of which is contributed by p65. Given that transcriptional up-regulation via NF-B is also necessary for induction of IL-6 (15) and IL-8 (48) by rhinoviruses and that in the case of IL-6 p65 is again the major component, this molecule also represents an important potential target for future therapeutic intervention in virus-induced asthma exacerbations.
In conclusion, we have demonstrated that rhinovirus infection of respiratory epithelial cells increases surface ICAM-1 expression via NF-B p65-mediated transcriptional up-regulation. We believe that these two molecules (ICAM-1 and NF-B p65) represent new targets for potential therapeutic intervention in virus-induced asthma exacerbations.