Expression of ICAM-1 in Nasal Epithelium and Levels of Soluble ICAM-1 in Nasal Lavage Fluid During Human Experimental Rhinovirus Infection

Most rhinovirus serotypes use intercellular adhesion molecule-1 (ICAM-1) as the receptor to enter cells, but ICAM-1 expression has not been detected on normal nasal epithelial cells. During experimental rhinovirus infection, expression of ICAM-1 on nasal epithelial cells was examined with immunohistochemical staining of nasal scrape biopsy specimens, and levels of soluble ICAM-1 in nasal lavage fluid were measured by sandwich enzyme-linked immunosorbent assay technique. Expression of ICAM-1 on nasal epithelial cells increased following inoculation in 20 of 23 infected subjects. The median number of ICAM-1–positive cells per 6.25-mm² area of stained biopsy specimen was 0 in control samples (day 20 or 33 after inoculation), and in those without infection, 6 on day 1 (P≤.05), 14.5 on day 3 (P≤.01), 1.5 on day 5, and 0 on day 9. In a different group of volunteers, soluble ICAM-1 in nasal lavage fluid was higher on days 1 and 3 compared with preinoculation levels (P≤.001), but only 11 of 23 infected subjects had a 2-fold or greater increase. Up-regulation of ICAM-1 receptor expression on nasal epithelial cells occurred within 24 hours after inoculation in experimental rhinovirus infections (prior to onset of symptoms) and declined promptly by day 5.

Most human rhinoviruses (HRVs) (the major receptor group) gain entrance to nasal epithelial cells by a specific cellular receptor that has been identified as intercellular adhesion molecule-1 (ICAM-1), a member of the immunoglobulin superfamily.^1-3 The natural ligand for the β₂ integrin leukocyte function-associated antigen-1 (CD11a/CD18), which is widely expressed on leukocytes and various other cells, ICAM-1 (CD54) can be up-regulated not only on endothelial cells but also on epithelial cells in the airways. In vitro, a 12-fold increase in ICAM-1 cell surface expression has been demonstrated in HRV-infected primary bronchial epithelial cells.⁴ Expression of ICAM-1 also occurs on cultured human nasal epithelial cells⁵ and on epithelial and endothelial cells of nasal biopsy specimens from subjects during inflammatory states, such as allergic rhinitis.^6,7 However, currently used methods have not detected ICAM-1 expression in noninflamed nasal epithelium.⁵ An immunohistochemical study of human adenoid epithelium has found that constitutive expression of ICAM-1 is not homogeneous but localized to the surface of a small number of cells.⁸

Shedding of an extracellular portion of ICAM-1, or soluble ICAM-1 (sICAM-1) occurs as evidenced by its detection in certain body fluids. The level of soluble ICAM-1 has been detected in sera of healthy individuals⁹ and is elevated in the peripheral blood of patients with bronchial asthma.¹⁰ The level of soluble ICAM-1 is also elevated in nasal lavage fluid (NAL) from allergic patients during allergy season¹¹ and has also been detected in NAL from patients with naturally acquired rhinovirus colds.¹²

The role of ICAM-1 during rhinovirus infection may be complex. Increased ICAM-1 expression may facilitate viral replication⁴; ICAM-1 enriched HeLa-I cells support viral growth to a greater extent than other cell types.¹³ Soluble ICAM-1 inhibits HRV replication in cell culture,^14,15 suggesting that endogenously produced sICAM-1 may exert an antiviral effect and modify the course of HRV infection. Increases in polymorphonuclear leukocyte concentrations in submucosal and nasal secretion are an integral part of the inflammatory response to rhinovirus infection¹⁶; blockade of cell-associated ICAM-1 by antibody has been shown to inhibit polymorphonuclear leukocyte and eosinophil migration through endothelial cells in vitro.^17-20 In addition, sICAM-1 could interfere with normal intercellular interactions and modify the immune response to HRV infection. In the present study, we examined ICAM-1 expression on nasal epithelial cells and concentrations of sICAM-1 in NAL during experimental HRV infection.

Healthy young adults with serum-neutralizing antibody titer of 1:2 or less to HRV serotypes 14 or 39 were recruited from the University of Virginia, Charlottesville, community. All volunteers provided written informed consent approved by the human investigation committee at the University of Virginia.

Levels of sICAM-1 were determined in 24 volunteers inoculated with HRV 39. Studies for ICAM-1 expression on nasal epithelial cells were conducted in separate HRV studies with 33 adult volunteers as previously reported.²¹ Because samples were collected in separate studies, contamination of NAL with serum ICAM-1 due to trauma from the biopsy was avoided.

Two different safety-tested virus pools were used: HRV 39 and HRV 14. The rhinovirus pools were cell culture harvests of clinical isolates grown in human embryonic lung fibroblast cultures²² and have been safety tested. Subjects were inoculated by nasal drops with a volume of 100 µL per nostril repeated once for a total inoculum of approximately 300 median tissue culture infective dose per subject. After virus inoculation, the volunteers were isolated in hotel rooms and monitored daily for illness on days 1 through 5.

For measurement of sICAM-1, NAL samples were obtained before virus inoculation and daily in the morning. Isotonic saline (5 mL per nostril) was instilled in the nasal cavity with the patient's head hyperextended. After 5 to 10 seconds the saline was expelled through the nose into a cup, with recovery of 6 to 8 mL. An aliquot (4 mL) was added to a 1-mL volume of concentrated (×5) viral transport medium. This sample was also used for isolation of virus by standard cell culture techniques.

For the nasal epithelial study, nasal secretions were blown onto plastic sheets. Swabs of these secretions and throat swabs were collected 1 day before inoculation with HRV and on each day thereafter and placed together in viral transport medium for HRV isolation. This was done to avoid any possible effect of the nasal-washing procedure on epithelial ICAM-1 expression.

Aliquots of NAL were frozen and stored at −70°C for up to 3 months and underwent 1 to 2 freeze-thaw cycles prior to sICAM-1 analysis. Soluble ICAM-1 was measured by a sandwich enzyme-linked immunosorbent assay technique (Bender Med Systems, Vienna, Va). The lower detection level was 3 ng/mL of NAL. Assays were done under masked conditions without knowledge of infection and illness status of the volunteers.

Four nasal scrape biopsies were performed with a plastic curette from each individual from the anterior part of the inferior turbinate on days 1, 3, and 5. The fourth biopsy was performed on days 9, 20, or 33 after inoculation. Biopsy specimens were immediately fixed in buffered 4% paraformaldehyde, included in a human plasma clot, and embedded in paraffin for sectioning.²¹

Tissue sections mounted on poly-L-lysine–coated slides were deparaffinized, treated with methanol peroxide to block endogenous peroxidase for 10 minutes, and then immersed in 10mM citrate buffer (pH 6.0) and boiled in the microwave oven for 30 minutes on the highest setting, replenishing the volume with distilled water every 5 minutes. The slides were allowed to cool for 10 to 15 minutes before being loaded onto the automated immunostainer (Ventana Enhanced System, Ventana Medical Systems, Tucson, Ariz). The diaminobenzidine immunohistochemistry procedure was done according to the instructions for the operation of the machine. Slides were incubated for 32 minutes with the primary monoclonal antibody anti-CD54 (Dako, Carpinteria, Calif) at a 1:20 dilution and then sequentially incubated with the following reagents of the diaminobenzidine detection system: endogenous biotin blocker, secondary biotinylated antibody, avidin-horseradish peroxidase, and diaminobenzidine complex, and a copper-enhancing solution for improved visualization of the diaminobenzidine reaction product. All washes between steps were done with the wash buffer provided by the manufacturer. Sections were counterstained with Harris hematoxylin. Expression of ICAM-1 on vascular endothelial cells in sections of human adenoid was used as a positive control. A section from each biopsy specimen was stained in parallel with the negative control irrelevant antibody provided by the instrument's manufacturer.

One section from each of the 4 nasal epithelial biopsy specimens from 33 subjects (132 sections) was analyzed by light microscopy under blinded conditions. The number of ICAM-1 expressing cells was counted in an area of 6.25 mm² using an eyepiece with a cell-counting grid. Only areas with a continuous sheet of epithelium that covered the entire area of the grid (6.25 mm²) were included in the analysis.

Nasal lavage fluid was inoculated onto monolayers of human embryonic lung fibroblast cells (MRC-5 and/or WI-38 strain) in screw-capped tubes for rhinovirus detection.²³ Isolates recovered were confirmed by immunologic methods (neutralization assay); paired serum samples collected before and 3 weeks after virus inoculation were used for measuring neutralizing antibodies to HRV.

A volunteer was considered infected if he/she shed the virus and/or had a 4-fold or greater rise in serum neutralizing antibody titer. A volunteer was considered to have illness if he/she had a minimum symptom score of 6, based on a modification of the Jackson respiratory illness scoring system.²⁴ Sham-inoculated volunteers who developed cold symptoms had nasal secretions tested for HRV by reverse transcriptase–polymerase chain reaction²⁵ and excluded from analysis if found positive.

Statistical analyses of sICAM-1 levels in NAL were performed using SAS 6.12 (SAS Institute Inc, Cary, NC). Soluble ICAM-l levels were modeled separately for the HRV-positive group using repeated-measures analysis of variance. The underlying assumptions and the impact of data transformation and outliers were investigated in sensitivity analyses. Analyses of sICAM-1 level to the −2 power are reported here because these residuals were the only ones that satisfied the normality assumptions. However, all repeated-measures analyses and, in addition, simple Wilcoxon signed rank tests of differences from day 0 led to the same qualitative results. The association between increases in sICAM-1 levels before and after inoculation and other measures (Jackson cold status, viral shedding, and seroconversion) was assessed by dichotomizing change in sICAM-1 level (below and above a 2-fold increase) and applying a Fisher exact test. The association between 2-fold increases in sICAM-1 level and the median total mucus weight was tested using a Wilcoxon rank sum test. Cellular expression of ICAM-1 in nasal biopsy specimens of a different group of HRV-infected subjects compared with noninfected sham-inoculated subjects was analyzed by the Mann-Whitney test. Statistical significance was indicated for P values of .05 or below without adjustment for multiple comparisons.

In the study of sICAM-1 elaboration, 21 (88%) of 24 volunteers became infected, 2 were not infected, and 1 had a wild rhinovirus detecg on the first study day. This volunteer was excluded from the analysis. Of the HRV-infected volunteers, 20 (95%) of 21 had colds by standard criteria. To evaluate ICAM-1 on epithelial cells, 23 (88%) of 26 HRV-inoculated subjects developed infection and 18 (78%) of 23 infected developed clinical colds. One sham-inoculated volunteer who was symptomatic and found positive for HRV by reverse transcriptase–polymerase chain reaction was excluded.

Levels of sICAM-1 appeared to increase following virus inoculation in the HRV-infected subjects (Figure 1). These increases were significant on day 1 (P<.001) and day 3 (P<.001) after inoculation compared with preinoculation (day 0) levels in the HRV group, while the increase at day 5 did not quite achieve significance (P = .06). Overall, 11 (52%) of 21 HRV-infected subjects showed a 2-fold or greater increase in NAL sICAM levels during the 5 days following HRV inoculation. There was no detectable association between a 2-fold increase in sICAM level in NAL and symptomatic colds by the Jackson criteria, mucus weight, viral shedding, or seroconversion.

Nasal epithelial cell expression of ICAM-1 (Figure 2) varied in relation to time of the infection (Table 1). Twenty of 23 HRV-infected volunteers had an increase in number of ICAM-1–positive nasal epithelial cells. The increase in ICAM-1 expression had already occurred on days 1 and 3 following inoculation, and the numbers of ICAM-1–expressing cells decreased by day 5. From the same volunteers, the control biopsy specimens obtained on day 20 or 33 were uniformly negative except for 1 (subject 11, Table 1). Unfortunately, no viral cultures were collected at this time. Only 3 volunteers (subjects 1, 4, and 7) did not have an increase above the range of the control biopsy specimens. The median numbers of cells expressing ICAM-1 per 6.25-mm² area of biopsy material obtained on days 1 and 3 were significantly increased (day 1, P≤.05; day 3, P≤.01) in subjects infected with HRV compared with the noninfected sham-inoculated subjects. No significant differences in the number of ICAM-1–positive cells were found between subjects with cold symptoms and those without colds. Few cells expressing ICAM-1 (median, 0; range, 0-8 per 6.25-mm² area) were detected in the nasal biopsy specimens from 5 of the 6 sham-inoculated subjects. One sham-inoculated subject who did not develop infection or cold symptoms had a higher number of ICAM-1–positive cells (22 per 6.25-mm² area) on day 1 (Table 1). Unfortunately, material from that subject was insufficient for analysis on days 3 and 5. On day 33 after infection, the nasal biopsy specimen from that subject still showed a high number of ICAM-1–positive cells (8 per 6.25-mm² area), suggesting a higher expression of ICAM-1, either constitutive or secondary to subclinical inflammation.

This study demonstrates that ICAM-1 expression on nasal epithelial cells was up-regulated within 24 hours following HRV inoculation as determined by immunohistochemical staining and decreased rapidly to baseline by day 9. This was mirrored by an increase in sICAM-1 levels in NAL, although the changes were modest and only significantly different on day 3 following HRV inoculation. Previous information on ICAM-1 in the upper airway mucosa during HRV infection is sparse. Hildebrandt and coworkers¹² found an increase in sICAM-1 levels in NAL during the acute phase of naturally acquired colds compared with convalescence. The magnitude of change observed in that study over time was comparable with the increase observed in the present study. It is of interest that the increase in sICAM-1 levels in our study did not correlate with occurrence of illness, nasal mucus weight, virus shedding, or seroconversion, and that it occurred prior to the full development of peak symptoms.

Expression of ICAM-1 in the airway mucosa is very low in the noninflamed nose,⁶ nasopharynx,⁸ and bronchial mucosa^26-29 under basal conditions. However, during inflammation ICAM-1 expression can be up-regulated not only on the endothelial cells but also on epithelial cells in the airway mucosa.^5,30-33 The present study also found low-level expression of ICAM-1 on nasal epithelial cells under baseline conditions but increased expression during the early phase of HRV infection. Expression of ICAM-1 in airway epithelium and endothelial cells in mice is up-regulated in response to various stimuli including ozone exposure,³⁴ interleukin 5,³⁵ tumor necrosis factor α, interleukin 1,³⁶ and activated CD8⁺ T cells.³⁷ Christensen et al³⁸ demonstrated a correlation between elevated circulating ICAM-1 levels in serum and virus-induced T-cell activation during lymphocytic choriomeningitis virus infection, which suggested release of ICAM-1 into the bloodstream. It is not clear if the source of sICAM-1 detection in NAL following HRV infection is from up-regulated ICAM-1 expression on nasal epithelial cells or serum transudation. In asthmatic subjects sICAM-1 levels in bronchoalveolar lavage fluid following experimental antigen stimulation were higher than what could be explained by serum transudation. The bronchoalveolar lavage to serum ratio of sICAM-1 was 250 times higher than the ratio of albumin, which suggested that sICAM was derived from additional sources in the lower respiratory tract.³⁹ Our demonstration of increased ICAM-1 expression on nasal epithelial cells also points to the mucosa as the likely source of sICAM-1 in nasal secretions.

The biological significance of increased ICAM-1 expression on nasal epithelial cells and sICAM-1 shedding in NAL fluid is not fully understood. Increased cell surface ICAM-1 expression could result in increased viral replication, whereas increased sICAM-1 in NAL conceivably may inhibit HRV receptor–binding in vivo, analogous to HRV inhibition by recombinant sICAM-1 in vitro. Exogenous application of sICAM-1 has been shown to reduce viral replication and illness following experimental HRV infection.⁴⁰ We did not find any clear relationship between the surface expression of ICAM-1 and expression of clinical colds of experimental HRV infection. However, the strength of this conclusion is limited by our relatively small sample size.

Cell surface expression of ICAM-1 may have a central role in recruitment of polymorphonuclear leukocyte traffic from the vascular bed through the airway epithelium and into nasal mucus. In the lower airways, ICAM-1 has been demonstrated on the alveolar surface in areas undergoing alveolar repair, suggesting that ICAM-1 may participate in controlling the movement of leukocytes onto and over the alveolar surface.⁴¹ Further studies are needed to evaluate how ICAM-1 expression relates to viral replication and the cytokine response in the upper airway tract mucosa to determine if up-regulation of HRV's "own" human cell surface receptor (ICAM) fosters virus replication in a feedback mechanism.

Accepted for publication September 19, 2001.

Corresponding author and reprints: Frederick G. Hayden, MD, University of Virginia Health System, Department of Internal Medicine, Box 800473, Charlottesville, VA 22908 (e-mail: fgh@virginia.edu).