INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Respiratory infections with human rhinoviruses (RV) are frequent triggers of cough and wheezing in children and adults with asthma. Soon after the onset of these illnesses, immune responses are triggered that usually eradicate viral replication within days. However, since RV does not destroy appreciable amounts of airway epithelium, it is presumed that the immune response to the virus also contributes to the pathogenesis of respiratory symptoms, including dysfunction of the lower respiratory tract in asthma.

This hypothesis is supported by several observations in volunteers after experimental infection with RV. First, antiviral medications are effective only if treatment is started in the first 24 to 48 h after the onset of clinical illness (1), suggesting that the immune response to the virus is thereafter responsible for respiratory symptoms. In addition, many studies have linked the intensity of the immune response to RV to the severity of cold symptoms. For example, neutrophil counts in nasal secretions, as well as quantities of neutrophil chemoattractants such as interleukin (IL)-8, correlate with cold symptoms (2, 3) and, in subjects with asthma, correlate with changes in airway responsiveness (4).

Lymphocyte responses have also been related to the outcome of RV infection. For example, the degree of T lymphopenia during the acute infection is inversely related to cold symptoms (5), and ex vivo secretion of IL-2 by peripheral blood mononuclear cells (PBMC) correlates with the severity of experimental infection (6). Moreover, when PBMC from volunteers were evaluated ex vivo before experimental infection with RV16, strong virus-induced proliferative responses and interferon (IFN)- secretion were associated with reduced viral shedding and with cold symptoms, respectively, after inoculation (7).

These data suggest that variations in the generation of cytokines that regulate neutrophils, and the character of T-helper cell type 1 (Th1) and Th2 responses to RV infection, could be responsible for greater severity of illness in some individuals. To test this hypothesis, we experimentally infected 22 seronegative subjects who had either allergic asthma or rhinitis with RV16, and compared cytokine production and cell recruitment in the subjects' nasal lavage fluid and sputum with outcomes of experimental infection as measured through viral titers in nasal secretions, symptom scores, and changes in pulmonary function. The cytokines measured were granulocyte colony-stimulating factor (G-CSF) and IL-8, which have been linked to the increase in blood and airway neutrophils, respectively, in RV16 infection (8, 9), and IFN- and IL-5, as examples of Th1- and Th2-like cytokines.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Subjects

Twenty-two subjects with allergic rhinitis or mild allergic asthma participated in the study; the subjects' characteristics are listed in Table I. Only subjects with respiratory allergies or mild allergic asthma were included in the study because these individuals are more likely to develop changes in lower airway physiology and/or lung function after experimental or natural infections (10, 11). Subjects with moderate or severe asthma, or those with a history of severe bronchospasm during viral colds, were excluded from this initial trial because of safety concerns. Asthma was defined by a clinical history of (1) physician diagnosis; (2) prescription for asthma medication; (3) cough, wheezing or shortness of breath and relief with appropriate asthma medication; and (4) evidence of a 12% or greater improvement in FEV₁ after inhalation of a -adrenergic agonist. In addition, methacholine testing was performed (Table 1), and for the purposes of this study, subjects with cumulative values for the provocative concentration of MCh needed to reduce FEV₁, by 20% (PC₂₀) > 64 mg/ml were classified as having allergic rhinitis and not asthma.

At the screening visit, all subjects underwent a physical examination, skin prick tests to 12 common aeroallergens (including cat, house dust mite, and an assortment of local pollens and molds), and spirometry. Atopy was defined as one or more positive skin tests (wheal size  histamine control). None of the subjects had neutralizing antibody to RV16 at the time of the screening visit (1 to 3 mo before inoculation), although seven subjects had low titers (median: 1.4) of RV16-neutralizing antibody detected in the serum obtained just before inoculation. Use of topical nasal corticosteroids was withheld for 30 d before the study, and acetaminophen was the only medication allowed for the treatment of cold symptoms during the study. The experimental protocol was approved by the University of Wisconsin Hospital and Clinics Human Subjects Committee, and informed consent was obtained from the study subjects before they were enrolled in the protocol. Data related to peripheral blood responses to viral infection in the study subjects have previously been published (7).

Overview of Study Design

Following the establishment of baseline values, subjects were inoculated with a standardized RV16 suspension on two successive days as previously described (7). Nasal lavage and sputum induction were performed at baseline and at 1, 2, 7, and 14 d after inoculation. Nasal lavage fluids were subjected to quantitative viral culture as previously described (10). Each of the subjects developed a respiratory infection, as demonstrated either by virus cultured from nasal secretions or by a rise in titer of RV16-specific neutralizing antibodies (Table 1). Symptoms (cough, nasal discharge, sneezing, stuffy nose, sore throat, headache, malaise, chilliness, shaking chills, fever, hoarseness, aching muscles or joints, and watering or burning of the eyes) were quantitated (0 = not present, 1 = mild, 2 = moderate, 3 = severe) daily by each subject and recorded in a diary. The maximum possible symptom score was 39. In addition, all subjects were questioned at each clinic visit about the presence of respiratory symptoms including cough and wheezing, and were asked about increased use of asthma rescue medications during the preceding phase of the study.

Sputum Induction

The method for obtaining sputum samples was based on that of Fahy and colleagues (12). At the beginning of the sputum induction protocol, subjects were given 2 puffs (180 µg) of albuterol, and spirometry was performed after 10 min. The subjects were then given general instructions to produce a deep cough, to ignore the natural inclination to swallow sputum, and to quickly blow the nose and rinse the mouth before the sputum was collected. Subjects next breathed a 3% buffered saline solution mist from an DeVilbiss (Jackson, TN) 65 ultrasonic nebulizer for 20 min, with a deep breath taken once per minute. After 5 and 10 min, spirometry was repeated to screen for bronchoconstrictive responses. The subjects were instructed to try to produce sputum whenever they felt the need to expectorate and at the 10-, 15-, and 20-min time points. To minimize contamination from the noses and mouths, the subjects were asked to blow their noses, vigorously rinse their mouths, and gargle with sterile water, followed by expectoration into a clean specimen container. If the subject's FEV₁ fell below 80% of the postbronchodilator FEV₁ at any time during the induction, the procedure was interrupted until the FEV₁ returned to within 20% of the postbronchodilator baseline value.

Nasal Lavage

Nasal lavage was performed on each subject both before RV16 inoculation and during RV infection, as previously described (8). Five milliliters of prewarmed Hanks' balanced salt solution (HBSS) with 0.5% gelatin was instilled into each nostril. The lavage fluid was then expelled consecutively from each nostril by having the subject blowing gently from one nostril while the other nostril was closed by compression.

Processing of Nasal Lavage and Sputum

To release cells from mucus, sputum and nasal lavage samples were mixed with an equal volume of Sputolysin (Calbiochem, La Jolla, CA; 0.1% dithiothreitol prepared in phosphate-buffered saline). This mixture was vortexed, rocked (20° C for 20 min), vortexed again and centrifuged (at 750 × g for 10 min at 4° C). Aliquots of supernatant fluids were saved at 80° C pending analysis for cytokines. The cell pellet was washed with HBSS (4° C) containing 2% newborn calf serum and was then filtered through 50-µm nylon mesh before counting of the total cell number. Both cell counts and differential counts were made.

RNA was isolated from samples of cells (5 × 10⁵) with a reagent containing phenol and chloroform (Trizol; Life Technologies, Rockville, MD), and the RNA extracts were frozen (80° C) until they were analyzed. Enzyme-linked immunosorbent assays (ELISAs) for IL-5, IL-8, and IFN- were performed as previously described (13, 14), and G-CSF was quantitated with a commercial kit (R&D Systems, Minneapolis, MN). The sensitivities of the ELISAs were as follows: IL-5 and IFN-, 5 pg/ml; IL-8, 15 pg/ml; and G-CSF, 20 pg/ml.

Cytokine Messenger RNA

The amount of IL-5 and IFN- messenger RNA (mRNA) in sputum cells was quantitated with a semiquantitative reverse transcription- polymerase chain reaction (RT-PCR) assay as previously described (13). Briefly, after RT of total cellular RNA, IL-5, IFN-, and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) sequences were amplified in 25 to 33 cycles of PCR (45 s at 94° C, 45 s at 60° C, 2 min at 72° C, and 7 min at 72° C for terminal extension) with specific primers (IFN-: 5'-GCA TCG TTT TGG GTT CTC TTG GCT GTT ACT GC-3' and 5'-CTC CTT TTT CGC TTC CCT GTT TTA GCT GCT GG-3'; IL-5: 5'-G CTT CTG CAT TTG AGT TTG CTA GCT-3' and 5'-TG GCC GTC AAT GTA TTT CTT TAT TAA G- 3'; GAPDH: 5'-TGA AGG TCG GAG TCA ACG GAT TTG GT- 3' and 5'-CAT GTG GGC CAT GAG GTC CAC CAC-3', (Clontech Laboratories, Palo Alto, CA). A standard curve generated from dilutions of a highly positive reference sample was included in each run of PCR. Negative controls included in each PCR run were samples containing reagents but no complementary DNA (cDNA). The PCR products were resolved by gel electrophoresis and transferred to a nylon membrane for Southern blotting with cDNA probes. Concentrations of cytokine mRNAs are expressed in units relative to a highly positive reference standard, which was arbitrarily assigned a value of 1,200 U. Detection of GAPDH mRNA was done to ensure that each sample contained cellular mRNA and also as a positive control for RT and PCR. Concentrations of the mRNAs for IFN- and IL-5 were not corrected for GAPDH mRNA because of the large shifts in the number and relative proportions of cells in the airway and were instead expressed as a ratio.

Detection of RV by RT-PCR

Picornavirus RNA was detected with RT-PCR (35 cycles) as previously described (15), using primers OL26 (5'-GCA CTT CTG TTT CCC C-3') and OL27 (5'-CGG ACA CCC AAA GTA G-3') (16). PCR products were resolved by gel electrophoresis, stained with ethidium bromide, and photographed. The presence of a 380-bp band was interpreted as a positive test for RV.

Statistical Analysis

Of the 22 subjects in the study, 15 had mild asthma. Subgroup analyses of subjects with allergic asthma versus those with allergic rhinitis revealed no group-specific significant differences in values for G-CSF or IL-8 protein, or for IFN- or IL-5 mRNA levels. Therefore, the cytokine values for these subgroups were pooled in subsequent statistical analyses. Spearman's rank correlations test was used to evaluate correlations of cytokine secretion or mRNA, symptom scores, and viral shedding with FEV₁. Cold-related changes in these parameters were evaluated with Wilcoxon's signed ranks test. All analyses were done with the aid of computer software (SigmaStat 2.0; Jandel Scientific, San Rafael, CA).

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Clinical Effects of RV16 Inoculation

Of the 22 subjects who had no detectable antibody to RV16 at the time of screening, seven had detectable neutralizing antibody to RV16 (median titer: 1.4, Table 1) approximately 1 mo later, at the time of inoculation. After inoculation with RV16, 21 subjects developed symptoms of upper respiratory infection and had infection verified by viral culture. The remaining subject neither developed cold symptoms nor shed virus after inoculation, but had an increase in the titer of neutralizing antibody to RV16 and was included in the analysis. The mean baseline symptom score was 0.95 (range: 0 to 3), and the mean peak symptom score, which occurred from 1 to 3 d after inoculation, was 10.2 (range: 1 to 23, Table 1). The mean peak symptom score (11.3) in the subgroup of subjects who had low titers of RV16-neutralizing antibody in their serum was similar to that of the group as a whole. There were significant correlations between viral shedding in nasal secretions and both symptom scores (r_s = 0.58, p < 0.005, Day 2) and peripheral blood neutrophil count (r_s = 0.61, p < 0.005, Day 2). None of the study subjects reported wheezing or increased use of albuterol associated with the induced colds, and there were no cold-related changes in mean FEV₁ for the group as a whole.

Cells and Cytokines in Nasal Secretions and Peripheral Blood

Sufficient amounts of nasal lavage fluid for both quantitative viral culture and cytokine analysis were obtained from 18 of the 22 study subjects. By the first day after inoculation, RV infection significantly increased nasal lavage neutrophil counts (median before RV16-induced colds: 32,200/ml [interquartile range (IQR): 5,344 to 94,600/ml]; median during colds: 363,000/ml [IQR: 13,700 to 2,830,000/ml], p < 0.05), and this coincided with increased concentrations of G-CSF (median before colds: 53 pg/ml [IQR: 33 to 114 pg/ml], median during colds: 868 pg/ml [IQR: 146 to 3,960] pg/ml p < 0.05) and IL-8 (median before colds: 288 pg/ml [IQR: 128 to 522]; median during colds: 519 pg/ml [IQR: 366 to 827 pg/ml], p < 0.05). Levels of G-CSF and IL-8 correlated closely with the amount of viral shedding in nasal secretions (r_s = 0.89 and r_s = 0.72, respectively, p < 0.001). In addition, there was a strong correlation between the amount of G-CSF in nasal secretions at 1 d after inoculation and the peripheral blood neutrophil count on the following day (Figure 1A). In contrast, the correlation between IL-8 and blood neutrophils was relatively weak (Figure 1B).

View larger version (14K): [in this window] [in a new window]   Figure 1.   Relationship between nasal cytokines and peripheral blood neutrophils. One day after inoculation with RV16, levels of (A) granulocyte colony-stimulating factor (G-CSF) and (B) interleukin (IL)-8 in nasal lavage fluid were compared with peripheral blood neutrophil counts 24 h later.

By the second day after inoculation, the relationships between nasal IL-8 and G-CSF and blood PMN were weaker (r_s = 0.55 and r_s 0.32, respectively), but the amounts of these cytokines in nasal secretions correlated closely with nasal PMN (Figure 2). In addition, there was a significant correlation between G-CSF in nasal secretions and cold symptom scores (r_s = 0.58, p = 0.01, n = 18), and a trend toward a positive association between nasal IL-8 and symptom scores (r_s = 0.41, p = 0.09, n = 18). There was no correlation between nasal IL-8 or G-CSF and changes in FEV₁ (data not shown).

View larger version (14K): [in this window] [in a new window]   Figure 2.   Relationship between cytokines and neutrophils in nasal lavage fluid after RV16 inoculation. Two days after inoculation with RV16, the levels of (A) G-CSF and (B) IL-8 were compared with neutrophil counts in nasal lavage fluid.

Sputum Analysis

To determine the kinetics of cell influx into the lower airway after RV infection, we analyzed induced sputum for its content of cells and cytokines at 2, 7, and 14 d after RV inoculation, and compared the results with baseline values. The quantities of total cells, lymphocytes, and eosinophils in sputum did not change significantly after RV inoculation (data not shown). Neutrophils in the sputum were significantly increased at 7 d after inoculation, with numbers after the passing of a further week being similar to those before cold induction (Figure 3).

View larger version (25K): [in this window] [in a new window]   Figure 3.   Cold-related changes in sputum neutrophils. Neutrophils were quantitated in specimens of induced sputum before (pre-cold) and 2, 7, and 14 d after inoculation with RV16.

To identify factors that could contribute to virus-induced recruitment of neutrophils into the airway, we measured IL-8 and G-CSF levels in sputum supernatant fluids. Sputum G-CSF was increased at 2 d after inoculation, and gradually drifted back toward its baseline value (Figure 4A). No significant changes were noted for sputum both IL-8 (data not shown). The levels of both IL-8 and G-CSF correlated with numbers of sputum neutrophils during the acute cold (Figures 4B and 4C). Neither IL-8 nor G-CSF levels in sputum correlated with cold symptom scores or changes in FEV₁ (data not shown).

View larger version (16K): [in this window] [in a new window]   Figure 4.   Relationship of sputum cytokines to influx of neutrophils. (A) Sputum G-CSF was quantitated before (pre-cold) and 2, 7, and 14 d after inoculation with RV16. The boxes represent the 25th to 75th percentile values, with the medians indicated by horizontal lines, and error bars designate the 10th and 90th percentiles. Seven days after inoculation, levels of sputum (B) G-CSF and (C ) IL-8 were compared with neutrophil counts in the sputum.

We measured protein and mRNA for both IL-5 and IFN- to determine the relationship between virus-induced Th1 and Th2 cytokine production and the effects of viral infection. Little or no IL-5 or IFN- protein was detected in sputum supernatants, and "add-back" experiments established that one or more factors (e.g. sputum proteoglycan) masked the detection of IL-5 and IFN- by ELISA (data not shown). In contrast, both IFN- mRNA and IL-5 mRNA were detectable both before colds and after inoculation. Expression of both IL-5 mRNA and IFN- mRNA was enhanced at 2 d after inoculation, and remained slightly increased at 7 d after inoculation (Figure 5).

View larger version (20K): [in this window] [in a new window]   Figure 5.   Changes in (A) interferon (IFN)- and (B) IL-5 during acute rhiovirus (RV) colds. Just before inoculation and 2 d and 7 d after inoculation, sputum cells were analyzed for IFN- and IL-5 mRNA. The horizontal lines represent median values. * p < 0.05 versus values before colds.

The mRNAs of IFN- and IL-5 were also expressed as a ratio. An increase in the IFN-/IL-5 ratio would represent a deviation toward a Th1-like pattern, whereas lower ratios would suggest a drift toward a Th2-like pattern. Although RV infection increased the mRNA concentrations for both IL-5 and IFN-, the ratio of IFN- mRNA to IL-5 mRNA did not change during colds (data not shown). However, during the acute phase of the illness there was an inverse relationship between the IFN- mRNA IL-5 mRNA ratio and peak cold symptoms (Figure 6A). There was no significant relationship between the IFN- mRNA/IL-5 mRNA ratio and changes in airway eosinophils during the acute cold (data not shown).

View larger version (15K): [in this window] [in a new window]   Figure 6.   Association of IFN- mRNA/IL-5 mRNA ratio to outcomes of respiratory infection. The IFN- mRNA/IL-5 mRNA ratios were calculated during the acute cold (48 h after inoculation) for each individual. These ratios were (A) compared with the peak symptom score during the acute cold, and (B) grouped according to viral clearance at 14 d after inoculation with RV16.

In addition, we compared the IFN-/IL-5 ratio with the presence of RV in nasal secretions and sputum. The percentage of subjects with picornavirus RNA detectable by RT-PCR in sputum decreased with the time after inoculation: virus was detected in three (14%) subjects before inoculation, in 21 (95%) subjects at 2 and 7 d after inoculation, and in 11 (50%) subjects at 14 d after inoculation. Subjects without detectable virus in their sputum at 14 d after inoculation had a significantly higher IFN- mRNA/IL-5 mRNA ratio than did the rest of the group (Figure 6B). There was no relationship, however, between the IFN- mRNA/IL-5 mRNA ratio in the sputum and peak viral titers in nasal lavage fluid specimens (data not shown).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

In this study of the immune response to induced colds, the severity of symptoms and amount of viral shedding were quite variable, despite careful control of the inoculation procedure and use of the same lot of virus for all subjects. Since the subjects were seronegative at the time of screening (although a few individuals had low titers of neutralizing antibody at the time of inoculation), the severity of the viral infection was largely determined by host immune factors other than the quantity of virus-specific antibody. Analysis of cells and cytokines in the blood, nasal lavage fluid, and sputum suggested that two different types of immune responses were related to the clinical and virologic outcome of the infection. First, our findings support previous studies suggesting that airway neutrophilia and factors such as IL-8 and G-CSF, which regulate neutrophil recruitment and activation, may be important in the pathogenesis of virus-induced respiratory symptoms and changes in airway responsiveness (3, 4, 9, 17). In addition, a novel finding of our study was the association between relatively weak Th1-like responses (low IFN-/IL-5 ratio) and more severe respiratory symptoms as well as a longer duration of viral shedding in the sputum.

In the nose, IL-8 and G-CSF were rapidly induced after viral inoculation, and appeared to be related to neutrophil trafficking in the airway. On the first day after inoculation, increases in nasal G-CSF correlated with increases in blood neutrophils, suggesting that G-CSF produced in the nose enters the systemic circulation and activates hematopoeisis by the bone marrow. On the second day, there was a close relationship between IL-8 and nasal neutrophils. Although an association between IL-8 and neutrophils or neutrophil myeloperoxidase activity has been previously observed (3, 9) and is consistent with the activity of IL-8 as a neutrophil chemoattractant, we observed an even closer correlation between G-CSF and neutrophils in nasal lavage fluid. This novel observation suggests either that G-CSF contributes to neutrophil recruitment to the airway or that airway neutrophils are a source of G-CSF during viral infection. Additional studies will be required to determine the mechanism responsible for this effect.

Measurement of cells and cytokines in the sputum was done in an effort to further define the lower airway response to RV infection and the timing of its development. Although the changes in cellularity were small, increased neutrophils were noted 1 wk after inoculation, suggesting that sputum neutrophilia follows the recruitment of neutrophils into the upper airway, perhaps indicating that RV infection first occurs in the upper airway and then gradually moves into the chest. Furthermore, changes in lower airway neutrophil numbers correlated with changes in G-CSF and IL-8, suggesting that the same factors are responsible for regulating neutrophil recruitment in the upper and lower airways. However, we found no relationship between changes in sputum neutrophils, G-CSF, or IL-8 and alterations in lung function. The lack of correlation could have been due to the relatively small changes in lung function, sputum cellularity, and cytokines, and our findings were similar to those reported in two previous studies of sputum physiology after experimental infection with RV16 (9, 18). One limitation common to studies utilizing RV inoculation is that few study subjects actually develop exacerbations of asthma, even when small changes in lung function are noted (9, 19). Examination of sputum during naturally occurring virus-induced exacerbations of asthma may reveal more dramatic changes in sputum physiology, and thus may have greater potential for establishing links with the pathogenesis of acute airway obstruction.

A novel aspect of our study was its comparison of the pattern of Th1- and Th2-like cytokine responses, as indicated by changes in sputum IFN- mRNA and IL-5 mRNA, with the clinical and virologic outcome of RV infections. The ratio of IFN- mRMA to IL-5 mRNA was determined in order to identify tendencies toward a Th1-like (high ratio) or Th2-like (low ratio) immune response to viral infection. RV infection increased IFN- mRNA, which is expected of a virus infection, but IL-5 mRNA was also increased. This indicates that RV infection causes activation of both Th1- and Th2-like cytokines, and this is supported by the absence of a net shift in the IFN- mRNA/IL-5 mRNA ratio during acute colds.

The IFN- mRNA/IL-5 mRNA ratio during the acute phase of infection was inversely related to two different outcome measures: peak cold symptom scores and time to virus clearance from the sputum. These findings suggest that strong Th1-like cytokine responses in the airway may play an important role in limiting the amount of viral replication and respiratory symptoms in colds caused by RV. This hypothesis is consistent with our previous observations that vigorous increases in RV-induced IFN- secretion by PBMC incubated with RV ex vivo were associated with reduced viral replication after experimental inoculation (7). Studies with rodents also suggest that Th1-like cytokines such as IFN- are important in the clearance of viral infections (20, 21) and that administration of exogenous IFN- can reduce chronic airway dysfunction after Sendai virus infection in a susceptible strain of weanling rats (22).

These findings raise questions about the origin of the IFN- mRNA and IL-5 mRNA in sputum, and the mechanism(s) for their generation. A likely source of these cytokine mRNAs are lymphocytes. RV-specific T-cell clones have been shown to secrete large amounts of IFN- in vitro, but some clones also secrete Th2-like cytokines such as IL-4 and IL-5 (14), and mouse models of viral infection also support the concept that virus-activated lymphocytes can produce a broad range of cytokines (23). In addition to antigen-specific activation of lymphocytes, pathways have been described for antigen-independent activation of lymphocytes through the virus-induced generation of soluble cytokines or mediators that secondarily stimulate lymphocyte IFN- production (24, 25). Regardless of the mechanism of their production, concurrent viral stimulation of IFN- and IL-5 could be a potent inflammatory stimulus. For example, in a mouse model of allergen-induced inflammation, passive transfer of Th1 clones enhanced, rather than counteracted, allergen-induced inflammation and airway responsiveness mediated by Th2 lymphocyte activation (26).

In summary, in the group of allergic individuals in our study, experimental infection with RV16 caused rapid increases in nasal G-CSF and IL-8, and these upper airway events were closely associated with neutrophilic responses in the blood (G-CSF) and nasal secretions (IL-8 and possibly G-CSF). Although corresponding changes in sputum neutrophils and G-CSF were relatively modest, analysis of cytokine mRNA in the sputum suggests that the balance of Th1- to Th2-like mRNA may play an important role in controlling viral clearance and respiratory symptoms in RV infections. There is accumulating evidence that markers of atopy are risk factors for more severe clinical manifestations during RV infections (10, 27, 28). Since allergy and asthma are associated with an immune response that is skewed toward Th2-like responses, the resulting cytokine imbalance may not only produce a distinct response to allergens but may also impair immune responses to viral infections. Additional controlled studies, to compare Th1- and Th2-like immune responses with measures of clinical outcome during naturally occurring RV infection, will be needed to determine whether this could contribute to the greater lower airway effects of RV infections in patients with asthma.