INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Although considerable advances in understanding the immunopathology of adult asthma have occurred, little knowledge is available on pathophysiologic mechanisms in asthma of childhood (1). In adults, the eosinophil is believed to be a key effector cell in asthma (2). We (3), and others (4), have shown that the same may be true in childhood asthma but not in infantile wheeze (5). Infantile wheeze is now recognized as a heterogeneous syndrome, some remitting in early childhood and others often with immunoglobulin E (IgE)-mediated sensitization who have persistent asthma (6). We recently found a neutrophilic rather than eosinophilic inflammation in infantile wheezers (3). We also reported that the neutrophilic inflammation was related to disease severity in both children with asthma and infantile wheeze. Intercellular adhesion molecule-1 (ICAM-1) is one of the molecules that contribute to both eosinophil and neutrophil influx into the airway lumen (7). ICAM-1 belongs to the immunoglobulin supergene family and is expressed by many airway cells including bronchial epithelium cells, endothelial cells, T cells, mast cells, eosinophils, and alveolar macrophages (8). It facilitates cell-to-cell interaction, which could potentiate chronic inflammation and is particularly important in neutrophil and epithelial cell adhesion (11). In addition, this adhesion molecule may contribute to susceptibility to viral infection by being a major receptor of rhinoviruses, which have been implicated as a principal cause of asthma exacerbations both in children and infants (12). Thus, up-regulation of ICAM-1 on epithelial and endothelial cells is believed to be a hallmark of asthma in adults (13, 14), although some studies failed to show any overexpression (8).

The identification of two lymphocyte subsets, Th-1 and Th-2, with their own cytokine pattern has shed new light on the allergic origins of asthma (15). Among the cytokines, interferon-gamma (IFN-) and interleukin-4 (IL-4) were shown to predominate in Th-1 and Th-2 cells, respectively, and the ratio IFN-/IL-4 is used as a marker of a Th-1/Th-2 switch (16). In addition, IFN- is known to have an antiviral effect and therefore was expected to be a protective cytokine in asthma. However, IFN- is a pleiotropic cytokine and is strongly involved in the up-regulation of ICAM-1 and other cytokines, suggesting a proinflammatory effect (17). The complexity of its regulation in asthma was recently highlighted. Contrasting with the commonly decreased expression of IFN- in groups with atopic asthma (16), either elevated serum INF- levels (18) or increased release from lung cells in adults with asthma have been reported (19). Similar apparent discrepancies in the regulation of IFN- have also been observed in young children. Thus, studies have shown a relatively lower release of IFN- from peripheral blood-stimulated mononuclear cells in childhood atopic (20) and infantile asthma (21), but increased levels of IFN- in respiratory secretions from infants with persistent wheezing (22).

To make progress in understanding the mechanisms of lung inflammation in childhood asthma, we evaluated the levels of both sICAM-1 and IFN- and their relationships in the bronchoalveolar lavage fluid (BALF) from children with asthma and infantile wheezers. We hypothesized that ICAM-1 could play a pivotal role in both childhood asthma and infantile wheeze as it is involved in recruitment of both eosinophils and neutrophils. Thus we analyzed the correlation between sICAM-1 in BALF, in both syndromes, with disease activity, and with IFN- which up-regulates ICAM-1.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Fiberoptic bronchoscopy was performed on 68 infants and children and four groups were specified according to the diagnosis at the time of the investigation. Sixteen children aged between 4 and 15 yr had established asthma (A). These children had a history of recurrent wheezing associated with evidence of ₂-agonist reversibility of airflow limitation. The indications for bronchoscopy were persistent radiological abnormalities (seven atelectasis and two asymmetrical overinflation of lung fields), two suspected extrinsic allergic alveolitis, one hemoptysis, and four unexplained failure of steroid treatment. The second group was comprised of 12 children (10 mo to 13 yr) with isolated chronic persistent cough not associated with wheezing (CC). They were investigated because of persistent radiological abnormalities (three persistent atelectasis and one marked bronchial thickening), hemoptysis (n = 1), and uncertain diagnosis (n = 7). The third group of 30 children aged between 5 and 46 mo were infants with recurrent wheezing (W). All had at least three episodes of wheeze and cough during presumed viral infection. Bronchoscopy was justified by persistent atelectasis (n = 10), asymmetrical lung fields on radiograph (n = 5), suspected lung malformation (n = 1), unexplained failure of steroids (n = 7), uncertain diagnosis (n = 5), and associated stridor (n = 2). Children aged 1.5-14 yr acted as controls with 10 children without any known chronic lower airway disease. They had tracheal or lung malformation (n = 5), stridor (n = 1), psychogenic cough (n = 1), a suspicion of lung malformation (n = 2), or primary tuberculosis (n = 1). Standard clinical histories and examination were performed with the addition of chest radiographs. Allergy prick tests were routinely performed on the A, CC, and W groups with total serum IgE and IgE antibodies available on 80% of these patients. The characteristics of the patients are summarized in Table 1.

Thirty-three of the children had received inhaled steroids (IS), all of them for at least 1 mo (Table 1). We classified patients of the groups A, CC, and W in accordance with existing symptoms (night cough, wheezing, or dyspnea) at the time of bronchoscopy (Table 1). The 9 asymptomatic patients (S0) were defined as children without any symptoms for at least 1 mo before and after the bronchoscopy; there were 14 children who suffered from at least weekly respiratory symptoms (S2) significantly affecting their quality of life (sleep, behavior, exercise tolerance, and school attendance) and/or causing failure to thrive. The remaining 35 symptomatic children (S1) had exercise-induced symptoms or mild to moderate exacerbations with little effect on their quality of life. Of children treated with inhaled steroids, there were 6 in S0, 16 in S1, and 10 in S2. Those children with asthma, cough or infantile wheeze having persistent radiological abnormalities are listed in Table 1. Chest X-ray features were classified in three groups: normal (group R0, n = 12), persistent parenchymal shadowing (group R2, n = 20), and others with mainly peribronchial thickening (group R1, n = 25).

Bronchoscopy Procedure

All children were admitted to the day ward before the procedure. A transnasal fiberoptic bronchoscopy (Olympus BF 3C20 [Keymed, Southend-on-Sea, UK], Olympus BF P10 in the teenagers) was performed under monitoring with continuous oximetry assessment and clinical evaluation. Premedication consisted of midazolam, pethidine, and in some atropine sulfate (0.01 to 0.02 mg/kg, and less than 1 mg). Topical anesthesia consisted of lignocaine 2% in the upper airways and 1% in the bronchial tree. Preventive salbutamol nebulization was prescribed in 23 patients, and 4 infants were investigated under oxygen supplementation via nasal canulae. The bronchoscope was wedged in a segmental bronchus. The site depended on the localization of abnormality from chest radiographs but if there was no local disease, then the right middle lobe was selected. Aliquots of 10 ml of warmed sterile 0.9% saline solution were injected and aspirated. The total volume ranged between 2 and 3 ml/kg and the specimens collected were 49 ± 19% of the original injected volume. The BALF recovered was pooled in fresh sterile containers and divided into aliquots for the various investigations. To reduce the risk of contamination from the upper airways, no suction was used until the tip of the bronchoscope had been wedged into the bronchial lumen. Written consent was obtained from the parents for the procedure, which in all cases was based on a clinical imperative. However, the hospital ethics committee approved additional use of the lavage specimens for research purposes.

Cell Count

The lavage fluid was not filtered but centrifuged for 10 min at 1,800 rpm at 4° C, and the pellet was resuspended using phosphate-buffered saline (PBS, 5 ml) containing 10% fetal calf serum. The total cell count was performed using a hemocytometer. Viability was assessed by trypan blue exclusion test. Differential cell counts were carried out on cytospin slide preparations stained with May-Grünwald-Giemsa. At least two slides and 200 cells were counted for each subject. Total cells and differential counts were detailed in a previous paper (3) and are displayed in Table 2.

Mediator Assays

sICAM-1 was measured using a commercially available immunoassay (R&D System, Abingdon, Oxon, UK). The assay range was between 2 and 45 ng/ml with sensitivity less than 0.35 ng/ml. All samples were diluted 1 in 20. IFN- was measured using a commercially available ELISA (AMS Biotechnology, Abingdon, Oxon, UK) with a detection level of 10 pg/ml and a detection range between 10 and 6,250 pg/ml.

Microbiological Analysis

Quantitative cultures were carried out for growth of the most common aerobic organisms, Haemophilus influenzae (Hi), Streptococcus pneumoniae (P), Moraxella catarrhalis (MC), Streptococcus sp. (S) and Escherichia coli. A significant result was defined with a growth > 10³ CFU/ml.

Data Analysis

All results were presented as medians and interquartile 25th-75th ranges. The results of sICAM-1 and IFN- levels were expressed per milliliter of the pooled retrieved liquid and no dilution markers were used as previously discussed (3).

Statistical Analysis

Statistical analysis employed Spearmans two-tailed ranked correlation (rho) in addition to Pearson linear regression two-tailed analysis (r). When applicable, differences between groups were determined using Mann-Whitney U test or unpaired t test. Fisher exact test was used for the qualitative analysis between the groups. A p value of < 0.05 was considered as significant while trends were considered with p values < 0.1.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

sICAM-1 and INF- Concentrations in BALF

Asthma was characterized by a higher median level of sICAM-1 than control and coughing children (Figure 1) and higher INF- levels than coughers (Figure 2). There was a trend to higher INF- levels in children with asthma compared with infantile wheeze. An increased level of sICAM-1 was measured in BALF from chronic coughers on steroid treatment in comparison with those with no steroids (113 [96-183] versus 55 [27- 84] ng/ml, p = 0.03) but this effect was not observed in asthma (106 [94-257] versus 145 [84-190] ng/ml, p = 0.9) or infantile wheeze (85 [51-181] versus 87 [32-158] ng/ml, p = 0.58). The steroids influenced INF- production with significantly higher levels in the treated infantile wheezers and a trend in chronic coughers, whereas no difference was observed in children with asthma (Figure 3). sICAM-1 was rank correlated to the number of lymphocytes in asthma (sICAM-1 levels: 108 [89-232] ng/ml, lymphocytes: 0.03 [0.02-0.056] 10⁶/ml; rho = 0.69, p = 0.05) and infantile wheeze groups (sICAM-1 levels: 86 [46-172] ng/ml, lymphocytes: 0.025 [0.006-0.092] 10⁶/ml; rho = 0.69, p = 0.01). No significant relationship was established between IFN- and any cell counts. sICAM-1 or INF- levels were similar from children with either positive or negative BALF culture. sICAM-1 levels were for group A: 190 [99-257] versus 150 [101-245] ng/ml, p = 0.77; group CC: 77 [55-135] versus 91 and 42 ng/ml, p = 0.56; group W: 86 [45-172] versus 90 [19- 181] ng/ml, p = 0.88. INF- levels were for group A: 120 [50- 205] pg/ml, p = 0.52; group CC: 19 [12-43] versus 48 and 23 pg/ml, p = 056; group W: 25 [20-58] versus 37 [9-283] pg/ml, p = 0.46. 

View larger version (13K): [in this window] [in a new window]   Figure 1.   sICAM-1 levels in bronchoalveolar lavage fluid (BALF) from children with asthma (A, n = 16), chronic cough (CC, n = 12), and infantile wheeze (W, n = 30) and in BALF from control children (C, n = 10). Horizontal bars indicate the median for each group of patients. Significant differences and trends are displayed on the figure and the remaining groups are comparable with each other.

View larger version (12K): [in this window] [in a new window]   Figure 2.   IFN- levels in BALF from children with asthma (A, n = 16), chronic cough (CC, n = 19), and infantile wheeze (W, n = 29) and in BALF from control children (C, n = 10). Horizontal bars indicate the median for each group of patients. Significant differences and trends are displayed on the figure and the remaining groups are comparable with each other.

View larger version (24K): [in this window] [in a new window]   Figure 3.   Influence of steroid treatment on IFN- levels in asthma (A: open bar), infantile wheeze (W: shaded bar), and chronic cough (CC: hatched bar). IFN- levels are expressed as box and whiskers plots with range and values displayed. No difference in IFN- levels was found between treated and untreated children with asthma, but significant differences are seen in the infantile wheezers.

sICAM-1 and IFN- Levels and Severity Scores

sICAM-1 increased with the severity of the disease in children with asthma, chronic coughers, and infantile wheezers (Figure 4). Asymptomatic children with asthma and infantile wheeze had similar, and children with severe disease (S2) had higher sICAM-1 levels compared with those measured in control infants. Among the patients with a moderate disease (S1), children with asthma have a more elevated sICAM-1 level than coughers, and controls (Figure 4). A comparable shift of IFN- levels was observed with the symptom score in infantile wheezers: 26 [20-29] for S0, 17 [11-32] for S1, and 68 [23-359] pg/ml for S2 with a significant difference between S1/S2, p = 0.02 but not between S0/S1, p = 0.56 and S0/S2, p = 0.14. On the other hand, variations in the IFN- levels were found neither in children with asthma: 79 [42-99] for S0, 120 [33-260] for S1, and 20, 41, and 815 pg/ml for S2 (p = 0.663) nor in chronic coughers: 23 [13-36] for S1 and 11, 51, and 93 pg/ml for S2 (p = 0.4). According to the radiographic score, peribronchial thickening was associated with the highest levels of sICAM-1, whereas IFN- levels were increased irrespective of the chest X-ray abnormalities (Figure 5).

View larger version (28K): [in this window] [in a new window]   Figure 4.   sICAM-1 levels in relation to the severity of symptoms in children with asthma (A: open bar, n = 16), infantile wheeze (W: shaded bar, n = 30), and chronic cough (CC: hatched bar, n = 12) and sICAM-1 in control (vertically hatched bar). S0, asymptomatic in the month before and after bronchoscopy lavage; S1, mild to moderate symptoms; S2, severe symptoms. Values are displayed and results are expressed as box and whisker plots with range or high-low values. Significant differences and trends are displayed on the figure, and the remaining groups are comparable with each other.

View larger version (23K): [in this window] [in a new window]   Figure 5.   sICAM-1 (open box) and IFN- (shaded box) levels in relation to the radiographic features in 58 children having asthma, chronic cough, or infantile wheeze. R0 (n = 12) had a normal chest X-ray, R1 (n = 25) had mainly bronchial thickness, and R2 (n = 20) had persistent atelectasis. Results are expressed as box and whiskers plot with ranges. Only significant differences and trends are displayed on the figure.

Relationships between sICAM-1 and IFN- Levels in BALF

A significant linear correlation was established between sICAM-1 and IFN- in infantile wheeze (r = 0.549, p = 0.002) but not in asthma (r = 0.380, p = 0.14) or in chronic cough (r = 0.314, p = 0.052) (Figure 6).

View larger version (11K): [in this window] [in a new window]   Figure 6.   Correlation between sICAM-1 and IFN- levels in BALF from infantile wheezers (n = 29) (r = 0.549, p = 0.002). No correlation was found either in asthma or chronic cough groups.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

We hypothesized that ICAM-1 and IFN- levels were involved in the generation of airway inflammation in both infantile wheeze and childhood asthma. Although not specific, there is evidence that ICAM-1 has a role in the recruitment of inflammatory cells in asthma in adults. Up-regulation of ICAM-1 was demonstrated on epithelial cells and T cells in the bronchial mucosa or airway secretions from children with asthma of varying severity (13, 14, 23) but was lacking in a homogeneous cohort of children with mild asthma (8). On the other hand, overexpression of ICAM-1 in the airway or circulation correlated with the severity of asthma (24). We can confirm that high airway levels of sICAM-1 characterized active asthma in childhood being higher in those with recent disease activity. Although the pathophysiological function of the soluble form of ICAM-1 remains to be specified (10), a shift of its lung concentration is likely to reflect increased cellular ICAM-1 expression (23). Although we believe that ICAM-1 is involved with inflammatory cell recruitment, we were not able to link sICAM-1 levels with either eosinophil or neutrophil cell counts. However, one spot measurement of sICAM-1 is unlikely to reflect patterns of production over the time frame when the cells were recruited. We did, however, show a correlation between sICAM-1 levels and lymphocyte numbers that might reflect its expression on airway CD4⁺ T cells (24).

ICAM-1 is overexpressed on epithelial cells from atopic patients (25) and is up-regulated after a bronchial allergen challenge (26) in adults. But such a role of atopy was not confirmed in children (27) and many of our patients were nonatopic. Viruses and mainly rhinoviruses (11), as well as bacteria such as Haemophilus influenzae (28), are able to increase ICAM-1 expression in lung airways. We did not investigate the presence of viruses in our study but found no influence of bacterial colonization on sICAM-1 levels. There was, also, no difference in the presence or absence of focal lung abnormalities. Therefore, we do not believe that our findings relate to secondary disease, but are representative of the global condition of the airways.

Several cohort studies have now shown that infantile wheeze is a syndrome, which differs from asthma per se (6). In addition, we recently reported that neutrophil but not eosinophil-mediated inflammation characterized persistent symptoms in infantile wheeze (3) and now provide evidence for the contribution of ICAM-1 to the severity of infantile wheeze. However, neutrophils and sICAM-1 levels were not correlated, implying at least partial ICAM-1-independent neutrophil recruitment in the lung (29). We have observed a nonspecific increase of ICAM-1 with the severity of diseases because we found a similar shift was observed in chronic coughers who, on bronchoalveolar cell profile, differed from those with asthma and infantile wheeze, and more resembled control subjects without lower airway disease (3). sICAM-1 levels were correlated to lymphocyte numbers in infantile wheeze and asthma, but not chronic cough. Because the percentage of CD8⁺ was significantly higher in asthma than infantile wheeze (3), these results suggest that up-regulation of the adhesion molecule on CD4⁺ subsets might be a common mechanism in asthma and infantile wheeze.

Our study highlights some differences in the regulation of sICAM-1 between children with asthma and wheezers. Thus, ICAM-1 levels were elevated only in BALF from infants with severe but not mild or moderate disease, indicating that the contribution of this adhesion molecule to the pathophysiology of infantile wheeze would be less characteristic than in chronic asthma. IFN- has been shown to be one of the most potent stimuli of ICAM-1 expression at the molecular level (7), but surprisingly was only correlated in infantile wheeze. Furthermore, the highest levels of IFN- were observed in infants having severe disease, although a wide range was measured in this group. Our findings sustained the hypothesis that INF- might act as a proinflammatory cytokine by underlying cell recruitment in infantile wheeze through an up-regulation of ICAM-1 expression. Hereby, the suggested mechanisms of lung inflammation in infantile wheeze remain consistent with the contribution of microorganisms to its pathogenesis (12). Viruses that are clearly involved in provoking wheezing have potent effects on IFN--release by T cells, which in turn influence expression of ICAM-1 airway cells (11). This was highlighted by recent pediatric studies. Grigg and coworkers (30) showed that various viruses were able to increase sICAM-1 levels and lymphocyte percentages in BALF from nonwheezy infants with colds. Van Schaik and coworkers (22) evaluated IFN- levels in nasopharyngeal secretions from infants mostly infected with respiratory syncytial virus. The proinflammatory role of IFN- appeared to depend on the site of the infection: the "TH1" cytokine being up-regulated when infants suffered from lower but not from upper respiratory infection. Passive smoking also aggravates recurrent wheeze (6) and is able to up-regulate airway sICAM-1 (31) but not circulating IFN- (21) production.

The consequences of a rise in IFN- secretion in the airways of young children remain difficult to analyze. Reduced allergen-induced production of IFN- in newborn infants is associated with the subsequent development of allergy (20). On the other hand, among the infants with bronchiolitis, those with lower IL-2-stimulated lymphocyte secretion of IFN- had an increased risk of recurrent wheezing (21). Thus, a low or normal IFN- level in BALF from wheezy infants might reflect either a weak inflammatory process or an alteration in the response to viral infections. Furthermore, Bianco and coworkers (11) emphasized the complex regulation of virus- mediated inflammation, with up- or down-regulation of ICAM-1 by IFN- depending on whether epithelial cells were or were not infected. However, our finding of a positive correlation between IFN- and sICAM-1 in BALF from infant wheezers but not children with asthma, suggests that IFN- may be important in the early phases of development of airway inflammation. In established asthma, many other factors become more important in sustaining ICAM-1 up-regulation.

In this study, steroids seemed to enhance IFN- production in infantile wheezers and chronic coughers but not in children with asthma. The interpretation of this result must be limited because we did not perform BAL before and after commencing inhaled steroids. In addition, steroids could act either on IL-10 (32) or IL-12 (33), which have been respectively involved in negative or positive regulation of IFN-. However, it is consistent with the finding of increase in IFN- mRNA expression in pooled airway cells from children with atopic asthma after steroid treatment (34). Conversely, several studies have demonstrated a suppressive effect of either inhaled or oral steroid on IFN- production at the cell level (32, 35). Therefore, we need to employ caution when considering whether the effect of steroids on IFN- is either mechanistically relevant or indeed beneficial. The data sound a qualified note of caution in employing steroids for virus-associated wheezing.

In conclusion, our study suggests that ICAM-1 and IFN- play a role in the activity of the inflammatory process in asthma in childhood and probably in infantile wheeze. Furthermore, raised levels of the cell adhesion molecule in BALF were related to peribronchial thickening on chest radiograph, which suggests diffuse airway inflammation, but not persistent atelectasis. Our results were consistent with the high circulating sICAM-1 measured in infants and children with acute symptoms (36). Therefore, sICAM-1 might be a target for future therapy in childhood. Although asthma in childhood presents a profile similar to that previously reported in adults, in some respects infantile wheeze had different profiles. In infantile wheeze, we found a close relationship between sICAM-1 and IFN- levels. The latter may initiate the increased expression of the cell adhesion molecule. Moreover, these alterations suggest a role of viruses. As some of the infantile wheezers will develop asthma, our results pose the question of whether elevated sICAM-1 is a predictor of asthma. We have in the past found that high levels of sICAM-1 in blood and bronchial secretion from premature babies were shown to be predictive of chronic lung disease (37). Enhanced by steroids, the significance of elevated levels of the "Th-1" cytokine, IFN-, remain to be specified. Last, no specific inflammation process was found in the group of chronic coughers in accordance with their bronchoalveolar cell profile, although increasing sICAM-1 was found in those with severe cough.