Dr Shigemi Yoshihara, Department of Pediatrics Dokkyo University School of Medicine 880 Kitakobayashi, Mibu, Tochigi 321–0293, Japan. E-mail: Shigemi@dokkyomed.ac.jp
We examined whether epithelial damage is associated with mobilization of neutrophils or eosinophils in the airway lumen during acute exacerbations of paediatric asthma. Aspirated sputum samples were harvested from 65 paediatric patients (mean age 3·4 ± 0·4 years) during acute exacerbations of asthma. Patients with signs of infection were excluded. The presence of conglomerates of epithelial cells (i.e. ‘Creola bodies) in the aspirated sputum was utilized as a marker of epithelial damage. Among the paediatric asthma patients, 60% displayed Creola bodies (CrB+: n = 39) in their sputum samples whereas the remaining patients did not (CrB–: n = 26). CrB+ patients displayed more than a 20-fold increase in the concentration of the neutrophil-mobilizing cytokine interleukin (IL)-8 (pg/ml) and of the neutrophil product neutrophil elastase (NE, g/l), respectively, compared with CrB– patients (IL-8: 7468·2 ± 1953·6 versus 347·9 ± 72·6, P < 0·01; NE: 2072·4 ± 419·0 versus 438·5 ± 125·7, P < 0·01). Even though not statistically significant, a corresponding trend was observed for the relative number of sputum neutrophils. In contrast, the concentration of the eosinophil-mobilizing cytokine IL-5 and the esoinophil product ECP tended to be lower in CrB+ than in CrB– patients (P > 0·05). In conclusion, as indicated by the analysis of aspirated sputum, epithelial damage is associated with a locally enhanced chemotactic signal for and activity of neutrophils, but not eosinophils, during acute exacerbations of paediatric asthma. It remains to be determined whether these indirect signs of neutrophil mobilization in the airway lumen mirror an increased number of neutrophils in the surrounding airway tissue.
Asthma is a chronic inflammatory disease that has previously been associated with the local accumulation and/or activation of eosinophils and subsequent epithelial damage in the airways [1–4]. Epithelial damage, in turn, has been linked to airway hyper-responsiveness [5–8]. Recent clinical studies have questioned the exclusive role of eosinophils in paediatric asthma. These studies show that the local accumulation and/or activation of neutrophils in the airways is a characteristic feature during acute exacerbations of asthma in paediatric patients and there is corresponding data from adult patients with acute, severe asthma [8–12]. Until now, the relationship between local epithelial damage, on one hand, and neutrophil mobilization, on the other hand, has not been characterized in acute exacerbations of paediatric asthma.
The presence of Creola bodies (CrB+) in the induced sputum or bronchoalveolar lavage (BAL) fluid from adult patients with acute exacerbation of asthma has been recognized as a correlate to epithelial damage in the airways [5–7,13–15]. In the referred studies, the presence of CrB+ has correlated with eosinophils in induced sputum and also with bronchial reactivity [5–7]. It is not known whether patients with paediatric asthma resemble adult patients in this sense.
We hypothesized that epithelial damage is associated with local mobilization of granulocytes in the airways during exacerbations of acute, paediatric asthma. Thus, in the present study, we evaluated this hypothesis by collecting aspirated sputum from patients with acute exacerbations of paediatric asthma. We related the presence of CrB+ with cytokines known to selectively accumulate either eosinophils or neutrophils, with specific activity markers and with the relative number of these two types of granulocytes.
Materials and methods
We consequtively recruited patients with paediatric asthma who had been hospitalized in the Department of Paediatrics of the University of Dokkyo during the period April 1999 to March 2000. All recruited patients had a history of three or more different episodes of wheezing and documented evidence of wheezing by auscultation and an effect of inhaled beta-agonists upon admission (> 3% increase in SpO2) . The exclusion criteria were as follows: congenital respiratory anomaly or cardiac disease or presence of foreign body or signs of severe infection. A positive X-ray of the chest or C-reactive protein > 0·4 mg/dl or a positive rhinopharyngeal smear culture for bacteria or a positive test for respiratory syncytial virus (RSV) antigen in nasal smear  was regarded and utilized as indicators of infection. The utilized bacterial culture normally detects the most common bacterial pathogens in the airways . We also examined Gram's staining in rhinopharyngeal smear to rule out bacterial pathogens.
Examination of sputum
Sputum was obtained from paediatric patients with asthma by using an aspirator [16,17] at the time of admission. The patients produced sputum easily when the following procedure was used: a suction device consisting of a two-way tube with a trap. One end of this tube was connected to an aspirator machine. As coughing will drive sputum up the trachea, coughing was then induced by rubbing the anterior neck of the patients, over the upper section of the cricoid cartilage. A tube was subsequently inserted orally, and a specimen taken quickly during coughing. In exceptional cases, the patient did cry, and neck stimulation was then not effective. It was then required that the sputum was harvested rapidly at the moment when coughing was triggered by the crying or the aspiration. The harvested sputum was stored on ice for immediate processing. Smears of sputum (100 µl) were spread on pairs of glass slides and the slides were subsequently fixed with ethanol (95%). One slide was stained with Papanicolaou's stain [5,6,16] to detect CrB+. CrB+ was identified by the presence of clusters of more than 10 desquamated respiratory epithelial cells  (see Fig. 1). A second slide was stained with Hancel's stain  and analysed for subtypes of inflammatory cells. A total of 300 cells per power field Line was counted in each preparation. The presence of alveolar macrophages in the specimen was considered evidence that the secretions came from the lower respiratory tract. Specimens with diffuse or massive squamous epithelial cells or without alveolar macrophages were excluded from the study, because these findings indicated a specimen originating from the upper respiratory tract only.
Characterization of patients
Characteristics (age, sex, serum-IgE, RAST score, history of atopy in family) were compared for CrB+ and CrB– patients. We also measured the concentrations (pg/ml) of the specific neutrophil-mobilizing cytokine interleukin (IL)-8 and the specific eosinophil-mobilizing cytokine IL-5 using enzyme-linked immunoassay (EIA) . The concentration of the neutrophil activity marker neutrophil elastase (NE, µg/l) was determined using the Latex concentration method  and the concentration of the eosinophil activity marker eosinophilic cationic protein (ECP, µg/l) was determined using a radioimmuno assay (RIA) . We also measured tumour necrosis factor (TNF-α, pg/ml) using EIA and, finally, bradykinin (pg/ml) using RIA .
Descriptive statistics were expressed as mean ± SD for basal characterization. Data utilized for analytical statistics were presented as mean ± SEM and evaluated utilizing the Mann–Whitney's U-test. In all cases, a P-value < 0·05 was considered to be significant; n represents the number of patients (independent observations).
The study was approved by the Regional Ethics Committee for Human Research at the Hospital of Dokkyo University School of Medicine. The parents of all patients participating in this study gave their informed oral and written consent.
Among the 72 paediatric patients with asthma that were originally recruited for this study, 70 obtained written informed consent from their parent or guardian. Among these 70 patients, 5 displayed diffuse massive squamous epithelial cells or lack of alveolar macrophages in their aspirated sputum and were therefore excluded. Thus, finally, 65 patients with paediatric asthma were included (mean age, 3·4 ± 0·4 years) in this study.
Among the included patients with paediatric asthma, the mean age was somewhat lower for CrB+ than for CrB– and this difference was statistically significant (Table 1). A typical sputum sample from each subgroup is presented in Fig. 1. In contrast, the distribution of sex, concentration of serum-IgE, RAST score and history of atopy in the family did not display any statistically significant difference for CrB+ and CrB– patients (Table 1). For comparison, the aspirated sputum of five paediatric patients without asthma was examined but these paediatric patients were all CrB– (data not shown).
Table 1. Clinical characteristics of CrB+ and CrB– patients.
Approximately 60% of the patients with exacerbation of paediatric asthma displayed CrB+ in aspirated sputum (Table 1). The concentration of IL-8 and NE, respectively, was more than one order of magnitude higher among CrB+ patients than among CrB– patients; and these differences were statistically significant (Table 3). Moreover, there was a trend towards a corresponding difference in the relative number of sputum neutrophils for CrB+ and CrB– patients but this difference did not reach statistical significance (P = 0·09) (Table 2). The concentration of TNF-α and bradykinin, respectively, tended to be higher in CrB+ than in CrB– patients but these differences were not statistically significant (Table 3). Also, the concentration of IL-5 and ECP, respectively, tended to be lower in the CrB+ than in CrB– group but, again, these differences did not reach statistical significance (Table 3). Finally, in line with these observations, there was a trend towards a smaller relative number of sputum eosinophils for CrB+ compared with the CrB– patients, even though this difference was not statistically significant (P = 0·27) (Table 2).
Table 3. Cytokine and mediator concentrations in aspirated sputum.
The key findings of the present study were that in paediatric patients with acute exacerbation of asthma, and no detectable infection, CrB+ was associated with markedly increased sputum concentrations of the specific neutrophil-mobilizing cytokine IL-8 plus the specific neutrophil activation product NE. In contrast, the sputum concentrations of the specific eosinophil-mobilizing cytokine IL-5 and the specific eosinophil product ECP tended to be lower among CrB+ patients, even though these latter differences were not statistically significant. There was a corresponding trend for the relative number of neutrophils (higher in CrB+) but it failed to prove statistically significant. Taken together, these findings imply that epithelial damage is associated with a local increase in chemotactic signal and activity of neutrophils, but not eosinophils, in the airways during acute exacerbations of noninfectious paediatric asthma. Our findings forward the possibility of inherent neutrophil mobilization as an important factor for the clinical course in paediatric asthma, even though we cannot completely rule out a mild infection due to either viruses other than RSV or rare bacteria as confounding factors in our study material.
It may seem paradoxical that we were unable to demonstrate a statistically significant increase in the relative number of neutrophils in the aspirated sputum of CrB+ patients, even though the concentrations of IL-8 and NE were markedly increased. We believe that this may be due our moderate sample size and to the method of aspirated sputum per se. To us, it seems likely that the method of aspirated sputum is more effective in collecting water soluble proteins such as IL-8 and NE, than it is for retrieving inflammatory cells that may reside in the airway mucosa or even deeper in the airway tissue. Thus, it is possible, but not proven, that there is a substantially increased number of neutrophils in the airway tissue of CrB+ patients, compared with CrB– patients, and that aspirated sputum only mirrors this tissue inflammation by collecting proteins diffusing from the tissue into the airway lumen.
Apart from the support obtained in the current study, at least two additional clinical studies support the idea that inherent neutrophil activation is of importance for the clinical course in childhood asthma [10,11]. In essence, these studies also indicate an increase of the concentration of the neutrophil-mobilizing cytokine IL-8 and the neutrophil product NE, respectively, locally in the airway lumen; the evidence was obtained utilizing analysis of BAL fluid or induced sputum harvested during acute exacerbations of paediatric asthma.
Indeed, several clinical studies have shown that acute exacerbations of asthma among adult patients are associated with neutrophil mobilization, even when there is no (detectable) infection . The exact mechanisms that lay behind this temporal association remain unknown, however, although it seems mechanistically plausible that neutrophil products such as free oxygen radicals and proteolytic enzymes including NE can contribute to the clinical deterioration of asthma . This is because these compounds can increase the secretion from airway glands, cause unspecific hyperreactivity in airway smooth muscle and degrade elastin in lung tissue during certain conditions .
The results of our current study are compatible with local neutrophil activation leading to epithelial damage, even though our results do not prove a causative relationship and the sequence of mechanisms over time remain to be confirmed. A link between damage of the respiratory epithelium and airway responsiveness has previously been indicated by several studies on airways in vitro and in vivo (reviewed in [12,22]). Of particular interest in this context, a very recent study of asthma in paediatric patients with RSV infection demonstrates that CrB+ is indeed associated with a higher symptom score after discharge from acute medical treatment . This finding thus suggests CrB+ as a marker of potential clinical utility for monitoring severe paediatric asthma.
In our current study, there was a difference in mean age between CrB+ and CrB– patients, with the CrB+ patients being somewhat younger by average. The relative number of sputum eosinophils also tended to be somewhat lower among younger patients (data not shown) but, importantly, we have no reason to suspect that this can explain why CrB+ patients have signs of neutrophil mobilization and CrB– patients do not, as shown in our current study. This is because, in a few cases, when individual CrB+ patients could be matched for age with individual CrB– patients, the CrB+ patients still displayed the higher concentration of IL-8 and NE (data not shown). Also, noteworthy, we did not detect any CrB+ among five paediatric patients without asthma that were investigated separately. Moreover, several clinical studies have shown signs of neutrophil mobilization in acute exacerbations of severe asthma among adult patients as well . It has also been reported that neutrophil mobilization is associated with severe and recurrent wheezing in children less than 3 years old, based upon the examination of inflammatory cells and mediators in BAL fluid . Taken together, these findings all support that neutrophilic inflammation is associated with epithelial damage in the airways among patients with paediatric asthma.
In conclusion, our current study of acute exacerbations of paediatric asthma where there are no signs of infection, does produce indirect evidence that epithelial damage is associated with local mobilization of neutrophils, implying that this is a sign of more severe airway inflammation. Our study did not find any corresponding evidence for eosinophils. The current results prompt new studies to evaluate how concurrent epithelial damage (as indicated by CrB+) and neutrophil mobilization relate to the clinical course of childhood asthma in the long term.