The studies we describe here provide novel insights into the differences between children whose asthma has resolved spontaneously and those in which it remains symptomatic. Increased levels of sputum eosinophils and sputum IL-5, increased peripheral blood mononuclear cell-derived TNF-α, IL-12, and decreased IFN-γ levels were associated with on-going asthma symptoms and airway hyperresponsiveness. We did not demonstrate any clear differences in T or B cell subsets, memory cells or activation markers between children with ongoing asthma symptoms and those in whom symptoms had completely resolved. However, where subtle differences were noted between the CA and RA groups, the RA group reflected values closer to those of the control subjects.
Initial clinical examination revealed that many of the asymptomatic subjects remained hyperresponsive as defined by methacholine challenge tests. In a sample of 551 adolescent and young adult subjects Kolnaar et al. observed that 42% of subjects exhibited airways hyper responsiveness, of which 70% were asymptomatic. These data would suggest a slightly lower rate in the population than our observation of hyper responsiveness in 8 of 15 asymptomatic RA subjects. A much larger study would be required to determine whether the rate of asymptomatic AHR in RA subjects was increased relative to the levels in the age matched general population. Other studies in childhood have suggested that asymptomatic bronchial hyperresponsiveness, in this case observed after allergen challenge, can predict the development of asthma. The lack of AHR in many of the RA subjects suggests that the AHR resolves in many subjects in which asthma symptoms resolve. It is interesting that there was no evidence of an increase in FEV1 reversibility in the RA group compared with the control group despite the high rate of AHR.
Analysis of induced sputum provides a unique window on the inflammatory status of the airways of the RA subjects and their CA counterparts. The most striking difference observed between these groups was in the levels of eosinophils and ECP observed. The high levels of these markers observed in CA were expected, based on extensive literature reports of eosinophilic airways inflammation associated with symptomatic asthma. Notably, despite the presence of positive skin tests and AHR in over half the RA group, both the ECP and sputum eosinophil levels were comparable with those of non-atopic control subjects and significantly lower than those in the CA group. The increase in eosinophils was highly specific, as observed in other asthmatic populations[33, 36] since no alterations were noted in the percentages of other cell types in the sputum. The local eosinophilia in the airways did not extend to significant changes in peripheral blood eosinophil numbers, although a peripheral eosinophilia has been described in some groups of asthmatics, especially in childhood[37, 38].
The levels of cytokines in the sputum supernatants reflected the eosinophilic nature of the ongoing inflammatory process associated with disease. The IL-5 levels were significantly elevated in the CA group compared with the control group. In contrast, the predominately anti-inflammatory cytokine IL-10 was found in similar amounts in all subject groups. This finding does not support the concept that subjects in which asthma has resolved have enhanced local production of IL-10 compared with those in which the disease persists. The intermediate levels of IL-5 and TNF-α observed in RA subjects suggest that in some individuals elevated local concentrations of these cytokines remain, although in the absence of substantial eosinophilia.
The relationship between eosinophilic inflammation and AHR in asthma is not entirely clear. In murine models of asthma, some have shown that AHR and eosinophilic inflammation can be dissociated[39–41] while others have demonstrated a link between them via IL-5[42–45]. Among the RA subjects, of whom 8 of 15 had AHR in the absence of symptoms, none had substantial levels of eosinophils in sputum, which seems to indicate that AHR can be present in certain subjects without a high level of eosinophilic inflammation. However, several of the RA subjects had increased IL-5 in their sputum in the absence of substantial eosinophilia, suggesting that perhaps this residual IL-5 may be associated with the residual AHR.
Our observations of elevated type 1 cytokine production by PBMC from CA subjects compared with control subjects were surprising in the context of a disease which is traditionally associated with inflammation driven by type 2 cytokines. There is recent information that supports the role of IFN-γ in normalizing asthma symptoms in childen. However, several pieces of information from the existing literature support the concept of an important role for the type 1 cytokine cascade in symptomatic asthma. In early studies, TNF-α was observed as increased in the BAL fluid of subjects with “symptomatic” rather than quiescent disease. Elevated production of TNF-α and IFN-γ by bronchoalveolar leukocytes from a series of 11 patients with allergic asthma compared with control subjects. Indeed, anti-TNF treatment has been reported to be beneficial in a subset of patients with severe asthma. More recently, using whole blood cultures Magnan et al. observed an overproduction of IFN-γ by CD8+ T-cells in the blood of asthmatic subjects which was related to asthma severity and suggested a role for IL-12 in inducing this response. Observations that IFN-γ inducing factor (IL-18) is increased in the airways of patients with atopic asthma support the concept that enhancing the type 1 cytokine cascade can be associated with the development of airways eosinophilia. Interestingly an elevated type-1 cytokine response, in atopic subjects, has also been implicated in studies of atopic dermatitis. Additional recent studies in human subjects and mice have reported the positive contribution of TH1 lymphocytes and IFN-γ to allergic lung inflammation[53–55]. It is also been reported that eosinophil-derived IFN-γ contribute to the persistence of airway hyperresponsivness and structural changes. Differences in type 1 cytokine production between asthmatic and control subjects might be missed experimentally if either the subjects were not withdrawn from corticosteroid therapy appropriately or complex stimuli such as whole bacteria were employed. Our data suggest that results obtained using SAC as a cytokine-inducing stimulus for PBMC does not reflect the baseline levels of cytokine expression or the response to more direct stimulation of the T-cell or Toll-like receptor systems.
The CA and RA subjects did not differ substantially in their production of any of the type 2 cytokines we measured in PBMCs. Other RA and CA groups in this study were predominately atopic and several studies have suggested that enhanced ability to produce these cytokines reflects atopic status[57, 58] rather than asthma specifically. Under appropriate stimulation conditions both IL-4 and IL-5 levels were elevated in supernatants from the CA and RA PBMC compared with the non atopic control group PBMC. This is in contrast with both IL-6 (data not shown) and IL-10 which were unchanged between groups regardless of the stimuli employed.
Flow cytometric studies of peripheral blood cells from all of the donors revealed few consistent or significant differences. All of the groups had similar proportions of T-cell subtypes, B-cells and monocytes suggesting that the changes we observed in cytokine expression were unlikely to be due to major alterations in cell distribution. The observation of elevated HLA-DR expression on the CD8 + ve T-cells from asthmatic subjects is consistent with the concept that these cells may be an important effector cell population or a source of type 1 cytokines in asthma. The increased expression of CD40 ligand that we observed in the asthma group compared with the control group could also relate to the presence of elevated type 1 cytokines since IL-12 has been demonstrated to be produced as a result of enhanced CD40/CD40L interactions.
This study is a unique attempt to evaluate the differences between carefully matched atopic subjects in whom asthma symptoms have or have not resolved. Previous studies of the immune parameters associated with growing out of asthma have concluded that lower serum IgE or fewer positive skin tests, male gender as well as less severe disease increase the likelihood that asthma will resolve spontaneously[9, 10, 20]. There has been some discussion that the resolution of asthma in late childhood/adolescence is temporary, although good prospective studies have not been performed in this area. The RA subjects examined by us have been surveyed 14–18 months after the completion of the study, and none of the subjects contacted had developed symptomatic asthma over this time period. Strategies aimed at the IL-5 dependent eosinophil recruitment into the airways or TNF-dependent airway hyperresponsiveness may hasten the resolution of asthma symptoms. Long-term prospective studies of children are necessary to investigate these strategies.
In conclusion, our studies of subjects who have “grown out” of asthma symptoms compared with those in whom disease continued has revealed that while many of these subjects retain some degree of AHR, there is little or no associated airways eosinophilia. As a group, those with continuous asthma demonstrated elevated production of type 1 cytokines compared with RA subjects. It is well recognised that corticosteroids are potent inhibitors of type 1 cytokine production, although this inhibition is temporary. The ongoing challenge is to develop new approaches to therapy which could induce the changes which take place during spontaneous disease remission. Further studies of subject groups such as the RA group we describe could provide important clues to this critical process.