Differences in Airway Cytokine Profile in Severe Asthma Compared to Moderate Asthma^* FREE TO VIEW

Severe asthma affects only a small proportion of patients with asthma, but it is responsible for an substantial proportion of the morbidity and costs related to treating this condition.¹² Severe or treatment-resistant asthma is likely a heterogeneous condition, and multiple factors have been proposed to explain the difficulty in managing these patients. These factors include poor compliance with treatment, lack of recognition of associated conditions, or true biological resistance to therapy.³ Distinct patterns of airway remodeling⁴ and inflammatory marker profiles⁵ have emerged as important areas of investigation in the attempt to better understand and define severe asthma.

Asthmatic airway inflammation is associated with an exuberant T-helper (Th) type 2 immune response that results in sputum and tissue eosinophilia. These processes are usually responsive to treatment with corticosteroids.⁶ However, some asthmatic patients are resistant to usual doses of steroids by mechanisms that have been extensively studied.⁷⁸⁹ One of the principal causes appears to be a failure to suppress T-cell cytokines such as interleukin (IL)-2, IL-4, and IL-5. It is therefore reasonable to anticipate that the inflammatory process in severe asthmatics might show a different pattern of cytokines to well-controlled moderate asthmatics. In order to address this issue, we recruited patients with severe asthma and a comparison group with moderate asthma and compared their clinical characteristics and several biomarkers, including induced-sputum inflammatory cells and exhaled nitric oxide. Furthermore, those subjects who consented underwent bronchoscopy, and bronchial biopsy specimens were obtained to compare prototypical Th-1 and Th-2 cytokines and neutrophil- or eosinophil-associated chemokines in the airway walls in order to help elucidate differences in the pattern of inflammation associated with severe asthma. We chose patients with moderate persistent asthma, well controlled on usual maintenance therapy, as a comparison group in order to limit the extent to which the results may be confounded by treatment. Both groups were treated with inhaled corticosteroids. The data were presented in abstract form at the Annual Meeting of the American Thoracic Society (ATS), May 2007.

Physicians at two tertiary care centers were asked to identify patients with severe asthma who met the following criteria proposed by the ATS workshop on refractory asthma.¹⁰ For severe asthma, subjects were required to have received treatment with daily oral steroids for > 50% of the previous 12 months or to have been prescribed high-dose inhaled-steroid equivalent to > 880 μg/d of fluticasone, (ex-actuator) with at least one other add-on therapy (long acting β-agonist, leukotriene receptor antagonist, or theophylline) continuously over the prior 12 months (major criteria), as well as fulfilling at least two of the following six minor criteria: need for daily short-acting β-agonist; persistent airflow obstruction (prebronchodilator FEV₁ < 80% and FEV₁/FVC < 70% of predicted); one or more urgent-care visits in the last 12 months; three or more steroid bursts in the last 12 months; prompt deterioration with ≤ 25% dose reduction of oral corticosteroids; or near-fatal asthma event in the last 3 years.

For comparison, we recruited a group of patients with moderate asthma by asking clinicians who had referred patients with severe asthma to identify subjects of similar age and gender distribution whose asthma was persistent but not severe according to the following criteria: asthma controlled on a minimum of 176 μg/d of fluticasone (or equivalent) and not > 880 μg/d of fluticasone (or equivalent) with or without concomitant therapy with a long-acting β-agonist, long-acting β-agonist leukotriene receptor antagonist, or theophylline; no more than two steroid bursts in the past year, and none in the past 3 months with total days on oral steroids < 30 days in the prior 12 months; FEV₁ > 70% of predicted and > 90% of personal best from the past 2 years; and a maximum of one unscheduled visit for asthma in the prior 12 months.

All subjects met ATS criteria for the diagnosis of asthma, were between the ages of 18 and 69 years, were not current smokers (ex-smokers who had quit for at least a year were included), did not have any other known pulmonary disease including COPD, and had no major comorbid diseases such as HIV infection, cancer, or congestive heart failure. The study was approved by the research ethics committee of each participating center. All subjects gave their written consent.

Subjects recruited with severe asthma underwent a 1-month run-in period for the assessment of compliance. These subjects were provided with a combination metered-dose inhaler containing fluticasone, 250 μg, and salmeterol, 25 μg, in each inhalation, and were instructed to take two inhalations bid. An electronic device (MDILog; Medtrac Technologies; Lakewood, CO) was attached to each subject to accurately monitor use.¹¹ We included only subjects who inhaled at least the four inhalations prescribed per day on at least 70% of days. This criterion led to the exclusion of 25% of subjects, approximately. Furthermore, subjects with severe asthma underwent evaluation and coaching by a certified asthma educator for the assessment of inhaler technique and environmental control measures. The current report pertains to 24 patients with severe asthma and 26 patients with moderate asthma who agreed to undergo bronchoscopy and from whom adequate bronchial biopsy specimens were obtained. Tissues from some of the subjects have been used for the comparison of sputum and tissue neutrophilia and eosinophilia, and have formed the basis of a separate report.¹²

Study subjects completed the Asthma Control Questionnaire as described by Juniper and colleagues.¹³ Spirometry was performed according to ATS recommendations.¹⁴ Exhaled nitric oxide analysis was performed according to ATS guidelines.¹⁵ An expiratory flow rate between 0.315 L/s and 0.385 L/s was used during collection of exhaled gas. Sputum induction was initiated using aerosolized normal saline solution. If well tolerated, hypertonic saline solution was used at increasing concentrations (3%, 4%, and 5%) delivered by an ultrasonic nebulizer, as described by Pin et al.¹⁶ After each period of inhalation, FEV₁ was measured for safety. Each subject was asked to blow his nose, rinse his mouth, and swallow the water to minimize contamination with postnasal drip and saliva. Subjects were instructed to cough into a sterile container, and the expectorated sputum was processed within 2 h. The samples were processed as described previously.¹⁷ Cytospins were prepared (Shandon Somerset; Cheshire, UK) on poly-L-lysine coated slides and fixed using acetone/methanol (60:40) for 7 min followed by air drying. A sample was considered adequate if the slides showed < 20% squamous epithelial cells. The cytospins were immunostained for major basic protein (MBP) and neutrophil elastase using the alkaline phosphatase anti-alkaline phosphatase method and monoclonal primary antibodies for MBP (a generous gift from Dr. R. Moqbel, University of Alberta, Edmonton, AB, Canada) and elastase (DakoCytomation; Mississauga, ON, Canada).

Bronchoscopy was performed under light sedation with midazolam. Endoscopic bronchial biopsy samples were obtained from various segmental and subsegmental carinae of the right lung (Olympus Biopsy Forceps 35C; Olympus Medical Systems Corporation; Tokyo, Japan), fixed in formaldehyde and embedded in paraffin prior to cutting. A peroxidase-based method of staining was used as previously described¹⁸ using 5-μm sections. Immunostaining was performed using monoclonal antibodies for IL-4, IL-5, interferon (IFN)-γ, IL-8, and eotaxin (R&D Systems; Minneapolis, MN). Mouse IgG₁ served as the isotype control for IL-4, IL-5, IL-8, and eotaxin, while mouse IgG₂A was the isotype control for IFN-γ. StreptABComplex/HRP and DAB substrate (DakoCytomation) were used to develop the reactions according to manufacturer guidelines. Slides were examined and images acquired by a BX51 Olympus microscope attached to a CoolSNAP-Pro color digital camera (Carsen Group; Markham, ON, Canada) using Image Pro-plus 4.0 (Media Cybernetics; Silver Spring, MD). Scoring of epithelial staining for IL-8 and eotaxin, based on the percentage of the epithelium that stained positively, was performed as follows: 0 = no staining, 1 = 0 to 12.5%, 2 = 12.5 to 25%, 3 = 25 to 37.5%, 4 = 37.5 to 50%, 5 = 50 to 62.5%, 6 = 62.5 to 75%, 7 = 75 to 87.5%, and 8 = 87.5 to 100%. Subepithelial immunoreactive cells are reported as the number of cells that stained positively per field with a given antibody. The analysis was performed by an observer blinded as to group status.

Subjects underwent allergy skin testing to 13 common aeroallergens (Dermatophagoides pteronyssinus, Dermatophagoides farinae, cat, dog, horse, ragweed, grass mix, tree mix, weed mix, feather mix, Alternaria, Aspergillus, Cladosporium) as well as positive (histamine, 10 mg/mL) and negative (diluent) controls (Medic-Savoure; Dutton, ON, Canada). A test result was considered positive if the larger of two perpendicular diameters was ≥ 3 mm greater than the control.

Mean results of each group were compared using t tests. Medians and interquartile ranges were also calculated for continuous variables and differences between patients with moderate and severe asthma compared using the Wilcoxon rank-sum test. The results were similar to those obtained with t tests such that only these are presented. Proportions were compared by χ². Correlations were performed using the Spearman rank test, and ρ values that were significant at p < 0.01 level of significance are reported.

Characteristics of the 26 subjects with severe asthma and 24 subjects with moderate or treatment responsive asthma are provided in Table 1 . Subjects were of similar age, and there was a trend toward a higher proportion of women in the severe group. Four subjects in each group had smoked > 10 pack-years, and a similar proportion was obese as judged by a body mass index (BMI) ≥ 30 kg/m². Among those in whom the age of onset could be determined, 18 of 21 subjects in the moderate group and 20 of 23 subjects in the severe group reported onset of symptoms or diagnosis at ≥ 12 years of age. Sixteen of 22 subjects in the severe group in whom allergy skin tests were interpretable had two or more positive reactions. Allergy skin testing was only available in 21 subjects in the moderate group, 19 of whom had two or more positive reactions to 13 common aeroallergens. As expected, the severe asthma group was more symptomatic, had a lower FEV₁, and was receiving higher doses of inhaled corticosteroids, expressed in Table 1 as micrograms of fluticasone. All subjects in the severe asthma group were receiving concomitant long-acting β-agonists, and 9 of 24 subjects with treatment-resistant asthma were receiving oral corticosteroids on a regular basis.

Expired nitric oxide was similar in the two groups: 11.9 ± 4.1 parts per billion in the severe subjects, and 13.3 ± 8.5 parts per billion in the moderate subjects. The severe group had a higher percentage of both eosinophils (12.1 ± 7.01%) compared to the moderate subjects (6.4 ± 3.91%; p = 0.0014) and neutrophils (51.7 ± 18.57% and 38.5 ± 14.01%, respectively; p = 0.0073) in their sputum. There were no significant correlations between nitric oxide and sputum eosinophils and FEV₁ (ρ = − 0.036 and ρ = 0.029, respectively). However, sputum neutrophil counts showed a modest inverse correlation with FEV₁ (ρ = − 0.529; p < 0.01).

We assessed the number of immunoreactive cells in the airways to assess IL-4 and IL-5 expression as representative Th-2 cytokines, whereas we assessed IFN-γ as the prototypical Th-1 cytokine. Subepithelial IL-5 expression was similar in the severe and moderate asthmatic subjects. In contrast, IL-4 expression was significantly less in the airway subepithelial compartment of subjects with severe compared to moderate asthma. Figure 1 (top, A) shows an illustrative example of IFN-γ expression in subepithelial cells that are predominantly inflammatory cells. Subepithelial IFN-γ–expressing cells were present in larger numbers in the severe group than in the moderate group. There was no evidence of cytokine expression by airway smooth muscle in either group.

Illustrative examples of immunoreactivity for IL-8 are shown in Figure 1 (center, B, and bottom, C). IL-8 immunoreactivity was clearly seen in the epithelial and subepithelial compartments and was greater in the severe asthma subjects. Eotaxin 1 expression was not different between groups. IL-8 and eotaxin immunoreactivity was detected in airway smooth muscle in both groups, but we did not assess this tissue for possible differences in expression levels. The results of the morphometric analysis of bronchial biopsies are provided in Table 2 and in Figures 2, 3 .

None of the cytokines or chemokines examined showed any correlation with FEV₁. However, epithelial and subepithelial expression of IL-8 was inversely correlated with sputum eosinophilia (ρ = − 0.516 and ρ = 0.479, respectively; p < 0.01).

We studied subjects with severe poorly controlled asthma, compliant with medication, and compared them to patients whose disease was moderate and managed successfully with regular inhaled steroids in conjunction with long-acting β-agonists in many instances. Despite high doses of inhaled or supplemental doses of oral corticosteroids, the patients with severe asthma had more airway inflammation as reflected in sputum eosinophilia and neutrophilia. Furthermore, these subjects expressed a chemokine (IL-8) and the cytokine (IFN-γ) in their bronchial walls, which are thought of as Th-1 inflammation-associated molecules more typical of the neutrophilic inflammatory response of patients with COPD and cystic fibrosis. There was lower expression of IL-4, a typical Th-2 cytokine. Bronchial IL-5 and eotaxin expression were similar among patients in the severe and moderate groups.

Sex distribution and BMI were similar between the subjects with severe and those with moderate asthma. This is in contrast to the much larger multicenter European series,¹⁹ in which women predominated and, among women, those with severe asthma were heavier. The subjects with severe asthma in our series were symptomatic, had a low mean FEV₁, and were receiving high doses of corticosteroids. In these respects, our subjects with severe asthma are quite comparable to the group of patients with treatment-resistant asthma reported by Heaney and colleagues.²⁰ There was no difference in the age of onset of asthma among subjects with severe asthma as compared to moderate, responsive asthma. In this respect, our severe asthma group of subjects differed from that reported on by Miranda and colleagues²¹ in that most of our subjects reported the onset of asthma after the age of 12 years. The moderate group was receiving long-acting β-agonists in 61% of cases, which may mean that the group had somewhat more severe asthma than moderate asthmatics as defined by ATS criteria. However, the use of this controller medication may have reduced the observed differences between the groups and strengthens any conclusions related to observed differences in inflammation. Our series of patients has two important characteristics. Firstly, the diagnosis of asthma and the exclusion of competing diagnoses were ensured by prior evaluation and prolonged follow-up by expert physicians. Secondly, patients in the severe asthma group were objectively monitored for compliance during a 1-month period prior to being accepted into the study, thereby excluding patients whose disease severity may have resulted from inadequate therapy.

The pattern of inflammation in the sputum differed between the severe asthma and moderate asthma groups. Subjects with severe asthma had a greater proportion of eosinophils and neutrophils in induced sputum than subjects with moderate, consistent with other reports²² of persistent inflammation despite usual treatment. We did not find a subgroup without eosinophilia; the lowest sputum eosinophil count in the severe group was 2%. Immunocytochemistry for the eosinophil granule MBP as opposed to usual histochemical stains of sputum may have enhanced the sensitivity of detection of low levels of eosinophilia. There were also differences in some of the important mediators of eosinophilia and neutrophilia in the airway wall biopsy findings. Notably, IL-8–expressing cells were increased, particularly in the subepithelial compartment of the airway wall, and this potent chemoattractant for neutrophils²³²⁴ may have accounted for the excess of neutrophils that was seen in the severe group as compared to those with more easily controlled or moderate asthma. Not surprisingly, there was an inverse relationship between sputum eosinophilia and epithelial IL-8 expression. There is increasing recognition of the association of sputum neutrophilia in the sputum with severe asthma and with virally induced exacerbations of asthma.²⁵²⁶ However, the effects of IL-8 may not be limited to actions on neutrophils. IL-8 may act on airway smooth muscle itself to cause contraction or may serve as a chemoattractant, potentially contributing to airway remodeling through promoting migration of smooth-muscle cells within the airway wall.²⁷

The stimuli that trigger the expression of IL-8 are not entirely clear. Endotoxin exposure and viral infection could result in increased expression of IL-8.²⁸²⁹ It seems possible that refractory asthma might be associated with persistent viral infection, as has been described in COPD.³⁰ IL-8 is increased also by smoking, but there were no current smokers among the patients we studied and an equal number of ex-smokers among patients with moderate and severe asthma. None of the above mechanisms is therefore likely to be pertinent to the current study. Leukotriene D₄, acting through the cysteinyl-leukotriene-1 receptor, was shown to transcriptionally regulate the synthesis of IL-8 through the transcription factors nuclear factor-κB and activating protein 1.³¹ Corticosteroids appear to favor IL-8 expression in vivo. An increase in the expression of IL-8 in the epithelium and an increase in airway neutrophils have been observed in patients receiving oral corticosteroids.³² IL-8 was increased in both the epithelium and subepithelium of our group with severe asthma, approximately a third of whom were receiving regular oral corticosteroids as well as a significantly higher average dose of inhaled corticosteroids as compared to the moderate treatment-responsive group. However, the apparent up-regulation of IL-8 by steroids is likely to be an indirect mechanism because there is good evidence that corticosteroids directly inhibit IL-8 production by human epithelial cells in vitro.³³

Subepithelial IFN-γ was increased and IL-4 was decreased in patients with severe asthma as compared to the control patients with moderate asthma. This interrelationship between these two cytokines is expected as they are counter regulatory.³⁴ IFN-γ in the sputum of patients with asthma has been related to asthma severity.³⁵³⁶ The source of the IFN-γ in the sputum appears to be CD8⁺ lymphocytes³⁵; however, we are not certain as to the cellular source in the airway tissues. Serum levels of IFN-γ have been found to relate to a greater decline in FEV₁ over time in patients with asthma.³⁷ The cells expressing IL-4 are also uncertain, but mast cells are a likely source.³⁸ One possible interpretation of the findings for the airway expression of IL-4 and IFN- γ is that treatment may have been adequate to suppress Th-2 inflammation to a greater extent than in moderate subjects and that other factors must then have accounted for the clinical manifestations of asthma that caused it to be more severe.

IL-5 and eotaxin in the airway wall were similar in patients with severe and moderate asthma despite the fact that sputum eosinophilia was more marked in the severe group. The discrepancy is possibly due to the fact that sputum is representative of both the proximal and distal airways while endobronchial biopsies sample the proximal airways alone. It is likely that there is more distal inflammation in severe asthma compared to moderate disease.³⁹ It is also possible that the trafficking of eosinophils through the airway wall is increased as a result of other eosinophil chemoattractants such as leukotriene B₄, the cysteinyl leukotrienes, or 5-OXO-eicosatetranenoic acid,⁴⁰ all products of the 5-lipoxygenase pathway. As reported by others,¹⁹ expired nitric oxide did not differentiate between subjects with moderate and severe asthma, perhaps because all subjects were receiving relatively high doses of corticosteroids.

In conclusion, patients with severe treatment-refractory asthma have a different cytokine profile in bronchial biopsy findings than a group with moderate asthma, demonstrating an increase in IL-8– and IFN-γ–expressing cells. The expression of the principal Th-2 cytokines was not more elevated in the group with severe asthma and suggests that persistence of Th-2 inflammation does not account for differences in the severity of disease. A sample of subjects with severe asthma fulfilling the same selection criteria showed evidence of more extensive airway remodeling,⁴¹ in particular in airway smooth-muscle mass. Differences in remodeling may therefore provide an alternative or additional mechanism to account for disease severity in this population.