Airway TGFβ₁ and oxidant stress in children with severe asthma: Association with airflow limitation

INTRODUCTION

Severe asthma in children is a complicated disorder encompassing children with both difficult-to-treat asthma and severe, refractory asthma¹ that is characterized by poor symptom control despite best attempts at treatment with high doses of inhaled corticosteroids (ICS) and other asthma controller medications.^2–4 The symptoms that accompany severe asthma in children are further associated with marked deficits of pulmonary function, including airflow limitation, that are not completely reversible with bronchodilator therapy and worsen significantly over time.^{5, 6} Although the specific mechanisms underlying pulmonary function derangements in children with severe asthma are yet unclear, structural changes in the airway wall (i.e., “remodeling”) are likely responsible.^{7, 8}

Transforming growth factor beta 1 (TGFβ₁) is a pro-fibrotic cytokine that is associated with airway fibrosis.⁹ TGFβ₁ is synthesized as a pre-propeptide precursor and is secreted as an inactive heterodimer bound to a latent complex that is targeted to TGFβ₁ binding proteins within the extracellular matrix. While there are several molecular mechanisms involved in the activation of TGFβ₁, exposure to reactive oxygen species has been shown to induce rapid cleavage of TGFβ₁ from the latent complex, presumably through alteration of the latent associated protein.¹⁰ Once TGFβ₁ is cleaved to the active (free) form, it binds to TGFβ₁ receptors on target cells in the extracellular matrix. While this process is essential for airway wound repair and healing, sustained activation of TGFβ₁ inhibits beneficial TGFβ₁ signaling¹¹ and also activates fibroblasts and the epithelial-mesenchymal trophic unit, resulting in further release of reactive oxygen species and ultimately airway remodeling.^{12, 13} The end result is a vicious cycle of oxidant stress whereby TGFβ₁ is chronically activated.

Although airway oxidant stress is a prominent feature of severe asthma in children, the relationship of oxidant stress and TGFβ₁ in this population has not been described. Compared to children with mild-to-moderate asthma, children with severe asthma have marked depletion of the airway antioxidant, glutathione, which corresponds to the magnitude of airflow limitation.^{14, 15} This disturbance is associated with increased formation of lipid peroxidation byproducts (8-isoprostanes and malondialdehyde),¹⁵ greater DNA nucleoside¹⁴ and nitric oxide¹⁶ oxidation, and sustained expression of pro-inflammatory cytokines and chemokines.¹⁷ Whereas key transcription factors such as nuclear factor (erythroid-derived 2)-like 2 (Nrf2) should stimulate glutathione and antioxidant defenses under these circumstances, Nrf2 is not active in these children and thus the oxidant burden is intensified.¹⁸ While the mechanisms underlying these observations remain poorly understood, TGFβ was recently shown to inactivate Nrf2 in airway smooth muscle cells from subjects with severe asthma.¹⁹ Thus TGFβ₁ may be a key mediator of the airway oxidant burden in severe asthmatic children. Given the relative paucity of pediatric severe asthma bronchoscopy studies and the magnitude of symptoms in this population, we sought to: 1) quantify TGFβ₁ protein and mRNA gene expression in the airways of children with mild-to-moderate and severe atopic asthma, and 2) determine the relationship of airway TGFβ₁ concentrations to oxidant burden (i.e., lipid peroxidation), Th2-mediated eosinophilic inflammation, and airflow limitation. We hypothesized that TGFβ₁ expression and activation would be increased in children with severe asthma and would be associated with increased 8-isoprostane, malondialdehyde and IL-13 concentrations, increased exhaled nitric oxide concentrations, and a greater magnitude of airflow limitation.

METHODS

RESULTS

Sixty-eight children with atopic asthma (severe asthma, n = 28) and 12 atopic adult controls were recruited. Indications for bronchoscopy in children are shown in the online repository (Table E1). Bronchoscopy was well tolerated in all subjects with no adverse events. Features of the groups appear in Table I. Within the mild-to-moderate group, 19 had mild asthma treated with low doses of inhaled corticosteroids (≤ 220 mcg inhaled fluticasone/day for age <12 years; ≤ 440 mcg of inhaled fluticasone/day for age ≥12 years). Children with severe asthma, by definition, were prescribed higher doses of ICS and additional controller medications but had greater healthcare utilization for asthma in the preceding year. Children with severe asthma also had higher exhaled nitric oxide concentrations and increased airway lavage eosinophils that corresponded to elevations of total serum IgE (control: 144 ± 61 kU/L; mild-to-moderate: 123 ± 33 kU/L; severe: 549 ± 143 kU/L, p < 0.001). Airflow limitation both at baseline and after bronchodilation were also features of the severe asthma group. However, at the same time, more children with severe asthma had FEV₁ bronchodilator reversibility >12% compared to children with mild-to-moderate asthma (57% vs. 16%, p = 0.001), most likely due to a higher baseline FEV₁ percent predicted value in the mild-to-moderate group. Other features of the airway lavage such as the protein and urea content did not differ (control vs. mild-to227 moderate vs. severe, protein: 134 ± 84 vs. 177 ± 116 vs. 187 ± 74 µg/mL, p = 0.144; urea: 0.57 ± 0.41 vs. 0.51 ± 0.48 vs. 0.40 ± 0.59 mg/dL, p = 0.104).

DISCUSSION

To our knowledge, this is the first study to comprehensively assess airway TGFβ₁ expression and activation and the relationship of airway TGFβ₁ to Th2-mediated inflammation, oxidant stress and functional assessments of airway remodeling in a heterogeneous sample of children with atopic asthma. A unique strength of the study is the inclusion of children with severe asthma, who, when compared to children w ith mild-to-moderate asthma and atopic adult controls, had higher airway TGFβ₁ concentrations and increased protein and gene expression of TGFβ₁ in airway macrophages. Moreover, the airway oxidative burden reflected by the lipid peroxidation biomarkers 8-isoprostanes and malondialdehyde was also greater in children with severe asthma and was associated with increased TGFβ₁ cleavage and activation. Furthermore, within the combined sample of children with mild-to-moderate and severe atopic asthma, airway TGFβ₁ concentrations were associated with airflow limitation. Together, these data suggest that increased expression and activation of airway TGFβ₁, along with an increased airway oxidant burden, may underlie the airway physiological alterations that accompany severe asthma in children.

TGFβ₁ is a pleiotropic cytokine that is widely expressed by airway cells and is thought to play a key role in airway remodeling in asthma.⁹ While increased TGFβ₁ protein and mRNA gene expression have been previously described in the airways of asthmatic adults at baseline and after segmental allergen challenge,^23–27 airway TGFβ₁ expression and activation have not been well quantified in children and no studies to date have focused on children with severe asthma. In two recent studies of predominantly mild asthmatic children who underwent endobronchial biopsy for clinical indications, no qualitative differences in the number of TGFβ₁-expressing subepithelial cells were noted as compared to controls.^{28, 29} Thus it was concluded that TGFβ₁ may be downregulated in childhood asthma.²⁹ While we also failed to observe differences in TGFβ₁ expression between mild-to-moderate asthmatic children and atopic adults, our finding of increased TGFβ₁ protein and mRNA gene expression in children with severe asthma refutes this claim. Moreover, we observed increased cleavage and activation of TGFβ₁ in children with severe asthma associated with a significant (i.e., 2-fold) increase in the lipid peroxidation biomarkers 8-isoprostanes and malondialdehyde. While the causal relationship between TGFβ₁ and lipid peroxidation i s not clear, the 8-isoprostane and malondialdehyde concentrations observed here are similar to what we have previously reported in this population.^14–16 Although additional studies are needed, these findings support previous findings of TGFβ₁-mediated inactivation of Nrf2¹⁹ and suggest that an ongoing cycle of TGFβ₁ activation from sustained generation of reactive oxygen species may underlie irreversible airway injury and remodeling in these children.

While the airway epithelium is the predominant site of airway TGFβ₁ expression in the lung,³⁰ other cells such as fibroblasts³¹ and airway smooth muscle cells³² also contribute to airway TGFβ₁ pools in asthmatic subjects. In the present study, although airway epithelial cells were not available for analysis, we observed increased TGFβ₁ protein and mRNA expression in airway macrophages from children with severe asthma similar to previous findings in asthmatic adults.³³ These observations are also in keeping with our previous observations of increased airway macrophage activation and pro-inflammatory cytokine release in this population of children.¹⁷

In the present study, we did not observe a clear association between TGFβ₁ expression and airway lavage eosinophils in our sample of asthmatic children. However, others have shown that the number of airway lavage eosinophils does not necessarily correlate with mucosal eosinophilia in children with severe asthma.³⁴ While airway eosinophils are known to be a significant source of TGFβ₁ in patients with asthma,²⁴ insufficient eosinophils counts in the present study prohibited isolation of these cells and thus their specific contribution to airway TGFβ₁ pools is unclear. However, the role of the eosinophil in these findings warrants further study. In previous investigations involving both animals³⁵ and humans,³⁶ inhibition of IL-5, a key mediator of eosinophil activation, attenuated TGFβ₁ expression and subsequent airway remodeling.

We also failed to observe significant associations between airway TGFβ₁ and exhaled nitric oxide concentrations. This is in contrast to previous work by Mahut et al. which demonstrated correlations between alveolar nitric oxide and BAL TGFβ₁ concentrations in asthmatic children.³⁷ However, that study did note a strong association between maximal conducting exhaled nitric oxide output and reticular basement membrane thickening.³⁷ Although the clinical relevance of reticular basement membrane thickening remains controversial,³⁸ reticular basement membrane thickness has been associated with airflow limitation in children with asthma^{7, 8} and therefore supports our observed associations between exhaled nitric oxide and airflow limitation in this sample.

Because increased TGFβ₁ is associated with allergic sensitization independent of asthma,³⁹ all of the subjects included in this study were atopic. While the inclusion of atopic subjects was important for our TGFβ₁ analyses, the exclusion of non-atopic subjects may have decreased the variability in our other measures. For instance, although airway lavage eosinophils and IL-13 concentrations were increased in both groups of asthmatic children, they did not differentiate asthma severity. Given that IL-13 has been shown to induce TGFβ₁ expression,⁴⁰ our observation of increased IL-13 in children with mild-to-moderate asthma may appear counterintuitive since TGFβ₁ expression was not significantly different in this group. However, we did observe correlations between IL-13 and free TGFβ₁ concentrations in the ELF, suggesting that IL-13 may be more relevant to the process of TGFβ₁ activation. Furthermore, TGFβ₁ and IL-13 are also synergistic⁴¹ and thus TGFβ₁ is essential for IL-13-mediated airway remodeling and fibrosis.¹³ This synergism between IL-13 and TGFβ₁ likely explains why airflow limitation was less pronounced in our sample of children with mild-to-moderate asthma despite higher expression of IL-13 in the airways.

While severe asthma in children is clearly a heterogeneous condition associated with a number of clinical phenotypes,⁴ airflow limitation is common feature of the disorder that is identifiable during early childhood.^{2, 3} Although airflow limitation is less pronounced in children as compared to adults with severe asthma, it persists across the lifespan and declines further as a function of the normal aging process.⁴² A subset of children with severe asthma also experiences a rapid worsening of airflow limitation during the early adolescent years⁵ which further increases the risk of severe, life-threatening asthma exacerbations and other adverse outcomes.⁴³ While the mechanisms underlying the progression of airflow limitation in children with severe asthma are not clear, our findings suggest that TGFβ₁ may play a role in this process. Importantly, airway TGFβ₁ expression in children with severe asthma was increased to a similar magnitude as what has been previously reported in severe asthmatic adults.²⁵ Furthermore, in this cross-sectional sample, airway TGFβ₁ concentrations were associated with airflow limitation, even after controlling other relevant factors. However, whether airway TGFβ₁ concentrations predict loss of lung function over time is unclear and warrants further study.

This study does have a number of limitations. Because bronchoscopy could only be performed for clinical indications in children, our control group was limited to atopic adults. Whether there are developmental differences in airway TGFβ₁ expression between adults and children is not clear. Also the utilization of clinical bronchoscopy samples from children resulted in relatively small group sizes wherein a number of comparisons were made, potentially increasing the risk of Type I statistical error. Utilization of clinical bronchoscopy samples also prohibited the collection of endobronchial biopsies for histological analysis and the isolation of other cell types, such as airway eosinophils and fibroblasts, which are important sources of TGFβ₁.^{9, 24} While it is also unclear how asthma medications may have affected the results, comparisons between children with severe and mild-to-moderate asthma help to eliminate potential confounding by asthma medication exposure and other key variables in the medical history. Although all children in this study were prescribed treatment with ICS, another potential shortcoming is related to adherence with ICS therapy since adherence was not rigorously assessed. The striking racial disparity between our asthmatic groups is also a potential source of confounding that warrants additional investigation. However, despite the relationship between race and airflow limitation that we observed here, we failed to observe consistent associations between race and airway TGFβ₁ expression (Online repository Figure E2).

In summary, we have demonstrated increased TGFβ₁ protein and mRNA gene expression in the airways of children with severe asthma associated with increased concentrations of the lipid peroxidation biomarkers 8-isoprostanes and MDA. Contrary to our hypothesis, we failed to observe consistent associations between airway TGFβ₁ and Th2-mediated eosinophilic inflammation and exhaled nitric oxide. However, TGFβ₁ activation as measured by free TGFβ₁ did correlate with concentrations of IL-13 in the ELF. While the mechanisms underlying these observations are not clear, these findings suggest an altered feedback mechanism in the airways of children with severe asthma resulting in a vicious cycle of airway oxidant stress and persistent TGFβ₁ activation. Given that TGFβ₁ concentrations were associated with airflow limitation in this sample of asthmatic children, additional studies are now warranted to determine how TGFβ₁ modulates longitudinal derangements of lung function and other outcomes in asthmatic children. Ultimately, interventions to attenuate TGFβ₁ expression and activation may be warranted in children with severe asthma to the structural airway changes and resulting symptoms associated with the disorder.