Given advances in our understanding of the molecular pathways involved in asthma and in particular the key cytokines, there is a major advantage to using modern molecular techniques to characterize asthma heterogeneity. Creating subgroups of asthma based on the activity (or lack thereof) of specific cytokine pathways immediately identifies mechanisms underlying clinical phenotypes, new pharmaceutical targets, biomarkers for clinical trials of targeted pharmaceuticals, and has the potential to predict treatment response in the clinic.
Epithelial cells are anatomically well positioned to be major drivers of airway remodeling, including sub-epithelial fibrosis (44
) and smooth muscle hyperplasia (48
). Based on the knowledge that airway epithelial cells respond to the cytokine milieu to contribute to airway immune responses (51
), our group has studied the gene expression profiles of epithelial brushings obtained through research bronchoscopy in a well-characterized cohort of asthmatics and healthy controls as a means of identifying distinct molecular phenotypes (52
). For characterization of molecular phenotypes based on native disease pathophysiology, we have focused first on mild to moderate asthmatics who are not using inhaled or systemic corticosteroids, as steroids likely alter epithelial gene expression. Among the most highly induced genes in mild asthmatics as compared to healthy controls were chloride channel, calcium-activated, family member 1
, and serine peptidase inhibitor, clade B (ovalbumin), member 2
, also known as plasminogen activator inhibitor-2
). Induction of periostin
, an integrin ligand and extracellular matrix protein with roles in cell adhesion, cell motility, matrix remodeling, and fibrosis (54
), and serpinB2
a member of the serpin class of proteases that promote fibrin formation and deposition and may be involved in the regulation of immune responses (56
), was confirmed at the protein level as well. All three of these genes were found to be directly regulated by IL-13 and IL-4 in isolated human airway epithelial cells in vitro
). These data established that the expression of three genes, CLCA1, periostin
, and serpinB2
, serves as a surrogate marker for the effects of IL-13/IL-4, and hence Th2 inflammation, in the airway.
Among the top 10 most differentially expressed genes in asthmatics were two mast cell proteases, tryptase
and carboxypeptidase A3
. Mast cells have been proposed to have a critical role in airway hyper-responsiveness (59
), though their presence in the epithelial compartment in asthmatics has varied across studies, perhaps due to small sample sizes, inclusion of subjects using inhaled corticosteroids, and imprecise methods of quantification. Using design-based stereology (61
), we showed that asthmatics have an accumulation of intraepithelial mast cells, and through immunohistochemistry, that these mast cells express tryptase and CPA3 (63
) at the protein level. Stimulation of primary human bronchial epithelial cells at an air-liquid interface with IL-13 induced the expression of stem cell factor
, a growth factor and attractant for mast cells. This induction of stem cell factor
provides a mechanism for the increased numbers of intraepithelial mast cells in asthmatics. Additional studies of intraepithelial mast cells in asthma highlight their particular importance in severe asthma (64
Further analysis of these study subjects showed that nearly half of those with asthma were indistinguishable from healthy controls based upon the epithelial brushing expression of the three surrogate markers of IL-13 exposure, periostin, CLCA1
, and serpinB2
) (). This finding suggested that our population of asthmatics was heterogeneous; some had Th2-high and others Th2-low inflammation. Increased expression of IL-13
in Th2-high subjects, as assessed by quantitative polymerase chain reaction (qPCR) in bronchial biopsy specimens, provided further confirmation for this molecular phenotype (). Compared to healthy controls, both Th2-high and -low subjects had decreased FEV1
, bronchodilator responsiveness, and positive allergen skin-prick tests. However, Th2-high subjects were enriched for airway hyper-responsiveness (as measured by the concentration of methacholine required to effect a 20% decline in FEV1
methacholine]), serum IgE levels, and both blood and especially bronchoalveolar lavage eosinophilia. Th2-high subjects also had increased numbers of intra-epithelial mast cells (63
). Through rigorous application of design-based stereology, we found that reticular basement membrane thickness, a measure of sub-epithelial fibrosis and hence airway remodeling, was significantly increased in Th2-high but not Th2-low subjects. Furthermore, there were dramatic differences in the types of epithelial mucin genes expressed in Th2-high and Th2-low subjects, with an increased MUC5AC/MUC5B
ratio in Th2-high asthma. These data support the existence of a distinct asthma molecular phenotype as a result of increased Th2 activity, from which follow testable hypotheses about the molecular mechanisms of eosinophilic inflammation, airway hyper-responsiveness and airway remodeling.
Expression levels of three IL-13 induced genes in the airway define two subgroups of asthmatics
We performed a randomized placebo-controlled trial of inhaled corticosteroids in this same study showing that the Th2-low phenotype did not respond with an increase in lung function to inhaled corticosteroids as did the Th2-high group (). After inhaled corticosteroids, the Th2-high group had a reduction in expression of periostin, CLCA1, serpinB2, and the mast cell proteases tryptase and carboxypeptidase (CPA3), the latter suggesting a mechanism by which inhaled corticosteroids improved lung function in the Th2-high subgroup. Further investigation of the differences between the Th2-high and -low groups may improve our understanding of corticosteroid insensitivity and resistance, which as mentioned earlier is a significant unmet clinical need.
Responsiveness of ‘Th2-high’ asthma to inhaled steroids in a randomized placebo-controlled trial
Another question of interest is whether the Th2-high and -low molecular phenotypes are stable characteristics of individual subjects or whether they represent the effect of a changing environment. In our data, the airway epithelial gene expression signature is stable for one week, confirming technical reproducibility. Our finding of increased airway remodeling in Th2-high subjects suggests that the phenotype has stability in an individual long enough to lead to structural changes (53
). However, ongoing and future studies will be necessary to definitively test this hypothesis.