Molecular Heterogeneity, Biomarker Discovery, and Targeted Therapy in Asthma - Genomic Biomarkers for Pharmaceutical Development - page 73

other half of asthma patients interspersed with all of the non-asthmatic controls. Thus, on
the basis of the bronchial epithelial expression levels of three genes, two molecularly distinct
subsets of asthma patients emerged, one of which was, on the basis of expression of these
genes, indistinguishable from healthy control subjects, despite having clear clinical manifes-
tations of asthma [13] .
To determine whether the coordinate elevated expression of CLCA1, serpinB2, and
periostin was potentially due to the activity of IL13, we assessed the expression of type 2
cytokines by quantitative RT-PCR (qPCR) in endobronchial biopsies obtained contempora-
neously with the epithelial brushings. Endobronchial biopsies are collected with forceps and
sample multiple layers of bronchial mucosal tissue, including the epithelium as well as sub-
epithelial layers including the lamina propria and airway smooth muscle bundles. In some
asthma patients, the epithelium and lamina propria exhibit infiltration of inflammatory cells
such as lymphocytes, eosinophils, neutrophils, mast cells, macrophages, and other cell types
capable of producing inflammatory cytokines such as IL13. We found that, in endobronchial
biopsies from patients with high bronchial epithelial expression of CLCA1, serpinB2, and
periostin, there was significantly elevated expression of IL13 and IL5 compared to patients
with low bronchial epithelial expression of CLCA1, serpinB2, and periostin [13] . Since IL5
and IL13 were defined at the time as cytokines typically expressed in T-helper type 2 (Th2)
polarized CD4+ T cells, we called asthma patients in the cluster with high expression of the
three genes 'Th2-high' and patients in the cluster with low expression of the three genes
'Th2-low'.
While the gene expression patterns clearly identified distinct subsets of asthma patients,
the cross-sectional nature of the analysis could not rule out the possibility that the gene
expression observed was transient in response to environmental stimuli (e.g., aeroaller-
gen exposures) near the time of bronchoscopy. To determine whether the gene expres-
sion patterns bore any relationship to independent measures of asthma, we compared the
'Th2-high' and 'Th2-low' asthma patients on the basis of multiple other clinical and path-
ological assessments. Both subsets of asthma patients had significant airway obstruction as
measured by spirometry (forced expiratory volume in one second, FEV1) and reversibility
to β 2 -adrenergic agonists compared to healthy controls, and while the 'Th2-high' subset was
sensitive to lower provocative concentrations of methacholine (a measure of airway hyper-
responsiveness, AHR) than the 'Th2-low' subset of asthma patients, both asthma groups
were significantly more sensitive to methacholine than healthy controls. However, the two
asthmatic subsets differed significantly on other measures, with the 'Th2-high' asthmat-
ics exhibiting elevated eosinophils in bronchoalveolar lavage (BAL) and peripheral blood,
altered mucus composition, and thicker airway reticular basement membranes (a measure
of bronchial fibrosis) compared to 'Th2-low' asthmatics. Thus, while both groups of asthmat-
ics met a clinical definition of asthma (reversible airway obstruction and AHR), they were
distinct with respect to other asthma pathologies. Upon repeat bronchoscopy one week after
the initial bronchoscopy, 'Th2-high' and 'Th2-low' asthmatics continued to have similar gene
expression patterns, suggesting short-term temporal stability of the phenotype. Importantly,
when randomized to an eight week course of ICS or placebo, only the 'Th2-high' asthmat-
ics demonstrated significant improvements in lung function on ICS, while the 'Th2-low'
asthmatics did not exhibit any FEV1 changes relative to placebo-treated patients; this FEV1