Molecular Heterogeneity, Biomarker Discovery, and Targeted Therapy in Asthma - Genomic Biomarkers for Pharmaceutical Development - page 74

“Th2-low” “Th2-high”

type 2 inflammation

Lower epithelial expression
of periostin, serpinB2, CLCA1
Lower mucosal expression
of IL5 and IL13
Fewer eosinophils

Higher epithelial expression
of periostin, serpinB2, CLCA1
Higher mucosal expression
of IL5 and IL13
More eosinophils

Less subepithelial fibrosis
Lower MUC5AC:MUC5B ratio

More subepithelial fibrosis
Higher MUC5AC:MUC5B ratio

Lower serum periostin and FeNO Higher serum periostin and FeNO

FIGURE 4.1 Asthma patients exhibit a continuum of type 2 airway inflammation. While all asthma patients
have reversible airway obstruction and airway hyper-responsiveness, there is a variable level of type 2 airway
inflammation as defined by gene expression in the bronchial mucosa. The degree of gene expression related to type
2 inflammation is correlated to pathophysiological features of asthma such as eosinophilia, fibrosis, mucus compo-
sition, and levels of non-invasive biomarkers such as serum periostin and FeNO.

improvement was reversed within one week of cessation of ICS treatment [13] . Taken
together, these corroborative data suggest that the molecularly defined phenotypes are rela-
tively stable and clinically meaningful in mild-moderate asthma.
Using the bronchial epithelial gene expression designation as a discriminating factor, we
assessed gene expression in matched endobronchial biopsy tissue from a subset of asthma
patients and controls in the study (N = 27 and 13, respectively). Despite the smaller number
of subjects and greater variability of sample composition in biopsies as opposed to epithelial
brushings, by comparing 'Th2-high' asthmatics to 'Th2-low' asthmatics and healthy controls
(rather than 'asthma' vs. 'control'), we identified 93 differentially expressed probes corre-
sponding to 79 genes using similarly conservative analytical methods [44] . This broader
set of highly co-regulated genes in bronchial mucosa enabled a more granular quantitative
description of the molecular phenotype, which could be condensed to a single gene expres-
sion score. This 'Th2 signature' score varied continuously across the dataset and correlated
continuously with independent measures of type 2 inflammation including blood and air-
way eosinophil levels. In addition, the 'Th2 signature' was highly positively correlated with
the expression levels of the Th2 cytokines IL13 and IL5, as well as with eosinophil chem-
oattractants CCL26 and CCL13, while it was negatively correlated with the expression level
of the Th1 cytokine IL12A. This molecular definition of asthma heterogeneity along a con-
tinuum of type 2 inflammation appeared to correspond well with the cytologic definition of
asthma heterogeneity according to airway eosinophilia. The greater precision afforded by
the biopsy gene expression signature underscored the reality that while it is often a conveni-
ent shorthand to define asthma phenotypes as discrete entities, there is clearly a continuum
of type 2 inflammation across the population. The findings described in [13] and [44] are
summarized conceptually in Fig. 4.1 .