Molecular Heterogeneity, Biomarker Discovery, and Targeted Therapy in Asthma - Genomic Biomarkers for Pharmaceutical Development - page 76

those proteins could be detected in peripheral blood. For example, CLCA1 is expressed on
the apical aspect of bronchial epithelial cells near the airway lumen [49] , consistent with
its proposed role in mucus secretion [35] . Its expression is therefore likely to be densest at
the luminal surface of the bronchial mucosa and CLCA1 protein, if shed in soluble form, is
most likely detectable in samples taken from the airway such as induced sputum or BAL
[34] . Periostin, on the other hand, is secreted from the basolateral aspect of bronchial epi-
thelial cells in response to IL13 stimulation [40] and may also be expressed in subepithelial
fibroblasts in the lamina propria [39] , more proximal to blood vessels serving the bronchial
mucosa. For this reason, we hypothesized that periostin may be more likely to be detect-
able in peripheral blood than CLCA1. In addition, some reports have described relatively
high levels of circulating periostin protein (10s-100sng/ml) in patients with epithelially-
derived cancers [50-53] , suggesting that substantial levels of the protein might be detect-
able in peripheral blood. Thus we prioritized periostin for further investigation, including
assay development. Many other soluble blood protein candidates were evaluated along sim-
ilar lines but will not be discussed further here. Nitric oxide synthase 2 (NOS2), encoding
inducible nitric oxide synthase (iNOS), was among the genes most significantly associated
with the Th2 signature in both bronchial epithelial brushings and endobronchial biopsies
[44] . iNOS is an enzyme that catalyzes the production of nitric oxide (NO) from L-arginine,
and its expression can be induced in bronchial epithelial cells by IL13. NO is an exhaled gas
that is detectable in exhaled air, and FeNO has been used extensively in the diagnosis and
management of asthma and clinically validated instruments for its detection are in wide use
[54,55] . Accordingly, we advanced FeNO as another candidate marker of airway IL13 activ-
ity. The Th2 signature is highly correlated with the expression of IL5, a Th2 cytokine that is
an obligate factor for eosinophil hematopoiesis, activation, and survival [56] . IL13 induces
the expression of multiple CCR3-binding chemokines such as CCL11, CCL13, CCL24, and
CCL26, which serve as eosinophil chemoattractants [57-59] . Local expression of IL5 and IL13
in asthmatic airway tissue can contribute to airway eosinophilia via expansion and recruit-
ment of eosinophils and might contribute to increased trafficking of eosinophils through the
systemic circulation. Therefore, we advanced peripheral blood eosinophil counts as a third
candidate marker of airway IL13 activity. Each of these three markers (serum periostin,
FeNO, and blood eosinophils) is mechanistically linked to the activity of Th2 cytokines in the
airways and could be measured noninvasively. The relationships between IL5, IL13, bron-
chial mucosa, and the markers discussed here are depicted schematically in Fig. 4.2 .
FeNO could be detected using existing instruments [55] , and peripheral blood eosino-
phil counts could be derived from a complete blood count (CBC) with differential, which is a
widely available clinical laboratory assay. However, there were no validated assays for periph-
eral blood periostin. 'Research-grade' immunoassays often use polyclonal antibodies for either
capture or detection in a sandwich enzyme-linked immunosorbent assay (ELISA) format.
However, the use of polyclonal antibodies introduced at least two challenges for the devel-
opment of a serum periostin assay, one of which is specific to periostin and one of which is
a general challenge for immunoassay-based clinical diagnostic tests. Periostin protein exists in
multiple alternatively spliced variants [60] , and several of these variants are expressed simulta-
neously in bronchial epithelial brushings from asthma patients (G. Jia and J. R. Arron, unpub-
lished observations). While exons 1-16 are conserved in most expressed variants, exons 17-23
undergo alternative splicing. Thus certain epitopes from the C-terminal portion of periostin