Identification of surrogate prognostic biomarkers for allergic asthma in nasal epithelial brushing samples by WGCNA

• Published in: Journal of cellular biochemistry Vol. 120; no. 4; pp. 5137 - 5150
• Main Authors: Liu, Zhaoyu, Li, Ming, Fang, Xiangming, Shen, Lu, Yao, Wenxia, Fang, Zhiyuan, Chen, Jitao, Feng, Xiao, Hu, La, Zeng, Zicheng, Lin, Chunyi, Weng, Jinsheng, Lai, Yuxiong, Yi, Gao
• Format: Journal Article
• Language: English
• Published: United States Wiley Subscription Services, Inc 01.04.2019
• Subjects: allergic asthma; Asthma; Biomarkers; Brushing; CD44 antigen; Cell activation; Correlation analysis; Degranulation; Enrichment; Epithelial cells; Gene expression; Genes; Immune system; Lymphocytes T; Modules; Molecular modelling; Mucous membrane; Network analysis; Respiratory tract; Respiratory tract diseases; Substructures; WGCNA; allergic asthma; WGCNA; gene expression
• ISSN: 0730-2312; 1097-4644; 1097-4644
• DOI: 10.1002/jcb.27790

Background Allergic asthma is a lower respiratory tract disease of Th2 inflammation with multiple molecular mechanisms. The upper and lower airways can be unified by the concept of a united airway and, as such, gene expression studies of upper epithelial cells may provide effective surrogate biomarkers for the prognostic study of allergic asthma. Objective To identify surrogate biomarkers in upper airway epithelial cells for the prognostic study of allergic asthma. Methods Nasal epithelial cell gene expression in 40 asthmatic and 17 healthy control subjects were analyzed by weighted gene coexpression network analysis (WGCNA) to identify gene network modules and profiles in allergic asthma. Functional enrichment analysis was performed on the coexpression genes in certain highlighted modules. Results A total of 13 coexpression modules were constructed by WGCNA from 2804 genes in nasal epithelial brushing samples of the 40 asthmatic and 17 healthy subjects. The number of genes in these modules ranged from 1086 (Turquoise module) to 45 (Salmon). Eight coexpression modules were found to be significantly correlated (P < 0.05) with two clinic traits, namely disease status, and severity. Four modules were positively correlated ( P < 0.05) with the traits and these, therefore, contained genes that are mostly overexpressed in asthma. Contrastingly, the four other modules were found to be negatively correlated with the clinic traits. Functional enrichment analysis of the positively correlated modules showed that one (Magenta) was mainly enriched in mast cell activation and degranulation; another (Pink) was largely involved in immune cell response; the third (Yellow) was predominantly enriched in transmembrane signal pathways; and the last (Blue) was mainly enriched in substructure components of the cells. The hub genes in the modules were KIT, KITLG, GATA2, CD44, PTPRC, and CFTR, and these were confirmed as having significantly higher expression in the nasal epithelial cells. Combining the six hub genes enabled a relatively high capacity for discrimination between asthmatics and healthy subjects with an area under the receiver operating characteristic (ROC) curve of 0.924. Conclusions Our findings provide a framework of coexpression gene modules from nasal epithelial brushing samples that could be used for the prognostic study of allergic asthma. Gene expression data of nasal epithelial brushing from asthmatics were utilized to identify gene coexpression networks involved in asthma pathogenesis. Thirteen coexpression networks (modules) were constructed using the weighted gene coexpression network analysis (WGCNA) method, four modules were found to be significantly associated with asthma status and severity. Gene ontology (GO) enrichment analysis showed these significant modules were mainly enriched in mast cell activation, mast cell degranulation, etc, the most central (hub) genes were KIT, KITLG, GATA2, CD44, PTPRC, and CFTR.