GPNMB contributes to a vicious circle for chronic obstructive pulmonary disease

Abstract

Osteoporosis (OP) is significant and debilitating comorbidity of chronic obstructive pulmonary disease (COPD). We hypothesize that genetic variance identified with OP may also play roles in COPD. We have conducted a large-scale relation data analysis to explore the genes implicated with either OP or COPD, or both. Each gene linked to OP but not to COPD was further explored in a mega-analysis and partial mega-analysis of 15 independently collected COPD RNA expression datasets, followed by gene set enrichment analysis (GSEA) and literature-based pathway analysis to explore their functional linked to COPD. A multiple linear regression (MLR) model was built to study the possible influence of sample size, population region, and study date on the gene expression data in COPD. At the first step of the analysis, we have identified 918 genes associated with COPD, 581 with OP, and a significant overlap (P<2.30e-140; 210 overlapped genes). Partial mega-analysis showed that, one OP gene, GPNMB presented significantly increased expression in COPD patients (P-value = 0.0018; log fold change = 0.83). GPNMB was enriched in multiple COPD pathways and plays roles as a gene hub formulating multiple vicious COPD pathways included gene MMP9 and MYC. GPNMB could be a novel gene that plays roles in both COPD and OP. Partial mega-analysis is valuable in identify case-specific genes for COPD.

Keywords: Osteoporosis; chronic obstructive pulmonary disease; mega-analysis; partial mega-analysis; pathway analysis.

Figure 1. Diagram of the workflow

Figure 2. The effect size, 95% confidence interval and weights for the gene GPNMB

(A) Mega-analysis results; (B) Partial mega-analysis results. Both results were from fixed-effects model. The bar plot on the right of each figure represents the normalized weights for each dataset/study, ranged within (0, 1); the brighter (green) the color, the bigger the weight (labeled right next to the bar). The star (in red) and lines (in blue) on the left are the mean of effect size (log fold change) and 95% confidence interval (CI) of each dataset/study, respectively.

Figure 3. Functional analysis of the molecular pathways connecting COPD and the GPNMB

(A) Pathways linking GPNMB and COPD as reconstructed from analysis of previous literature. (B) Partial mega-analysis of the expression levels of the genes included in the reconstructed pathways. (C) Control mega-analysis of the expression levels of the genes included in the reconstructed pathways performed on datasets excluded from partial mega-analysis. The pathways in (A) was generated in Pathway Studio environment (www.pathwaystudio.com). Each relation (edge) in the figure has one or more supporting references. The colorbar is for the panels (B) and (C) only. Red represents increased expression level, and blue decreased.