Analysis of gene expression profiles and protein-protein interaction networks in multiple tissues of systemic sclerosis

Abstract

Background: Systemic sclerosis (SSc), a multi-organ disorder, is characterized by vascular abnormalities, dysregulation of the immune system, and fibrosis. The mechanisms underlying tissue pathology in SSc have not been entirely understood. This study intended to investigate the common and tissue-specific pathways involved in different tissues of SSc patients.

Methods: An integrative gene expression analysis of ten independent microarray datasets of three tissues was conducted to identify differentially expressed genes (DEGs). DEGs were mapped to the search tool for retrieval of interacting genes (STRING) to acquire protein-protein interaction (PPI) networks. Then, functional clusters in PPI networks were determined. Enrichr, a gene list enrichment analysis tool, was utilized for the functional enrichment of clusters.

Results: A total of 12, 2, and 4 functional clusters from 619, 52, and 119 DEGs were determined in the lung, peripheral blood mononuclear cell (PBMC), and skin tissues, respectively. Analysis revealed that the tumor necrosis factor (TNF) signaling pathway was enriched significantly in the three investigated tissues as a common pathway. In addition, clusters associated with inflammation and immunity were common in the three investigated tissues. However, clusters related to the fibrosis process were common in lung and skin tissues.

Conclusions: Analysis indicated that there were common pathological clusters that contributed to the pathogenesis of SSc in different tissues. Moreover, it seems that the common pathways in distinct tissues stem from a diverse set of genes.

Keywords: Common pathway; Functional analysis; Integrative gene expression analysis; Systemic sclerosis.

Fig. 1

Flowchart of methods

Fig. 2

Boxplots of data before and after batch effect correction. a) Boxplot of PBMC data before batch effect removal. b) Boxplot of PBMC data after batch effect removal. c) Boxplot of lung data before batch effect removal. d) Boxplot of lung data after batch effect removal. e) Boxplot of skin data before batch effect removal. f) Boxplot of skin data after batch effect removal. Different datasets were displayed in different colors

Fig. 3

Lung PPI network. Twelve distinct functional clusters were detected in lung tissue. Each cluster is a set of highly-connected nodes and is illustrated in a discrete color

Fig. 4

a) PBMC PPI network. Two distinct functional clusters were identified in PBMC. b) Skin PPI network. Four distinct functional clusters were detected in skin tissue. Each cluster is a set of highly-connected nodes and is illustrated in a discrete color