Bioinformatic meta-analysis reveals novel differentially expressed genes and pathways in sarcoidosis

Abstract

Introduction: Sarcoidosis is a multi-system inflammatory disease of unknown origin with heterogeneous clinical manifestations varying from a single organ non-caseating granuloma site to chronic systemic inflammation and fibrosis. Gene expression studies have suggested several genes and pathways implicated in the pathogenesis of sarcoidosis, however, due to differences in study design and variable statistical approaches, results were frequently not reproducible or concordant. Therefore, meta-analysis of sarcoidosis gene-expression datasets is of great importance to robustly establish differentially expressed genes and signalling pathways. Methods: We performed meta-analysis on 22 published gene-expression studies on sarcoidosis. Datasets were analysed systematically using same statistical cut-offs. Differentially expressed genes were identified by pooling of p-values using Edgington’s method and analysed for pathways using Ingenuity Pathway Analysis software. Results: A consistent and significant signature of novel and well-known genes was identified, those collectively implicated both type I and type II interferon mediated signalling pathways in sarcoidosis. In silico functional analysis showed consistent downregulation of eukaryotic initiation factor 2 signalling, whereas cytokines like interferons and transcription factor STAT1 were upregulated. Furthermore, we analysed affected tissues to detect differentially expressed genes likely to be involved in granuloma biology. This revealed that matrix metallopeptidase 12 was exclusively upregulated in affected tissues, suggesting a crucial role in disease pathogenesis. Discussion: Our analysis provides a concise gene signature in sarcoidosis and expands our knowledge about the pathogenesis. Our results are of importance to improve current diagnostic approaches and monitoring strategies as well as in the development of targeted therapeutics.