Gene expression profiling is a valuable tool to identify the altered cellular status of cells or tissues. Metaanalysis of available gene expression profiles from numerous sources provides a means to assess the molecular alterations characteristic of autoimmune diseases, such as systemic lupus erythematosus (SLE) and to develop insights into possible new therapies. Toward this end, all of the publicly available gene expression datasets from SLE blood and tissue were assembled and a variety of analytic algorithms were employed to identify differentially expressed genes, gene modules associated with clinical features, molecular pathways associated with SLE and potential upstream regulators. This information was employed to cross-reference a variety of databases to predict novel molecular pathways and potential drug candidates. Specifially, gene expression profiles obtained from lupus affected skin, synovium and kidney were obtained, compared to metaanalyzed data obtained from active lupus B cells, T cells and myeloid cells, and cross-referenced to various pathway analytic tools including Molecular Signature (MS©)-Scoring, Ingenuity Pathway Analysis© Upstream Regulator (IPA©-UR) and Library of Integrated Network Based Cellular Signatures (LINCS). More than 300 arrays from lupus patients and appropriate controls were analyzed to determine differentially expressed (DE) genes [8279 discoid skin, 5465 lupus synovium, 6381 lupus nephritis glomerulus, 5587 lupus nephritis tubulointerstitum]. Notably, the majority of lupus affected tissue DE genes were detected in more than one tissue and 439 were differentially expressed in all tissues. Tissue inflammatory/immune cell infiltration was documented by genes encoding specific cell markers as well as by unique gene expression signatures (Biologically Informed Gene Clustering, BIG-C©). Examination of curated predicted functional groups from the STRING (Search Tool for Retreaval of Interacting Genes/Proteins) output of common up-regulated transcripts in lupus tissue predicted therapeutic targets and drugs using STITCH (Search Tool for Interacting Chemicals) and IPA's BioProfiler© that were confirmed by connectivity using LINCS. Analysis of gene expression datasets for the presence of specific gene modules using GSVA (Gene Set Variation Analysis) documented the overall similarity of active lupus gene sets as well as the uniqueness of individual patient samples.These results demonstrate the value of comprehensive application of orthogonal curated bioinformatics tools in identifying the role of inflammatory/immune cells in lupus pathogenesis and tissue damage. This approach demonstrated that there are molecular pathways common to all lupus patients and tissues, and there are pathways involved in inflammatory response of some but not all patients and tissues. Further analysis should generate a molecular model of lupus immunopathogenesis and could identify therapies that may be useful in all lupus patients versus those with involvement of specific tissues and treatments specifically targeted to individual lupus patients.
Grammer AC et al. Lupus (2016) 25:1150–1170.
Disclosure of Interest None declared