Session Type: ACR Poster Session C
Session Time: 9:00AM-11:00AM
Background/Purpose: Rheumatoid arthritis (RA) is a common and chronic autoimmune joint disease. RA is pathologically heterogeneous with multiple contributing factors. While there has been a major emphasis in studies of the immune regulatory factors the gene networks regulating the synovial pathology remains incompletely understood. We aimed at establishing a RA synovial tissue gene network to identify new key regulators/driver of the synovial and fibroblast-like synoviocyte (FLS) pathology.
Methods: Several large RA synovial tissue gene expression data sets (GSE48780 and GSE21537, consisting of 83 and 62 synovial biopsies from RA patients, respectively) were used. We used a software suite that we developed, RIMBANet, to construct a Bayesian network model for RA. We computationally estimated cell composition in synovial biopsies (T cells, neutrophils, FLS, macrophages, others).
Results: Immune cell compositions in inflamed and non-inflamed synovial tissues were significantly different, with increased numbers of M1 macrophages in inflamed tissues than in non-inflamed tissues. We adjusted synovial tissue profiling data for cell composition, and combined the two data sets together used in network reconstruction. We identified 7,521 nodes representing expression levels of genes and 8,118 edges representing putative causal relationships between genes. We compared the constructed RA network with well-known interaction databases or canonical pathways, and showed that the RA network overlapped with existing network/pathways significantly better than random networks. Then, we collected a set of RA severity related gene sets from a variety of sources (FLS gene expression, rodent studies, etc) and projected them onto the RA network identified key regulators for disease severity. Top inferred key regulators included DLX4, SEMA3E, LCP1, TRIM22, and ZNF385B. To test whether the inferred key regulators have any impact on RA disease severity, we examined the role of DLX4 in FLS invasiveness, an in vitro phenotype that strongly correlates with joint damage in RA. DLX4 was expressed in RA FLS and its knockdown with siRNA in primary RA FLS significantly reduced FLS invasiveness by 80%.
Conclusion: We here describe a new RA synovial gene network. This study also provides encouraging validation of the new synovial gene network suggesting that it has the potential to lead to the discoveries and of new important genes involved in disease pathogenesis and potentially new targets for treatment.
To cite this abstract in AMA style:Wang W, Lahiri A, Laragione T, Zhu J, Gulko PS. A Molecular Bayesian Network for Rheumatoid Arthritis Reveals Multiple Candidate Key Regulators for Disease Severity [abstract]. Arthritis Rheumatol. 2018; 70 (suppl 10). https://acrabstracts.org/abstract/a-molecular-bayesian-network-for-rheumatoid-arthritis-reveals-multiple-candidate-key-regulators-for-disease-severity/. Accessed October 27, 2020.
« Back to 2018 ACR/ARHP Annual Meeting
ACR Meeting Abstracts - https://acrabstracts.org/abstract/a-molecular-bayesian-network-for-rheumatoid-arthritis-reveals-multiple-candidate-key-regulators-for-disease-severity/