Background/Purpose: The search for novel targets in RA requires novel computational methods and in silicosystems to identify non-obvious pathways that account for the diversity of responses to targeted agents. We developed and applied a novel integrative systems biology method that identifies transcription factors (TFs) central to regulatory patterns in RA fibroblast-like synoviocytes (FLS) and significant variations in gene networks between RA patients.

Methods: Our whole genome ATAC-seq and RNA-seq data (from 11 RA and 11 osteoarthritis [OA] FLS lines) were evaluated for overlaps between ATAC-seq peaks and known gene promoter regions (4kb upstream and 1kb downstream from the TSS). ATAC-seq peaks not assigned to promoters were considered as enhancer regions and linked with the nearest gene. TF binding motifs were curated from the CIS-BP database. 745 TFs had binding sites within 150-bp regions in ATAC-seq peak summits. 22 unique network topologies were constructed by forming directed edges between any parent node TF and child node gene or child node TF. Via the integration of RNA-seq expression data for each sample, the Personalized PageRank (PPR) algorithm was developed and run measure the global influence of each node and variance between RA patients.

Results: Initial analysis focused on differences between RA and OA. The mean PPR was calculated for TFs from RA and OA samples and were ranked based on the largest absolute difference between RA and OA (DPPR). Of these, 33 TFs were significantly different for RA vs. OA (p < 0.05). For example, the glucocorticoid receptor NR3C1 was the highest rank DPPR (p value = 0.03). BACH1, which regulates osteoclastogenesis, was the second highest rank DPPR (p = 0.0005). Other high DPPR genes included STAT1, YY1 (JAK-STAT signaling) and SP1. We then looked within RA for inter-RA patient differences to understand individual network profiles. The intersection of the 33 significant RA vs OA TFs with TFs that have the highest variance of normalized PPR within the RA networks, yielded 15 hits. MGA (MAX gene-associated protein), which regulates the expression of MYC-MAX and T-box family target genes, had the highest intra-RA variance with a high rank DPPR (p = 0.024) suggesting that this TF defines differences in RA pathogenesis between patients. Another high intra-RA variance gene within the intersection includes TBX2, which participates in mesenchymal cell differentiation. A smaller subset of RA patients (~30%) have high PPR values for ID1. ID family genes play a role in cell proliferation and angiogenesis in RA.

Conclusion: This systems biology approach not only defines disease specific TFs that contribute to the RA phenotype but also distinguishes patient-to-patient differences. The unique computational approach identifies novel targets and helps elucidate the mechanism of differential responses to highly targeted agents in RA. Key transcription factors, including well known genes such as NR3C1 and STAT1, along with novel patient-specific TFs targets like MGA emerge from this in silico method to individualize treatment.

« Back to 2017 ACR/ARHP Annual Meeting

ACR Meeting Abstracts - https://acrabstracts.org/abstract/integrative-systems-biology-approach-identifies-key-transcription-factors-and-novel-rheumatoid-arthritis-ra-and-individualized-therapeutic-targets/