Background/Purpose: Transcriptome profiling has expanded our ability to identify biomarkers and therapeutic targets and better understand disease progression in a wide variety of conditions. Serum C-reactive protein (CRP) is a marker of inflammation and is widely used to assess disease activity in RA. The aim of this study was to identify differentially expressed genes in CRP high and low patients with RA.

Methods: Whole blood samples were collected from participants in the Arthritis Internet Registry and FORWARD, The National Databank for Rheumatic Diseases. The biosamples are associated with demographic and clinical data from comprehensive biannual questionnaires. RNA was extracted and sequenced for 60 participants with RA (30 with high inflammatory levels and 30 with low inflammatory levels, assessed using blood concentrations of CRP). Version 1.2.3 of the bcbio pipeline was used to perform alignments (hisat2), read data qc (qualimap, fastqc), and transcript quantification (salmon). For differential expression analysis, DESeq2 was used in conjunction with the bcbioRNAseq R package. Student’s t tests and Χ² tests were performed to assess significant demographic and clinical differences between the high and low disease activity cohorts.

Results: After restricting the dataset to samples from participants with an associated comprehensive questionnaire fewer than 90 days from the date of sample collection, the high disease activity (CRP >0.8 mg/dL; CRP-high) cohort included 27 participants and the low disease activity (CRP< 0.1 mg/dL; CRP-low) cohort included 25 participants. There were no statistically significant differences in demographics or medications between the two groups, with the exception of BMI, which was higher in the CRP-high group. As expected, the CRP-high group had significantly higher scores on measures of disease activity. All participants in both groups were RF and/or anti-CCP positive (Table 1). Differential Gene Expression analysis with an alpha of 0.01 revealed 32 genes dysregulated between the CRP-high and CRP-low groups. Clustering with Ward’s method using these genes as input generated two distinct clusters: one included 30 samples (5 CRP-high, 25 CRP-low) and the other exclusively contained CRP-high samples (22).

Conclusion: Notable gene signatures resulted in two distinct clusters each comprised almost exclusively of the high and low disease activity groups. Significantly differentially expressed genes identified in this study include several that have previously been implicated in RA and other autoimmune diseases (KLRC1, RAP1GAP, IFI27, LY6E, ISG15, PIM2) as well as several genes not clearly associated with RA, including a group of genes related to cell signaling by receptor tyrosine kinases (EFNA1, SHC2, RHBDF1). RHBDF1, which encodes a catalytically inactive rhomboid protein, may be of particular interest due to its regulatory relationship with ADAM17, a key regulator of TNF that has been implicated in the development and progression of several autoimmune diseases, including RA.

Table 1. Demographic and clinical characteristics of high and low disease activity groups. Bold p values indicate statistical significance.

Figure 1. Heat map of dysregulated genes labeled by disease activity (CRP high vs low) and grouped using hierarchical clustering.

« Back to ACR Convergence 2020

ACR Meeting Abstracts - https://acrabstracts.org/abstract/gene-expression-signatures-in-c-reactive-protein-high-and-low-rheumatoid-arthritis/