Background/Purpose: Small RNAs (sRNAs), such as microRNAs, are important regulators of biological processes and serve as important biomarkers of disease. We found that approximately half of sRNAs in human plasma fail map to the human genome, but many map to microbial genomes. sRNAs, human and non-human alike, may alter cellular function. Thus, non-human plasma sRNAs may serve as an effector compartment of the human microbiome, an area of great interest in rheumatoid arthritis (RA). Our objective was to determine if sRNAs of microbial origin are altered in RA plasma.

Methods:  RNA was extracted from plasma from 165 patients with RA and 91 controls frequency-matched for age, race and sex. Purified plasma RNA was used to prepare sRNA cDNA libraries. Sequencing was performed using Illumina NextSeq500. Using TIGER, an in-house sRNA-seq analysis pipeline, high quality reads were first aligned to the human genome allowing for 1 mismatch. Remaining reads were aligned to 57 genomes of bacteria previously found to be altered in RA. Separately, remaining reads were aligned to representative genomes of 207 human microbiome bacteria, 8 fungi, and 167 environmental bacteria. Differential expression analysis of sRNA and microbial genome counts were performed using DESeq2, adjusting for age, race, sex and batch, with 5% false discovery rate. Benjamini and Hochberg method was used for multiple test correction.

Results: Three microbial sRNAs, of variable length of the same sRNA were increased approximately 2-fold in the plasma of RA patients compared to control subjects (P=0.001 to P=0.02). These sRNAs mapped to multiple bacterial genomes of the human microbiome and environment. The three microbial sRNAs have an identical seed region, which in the case of miRNAs, is the region that binds with complementarity to target mRNAs to alter translation. This microbial sRNA seed region is predicted to target several human genes which are altered in RA including CARD8, IL17RA, and TRAF1. Comparing genome counts of RA-specific bacteria, Lactobacillus salivarius was significantly decreased in RA vs control (P=0.03). Comparing genome counts of representative genomes of human microbiome, fungi and environmental bacteria, Anaerobaculum hydrogeniformans (P=0.01), Staphylococcus epidermidis (P=0.04), Sphingobacterium spiritivorum (P=0.04), and Staphylococcus aureus (P=0.04) were significantly decreased in RA compared to control subjects.

Conclusion: Microbial plasma small RNAs are altered in patients with RA. Three microbial plasma sRNAs were increased significantly in RA and are predicted to target genes associated with RA. Further studies will determine the role of these microbial sRNAs in RA.