Circulating microRNA profiles in early-stage osteoarthritis and rheumatoid arthritis

Abstract

Osteoarthritis (OA) and rheumatoid arthritis (RA) are prevalent joint diseases, yet early diagnosis remains challenging with existing methods. Circulating microRNAs are promising biomarkers for detection and differentiation of arthritis subtypes. This study aimed to profile plasma microRNAs from early OA (N = 22), early RA (N = 12), and non-OA/RA (N = 50) individuals using microRNA-sequencing. Principal component analysis revealed distinct clustering of early OA from both early RA and non-OA/RA, but not for early RA and non-OA/RA. A total of 170 differentially expressed microRNAs were identified in early OA versus the other groups, with no significant differences found between early RA and non-OA/RA. Stepwise filtering followed by RT-qPCR validation in independent samples identified six microRNAs: miR-16-5p and miR-29c-3p were upregulated in early OA compared to both early RA and non-OA/RA, while miR-744-5p, miR-382-5p, miR-3074-5p, and miR-11400 were upregulated in early RA compared to the other two groups. Additionally, three novel microRNAs were identified using bioinformatic tools-one enriched in early OA and two in early RA. Target prediction and pathway analyses revealed that early OA microRNAs were linked to extracellular matrix degradation pathways, and early RA microRNAs were linked to immune signaling. These findings highlight six known and three novel circulating microRNAs with potential as biomarkers to distinguish early OA from early RA.

Keywords: Biomarker; Epigenetics; Human; Novel microRNAs; Plasma; microRNA-sequencing.

Fig. 1

Overview of study design. (A) The study included three groups defined by predetermined inclusion and exclusion criteria. (B) MicroRNA-sequencing was performed on N = 62 plasma samples. (C) Data were processed using an established pipeline. (D) Exploratory analyses included principal component analysis (PCA). (E) Stepwise filtering was performed to shortlist microRNAs. (F) RT-qPCR validation experiments were performed for shortlisted microRNAs in N = 22 independent plasma samples. (G) Dysregulated known and novel microRNAs may be used to distinguish individuals with early OA or early RA.

Fig. 2

Distinct circulating microRNAs are present in early OA and early RA. (A–C) Principal component analyses showing clustering of samples when stratified by cohort. Each point represents a sample. Each axis represents the percentage variance for the principal component (PC). (D–F) Volcano plots showing differentially expressed microRNAs in each comparison, with FDR ≤ 0.05 (horizontal line). Each dot represents a microRNA.

Fig. 3

Stepwise filtering identified microRNAs dysregulated in early OA and early RA. (A) Four sequential filtering criteria were applied to identify dysregulated microRNAs. The brown box shows filtering for microRNAs in early OA versus early RA and non-OA/RA. The purple box shows filtering for microRNAs in early RA versus early OA and non-OA/RA. (B) Heatmap depicting plasma levels of 12 microRNAs dysregulated in early OA and early RA. Z-scores represent standardized microRNA counts, with green indicating higher levels and red indicating lower levels.

Fig. 4

RT-qPCR validation of six microRNAs in early OA and early RA. RT-qPCR was performed in independent N = 10 early OA, N = 6 early RA, and N = 6 non-OA/RA control plasma samples. Relative microRNA levels represent fold-change in early OA (brown) and early RA (purple) normalized to average levels in non-OA/RA (grey). Error bars = 95% confidence interval. (A,B) Two microRNAs were upregulated in early OA versus early RA. *p < 0.05 shows the significance for comparison between early OA versus early RA and non-OA/RA. (C–F) Four microRNAs were upregulated in early RA versus early OA. *p < 0.05 shows the significance for comparison between early RA versus early OA and non-OA/RA.

Fig. 5

Pathway network analysis of (A) early OA, and (B) early RA-associated microRNAs. Target genes of prioritized microRNAs in early OA and early RA were analyzed using pathDIP v.5 software. Networks were constructed using Cytoscape software and include only the genes targeted by our known and novel microRNAs within each pathway. Nodes represent genes and are color-coded by microRNA. Genes targeted by multiple microRNAs are shown with multi-color shading. Each edge (line) connects a target gene to its corresponding microRNAs.