The main characteristic of RA is chronic synovial inflammation, which leads to erosion and damage of joints. Early diagnosis and treatment of RA will effectively prevent joint damage and improve quality of life. However, early RA is difficult to diagnose due to the lack of effective biomarkers. It is crucial to identify new and effective biomarkers for the early diagnosis and treatment of RA.
In our study, we identified 275 DEGs, including 71 tissue/organ-specific expressed genes, by comparing genes expressed in RA and OA samples. GO enrichment analysis of all genes and DEGs indicated that the immune responses, such as the immune cell-mediated immune response and the regulation of humoral immune response, were stronger in RA samples than in OA samples. KEGG pathways that were enriched included cytokine-cytokine receptor interaction, primary immunodeficiency, JAK-STAT signalling pathway, Fc gamma R-mediated phagocytosis, and neuroactive ligand-receptor interaction. Reactome enrichment analysis also showed that DEGs were mostly enriched in the immune system and signal transduction. We observed that the results of these enrichment analyses were consistent with actual differences between RA and OA.
After the hub genes that were screened by the PPI network were validated using the GEO datasets, we identified three haematologic/immune system-specific expressed genes, namely, GZMA, PRC1, and TTK, as biomarkers for the diagnosis of early RA. In addition, we constructed an mRNA-miRNA co-expression network and ceRNA networks to elucidate the pathogenesis of RA at the transcriptome level.
GZMA, a member of the serine protease family, is secreted by cytotoxic cells such as cytotoxic T cells and natural killer (NK) cells and plays an important role in cell death, cytokine processing, and inflammation[17, 18]. Several studies have reported that compared with the expression level of GZMA in OA patients, the expression level of GZMA increases in plasma, synovial tissues, and synovial membranes in patients with RA[19, 20]. This indicates that GZMA plays a significant role in the pathogenesis of RA. Consistent with this research, our study found that GZMA was upregulated in the synovial membrane of RA, especially in early RA. In addition, the ROC curve of GZMA indicated that it has a very high diagnostic value in early RA (AUC = 0.906). We considered GZMA a very effective biomarker for the diagnosis of early RA.
PRC1 (also called ASE1), a human mitotic spindle-associated CDK substrate protein, is a key regulator of cell division[21]. According to BioGPS, PRC1 is specifically expressed in early erythrocytes, endotheliocytes, and B lymphocytes. At present, PRC1 has not been reported in RA-related studies. However, in our study, PRC1 was upregulated in the synovial membrane of RA, especially in early RA. Compared with OA, synovial inflammation and hyperplasia are obvious in RA. In addition, it has been reported that the metabolic level of the synovial membrane is elevated, similar to that of tumour tissue[22]. These results all reflect the increased proliferation of cells in the synovial membrane of RA to some extent. Therefore, PRC1 may play an important role in the proliferation of synovial cells and the disease progression of RA.
TTK (also called MPS1 and CT96), which encodes a dual specificity protein kinase that phosphorylates a variety of amino acids such as tyrosine and serine, is related to cell proliferation[23]. Similar to PRC1, TTK is also highly specifically expressed in early red blood cells and endothelial cells. A study by H Ah-Kim et al. reported that tumour necrosis factor-alpha (TNF-α) can increase TTK expression in human articular chondrocytes[24], suggesting that TTK is regulated by TNF-α in some biological processes. We know that TNF-α plays a very significant role in the pathogenesis of RA[25].
Thus, we hypothesized that TTK plays an important role in synovial cell proliferation and TNF-α-mediated pathogenesis. In addition, we identified that TTK was highly expressed in the synovial membrane of RA and has a high diagnostic value in early RA (AUC = 0.793). We considered TTK as a novel and effective biomarker for the diagnosis of early RA.
Furthermore, target miRNAs and the target lncRNAs and circRNAs of these miRNAs were predicted for GZMA, PRC1, and TTK, and a ceRNA network was constructed with Cytoscape.
This network reveals the mechanism by which selected genes are regulated at the transcriptome level. According to the ceRNA hypothesis, we performed a literature search to select downregulated miRNAs in RA for further analysis. Among the target miRNAs of GZMA, PRC1, and TTK, the expression of the following miRNAs was downregulated in RA: miR-129-5p (in RA synovial tissue and synovial fibroblasts), miR-132-3p (in RA synovial fibroblasts), miR-212-3p (in RA synovial tissue and synovial fibroblasts), and miR-25-3p (in peripheral blood mononuclear cells)[26–29]. In addition, it has been reported that the lncRNA NEAT1 is upregulated in peripheral blood mononuclear cells of patients with RA[30]. Therefore, we propose that NEAT1-miR-212-3p/miR-132-3p/miR-129-5p-TTK might be a potential RNA regulatory pathway to regulate the disease progression of early RA. Additionally, although lncRNA XIST has not been reported in RA, it has been reported to be upregulated in another autoimmune disease, Sjogren's syndrome[31]. We hypothesize that XIST-miR-25-3p/miR-129-5p-GZMA has an important regulatory role in RA. Regarding the prediction results of circRNAs, we found a circRNA (TTK_hsa_circ_0077158) predicted by target miRNAs of TTK, and its target was TTK. Hence, we proposed a circRNA-miRNA-mRNA pathway: TTK_hsa_circ_0077158-miR-212-3p/miR-132-3p/miR-129-5p-TTK; it might be a key regulatory pathway in the pathogenesis of early RA. Of course, these RNA regulatory pathways need to be further experimentally verified, which is also a limitation of our study.