A landscape of circulating long non‐coding RNA (lncRNA) expression profile and the predictive value of candidate lncRNAs for disease risk of knee osteoarthritis

BACKGROUND: This study aimed to investigate the plasma long non‐coding RNA (lncRNA) expression profile in knee osteoarthritis (KOA) patients and the value of candidate lncRNAs for predicting KOA risk. METHODS: Plasma was obtained for RNA sequencing (RNA‐seq) in eight KOA patients and eight healthy controls (Ctrls). Ten candidate lncRNAs were then selected from the differentially expressed (DE) lncRNAs according to the rank of absolute value of Log(2) (fold change). Afterward, RT‐qPCR was used to examine 10 candidate lncRNAs expressions in plasma of 100 KOA patients and 100 Ctrls. RESULTS: In eight KOA patients and eight Ctrls, principal component analysis and heatmap plots disclosed that lncRNA and mRNA expression profile could distinguish KOA patients from Ctrls. Then Volcano plot identified 418 upregulated lncRNAs, 347 downregulated lncRNAs, 521 upregulated mRNAs, and 333 downregulated mRNAs in KOA patients compared to Ctrls. Next, enrichment analyses revealed that DE lncRNAs were mainly enriched in biological processes, molecular functions, and signaling pathways related to inflammation and bone formation. In 100 KOA patients and 100 Ctrls, eight candidate lncRNAs were dysregulated in KOA patients compared to Ctrls, including lncRNA ABCF2P2, lncRNA RP13‐16H11.7, lncRNA CTC‐340A15.2, lncRNA RP4‐735C1.6, lncRNA RP11‐293G6‐B.8, lncRNA RP11‐1246C19.1, lncRNA RP11‐303E16.6, and lncRNA RP5‐882C2.2. Receiver operating characteristic curve analysis revealed that these eight candidate lncRNAs presented with values for predicting KOA risk. Furthermore, multivariate logistic regression elucidated that six candidate lncRNAs could independently predict KOA risk. CONCLUSION: We disclosed a landscape of circulating lncRNA expression profile in KOA patients, and discovered several specific lncRNAs which could assist in KOA management.