Identification of prognostic biomarkers for papillary thyroid carcinoma by a weighted gene co‐expression network analysis

Abstract

Abstract Aim Whole transcriptome analysis was conducted to identify differentially expressed RNAs and regulatory networks associated with papillary thyroid carcinoma (PTC). Methods A weighted gene co‐expression network analysis based on high‐throughput sequencing data for six pairs of PTC and adjacent tissue samples was conducted to understand the biological functions and regulatory networks involving long non‐coding RNAs (lncRNAs), circular RNAs (circRNAs), microRNAs (miRNAs), and messenger RNAs (mRNAs). Results We detected 131, 338, 31, and 556 differentially expressed circRNAs, lncRNAs, miRNAs, and mRNAs, respectively. We identified modules that were significantly positively and negatively related to cancer and lymph node metastasis. Gray and turquoise modules were positively correlated with cancer phenotypes (p < 0.05), whereas yellow, brown, and blue modules were negatively correlated with cancer (p < 0.05). Gray module was positively correlated with lateral lymph node metastasis (p = 0.02). Kaplan–Meier analyses revealed that the levels of transmembrane protein 63C (TMEM63C), lysyl oxidase‐like 1 (LOXL1), collagen type V alpha 1 chain (COL5A1), ADAM metalloproteinase with thrombospondin type I motif 2 (ADAMTS2), and LysM‐domain containing 3 (LYSMD3) were significantly associated with overall survival (p < 0.05). Significant increase in the expression of COL5A1 and LOXL1 in tumor tissues was validated by quantitative real‐time polymerase chain reaction (p < 0.05). COL5A1 and LOXL1 promoted PTC cell growth and invasion in vitro. Conclusions We identified COL5A1 and LOXL1 as potential prognostic biomarkers, providing new insights into the occurrence and progression of PTC.

Keywords

• biomarker
• overall survival
• papillary thyroid carcinoma
• weighted gene co‐expression network analysis