Abstract
MicroRNAs have been proposed as novel biomarkers for the diagnosis and treatment of many types of cancer. The levels of five candidate microRNAs (miRNAs) (miR-99a-5p, miR-31-5p, miR-138-5p, miR-21-5p, and miR-375-3p) in sera from oral cancer patients and paired tumor and normal tissues were detected by real-time qPCR. The diagnostic power of these miRNAs was analyzed by receiver operating characteristic (ROC) curves. Patient-derived xenograft (PDX) models of oral cancer were established and utilized to verify the potential therapeutic effect of miR-31-5p. Candidate miRNAs were screened from our previous studies and verified in 11 paired oral cancer and adjacent normal tissues. Only serum miR-31-5p levels were significantly different between oral cancer patients and healthy controls and between pre- and postoperative patients. Based on the logistic regression model, this panel of five miRNAs distinguished oral cancer patients from healthy control, with an area under the ROC curve (AUC) of 0.776 (sensitivity = 76.8% and specificity = 73.6%). Furthermore, a miR-31-5p mimic enhanced the proliferation of normal epithelial cells, and antagomiR-31-5p inhibited the proliferation of oral cancer cells in vitro. In vivo, antagomiR-31-5p significantly inhibited tumor growth in oral cancer PDX models. Our findings suggest that circulating miR-31-5p might act as an independent biomarker for oral cancer diagnosis and could serve as a therapeutic target for oral cancer.