Abstract
Anoikis is essential for the progression of many malignant tumors. However, the understanding of anoikis' roles in osteosarcoma remains scarce. This study conducted an extensive bioinformatics analysis to identify anoikis-related genes (ARGs), developed ARGs modeles for predicting OS and RFS, and evaluated the effect of these ARGs on osteosarcoma cell migration and invasion. The GSE16088 and GSE28425 datasets provided the differentially expressed genes (DEGs). The prognostic significance and functions of these DEGs were systematically investigated using several bioinformatics techniques. Transwell assays were conducted to determine the effect of OGT on osteosarcoma cell migration and invasion. Seven genes were identified as hub genes, including FN1, CD44, HRAS, TP53, PPARG, CTNNB1, and VEGFA, while 71 ARGs were identified as DEGs. Four ARGs-BRMS, COL4A2, FGF2, and OGT-were used to develop an RFS-predicting model, whereas seven ARGs-CD24, FASN, MMP2, EIF2AK3, ID2, PPARG, and PIK3R3-were used to develop an OS-predicting model in patients with osteosarcoma. In both the training and validation cohorts, high-risk group patients had significantly shorter OS and RFS duration than low-risk group patients. Furthermore, using the aforementioned ARGs, we developed clinically applicable nomograms for OS and RFS prediction. The proportion of tumor-infiltrating immune cells was significantly linked to risk scores. In vitro experiments revealed that knocking down OGT significantly inhibited the ability of MG63 and U2OS cells to invade and migrate. ARG-based gene signatures reliably predicted RFS and OS in osteosarcoma, and OGT showed promise as a potential biomarker. These findings contribute to a better understanding of ARGs' prognostic roles in osteosarcoma.