Identification of an Anoikis-Related Gene Signature to Predict the Prognosis in Patients With Oral Squamous Cell Carcinoma

ABSTRACT: Background: Anoikis plays an important role in cancer metastasis. Nevertheless, research that focuses on the prognostic values of anoikis-related genes (ARGs) in oral squamous cell carcinoma (OSCC) is still scarce. Methods: The ARG-related gene signature was developed via LASSO regression using the TCGA-OSCC cohort. OSCC patients in the TCGA_OSCC and GSE41613 datasets were grouped into high- and low-ARG score (H-score, L-score) groups, and the overall survival (OS) between the 2 groups was analysed. Meanwhile, functional enrichment, CIBERSORT and drug sensitivity were conducted based on the data from the TCGA_OSCC cohort. Loss-of-function assay was employed to explore the biological roles of SPINK1 in OSCC. Results: In the current research, we constructed a novel prognostic model based on 5 ARGs (SPINK1, PLAU, VEGFA, PLK1 and EGF) for OSCC. The OS of OSCC patients in the H-score group was worse than that in the L-score group. Moreover, patients with the H-score had low immune and ESTIMATE scores, suggesting a declined immune status in these patients. Additionally, some tumour-promoting signalling pathways (e.g. focal adhesion, PI3K-Akt and WNT signalling pathways) were notably activated in the H-score group. Functionally, downregulation of SPINK1 notably suppressed OSCC cell viability, migration and invasion and induced ferroptosis through inactivation of the Nrf2/GPX4 signalling pathway. Conclusion: We have developed a novel 5-gene signature related to ARG, which may serve as a predictor for OS in patients with OSCC. Our findings may open a new possibility of providing personalised treatment for OSCC patients.