Lymph node metastasis-related gene signature shows good performance in predicting prognosis and immune infiltration in cervical cancer

Aims: This study aimed to construct a lymph node metastasis-related gene signature to predict prognosis and immune infiltration in patients with cervical cancer.

Methods: Clinical and RNA sequencing data of 193 patients with cervical cancer, which were divided into lymph node metastasis (N1) and non-lymph node metastasis (N0) groups, were acquired from TCGA. Differentially expressed genes (DEGs) between the N1 and N0 groups were detected, and protein-protein interaction combined with LASSO analysis was conducted to further screen lymph node metastasis-related genes. Univariate and multivariate Cox regression analyses were performed to establish a predictive signature. The genetic features, potential biological behavior, and immune infiltration characteristics of the predictive signature were explored. Furthermore, the sensitivity of patients to chemotherapy drugs was estimated based on the predictive signature and the expression of TEKT2 and RPGR was investigated in the cervical cancer tissue samples.

Results: A total of 271 lymph node metastasis-related DEGs, including 100 upregulated and 171 downregulated genes, were identified. Two genes, TEKT2 and RPGR, were associated with lymph node metastasis and prognosis in cervical cancer, and were used to construct a lymph node metastasis-related predictive signature. Based on the predictive signature, patients with cervical cancer were divided into high- and low-risk groups. The high-risk group, characterized by a higher tumor mutation burden and somatic mutation rate, indicated a poor overall survival. The activation of immune infiltration and increased expression of checkpoint genes were observed in the high-risk group, indicating that they might benefit from immunotherapy. Cytarabine, FH535, and procaspase-activating compound-1 were estimated as reasonable chemotherapy options for patients in the high-risk group, whereas two taxanes and five tyrosine kinase inhibitors, including etoposide and vinorelbine, had therapeutic significance for patients in the low-risk group. The expression of TEKT2 and RPGR was significantly downregulated in cervical cancer tissues, especially in metastatic lymph node tissues.

Discussion: The lymph node metastasis-related predictive signature based on TEKT2 and RPGR showed good performance in predicting the survival outcomes of patients with cervical cancer. The risk score of the predictive signature was related to genetic variation and immune infiltration, which could guide immunotherapy and chemotherapy strategies.