Background: The persistent expression of high-risk human papillomavirus (hrHPV) E6 oncoproteins is a critical determinant in driving and maintaining the malignant phenotype of cervical cancer, a pathogenic process where autophagy serves as a key regulatory mechanism. This study aimed to identify autophagy-related genes as potential biomarkers for prognostic evaluation in cervical cancer.

Methods: This study established an HPV 16 E6-induced C33a cervical cancer cell model, which was treated with the autophagy activator rapamycin or inhibitor 3-methyladenine to modulate autophagy. The expression of the autophagy-related gene C-X-C motif chemokine ligand 8 (CXCL8) was analyzed by reverse transcription-quantitative PCR (RT-qPCR), western blotting, and enzyme-linked immunosorbent assay. Cell viability, migration, invasion, and apoptosis were assessed using the Cell Counting Kit-8 assay, wound healing assay, Transwell assay, and Terminal deoxynucleotidyl transferase–mediated deoxyuridine triphosphate nick-end labeling (TUNEL) assays, respectively. Furthermore, CXCL8 expression levels in tumor tissues from cervical cancer patients with favorable (n = 81) and unfavorable (n = 61) prognosis were examined by immunohistochemistry and RT-qPCR. Univariate and multivariate Cox proportional hazards regression analyses were conducted to identify independent risk factors influencing disease-free survival (DFS) in cervical cancer patients. Receiver operating characteristic (ROC) curve analysis was employed to evaluate the predictive value of CXCL8 for unfavorable prognosis risk.

Results: Mechanistic studies indicated the participation of CXCL8 in HPV 16 E6–induced malignant phenotypes. The autophagy activator rapamycin or CXCL8-neutralizing antibody could neutralize the oncogenic effects of HPV 16 E6. CXCL8 was highly expressed in tumor tissues with poor prognosis. It showed correlations with poor tumor differentiation, cervical infiltration depth ≥2/3, and lymph node metastasis, independent of clinical stage. CXCL8 was identified as an independent prognostic factor in cervical cancer [hazard ratio = 3.143, 95% confidence interval (CI): 1.519–6.507, p = 0.002] and was significantly correlated with inferior DFS (χ² = 34.905, p < 0.0001). Furthermore, a model combining CXCL8 with squamous cell carcinoma antigen and cytokeratin 19 fragment showed promising prognostic accuracy, achieving an area under the curve of 0.897 (95% CI: 0.835–0.942) with 90.16% sensitivity and 88.89% specificity.

Conclusion: The CXCL8 gene promotes cervical cancer progression by contributing to the regulation of autophagy mediated by HPV 16 E6. This functional role underpins its potential utility as a clinical prognostic biomarker.