CESC is one of the frequently detected malignancies and pose a health threat to women all over the world[1–3]. Clinical stage, pathologic type, response to therapy and long-term prognosis can vary considerably among CESC patients. It is important for CESC patients to obtain survival time individually. Research suggested that biomarkers, such as alternative mRNA splicing, might be an effective prognostic predictor[6]. AS is a vital biological process that can expand protein species diversity[9]. There is an increasing number of studies that AS plays a crucial role in the oncogenesis and development of cancer. Previous studies reported that AS may accurately evaluate prognostic outcomes in some cancers, including glioblastoma multiforme, colorectal cancer, prostate adenocarcinoma, bladder urothelial carcinoma, gastric cancer and so on[23, 24]. However, the relevant research of cervical squamous cell cancer is scarce.
In the present study, we constructed AS-associated prognostic signatures. AS-associated prognostic signatures were exactly independent prognostic factors in CESC. And prognostic signatures had high AUC values, proving prognostic signatures had reliable predictive efficacy. The findings highlight the clinical significance of AS-associated prognostic signatures in CESC. After integrating age, stage T, N, M and AS-associated prognostic signatures, we carried out a comprehensive analysis, that is nomogram, to quantitatively predict individuals’ survival probability. Compared with histopathological type or TNM stage, it had a more excellent performance in predicting survival time in CESC. These findings suggest that AS-associated prognostic signatures may eventually act as efficient prognostic biomarkers to assess survival time in CESC. This is the first and the most comprehensive analysis to construct a prognostic prediction model, which incorporate age, TNM stage and AS-associated prognostic signatures, in cervical squamous cell cancer.
Alternative splicing events are thought to be accurately regulated by hundreds of splicing factors, which specifically combine with the splicing sites and participate in subsequent splicing events. There is no doubt that aberrant splicing factors or abnormal expression of splicing factors directly leads to changes of splicing events pattern and splicing events expression[14, 15]. Some studies reported that there was correlation between splicing factors and poor prognosis in chronic lymphocytic leukemia, lung adenocarcinoma, pancreatic cancer, breast cancer and melanoma[33–35]. In our study, splicing factors were collected from the SpliceAid2 database, and correlation network between SASEs and SFs was explored. The most significant SASEs and SFs were SEC23A-27346-AP and EIF3A with the maximum correlation coefficient. EIF3A had a strongly positive regulation with SEC23A-27346-AP. Our findings revealed that EIF3A might be the upstream regulator of SEC23A-27346-AP. The in-depth correlation analysis of regulation network between SESAs and SFs provides a new insight into the mechanism of AS involved in the unfavorable prognosis of cervical squamous cell cancer.
Alternative splicing events not only can be regulated by splicing factors, but also can respond to multiple signaling pathways. As we all know, weakening or lacking tight cell-cell adherens junction has been considered to be a fatal hallmark of cancer[36, 37]. In normal physiological conditions, cell-cell adherens junction is the foundation of tissues architecture[38]. However, tumor cells tend to loss of adherens junctions, and display local dissemination and distant metastasis especially in advanced cancers. Cadherins and catenins are the core proteins of cell-cell adherens junction. Low or loss of expression of cadherins and catenins inevitably lead to tumor progression and unfavorable prognosis[39]. In the current study, we selected prognostic KEGG pathways in CESC by Gene Set Variation Analysis (GSVA). The correlation analysis between SASEs and prognostic KEGG pathways was carried out. The most significant SASEs and prognostic KEGG pathways were SEC23A-27346-AP and adherens junction pathway with the maximum correlation coefficient. Therefore, we supposed that adherens junction pathway might be one of the downstream pathways of SEC23A-27346-AP in CESC.
SEC23A, as a core component of coat protein complex Ⅱ vesicles, can not only transport proteins but also cause human diseases. Aberrant SEC23A or abnormal expression of SEC23A was reported to be relevant to the development and progression of human cancer[40]. A previous study implicated that RNA binding motif 5 (RBM5) participate in regulating mRNA splicing of SEC23A[41]. However, our results indicated that SEC23A-27346-AP was associated with poor prognosis in CESC patients, and splicing factor EIF3A positively regulated SEC23A-27346-AP. Above all, SEC23A regulated by EIF3A may facilitate the development of cervical squamous cell cancer via adherens junction pathway.
EIF3A is the largest subunit of eukaryotic initiation factor 3 (EIF3), which plays a crucial role in all steps of translation initiation. Numbers of studies have shown that EIF3A exerts an important influence on the regulation of cell cycle, cell growth and differentiation [42]. Recent studies have found that EIF3a is correlated with carcinogenesis, development and chemotherapy efficacy[43, 44]. However, it is unknown whether EIF3A is a proto-oncogene in CESC. In the study, our results indicated that EIF3A knockdown decrease cell proliferation and cell migration of CESC. Therefore, EIF3A can promote proliferation and migration of CESC.
There are some limitations in this study. Firstly, the number of CESC patients in TCGA database was relatively small. Clinical information of 308 CESC patients was extracted from the TCGA website. After excluding the patients who didn’t accord with the inclusive criteria, a total of 216 CESC patients were enrolled in our study. Secondly, we performed a nomogram to more precisely obtain individuals’ survival probability by integrating multiple risk factors. However, we failed to take into consideration other risk factors, such as pathological grade, therapeutic strategies and so on. Moreover, the mechanism of alternative splicing events in the development of CESC was unclear. We carried out correlation analysis to explore upstream regulators and downstream pathways of alternative splicing events. In the meanwhile, we merely validated the relationship between EIF3A and CESC proliferation and migration. In the follow-up study, the molecular mechanisms of SEC23A and adherens junction pathway in the occurrence and development of cervical squamous cell cancer need further investigations to verify.