Cellular senescence is a stable cell cycle arrest that remodels the TME through various cancer hallmarks, such as tumour proliferation, migration, invasion, angiogenesis, EMT, and tumour immune response, thereby affecting cancer patients’ prognosis [25, 26]. However, the correlation between cellular senescence and CC is still unclear. In this study, we sought to explore the potential role of cellular senescence in the pathogenesis of CC using bioinformatics and in vitro experiments to better predict the prognosis of CC and provide novel targets for its treatment.
Firstly, we explored the expression of senescence-associated genes in normal and CC samples using databases. We found a total of 23 senescence associated genes that were differentially expressed in CC samples. These 23 genes include AGT, SOX2, E2F1, GNG11, SERPINE1, IL-1α, IFNG, CAVIN1, CDKN2B, CXCL1, HEPACAM, HSPB2, ID1, IGFBP5, ING2, KL, MMP9, PMVK, SFN, SIX1, TNFSF15, VENTX, ZFP36. This result preliminarily illustrated the important role of cellular senescence in the disease occurrence and progression of CC.
With the development of high-throughput sequencing technology as well as computer algorithms, constructing gene sets for disease prediction has provided strong support for prognostic prediction of tumors, which is beneficial for optimizing treatment decisions in clinical practice. Numerous gene signatures have been developed for tumor prognosis prediction, including cell senescence gene sets. For example, a study has established four senescence-related genes (BAK1, DKK1, CDKN2A, and MIF) based prognosis model for predicting the patient’s survival rate in head and neck squamous cell carcinoma [27]. Lin et al. [28] identified three cellular senescence clusters associated with different patient prognoses by analysing 278 CSRGs in lung adenocarcinoma. However, cellular senescence-related gene sets associated with patient prognoses have not been identified and validated in CCs. Therefore, we established a risk model consisting of eight CSRGs like AGT, SOX2, E2F1, GNG11, SERPINE1, IL-1α, IFNG, and CAVIN1. We used the median risk score as a parameter to divide the CC patients into the HR-G and LR-G. The survival and Cox regression analyses revealed that the risk signature could independently predict patients’ survival outcomes in CC. Additionally, a nomogram was established by integrating the risk scores and tumour stage. This nomogram verified the predictive efficiency and clinical utility of the risk model. Collectively, these results demonstrate that our risk model is capable of predicting the patient’s prognosis, which would aid in identifying biological factors involved in CC development.
SASP is an important feature of cell senescence, which includes cytokines, chemokines, growth factors, and proteases. Different SASP molecules serve different functions in the TME. We therefore analyzed the expression of SASP molecules in the HR-G and LR-G, and we found that multiple SASP molecules, including IL-6, IL-8, IL-1β, and VEGFA, were increased in high-risk CC patients. This is consistent with earlier findings. Multiple studies have shown that IL-6 is important for cervical carcinogenesis. Pan et al. conducted an immunohistochemistry analysis of IL-6 expression in CC tissues, discovering a significantly elevated expression in tumor tissues [29]. Additionally, research has shown that IL-6 is abundantly expressed in invasive CC and is implicated in the pathogenesis of HPV-related CC [30]. All the above findings suggest that IL-6 is a detrimental factor for the development of CC. IL-8 is a pro-inflammatory factor that promotes tumor growth. Fujimoto et al. [31] found that CC patients with high levels of IL-8 had an extremely poor prognosis, while patients with lower levels had a better 24-month survival rate, which indicates IL-8 is a prognostic indicator of CC. In 2017, Jia et al. [32] found that IL-8 were associated with the tumorigenesis of CC, and exogenous IL-8 promotes the carcinogenic potential of HeLa cells. VEGFA is a key factor in blood vessel formation, and previous studies have shown that the expression amount of serum VEGFA is upregulated in CC, and targeting VEGFA is beneficial for the treatment of CC [33, 34]. It therefore has the potential to be an effective treatment modality for cervical cancer by modulating SASP molecules in the TME.
Another important finding in our study is the significant correlation between CSRGs and the composition of tumour-infiltrating immune cells. Many studies have shown that cellular senescence is associated with the TIME [35]. Senescent cells secrete numerous cytokines and chemokines to induce immune cells and promote the body's immune response. Multiple studies have confirmed that most of cervical cancer is HPV positive, and the body shows antiviral immune response after infection with HPV virus [35]. Therefore, the TIME is important for the development of cervical cancer, which deserves to be fully studied. In this work, GSEA analysis revealed the up-regulation of several pathways associated with immune responses like natural killer cell-mediated cytotoxicity, the B, T, and Toll-like receptor signalling pathways in patients in the LR-G. Our correlation analysis revealed that the stromal score had a positive correlation with the risk score, while the immune score had a negative correlation with the risk score. Based on this, we analysed the composition of immunocytes, and we found that most immune cells were up-regulated in the LR-G, including B cells, CD8 + T cells, NK cells, and neutrophils. These results illustrate a more active and complex immune response in patients in the LR-G, which also lays the foundation for immunotherapy in CC.
Immunotherapy is a rapidly developing therapeutic strategy which holds tremendous potential in clinical settings. Immunotherapy targets and eliminates tumour cells by activating the immune system of the patient. Studies have shown the efficacy of immunotherapy in treating various solid cancers like lung, breast, and renal [36–39]. Additionally, studies have demonstrated the benefits of immunosuppressive agents targeting PD-1 and CTLA-4 or its primary ligand PD-L1 in treating patients with advanced and metastatic CC [40, 41]. Cellular senescence-related immune remodelling could influence the efficacy of immune checkpoint blockade. Our results showed an increase in PD-1, CTLA-4, and PD-L1 expression in patients in the LR-G, thus indicating a higher sensitivity of these patients to immune checkpoint blockade therapy. Additionally, in the IMvigor210 cohort, patients with low-risk scores were highly sensitive to PD-L1 inhibitors. Therefore, the risk model could aid in screening patients who could benefit from combination therapy.
We established the risk model based on eight CSRGs. KM analysis showed an independent association between SERPINE1 and IL-1α expression and the prognosis of patients with CC. SERPINE1 negatively regulates the pericellular proteolytic pathway. Studies have shown a correlation between high SERPINE1 expression and poor disease as well as shorter disease-free survival outcomes in several cancers, like, breast and gastric [42, 43]. Hazelbag et al. used “multivariate COX regression analysis” and identified SERPINE1 as a strong independent prognostic factor for CC. Additionally, the study has shown an association between SERPINE1, poor survival, and disease recurrence in a subgroup of patients with CC without lymph node metastasis [44]. Interestingly, our results showed high SERPINE1 expression in patients in the HR-G. The clinical outcome of CC patients expressing high SERPINE1 levels was poor. In addition, SERPINE1 expression was higher in CC cells and tissues. IL-1α is a crucial cytokine involved in inflammatory processes and promotes cancer pathogenesis. However, IL-1α exert pro- and anti-cancer effects; hence its involvement in cancer progression is still controversial. Liu, S. et al. [45] showed that IL-1α promotes breast cancer progression by increasing the activation of the NF-kB and STAT3 signalling pathways. Interestingly, Dagenais, M. et al. [46] showed that IL-1α suppresses breast cancer by inhibiting cell proliferation via the IL-1α signalling pathway. However, no study has reported IL-1α expression and functions in CC. Our results showed an increase in IL-1α expression in patients in the HR-G. The prognosis of patients expressing high IL-1α levels was poor. Additionally, an increase in IL-1α expression was observed in the CC cells and tissues, thus indicating that IL-1α could promote malignant transformation of cells, thereby detrimental to the prognosis of patients with CC. In vitro as well as in vivo should be conducted to determine the involvement of SERPINE1 and IL-1α in CC.
However, our study has several limitations. First, we used data extracted from publicly available databases. Hence, prospective studies involving human subjects are required to validate our results. Additionally, cell-based and animal experiments should be performed to enhance our understanding of the mechanisms of CSRGs in the progression of CC.