Integrative transcriptional characterization of cell cycle checkpoint genes promotes clinical management and precision medicine in bladder carcinoma

Cell cycle checkpoint-related genes (CRGs) play pivotal roles in cell cycle progression (CCP), ensuring and control regulating cell cycle events (1). Generally, in eukaryotic cells, the mitotic cell cycle is composed of two stages, the interphase (G1, S, and G2) and the mitotic (M) phase. The gap phases of G1-to-S (2), S-to-G2 (3), and G2-to-M (4), likely as decision windows, can determine cell cycle entry and progression. In cancer cells, some CRGs preventing DNA damage are usually compromised, contributing to genetic alterations and genomic instability (5), but those CRGs involved in DNA replication stress are scarcely altered to endure the replication stress (6). On the other hand, cancer cells could potentiate DNA replication stress through transcriptional regulation of CRGs (7). Overall, CRGs and CRG-related signaling pathways play a key role in regulating the phase transitions, CCP, and cell cycle entry and exit in cancer cells (8).

In previous studies, it has been revealed that aberrantly expressed CRGs might be an essential prerequisite for cells to become cancerous, leading to tumor development and progression. For instance, CRG ataxia telangiectasia mutated (ATM) plays a core role in responding to DNA damage and stimulating DNA repair signaling pathways, and its absence is highly prone to giving rise to carcinogenesis (9). Of note, its downregulation or inactivation is associated with the highly accumulated genomic aberrations, which is one of the hallmarks of cancer (10). Ataxia telangiectasia and Rad3-related (ATR), another CRG encoding protein kinase that regulates DNA damage response (DDR), is correlated with the polarization of M2 tumor-associated macrophages, lymph node metastasis, and poor prognosis of patients with nasopharyngeal carcinoma (11). The abnormal expression of CRGs has been found to be associated with the development and progression of multiple kinds of cancers, including melanoma (12), lung cancer (13), colorectal cancer (14), and hepatocellular carcinoma (15). A growing body of evidence has found that the dysregulation of CRGs could render cells to be cancerous and promote cancer cell proliferation, but most studies only disclose the role of a single CRG in cell cycle, tumor carcinogenesis, and progression. Currently, the role of integrated CRGs representing checkpoint mechanisms in the regulation of cell cycle in tumor carcinogenesis and progression remains to be fully delineated.

Previous studies have already expounded that dysregulation of CRGs is correlated with increased genomic instability and malignant progression in bladder carcinoma (BC) patients (16–18), indicating that there is a great potential of CRGs to become prognostic or targeted biomarkers for BC patients. BC generally presents as non-muscle-invasive bladder cancer (NMIBC), muscle-invasive bladder cancer (MIBC), or metastatic BC, of which the MIBC subtype has a relatively worse prognosis and poor treatment responses (19). Moreover, BC is a molecularly heterogeneous cancer with divergent clinical outcomes (20). The heterogeneity of tumor is always a huge challenge for cancer management and could reduce the efficacy of molecularly targeted therapies; therefore, the dissection of molecular signatures is urgently needed. Based on transcriptomic profiling of CRGs, a novel signature was constructed in the present study with good performance in prognosis prediction and that could function as a biomarker for treatment response. Overall, this study was of guiding significance in the clinical management of BC patients and promoting precision treatment.

Bladder cancer, according to the latest cancer statistics worldwide, remains to be one of the most prevalent cancers, with approximately 550,000 new cases annually (27). Recently, next-generation sequencing technology has led to the significant advances in the research field of bladder cancer (28), discovering a number of genomic, transcriptomic, and proteomic biomarkers for predicting diagnosis or prognosis of patients, as well as promoting clinically favorable targeted therapeutics and effective immune therapies. Furthermore, the newly developed methods have been widely applied to explore the tumor immune microenvironment in various cancers, which expanded the repertoire of precision medicine, especially immunotherapy (29). In the present study, it was the first time to investigate the integrative transcriptional characterization of cell cycle checkpoint genes in BC progression, and eventually a novel CRG signature and CRG-based nomogram were established, with remarkably robust and stable capacity in prognosis prediction. The clinical association analysis underlying the CRG signature further disclosed that the high-risk group had more patients who were female, diagnosed at older age, and with more aggressive diseases. In addition, the molecular characterization, including functional differentiation analysis, genomic alteration profiling, and TME, between different subgroups promoted the potential strategies of precision treatment for BC patients.

The early retrospective study proposed a risk model, namely, CCP score, based on the expression signature of 31 CCP genes, which was created to predict the aggressiveness of prostate cancer (30). Afterwards, this 31-CCP gene expression signature exhibited significant prognostic values in various cancers, such as ductal carcinoma in situ of the breast (31), clear cell renal cell carcinoma (32), lung cancer (33), and bladder cancer (34). According to the study of bladder cancer, an optimized 12-CCP signature was established, of which the AUC values in predicting patient progression (non-progressor vs. progressor) were 0.70 in the Lindgren cohort and 0.68 in the CNUH cohort. In contrast, it seemed that the CRG signature might outperform the CCP score in prognosis prediction, with the AUC values at 1, 2, 3, 4, and 5 years for OS all beyond 0.75. Furthermore, recent studies have revealed that the regulation of CCP, such as the initiation of replication, the overall rate of replication, and the recovery and resumption of replication forks, was strictly controlled in cancer cells by checkpoints (35–38). In brief, in cancer cells, cell cycle checkpoints could control CCP and further regulate the cancer cell divisions, unless uncontrolled CCP can drive more mutations and genomic instabilities causing the existence of cell cycle and cell apoptosis.

In past decades, a large body of studies focused on the regulation of CCP in tumor progression (39), in which the important role of checkpoints regulating the entire cell cycle processing was often neglected. In the present study, a total of 52 CRGs in cell cycle were found to be significantly correlated with the clinical outcomes of BC patients; meanwhile, the novel CRG signature was constructed as a prognostic model consisting of further selected 23 CRGs. Among these 23 selected CRGs, the expression of PPP2CB has already been proven to be involved in promoting BC cell proliferation and migration (40). The expression of checkpoint CUL4A could mediate the degradation of BECN1 protein to alleviate cell autophagy and enhance the growth of BC (41). The overexpressed PLK2 was detected in BC, and urinary PLK2 protein level was highly correlated with bladder transitional cell carcinoma (42). The suppressed expression of THOC1 could mediate BC cell apoptosis (43). Except for the in-depth insight into the biological functions of the above-mentioned four CRGs, 19 other identified CRGs involved in the signature, namely, ANAPC4, MMAB, B9D2, NDEL1, SLC25A15, PSMA7, PSMB10, PSMB5, PSMB8, RAD9A, REXO2, ARID3A, CHMP4C, DDX39B, FBXO31, FBXO6, PRPF19, RGCC, and TIPIN, were first identified to play a potential role in BC progression, and their detailed functions remained enigmatic, which merited further experimental verification.

Underlying the CRG signature, correlation analysis revealed that signature-associated genes were mainly responsible for the collagen-containing ECM, and ECM-related and MAPK signaling pathways, indicating that the ECM-specific heterogeneity and related signaling pathways between different risk groups mainly caused the significant difference in OS. It has been revealed that COL1A1 protein was markedly upregulated in MIBC, which could activate the epithelial–mesenchymal transition and TGF-β signaling pathway contributing to the proliferation and invasion of bladder cancer cells (44), whereas the significantly downregulated expression of collagen type IV-α1 and α2 (COL6A1 and COL6A2) was found to promote tumor progression in both the NMIBC and MIBC tissue samples (45). Moreover, the tumor-related macrophages could secrete the type I collagen via activating the PI3K/AKT signaling pathway to stimulate the development of bladder cancer; meanwhile, the number of macrophages and the expression of M2 macrophage-associated genes (ARG-1, IL-10, and TGF-β) were remarkably elevated in malignant bladder tumor tissue samples (46). Furthermore, TGF-β could prompt the tumor immune escape and the resistance of immune therapy (47), which has been validated in the IMvigor210 cohort in which BC patients not responding to the therapeutic treatment of anti–PD-L1 agent (Atezolizumab) were highly correlated with the TGF-β signature in fibroblasts and commonly had fibroblast- and collagen-rich peritumoral stroma (48). In addition, a single-cell proteomic analysis further revealed the chaotic tumor-associated collagens in the TME of MIBC (49). Thus, therapeutic treatment targeting collagen-specific ECM and related signaling pathways could be a potential regimen for BC patients.

Genomic alteration analysis revealed that the high-risk group had higher prevalence of tumor suppressor genes RB1, FBXW7, and NFE2L2, consistent with the previous findings that the altered RB1 (50), FBXW7 (51), and NFE2L2 (52) were correlated with tumor progression and worse outcomes of BC patients. However, the potential contribution of frequently altered ASAP2 and PCSK6 in the progression of bladder cancer remains unclear, needing to be further investigated. As known, the FGFR3 alterations were mainly enriched in the luminal bladder cancer correlated with better prognosis; specifically, the low-risk group in the present study was also observed to have more patients with FGFR3 alterations. More impressively, it was uncovered that there was a significant difference in alteration frequency of only one DDR signaling pathway-related gene—POLE—between high-risk and low-risk groups, and Kaplan–Meier analysis revealed that MIBC patients with altered POLE exhibited a trend of better OS. Owing to the limited number of MIBC patients presenting POLE alterations in the present study, this result should be highlighted and verified in the future.

Taken together, the difference in TME, including tumor purity, T-cell viability, and the proportion of TILs, between high-risk and low-risk groups was further investigated. Macrophage M0 and M2 were consistently enriched in high-risk groups, which also had a lower tumor purity expressed by the higher ESTIMATE score and a higher level of exhausted T cells indicated by the higher TIDE score. Similarly, it has been reported that tumor-associated macrophages could restrain the infiltration level of T cells in tumor tissues (53). In addition, altered FGFR3 is highly enriched in luminal or its papillary subtypes, which have been characterized by immune-suppressive states (50). Controversially, a retrospective analysis revealed that there was no difference in immune checkpoint blockade response between patients with altered FGFR3 and those with no altered FGFR3 (54), and a recent study found the equivalent T-cell receptor diversity between these patients (26). Whether the status of FGFR3 could influence the treatment response by the immune checkpoint blockade is discussed heatedly. In the present study, it was found that the discrepancy of TME between high-risk and low-risk groups among BC patients with wild-type FGFR3 was a little small, except that low-risk patients had more activated T cells but a lower proportion of tumor-associated macrophages (especially for macrophage M0 and M2), while compared to the BC patients with wild-type FGFR3, high-risk patients with altered FGFR3 had nearly similar predicted infiltrated immune cell proportions and immune responses, but only the expression levels of TIM3 and CD47 were equal, such that these BC patients seemed to be more sensitive to immune checkpoint inhibitors targeting TIM3 and CD47, which needed further analysis. Additionally, the low-risk BC patients with altered FGFR3 exhibited the highest tumor purity and the most activated T cells, which were associated with better treatment responses of immunotherapies (23). However, in the present study, the expression levels of well-known immune checkpoint genes were found to be extremely low. Thus, in clinic, it was highly recommended that immunohistochemistry staining of immune checkpoint(s) should be detected first during immunotherapeutic treatment.

In addition, the different BC molecular subtypes were identified to have distinctive sensitivity to the chemotherapeutic drugs, such as cisplatin, docetaxel, paclitaxel, methotrexate, and doxorubicin, which were commonly used for BC patients (28). The CRG signature further demonstrated its predictive ability in therapeutic treatment response. Currently, cisplatin has been regarded as a typical chemotherapeutic strategy for UC patients, and it mainly functioned to induce cell cycle arrest and cell apoptosis (55, 56). Furthermore, a previous study has revealed that docetaxel could make antiproliferative and apoptotic effects on bladder cancer cells (57). In the present study, it was found that patients with a higher CRG score, regardless of FGFR3 status, seem to be more sensitive to cisplatin or docetaxel. Although clinical benefits of paclitaxel were limited owing to patients’ resistance, the failure of first-line combination treatment of cisplatin and gemcitabine for advanced and/or metastatic UC patients provided an opportunity for paclitaxel because of its apoptotic effects (58). Through the evaluation analysis of chemotherapy treatment response, it was recommended that low-risk BC patients with wild-type FGFR3 might benefit more from paclitaxel. Additionally, the CRG signature also exerted its influence on the patients’ sensitivity to methotrexate, especially suitable for low-risk BC patients with wild-type FGFR3; however, its applicability for BC patients in clinic was rare, which needed more verification. Doxorubicin has been proven to disturb the regulation of cell cycle and induce cell death, usually as a vital constituent of combination treatment for MIBC patients (59). Indeed, the alteration frequency of FGFR3 in MIBC patients was relatively lower, compared with NMIBC patients (60, 61). In the present study, we further found that high-risk BC patients with wild-type FGFR3 were likely to have higher sensitivity to doxorubicin. Collectively, the checkpoint mechanisms in the regulation of cell cycle greatly influenced the response of chemotherapeutic treatment.

The CRG signature also demonstrated its ability in predicting the response of immunotherapeutic treatment. Of note, the CRG signature exhibited better response prediction capacity for BC patients with wild-type FGFR3, of whom those with lower CRG score might be more sensitive to the immunotherapeutic treatment with anti-PD-1 or anti-PD-L1 drugs. Regarding patients with altered FGFR3, the CRG signature could not differentiate the potential CR/PR and SD/PD patients. As described before, the alteration frequency of FGFR3 in low-grade and/or early-stage bladder tumors was evidently higher comparatively, such that the CRG signature could perform better in the response prediction of immunotherapeutic treatment for higher-grade BC patients, especially for those with wild-type FGFR3, while only 17 BC patients with altered FGFR3 included in the GSE176307↗ dataset received immune therapy treatment hence, the ability of the CRG signature to predict the immunotherapeutic treatment response for BC patients with FGFR3 alterations remained to be explored.

The present study comprehensively delineated the integrative transcriptional profiling of cell cycle checkpoint genes in the BC progression. The novel prognostic CRG signature and nomogram exhibited good performance in prognosis prediction for BC patients; furthermore, the molecular characterization underlying the CRG signature provided deep insights into key risk factors leading to the aggressive bladder cancer, promoting the development of precision medicine. The predictive role of the CRG signature in treatment response offered potential precision treatment strategies for different BC molecular subtypes.

The original contributions presented in the study are included in the article/. Further inquiries can be directed to the corresponding authors. 2

The studies involving human participants were reviewed and approved by The ethics committee of the Chinese PLA General Hospital (S2019–302-01). The patients/participants provided their written informed consent to participate in this study.

W-WS, J-ZG, YH, and BY proposed and designed this study. W-WS and J-ZG collected raw data. Y-PL, QS, QX, and B-YQ pre-treated raw data and performed the statistical analyses. W-WS, J-ZG, Y-PL, Z-QM, and BY wrote the original manuscript. YH and BY revised the manuscript. All authors have read and approved this study to be published.

This study was supported by the Department of Medical Oncology, Senior Department of Oncology, The Fifth Medical Center of PLA General Hospital.

We are very grateful for the support of The Fifth Medical Center of PLA General Hospital and the great efforts made by all participants. Meanwhile, we appreciate the help offered by the Lifehealthcare Clinical Laboratories, Hangzhou, Zhejiang Province, People’s Republic of China.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.