Identification of prognostic cellular senescence-related DEGs in the TCGA cohort

A total of 498 PCa patients from the TCGA-PRAD cohort and 297 normal samples from the TCGA-PRAD and GTEx cohorts were ultimately included in the study. We identified 34 cellular senescence-associated core genes from previously published literature [28], excluding undetected genes (GUCY1B1), pseudogenes (WTAPP1),and noncoding RNAs (C1ORF147),leaving 31 cellular senescence-associated core genes for further analysis.Most of the cellular senescence-related genes (SRGs) (30/31, 96.78%) were differentially expressed between tumour tissues and adjacent nontumor tissues (Fig. 1A). Three of the SRGs were correlated with OS and ten were correlated with PFS in the Cox regression analysis (Fig. 1C, D). In addition, 2 of these 3 OS-related genes (DMC1, RRM2) were also among the 10 PFS-related genes. Further GSEA suggested that the differentially expressed genes were enriched in oxidative phosphorylation, the citrate cycle (TCA cycle) and other pathways (Fig. 1B).

Fig. 1

Analysis of cellular senescence-associated core genes in prostate cancer. A Differential expression analysis of cellular senescence-related genes (SRGs) between tumor tissues and adjacent non-tumor tissues. The heatmap displays the fold change in gene expression, with upregulated genes marked in red and downregulated genes marked in blue. B Gene Set Enrichment Analysis (GSEA) showing the enrichment of differentially expressed genes (DEGs) in pathways such as the (TCA cycle). The enrichment score and nominal P-value are provided. C Cox-regression analysis results demonstrating the association between three cellular SRGs and overall survival (OS) in prostate cancer (PCa) cases. The hazard ratio (HR) and P-value are shown. D Kaplan–Meier survival analysis illustrating the association between ten cellular aging-associated genes and progression-free survival (PFS) in PCa cases. The log-rank P-value and hazard ratio (HR) are provided

Bioinformatics analysis of gene signatures related to RRM2 in PCa patients

RRM2 is an enzyme of significant importance in the processes of DNA synthesis and repair [29]. Its crucial role lies in facilitating the conversion of ribonucleotides into deoxyribonucleotides, fundamental components essential for DNA replication and repair mechanisms [30]. The dysregulation of RRM2 has been firmly associated with diverse cancer types, rendering it an appealing candidate for cancer therapeutic interventions. Moreover, it is worth noting that the inhibition of RRM2 has consistently demonstrated its ability to induce cell cycle arrest and promote apoptosis in cancer cells, thereby firmly establishing RRM2 as a highly promising therapeutic target [31-34,32,33,] Additionally, the expression levels of RRM2 have emerged as valuable prognostic markers in numerous cancer types, underscoring their potential utility in predicting disease outcomes and guiding treatment decisions [35-37,36,]. Furthermore, noteworthy findings reveal that RRM2 inhibition can sensitize cancer cells to chemotherapy and radiation therapy [34,35,36], indicating that RRM2 inhibition may have potential utility as a combination therapy with existing cancer treatments.

Based on univariate and multivariate Cox-regression analysis, we identified RRM2 as a key prognostic risk factor in PCa (Fig. 2A–D). Additionally, we have established a corresponding prognostic model, leveraging the expression profiles of the aforementioned ten genes, employing classical Cox regression analysis. Following an extensive evaluation, we determined that an optimal threshold value of lambda yielded a predictive signature comprising four genes. Subsequent survival analysis unveiled that an elevated RRM2 expression mitigated the prognostic impact of the disease, as illustrated in Fig. 2E–G.

Fig. 2

Prognostic Risk Analysis and Survival Outcomes in Prostate Cancer (PCa). A–D Cox regression analysis was performed to assess the prognostic significance of RRM2 in prostate cancer (PCa): A Results from the Cox regression analysis revealed that RRM2 expression is a key prognostic risk factor in PCa, indicating a statistically significant association with both (B) Overall Survival (OS) and (C) Progression-Free Survival (PFS). D The analysis demonstrated the strength of this association by displaying hazard ratios and their confidence intervals (CI), highlighting the impact of RRM2 expression on patient outcomes. E–G Survival analysis results indicated that high RRM2 level was obviously related to worse prognosis. H To further explore the prognostic value of RRM2, the cohort of PCa cases was stratified into high and low-risk groups using the median expression value of RRM2 as the threshold. This categorization yielded two distinct patient groups comprising (insert number here) individuals each, allowing for a more detailed analysis of RRM2’s impact on PCa prognosis. I The predictive accuracy of the risk score for Overall Survival (OS) was evaluated using a Receiver Operating Characteristic (ROC) curve analysis. This analysis assesses the sensitivity and specificity of the risk score in predicting survival outcomes, providing valuable insights into its clinical utility as a prognostic marker. The area under the ROC curve (AUC) was calculated to quantify the predictive performance of the risk score

Presented below is the formulation of the prognostic risk model:

$$}\, = \,\left( } \right)*}\, + \,\left( } \right)*}\, + \,\left( } \right)*}\, + \,\left( } \right)*} (}.}\, = \,0.0)$$

According to the median expression values, the patients were grouped, namely the high and low risk groups, with equal numbers in both groups. The results showed a close correlation between high risk and poor prognosis (Fig. 2H). For the OS risk score obtained in the early stage, the prediction performance is mainly evaluated by the time-dependent ROC curve, so that the (AUC) values are 0.744,0.706 and 0.64 in 1 year, 3 years and 5 years respectively. (Fig. 2I).

RRM2 is upregulated in prostate cancer and associated with poor prognosis

To elucidate the pivotal role of RRM2 in prostate cancer, we meticulously acquired relevant data from comprehensive sources, including the Cancer Genome Atlas (TCGA) database and other pertinent repositories, to investigate the expression levels of RRM2 mRNA across these datasets. The results are comprehensively presented in Fig. 3A, B. Notably, we observed a marked upregulation of RRM2 expression in prostate cancer tissues as compared to their healthy counterparts. Subsequent in-depth analyses revealed substantially elevated RRM2 levels in metastatic lesions (GSE35988, GSE59745), reinforcing its association with disease progression. Furthermore, we scrutinized patients experiencing biochemical recurrence of prostate cancer (GSE120741) and identified notably heightened RRM2 expression in these cases (Fig. 3C). Notably, our investigation based on clinical data from the TCGA database revealed a significant correlation between elevated RRM2 levels and higher Gleason scores and advanced T stage, indicative of an unfavorable clinical prognosis (Fig. 3D–F).

Fig. 3

RRM2 is upregulated in prostate cancer and associated with poor prognosis. A, B Analysis of RRM2 mRNA expression levels in prostate cancer datasets from The Cancer Genome Atlas (TCGA) and Gene Expression Omnibus (GEO) databases. C Higher expression levels of RRM2 were observed in patients experiencing biochemical recurrence of prostate cancer (GSE120741). D–F Clinical data from The Cancer Genome Atlas database revealed a positive correlation between a high RRM2 expression level and a higher Gleason score, T stage, and serum PSA level, indicating an unfavourable clinical prognosis. G–I Higher expression levels of RRM2 were observed in prostate cancer tissue samples than in adjacent tissue samples. Additionally, RRM2 expression levels were significantly higher in castration-resistant prostate cancer (CRPC) tissues than in hormone-sensitive prostate cancer (HSPC) tissues. J Integrated analysis of multiple publicly available databases containing prognostic information showed that high expression of RRM2 indicated an unfavourable prognostic outcome in various survival cohorts

Similarly, in the clinical samples obtained from our research center, we observed elevated levels of RRM2 in prostate cancer tissue juxtaposed with lower expression levels in adjacent healthy tissue. A comparative analysis further underscored the heightened presence of this gene in castration-resistant prostate cancer (CRPC) tissue, while its expression was comparatively reduced in hormone-sensitive prostate cancer (HSPC) tissue. These findings align with the clinical data extracted from The Cancer Genome Atlas database, lending additional support to the observed patterns (Fig. 3G-I). Furthermore, we integrated and analysed multiple publicly available databases associated with prognostic data and found that high RRM2 level indicated worse prognosis in the survival cohorts (DKFZ, Belfast, Stockholm, TCGA) (Fig. 3J). This collective evidence underscores the pivotal role of RRM2 as a key regulatory molecule in the onset and progression of prostate cancer (PCA).

To elucidate the roles of RRM2 in PCa progression, we first detected the expression of RRM2 in different PCa cells (Fig. 4A, B), to silence or regain the expression of RRM2 with higher efficiencies, as well as providing reliable conclusion through validating the function of RRM2 in two cell lines, we chose the two cell lines with moderate RRM2 expression, PC3 and DU145,to perform further functional assays and transfected two independent siRNAs to knockdown RRM2 in PCa cells (Fig. 4C). Our results unequivocally demonstrate that the suppression of RRM2 leads to a substantial reduction in the clonogenic and proliferative capacities of prostate cancer cells (Fig. 4D-H). Moreover, following RRM2 knockdown, we observed a notable increase in the rates of necroptosis and apoptosis in prostate cancer cells, as depicted in Fig. 4I-L and Additional file 1: Figure S1A–D. Furthermore, transwell assays revealed a significant inhibition in the migration speed and the number of migrated cells in RRM2-silenced PCa cells, as illustrated in Additional file 1: Figure S1E–J. In conclusion, the preliminary in vitro results show that RRM 2 aggravates the progression of prostate cancer, which provides a new direction for the treatment of the disease.

Fig. 4

Functional analysis of RRM2 in prostate cancer progression. A, B Analysis of RRM2 mRNA and protein levels across different cell lines revealed heightened expression in PC3 and DU145 cells in comparison to RWPE-1 cell lines. C Effective downregulation of RRM2 in PC3 and DU145 cells, which originally exhibited elevated RRM2 expression, was achieved through the transfection of two independent siRNAs. D–H Notably diminished cloning and proliferation capacity of prostate cancer cells was observed consequent to RRM2 knockdown. I–L A significant increase in the rates of necroptosis and apoptosis observed in prostate cancer cells following RRM2 knockdown

RRM2 regulates sensitivity to docetaxel in prostate cancer cells

RRM2 is a multifaceted factor in chemotherapy resistance, affecting DNA repair, cell survival, proliferation, and drug response [34,35,36]. Understanding its role in specific cancer types and contexts is essential for developing targeted therapies to overcome chemotherapy resistance. Analyses in the Cancer Drug Sensitivity Genomics (GDSC) database showed that high expression of RRM2 significantly weakened the sensitivity of cancer cells to docetaxel, a widely used chemotherapeutic agent (Fig. 5A). Drug screening experiments on PC3 and DU145 cells revealed that their RRM2 content was significantly higher than other prostate cancer cell lines. Further studies showed that the RRM2 knockdown cells had significantly lower resistance to docetaxel, compared to the controls (Fig. 5B, C). Further validation indicates a synergistic effect between RRM2 knockdown and docetaxel therapy (Additional file 1: Figure S3J, K). To better understand the fundamental principle of RRM2 regulating docetaxel sensitivity, two groups of cells (PC3 and DU145) were analyzed after RRM2 silencing using RNA-seq analysis (Fig. 5D). The results showed that RRM2 silencing significantly hindered drug metabolism and chemoresistance pathways, along with the oxidative phosphorylation pathway (Fig. 5E). Upon further investigation, an unexpected discovery emerged regarding the behavior of LNCaP and 22RV1 cells, initially exhibiting a low baseline expression of RRM2. These cells exhibited an increase in RRM2 expression in vitro following treatment with docetaxel and demonstrated a concentration-dependent trend within a specific treatment range (Additional file 1: Figure S3I). This intriguing finding was further validated in the PC3 and DU145 cell lines (Fig. 5F).Considering the evident association between RRM2 expression at the translational level and clinical outcomes, it can be inferred that this gene is notably more abundant in docetaxel-resistant prostate cancer tissues as opposed to sensitive tissues. Additionally, a heightened presence of this gene was observed in the diseased tissue of patients experiencing biochemical recurrence, in contrast to those without biochemical recurrence (Fig. 5G).Furthermore,our subsequent findings demonstrated that silencing RRM2 significantly potentiated the antitumor effectiveness of docetaxel in vivo (Fig. 5H, I), the calculated synergistic effect index was 0.732, affirming the presence of a synergistic effect between RRM2 knockdown and docetaxel treatment. Collectively, based on the insights gathered from the preceding discussions, it can be deduced that the ectopic expression of RRM2 plays a pivotal role in contributing to the development of docetaxel resistance in the clinical treatment of prostate cancer.

Fig. 5

Mechanisms of RRM2-Mediated Docetaxel Resistance in Prostate Cancer. A Analysis in the GDSC database indicated that high level of RRM2 decrease the sensitivity of PCA to docetaxel therapy. B, C Drug screening assay in PC3 with RRM2 knockdown. The IC50 values of docetaxel were lower in RRM2 knockdown cells than in the control cells. C RNA-seq analysis in PC3 and DU145 cells after RRM2 knockdown. D RRM2 silencing inhibits the drug metabolism and chemoresistance pathway. F Increase in RRM2 translational expression level in PC3 and DU145 cells after docetaxel treatment in vitro. G Validation of the relationship between RRM2 expression at the translational level and clinical events in clinical samples. H, I RRM2 silencing enhanced the effects of docetaxel in vivo

Additionally, we conducted preliminary investigations into the role of RRM2 in the process of docetaxel-induced senescence in prostate cancer cells. The results provide evidence that RRM2 promotes senescence in prostate cancer cells induced by docetaxel (Additional file 1: Figure S4).

RRM2 interacts with ANXA1 to activate AKT signalling in prostate cancer cells

Further study found that the reduction in RRM2 gene expression was found to be directly correlated with a decrease in AKT phosphorylation levels, ultimately leading to a heightened anti-tumor effect of docetaxel (Figs. 5 B, C; 6A). These experimental findings highlight the pivotal role of the PI3K/AKT signaling pathway in directly modulating docetaxel resistance in PCa. Furthermore, a noteworthy observation was made as pretreatment of various prostate cancer cells with docetaxel resulted in increased RRM2 expression.This gene’s potential regulatory influence on drug resistance appears to involve the activation of the AKT signaling pathway.

Fig. 6

Mechanisms of RRM2-Mediated AKT Activation in Prostate Cancer. A RRM 2 knockdown somewhat reduced the phosphorylation level of AKT while improving the antitumor effect of docetaxel. B, C Immunoprecipitation, silver staining and mass spectrometry analysis shows that ANXA1 is a binding partner of RRM2 in PCa,and the Log2 ratio indicates the abundant presence of ANXA1 in the RRM2 immunoprecipitates. D Protein interaction network (PPI) analysis indicating the interaction relationship between RRM2 and ANXA1. E Coimmunoprecipitation (Co-IP) assay results confirming the interaction between RRM2 and ANXA1 in PC3 and DU145 cells. F ANXA1 silencing decreased the phosphorylation of AKT in both cell classes. G The expression level and colocalization of RRM2 and ANXA1 in PC3 and DU145 cells

To elucidate the mechanism by which RRM2 activates AKT in diseased tissue, Immunoprecipitation, silver staining and mass spectrometry were employed to identify potential binding partners of this gene (Fig. 6B, C).The analysis revealed ANXA1 as the primary binding partner in body tissue, with the highest log2ratio. Protein interaction network (PPI) analysis further indicated an interaction between RRM2 and ANXA1 (Fig. 6D). Previous studies have suggested that ANXA1 promotes PI3K/AKT signaling by regulating FPR1 and FPR2 in cancerous cells [11, 14, 38]. Therefore, this study focuses on investigating whether RRM2 regulates AKT signaling through ANXA.

Co-immunoprecipitation (Co-IP) tests were conducted, demonstrating a close association between RRM2 and ANXA1 in both PC3 and DU145 cells (Fig. 6E). Silencing of ANXA1 differentially inhibited AKT phosphorylation in these cell lines (Fig. 6F). Subsequent fluorescence staining experiments confirmed the expression and co-localization of RRM2 and ANXA1 in both cell types (Fig. 6G). Based on the aforementioned discussion, it is speculated that ANXA1 plays a crucial role in the activation of AKT.

RRM2 facilitated docetaxel resistance in PCa cells in an ANXA1-dependent manner

ANXA1 was proven to be associated with drug resistance to promote cancer development [16, 39]. Silencing ANXA1 significantly reduced docetaxel resistance in PCa cells (Additional file 1: Figure S1L). Next, we overexpressed ANXA1 in RRM2-silenced PCa cells and observed that the decrease in docetaxel resistance resulting from RRM2 suppression was largely reversed by ANXA1 overexpression (Fig. 7A). Furthermore, RRM2 knockdown partially reduced AKT phosphorylation in PCa, leading to an improvement in the antitumor effect of docetaxel. Conversely, the increased expression of ANXA1 contributed to the restoration of AKT phosphorylation, there by enhancing docetaxel sensitivity (Fig. 7B, Additional file 1: Figure S1K). Building upon the aforementioned discussions, it is plausible to speculate that ANXA1 plays a pivotal role in RRM2-mediated AKT activation within the context of prostate cancer. Furthermore, our investigations revealed that PCa cells exhibited an increase in ANXA1 expression in vitro upon treatment with docetaxel, displaying a concentration-dependent trend within a specific treatment range. Additionally, ANXA1 expression was notably elevated in docetaxel-resistant samples and prostate cancer tissues experiencing biochemical recurrence (Fig. 7C, Additional file 1: Figure S3I).

Fig. 7

ANXA1 Mediates RRM2-Induced Activation of AKT and Docetaxel Resistance in Prostate Cancer. A RRM 2 knockdown significantly inhibited the phosphorylation of AKT in diseased tissue, while improving the therapeutic effect of docetaxel, and overexpression of ANXA 1 enhanced the phosphorylation of AKT. ANXA 1 silencing inhibited RRM 2 changes, which in turn affects AKT phosphorylation and drug sensitivity. B Expression of ANXA1 increases in LNCaP and 22Rv1 cells upon treatment with increasing concentrations of docetaxel within a certain range. ANXA1 level is obviously higher in docetaxel-resistant tumour group compared to that in sensitive group. C The docetaxel resistance induced by RRM2 suppression is largely reversed by ANXA1 overexpression in prostate cancer cells. D–F RRM2 knockdown decreases the protein level of ANXA1, shortens its half-life, and increases its ubiquitination level without affecting its mRNA level. G UBE3A knockout promotes the protein expression of ANXA1 in PC3 cells, which reflected that UBE3A plays a role in ANXA1 degradation. H–I The ANXA1 protein level is positively correlated with the RRM2 level in PCA tissues

To substantiate the hypothesis that RRM2 influences ANXA1 content in PCa cells, we meticulously examined the protein and mRNA levels of ANXA1 following RRM2 knockdown in prostate cancer cells. Intriguingly, a substantial reduction in ANXA1 protein levels was observed, while mRNA levels remained unaffected (Additional file 1: Figure S3O). Furthermore, it was noted that the half-life of ANXA1 considerably shortened following RRM2 gene knockdown, and ANXA1 ubiquitination levels showed a discernible increase during this process (Fig. 7D–F). Previous experimental studies have shown that the E3 ligase UBE3A interacts with the C-terminal domain of ANXA1, leading to ANXA1 degradation [11, 40]. Knockdown of RRM2 resulted in decreased ANXA1 protein level in PCa cells, and UBE3A knockout increased the protein levels of ANXA1 (Fig. 7G). Based on the above discussion, it can be inferred that RRM2 stabilizes ANXA1 in Pca tissue by competing with UBE3A. Furthermore, we found a close relationship between RRM2 content and ANXA1 in PCa tissue samples, with higher RRM2 content associated with increased ANXA1 levels (n = 56, P < 0.001) (Fig. 7H, I).

Drawing upon the clinical translational significance of RRM2, we proceeded to assess the impact of COH29, an RRM2 inhibitor, on PC3 and DU145 cells in vitro. The outcomes demonstrated that COH29 alone exhibited a noteworthy capacity to inhibit the growth of PC3 and DU145 cells. When administered in combination with docetaxel treatment, COH29 displayed a synergistic effect alongside the latter (Additional file 1: Figure S3A–F).