Research ArticleExperimental Studies
Open Access

Ubiquitin-proteasome Pathway-linked Gene Signatures as Prognostic Indicators in Prostate Cancer

YASUO TAKASHIMA, KENGO YOSHII, MASAMI TANAKA and KEI TASHIRO
Anticancer Research May 2025, 45 (5) 1825-1841; DOI: https://doi.org/10.21873/anticanres.17562

Abstract

Background/Aim: Prostate cancer (PCa) is the most frequently diagnosed cancer in men and a leading cause of cancer-related death. While prostate-specific antigen is a widely used biomarker, its specificity is limited. This study investigated the prognostic significance of gene subsets associated with the ubiquitin-proteasome pathway in PCa.

Materials and Methods: We analyzed transcriptomic and clinical data of 94 early-onset (age <55) patients with prostate cancer using public dataset. Differentially expressed genes linked to the ubiquitin-proteasome system were identified across cancer progression stages. Kaplan–Meier survival analysis, Cox regression, and least absolute shrinkage and selection operator (LASSO) modeling were applied to assess their prognostic potential.

Results: Differential expression of IKBKB, UBQLN3, TMUB2, UBE2S, and BRCA1 was observed at relative-early stages of pT3a and Gleason 3+4. Similarly, HERPUD1, CDC20, UHRF1, PSMD7, PIAS3, MALT1, TNF, UBD, CD3E, CD247, SOCS1, UBE2C, CARD16, ZAP70, UBA7, and UBE3C expression levels also changed at pT3b and Gleason 4+3. At metastatic stages (pT4 and Gleason ≥8) OASL expression was up-regulated, whereas that of DDB1, RPN1, UBE3B, UBE2H, PPIL2, WWP2, and CDH1 was down-regulated. In addition, higher expression of PSMD2, CDC20, NFKB1, and STIP1 or lower expression of HERPUD2, NEDD4, ANAPC16, LNX1, and HERPUD1 was associated with poor prognoses according to the Kaplan–Meier or receiver operating characteristic analyses for biochemical recurrence-free survival. A LASSO-Cox model identified six gene candidates including LNX1, PSMD2, SUMO4, UBE2C, UBR5, and UHRF1.

Conclusion: The identified gene subset provides novel prognostic insights into PCa progression and survival. These findings highlight potential biomarkers and therapeutic targets within the ubiquitin-proteasome pathway, offering new avenues for personalized treatment strategies.

Keywords:

Introduction

Prostate cancer (PCa) is the most diagnosed cancer in men worldwide and ranks as the second leading cause of death (1, 2). Prostate-specific antigen (PSA) is a traditional marker for the diagnosis of PCa, but its specificity is limited (3). Androgens and androgen receptor signaling promote PCa progression. Subsequently, androgen deprivation therapy has become the primary treatment for patients with PCa at different stages of the disease (4). However, patients with PCa frequently exhibit resistance to androgen deprivation therapy, a condition known as castration-resistant PCa (4). Therefore, the identification of biomarkers for the early detection of PCa is essential. However, little is known about candidate cell signaling pathway markers that can be used as biomarkers and potential pathways that may serve as therapeutic targets.

The ubiquitin-proteasome pathway is required for the targeted degradation of proteins in cells. The attachment of polyubiquitin tags to proteins initiates their degradation process (5, 6), which occurs within proteasome complexes (7), a primary mechanism by which cells regulate protein concentration and degrade misfolded proteins (8). The complexity of the ubiquitin system is reflected in the wide range of processes it regulates, including key steps in cell cycle, processing of external proteins presented by the major histocompatibility complex class I molecules (9), and the regulation of cell growth and proliferation (10). However, promising biomarkers or protein degradation associated with the ubiquitin-proteasome system remain elusive in PCa.

This study aimed to identify novel biomarkers and intracellular signaling pathways associated with the ubiquitin-proteasome system to understand cancer development and prognosis prediction and to develop advanced targeted therapies for PCa through protein degradation strategies. We performed comprehensive analyses using transcriptome and clinical data of patients with PCa obtained from a public database, leading to a ubiquitin-proteasome pathway signature linked to prognosis prediction in PCa.

Materials and Methods

Dataset. A dataset containing gene expression and clinical information from patients with PCa was used (11). Gene expression values in terms of reads per kilobase of exon per million mapped reads (RPKM) were normalized and subjected to subsequent analyses. Of the 292 patients with PCa, gene expression data was included for 94 early-onset (EO) patients (age <55) in the Prostate Cancer, German Cancer Research Center (Deutsches Krebsforschungszentrum: DKFZ) dataset (Table I). Representative results were validated using the datasets of the Prostate Adenocarcinoma, The Cancer Genome Atlas (TCGA), PanCancer Atlas (494 patients) and Prostate Adenocarcinoma, Firehose Legacy (500 patients). Genes of interest were annotated online using GOstat2.5 (12) and Database for Annotation, Visualization, and Integrated Discovery6.8 (13). Gene expression values were adjusted based on variances from the mean levels in patient samples. Subsequently, graphs were generated using appropriate applications (14, 15).

View this table:
Table I.

Characteristics of the patients with prostate cancer examined in this study.

Survival analysis. Correlations between gene expression and survival time were evaluated using the Cox proportionality hazard regression analysis with R4.3.3 as described (16). Kaplan–Meier analysis was performed to estimate the survival distribution of the subgroups using R4.3.3 as described (16). Hazard ratios (HRs) and confidence intervals (CIs) were calculated using the Cox regression model according to patient survival times and assessed to compare subgroups as described (17). Biochemical recurrence (BCR)-free survival (BCRFS) was defined as the time from the date of surgery for PCa to the date of recurrence or last follow-up. Subgroups were divided by the median expression of genes of interest or median risk scores. The prognostic model genes were confirmed, and the risk scores were imputed as follows: Embedded Image

where ki,coef indicates the Cox regression coefficient, i indicates prognostic gene candidates, and mi,RPKM indicates the gene expression level as RPKM.

Graphical least absolute shrinkage and selection operator (LASSO) network analysis. Genetic interactions within hub networks among variables from gene expression were analyzed using the graphical LASSO estimation of Gaussian graphical models, such as a sparse inverse covariance matrix using a LASSO (L1) penalty, using the glasso package in R4.3.3 (18).

Statistical analysis. Statistical analyses were performed using R4.3.3. Statistical significance was set at p<0.05.

Results

Differential expression of the ubiquitin-proteasome pathway-related genes during cancer progression at relatively early stages. We used the transcriptome data and clinical information of 94 EO patients of 292 patients with PCa (Table I). The detailed expression profiles of 248 ubiquitin-proteasome system-related genes (Figure 1) (19-22) were classified based on TNM classification, Gleason score, and biochemical recurrence-free survival. Initially, the expression of the 248 ubiquitin-proteasome system-related genes was investigated during cancer progression and metastatic stages. First, differentially expressed genes (DEGs) were observed at a relatively early stage pT3a or Gleason score 3+4 compared to pT2a and Gleason 3+3, the earliest stages in the dataset used (Figure 2A and B). IKBKB expression was up-regulated in patients with pT3a (1.34-fold, p=0.049) and Gleason score ≥8 (1.45-fold, p=0.003). UBQLN3 expression was up-regulated in patients with Gleason score 3+4 (6.0-fold, p=0.001). BRCA1 expression was up-regulated in patients with Gleason score 3+4 (1.24-fold, p=0.048). TMUB2 expression was down-regulated in cases with Gleason score 3+4 (0.87-fold, p=0.013), 4+3 (0.83-fold, p=0.008), and ≥8 (0.81-fold, p=0.029). UBE2S expression was down-regulated in patients with Gleason score 3+4 (0.68-fold, p=0.042). These genes were associated with “the cellular response to tumor necrosis factor” (GO: 0071356, p=0.028).

Figure 1.
Figure 1.

A panel of the ubiquitin-proteasome pathway-related genes examined in this study. Genes were referred from references (19-22).

Figure 2.
Figure 2.

Differential expression of ubiquitin-proteasome pathway-related genes at relatively early stages of prostate cancer in relation to Gleason and TNM classifications. (A) Summary of differential expression of genes at progression and metastatic stages. Venn’s diagrams are shown for increased (left) or decreased (right) genes at pT3a and Gleason 3+4. (B) Differential expression of the representative genes is shown in box-whisker plots. Fold changes of average expression levels are indicated, compared with Gleason 3+3 or pT2a. p-Values were calculated using the Steel–Dwass multiple comparison test. RPKM: Reads per kilo base of transcript per million mapped reads.

Second, DEGs at the relatively early stage pT3b or Gleason score 4+3 are summarized in Venn diagrams (Figure 3A). In particular, at the Gleason score 4+3 and pT3b stages, the up-regulated genes were CDC20 (Gleason 4+3: 4.43-fold, p=0.001; pT3b: 5.02-fold, p=0.001), UHRF1, (Gleason 4+3: 4.04-fold, p<0.001; pT3b: 3.21-fold, p=0.006), and PIAS3 (Gleason 4+3: 1.28-fold, p=0.001; pT3b: 1.39-fold, p=0.029) (Figure 3A and B). UBE2C and UBE3C were up-regulated in patients with Gleason score 4+3 (Figure 3A). In contrast, HERPUD1 was down-regulated in patients with Gleason score 4+3 and pT3b stages as follows: Gleason score 4+3, 0.56-fold; p=0.003; pT3b, 0.52-fold; p=0.001 (Figure 3A and B). In addition, PSMD7 and MALT1 were down-regulated at pT3b, and UBD and UBA7 were down-regulated at Gleason score 4+3, respectively (Figure 3A). Notably, TNF, CD3E, CD274, SOCS1, CARD16, and ZAP70 levels decreased in patients with Gleason score 4+3 but increased for pT4 (Figure 3A). TNF (Gleason score 4+3: 0.39-fold, p=0.001; pT4: 3.67-fold, p<0.001), CD3E (Gleason score 4+3: 0.44-fold, p=0.001; pT4: 4.02-fold, p=0.036), CD247 (Gleason score 4+3: 0.52-fold, p=0.006; pT4: 5.09-fold, p<0.001), SOCS1 (Gleason score 4+3: 0.34-fold, p=0.006; pT4: 1.51-fold, p=0.049), CARD16 (Gleason score 4+3: 0.61-fold, p=0.034; pT4: 3.69-fold, p=0.001), and ZAP70 (Gleason score 4+3: 0.53-fold, p=0.042; pT4: 3.79-fold, p=0.001) demonstrated reciprocal changes in expression (Figure 3B). These genes were associated with several processes as “T cell receptor signaling pathway” (KEGG: hsa04660, p=2.3×10−5), “initiation and activation of TCR by Lck and Fyn tyrosine kinases” (BioCarta: h_tcraPathway, p=2.4×10−4), “T cell receptor signaling pathway” (GO: 0050852, p=1.0×10−4), and “activation of Csk by cAMP-dependent protein kinase”, which inhibits signaling through the “T cell receptor” (BioCarta: h_cskPathway, p=1.0×10−3), “positive regulation of cysteine-type endopeptidase activity involved in the apoptotic process” (GO: 0043280, p=5.4×10−4), and “positive regulation of I-kappaB kinase/NF-kappaB signaling” (GO: 0043123, p=5.5×10−4).

Figure 3.
Figure 3.

Differential expression of ubiquitin-proteasome pathway-related genes across progression stages in prostate cancer, analyzed in relation to Gleason and TNM classifications. (A) Summary of differential expression of the genes at progression stages. Venn’s diagrams are shown for increased at pT3b and Gleason 4+3 (left), decreased at pT3b and Gleason 4+3 (center), and reciprocally changed as decreased at Gleason 4+3 but increased at pT4 (right). (B) Expression of the representative genes are shown in box-whisker plots. Fold changes of average expression levels are indicated, compared with Gleason 3+3 or pT2a. p-values are calculated with the Steel–Dwass multiple comparison test. RPKM: Reads per kilo base of transcript per million mapped reads.

Differential expression of the ubiquitin-proteasome pathway-related genes at metastatic stages. Next, DEGs at metastatic stage pT4 or Gleason score ≥8 were summarized in Venn’s diagrams (Figure 4A). At both the pT4 and Gleason score ≥8 stages, the up-regulated gene was OASL (Gleason ≥8: 3.48-fold, p=0.001; pT4: 4.80-fold, p=0.037) (Figure 4A and B). TNFRSF1B, HECW2, PSMC1, TRIM32, and UBE2C were up-regulated in patients with pT4, and ISG15, SUMO4, and SIAH1 were up-regulated in patients with Gleason score ≥8 (Figure 4A). Moreover, several genes were down-regulated; these include DDB1 (Gleason ≥8: 0.85-fold, p=0.017; pT4: 0.72-fold, p<0.001), RPN1 (Gleason ≥8: 0.76-fold, p=0.014; pT4: 0.71-fold, p=0.005), UBE3B (Gleason ≥8: 0.77-fold, p=0.001; pT4: 0.68-fold, p=0.001), UBE2H (Gleason ≥8: 0.80-fold, p=0.025; pT4: 0.69-fold, p=0.001), PPIL2 (Gleason ≥8: 0.74-fold, p=0.006; pT4: 0.59-fold, p=0.001), WWP2 (Gleason ≥8: 0.74-fold, p=0.023; pT4: 0.51-fold, p<0.001), and CDH1 (Gleason ≥8: 0.59-fold, p=0.003; pT4: 0.25-fold, p<0.001) (Figure 4A and B). Notably, several genes showed unique expression changes, such as down-regulation at Gleason 4+3 but up-regulation at pT4. These genes included SOCS1, TNF, CARD16, ZAP70, CD3E, and CD247 (Figure 3A, B, and Figure 4C), considered as potential prognostic gene marker candidates for predicting the progression of PCa to the metastasis stage. The ontology terms derived from these genes were enriched in “regulation of potassium ion (K+) transmembrane transporter activity” (GO: 1901016, p=4.2×10−5) and “negative regulation of sodium ion (Na+) transmembrane transporter activity” (GO: 2000650, p=1.0×10−4).

Figure 4.
Figure 4.

Differential expression of ubiquitin-proteasome pathway-related genes at metastatic stages of prostate cancer. (A) Summary of differential expression of the genes at metastatic stages. Venn’s diagrams are shown for increased (left) or decreased (right) genes. (B) Expression of the representative genes are shown in box-whisker plots. Fold changes of average expression levels are indicated, compared with Gleason 3+3 or pT2a. (C) Composite evaluation of cases with Gleason 4+3, pT3b, and pT4. Venn’s diagram is shown for genes decreased at Gleason 4+3 and increased at pT4. p-Values were calculated using the Steel–Dwass multiple comparison test. RPKM: Reads per kilo base of transcript per million mapped reads.

Single prognostic factor candidates of ubiquitin-proteasome pathway-related genes. In addition to differential expression analyses during cancer stages, we estimated BCRFS to examine single prognostic factor candidates with univariate Cox hazard regression and Kaplan–Meier survival analyses, which were re-evaluated using receiver operating characteristic (ROC) analyses. As shown in Venn’s diagrams (Figure 5A), PSMD2, CDC20, NFKB1, and STIP1 were identified as potential markers for poor prognosis, with HR >1.0 (p<0.05) and the lower limit of 95%CI of area under the curve (AUC) >0.5 (Figure 5B), while UHRF1, UBE2S, CBL1, CDC34, SUMO4, BRCA1, HECW2, UBE2C, UBFD1, and ANAPC11 were sufficient with HR >1.0 (p<0.05) (Figure 5A). In contrast, PARK2 was assessed as a positive prognostic factor with an HR <1.0, (p <0.05) and a lower 95%CI of AUC >0.5 (Figure 5A and B). Meanwhile, HERPUD2, NEDD4, ANAPC16, LNX1, and HERPUD1 demonstrated a lower 95%CI of AUC >0.5 (Figure 5A). Therefore, correlations between the expression levels of these genes may serve as prognostic indicators in PCa.

Figure 5.
Figure 5.

Candidate of prognostic factors of ubiquitin-proteasome pathway-related genes with univariate Cox hazard and receiver operating characteristic (ROC) analyses in prostate cancer. (A) Summary of prognostic gene candidates with Kaplan–Meier and ROC analyses. Venn’s diagrams show the genes associated with poor prognosis (left) or better prognosis (right). (B) (left panel) Biochemical recurrence (BCR)-free survival rates were estimated using univariate Cox hazard analysis. High and low indicate the subgroups with higher and lower expression levels, respectively, compared to the median expression level of the gene. HR: Hazard ratio; CI: confidential interval. (right panel) ROC analysis for BCR-free survival times. Area under the curve (AUC) is calculated.

Approximate composite classification using expression correlation and graphical LASSO. Furthermore, we performed Spearman’s rank correlation coefficient test and graphical LASSO analysis to investigate co-expression patterns of 39 of the ubiquitin-proteasome pathway-related genes with differential expression in specific stages or survivals derived from the results above. The Spearman’s rank correlation coefficient test was constituted of a matrix with two clusters (Figure 6A). This result indicates a combination of reciprocal gene expressions and suggests a possibility of predicting the expression patterns of other genes from some gene expression patterns.

Figure 6.
Figure 6.

Correlation coefficient and graphical LASSO model using representative genes related to the ubiquitin-proteasome pathway in prostate cancer. (A) Co-expression patterns of the representative genes involved in ubiquitin-proteasome pathway. Color configurations represent contributions with Spearman’s rank correlation coefficient as direct (red) to inverse (blue). (B) A ubiquitin-proteasome pathway-related gene network generated from the graphical LASSO model based on the co-expression patterns of the genes. Thick and thin lines represent relatively strong and weak interaction, respectively, with direct (green) or inverse correlation (red). The numbers in the parentheses indicate the numbers of edges of the nodes. Highlighted circles represent candidates in the hubs harboring edge numbers ≥4 and edge weight >0.15.

However, the graphical LASSO model constituted a dense hub network containing the genes ANAPC16, CBL, DDB1, NFKB1, SF3A1, UBA1, UBE3C, UBR5, BRCA1, LNX1, NAE1, PSMD2, RNF4, and TCEB1 (edge number ≥4, edge weight >0.15) (Figure 6B). Considering the gene correlations with |edge weight| ≥0.2, positive correlations were observed for ANAPC16-NAE1 (edge weight=0.28), ANAPC16-TCEB1 (0.28), CBL-NFKB1 (0.28), CDC20-UBE2C (0.58), DDB1-UBA1 (0.32), RAD23B-UBFD1 (0.3), RAD23B-TCEB1 (0.29), STUB1-UBE2S (0.27), UBE2C-UHRF1 (0.24), EDARADD-NFKB1 (0.22), CDC20-UHRF1 (0.22), HERPUD1-NAE1 (0.21), SF3A1-UBE3C (0.21), RNF4-UBFD1 (0.2), and HECW2-UBR5 (0.2) (Figure 5B). Similarly, negative correlations were observed for ANAPC16-CBL (−0.23), HERPUD2-STIP1 (−0.26), STUB1-UBR5 (−0.27), TCEB1-UBA1 (−0.23), BRCA1-UBTD1 (−0.2), SF3A1-UBR5 (−0.2), and SF3A1-TCEB1 (−0.2) (Figure 6B). Therefore, part of the gene network suggests the possibility of conflicting biological functions within the protein degradation pathway in PCa.

Construction of the LASSO-Cox prognostic prediction model with ubiquitin-proteasome pathway-related genes. The survival time data contained in the datasets used were different in each dataset. The DKFZ included BCRFS only, the TCGA Firehose Legacy included disease-free survival (DFS) only, and the TCGA PanCancer Atlas included DFS, progression-free survival (PFS), and overall survival (OS). Therefore, based on the information contained in the datasets, we conducted survival time analyses for BCRFS, DFS, PFS, and OS for EO, all cases, and specific stages as T3 and T4. The 39 genes were associated with differential expression, single prognostic marker potential, or co-expression correlations. Therefore, the gene subset was further subjected to reduce the number of sparse groups in the LASSO model to construct a prognosis prediction formula, resulting in 6 genes, UHRF1, SUMO4, UBR5, PSMD2, UBE2C, and LNX1 which underwent multivariate Cox hazard regression analysis (Figure 7A left). Finally, using coefficient values (p<0.1) in the Cox model, a prognosis prediction formula was obtained as follows: Risk score=−8.4×10−1 LNX1 + 4.3×10−1 SUMO4 + 6.2×10−1 UHRF1 (Figure 7A right). The subgroup with risk scores higher than the median exhibited shorter BCRFS times than the subgroup with lower scores (HR=5.54, 95%CI=1.59-19.31, p=0.007) (n=94) (Figure 7A right). An alternative risk score model, using 1.2×10−1 SUMO4 + 3.5×10−4 PSMD2 + 2.7×10−3 UBE2C, estimated an HR=2.43 (95%Cl=1.56-3.77) with p<0.001 for DFS in an independent dataset of prostate adenocarcinoma (TCGA, Firehose Legacy) (n=491) (Figure 7B). These results suggested that the three genes identified in the LASSO-Cox model could be effective in predicting survival under certain conditions of PCa progression. Furthermore, in a separate cohort of prostate adenocarcinoma (TCGA, PanCancer Atlas), the combined expression of SUMO4 and UBR5 was associated with significant HRs in DFS across all samples (n=186; HR=2.39, 95%CI=1.10-5.20, p=0.028) (Figure 7B). Similarly, the use of UBR5 and UBE2C as markers resulted in significant HRs in DFS for the samples with T3 and T4 (n=88; HR=3.55, 95%CI=1.16-10.82, p=0.026) (Figure 7B). Furthermore, the use of SUMO4 and PSMD2 resulted in significant HRs in OS for the samples with T3 and T4 (n=251; HR=1.57, 95%CI=1.11-2.23, p=0.011) (Figure 7B). Finally, the combination of UBR5 and PSMD2 resulted in significant HRs in PFS for all samples (n=407; HR=1.65, 95%CI=1.22-2.24, p=0.001) (Figure 7B). Thus, selection of appropriate genes from LNX1, PSMD2, SUMO4, UBE2C, UBR5, and UHRF1 derived from the LASSO model enabled the prediction of survival outcomes in a limited subgroup of patients with PCa (Figure 7C).

Figure 7.
Figure 7.
Figure 7.

Alternative prognostic prediction using the LASSO-Cox model. (A) The 6 genes as UHRF1, SUMO4, UBR5, PSMD2, UBE2C, and LNX1 were selected with the LASSO in the early-onset (EO) patients in the Prostate Adenocarcinoma, dataset (n=94, DKFZ) (left panel). The prognosis prediction (risk score) formula developed based on the LASSO-Cox model with LNX1, SUMO4, and UHRF1 (p<0.05) for BCR-free survival (BCRFS). The Kaplan–Meier survival curves are shown (right panel). (B) Alternative model of the prognosis prediction formula with SUMO4, PSMD2, and UBE2C from the LASSO-Cox model for disease-free survival (DFS) in all samples (n=491) in the Prostate Adenocarcinoma (TCGA, Firehose Legacy) dataset. Similarly, a model of the prognosis prediction formula with SUMO4 and UBR5 derived from the LASSO-Cox model for DFS in all samples (n=186), the formula with UBR5 and UBE2C for DFS in the samples with tumor stages T3 and T4 (n=88), the formula with SUMO4 and PSMD2 for overall survival (OS) in the samples with tumor stages T3 and T4 (n=251), and the formula with UBR5 and PSMD2 for progression-free survival (PFS) in all samples (n=407). The Kaplan–Meier survival curves are shown. (C) Summary of selections of candidate genes and target survival times. HR: Hazard ratio; CI: confidence interval. High and low subgroups were divided by the median risk scores from the models.

Discussion

In this study, we identified differentially expressed genes (DEGs) such as IKBKB, NFKB1, CDC20, CD3E, CD247 (PD-L1), CDH1 (E-cadherin), and BRCA1 as potential upstream regulators involved in the ubiquitination process during PCa progression and metastasis. Additionally, CUL4A, DDB1, VCP, NEDD4L, WWP2, and SIAH1 were found to be DEGs implicated in catabolic processes associated with PCa. Notably, NEDD4 and NEDD4L positively regulate the K+/Na+ transport system, whereas WWP2 antagonizes NEDD4L, leading to destabilization of ion transport. Importantly, WWP2, NEDD4L, and NEDD4 encode E3 ligases containing a homologous to E6-AP carboxyl terminus domain, suggesting their potential involvement in protein catabolism and critical roles in cancer development and metastasis (29). For metastasis, TNF, CD3E, CD274 (PD-L1), SOCS1, CARD16, and ZAP70 were up-regulated at the pT4 stage after down-regulation at Gleason score 4+3, suggesting that these genes may become markedly up-regulated at the initiation of the metastatic process, playing crucial roles as T-cell costimulatory signaling, essential for T-cell proliferation and cytokine productions including interleukin-10 and interferon-γ. Additional multivariate gene network analysis identified 14 genes, including ANAPC16, CBL, DDB1, NFKB1, SF3A1, UBA1, UBE3C, UBR5, BRCA1, LNX1, NAE1, PSMD2, RNF4, and TCEB1. Of these, BRCA1 and NFKB1 were detected as both DEGs and biomarker candidates. In addition, based on the LASSO-Cox model, we proposed that the combined expression of UHRF1, LNX1, SUMO4, UBR5, PSMD2, and UBE2C could be a promising gene signature candidate for the prognosis prediction of survival time under specific conditions in PCa.

During the late cancer progression and metastatic stages, p53 activation and NF-κB inhibition induce apoptosis (30-32), while ubiquitin activation of anaphase-promoting complex/cyclosome (APC/C) facilitates the progression from the M phase to the G1 phase by cleaving chromosomal cohesin at specific stages of progression (33-35). In these signaling pathways, δ-catenin serves as a key upstream factor (36). However, E-cadherin reduction and APC/C suppression might arrest the M phase of the cell cycle at metastatic stages (37-39), considering that intracellular reconstitution and plasticity may be required for invasion, metastasis, and epithelial-mesenchymal transition. Differential expression and survival predictor candidates may also be closely implicated in tumor growth and molecular pathways associated with apoptosis and cell cycle facilitated by protein degradation via the ubiquitin-proteasome system (Figure 8A and B).

Figure 8.
Figure 8.

A model for the ubiquitin-proteasome pathway associated with apoptosis and cell cycle regulation in prostate cancer. (A) A ubiquitin regulation model for intracellular signaling pathways. (B) A ubiquitin regulation model for cell cycle control in prostate cancer cells. Thick arrows: activated; thick T signs: inhibited. Gxx and pTxx indicate Gleason and TNM classifications, respectively, with increased expression (upward arrow) and decreased expression (down arrow). Poor and better: poor and better prognosis with higher expression of the gene, respectively.

The ubiquitin-proteasome pathway has substantial potential for the treatment of tumors, including prostate adenocarcinomas. Focused on the ubiquitin-proteasome pathway, we identified prognostic gene signature candidates, such as UHRF1, LNX1, SUMO4, UBR5, PSMD2, and UBE2C. Of these, SUMO4, PSMD2, UBE2C, and UBR5 contributed to the estimation of DFS, PFS, and OS in certain combinations, whereas LNX1 and UHRF1 may assist in predicting BCRFS coupled with SUMO4. However, the underlying mechanism of the ubiquitin-proteasome pathway and relevant phenomena are not well understood. Instead of the traditional fine marker, PSA, it may be important to predict prognosis and tumor cell proliferation by analyzing the gene expression in blood-circulating tumor cells. As the ubiquitin-proteasome pathway may reveal unexplored mechanisms of cancer proliferation and metastasis, understanding the detailed mechanisms by which the ubiquitin-proteasome system is regulated by intra/extracellular signaling in tumors is crucial and the gene subset candidates detected in this study require further studies. Furthermore, additional cohort studies should also be performed in the future.

Acknowledgements

The Authors would like to thank The German Cancer Research Center and the National Institutes of Health.

Footnotes

  • Authors’ Contributions

    Y.T.: designed the study. Y.T., K.Y., and M.T.: performed the experiments. Y.T., K.Y., and M.T.: analyzed the data. Y.T., K.Y., M.T. and K.T.: wrote the manuscript. All Authors read and approved of the final manuscript.

  • Conflicts of Interest

    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.

  • Funding

    This study was supported in part by MEXT/JSPS KAKENHI (23K08528) to Y.T. The funder had no role in the study design, data collection and analysis, decision to publish, or preparation of the manuscript.

  • Received March 3, 2025.
  • Revision received March 20, 2025.
  • Accepted March 21, 2025.

This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY-NC-ND) 4.0 international license (https://creativecommons.org/licenses/by-nc-nd/4.0).

References

  1. ↵
    1. Torre LA,
    2. Bray F,
    3. Siegel RL,
    4. Ferlay J,
    5. Lortet-Tieulent J,
    6. Jemal A
    : Global cancer statistics, 2012. CA Cancer J Clin 65(2): 87-108, 2015. DOI: 10.3322/caac.21262
    OpenUrlCrossRefPubMed
  2. ↵
    1. Calcinotto A,
    2. Spataro C,
    3. Zagato E,
    4. Di Mitri D,
    5. Gil V,
    6. Crespo M,
    7. De Bernardis G,
    8. Losa M,
    9. Mirenda M,
    10. Pasquini E,
    11. Rinaldi A,
    12. Sumanasuriya S,
    13. Lambros MB,
    14. Neeb A,
    15. Lucianò R,
    16. Bravi CA,
    17. Nava-Rodrigues D,
    18. Dolling D,
    19. Prayer-Galetti T,
    20. Ferreira A,
    21. Briganti A,
    22. Esposito A,
    23. Barry S,
    24. Yuan W,
    25. Sharp A,
    26. de Bono J,
    27. Alimonti A
    : IL-23 secreted by myeloid cells drives castration-resistant prostate cancer. Nature 559(7714): 363-369, 2018. DOI: 10.1038/s41586-018-0266-0
    OpenUrlCrossRefPubMed
  3. ↵
    1. Alexander EE,
    2. Qian J,
    3. Wollan PC,
    4. Myers RP,
    5. Bostwick DG
    : Prostatic intraepithelial neoplasia does not appear to raise serum prostate-specific antigen concentration. Urology 47(5): 693-698, 1996. DOI: 10.1016/s0090-4295(96)00004-0
    OpenUrlCrossRefPubMed
  4. ↵
    1. Watson PA,
    2. Arora VK,
    3. Sawyers CL
    : Emerging mechanisms of resistance to androgen receptor inhibitors in prostate cancer. Nat Rev Cancer 15(12): 701-711, 2015. DOI: 10.1038/nrc4016
    OpenUrlCrossRefPubMed
  5. ↵
    1. Connor MK,
    2. Seth A
    : A central role for the ring finger protein RNF11 in ubiquitin-mediated proteolysis via interactions with E2s and E3s. Oncogene 23(11): 2089-2095, 2004. DOI: 10.1038/sj.onc.1207380
    OpenUrlCrossRefPubMed
  6. ↵
    1. Adams J,
    2. Kauffman M
    : Development of the proteasome inhibitor Velcade™ (Bortezomib). Cancer Invest 22(2): 304-311, 2004. DOI: 10.1081/cnv-120030218
    OpenUrlCrossRefPubMed
  7. ↵
    1. Hochstrasser M
    : Ubiquitin, proteasomes, and the regulation of intracellular protein degradation. Curr Opin Cell Biol 7(2): 215-223, 1995. DOI: 10.1016/0955-0674(95)80031-x
    OpenUrlCrossRefPubMed
  8. ↵
    1. Sengupta S,
    2. Wasylyk B
    : Physiological and pathological consequences of the interactions of the p53 tumor suppressor with the glucocorticoid, androgen, and estrogen receptors. Ann N Y Acad Sci 1024(1): 54-71, 2004. DOI: 10.1196/annals.1321.005
    OpenUrlCrossRefPubMed
  9. ↵
    1. Gaczynska M,
    2. Rock KL,
    3. Goldberg AL
    : Role of proteasomes in antigen presentation. Enzyme Protein 47(4-6): 354-369, 1993. DOI: 10.1159/000468693
    OpenUrlCrossRefPubMed
  10. ↵
    1. Pramanik J,
    2. Chen X,
    3. Kar G,
    4. Henriksson J,
    5. Gomes T,
    6. Park JE,
    7. Natarajan K,
    8. Meyer KB,
    9. Miao Z,
    10. McKenzie ANJ,
    11. Mahata B,
    12. Teichmann SA
    : Genome-wide analyses reveal the IRE1a-XBP1 pathway promotes T helper cell differentiation by resolving secretory stress and accelerating proliferation. Genome Med 10(1): 76, 2018. DOI: 10.1186/s13073-018-0589-3
    OpenUrlCrossRefPubMed
  11. ↵
    1. Gerhauser C,
    2. Favero F,
    3. Risch T,
    4. Simon R,
    5. Feuerbach L,
    6. Assenov Y,
    7. Heckmann D,
    8. Sidiropoulos N,
    9. Waszak SM,
    10. Hübschmann D,
    11. Urbanucci A,
    12. Girma EG,
    13. Kuryshev V,
    14. Klimczak LJ,
    15. Saini N,
    16. Stütz AM,
    17. Weichenhan D,
    18. Böttcher LM,
    19. Toth R,
    20. Hendriksen JD,
    21. Koop C,
    22. Lutsik P,
    23. Matzk S,
    24. Warnatz HJ,
    25. Amstislavskiy V,
    26. Feuerstein C,
    27. Raeder B,
    28. Bogatyrova O,
    29. Schmitz EM,
    30. Hube-Magg C,
    31. Kluth M,
    32. Huland H,
    33. Graefen M,
    34. Lawerenz C,
    35. Henry GH,
    36. Yamaguchi TN,
    37. Malewska A,
    38. Meiners J,
    39. Schilling D,
    40. Reisinger E,
    41. Eils R,
    42. Schlesner M,
    43. Strand DW,
    44. Bristow RG,
    45. Boutros PC,
    46. von Kalle C,
    47. Gordenin D,
    48. Sültmann H,
    49. Brors B,
    50. Sauter G,
    51. Plass C,
    52. Yaspo ML,
    53. Korbel JO,
    54. Schlomm T,
    55. Weischenfeldt J
    : Molecular evolution of early-onset prostate cancer identifies molecular risk markers and clinical trajectories. Cancer Cell 34(6): 996-1011.e8, 2018. DOI: 10.1016/j.ccell.2018.10.016
    OpenUrlCrossRefPubMed
  12. ↵
    1. Ashburner M,
    2. Ball CA,
    3. Blake JA,
    4. Botstein D,
    5. Butler H,
    6. Cherry JM,
    7. Davis AP,
    8. Dolinski K,
    9. Dwight SS,
    10. Eppig JT,
    11. Harris MA,
    12. Hill DP,
    13. Issel-Tarver L,
    14. Kasarskis A,
    15. Lewis S,
    16. Matese JC,
    17. Richardson JE,
    18. Ringwald M,
    19. Rubin GM,
    20. Sherlock G
    : Gene ontology: tool for the unification of biology. The Gene Ontology Consortium. Nat Genet 25(1): 25-29, 2000. DOI: 10.1038/75556
    OpenUrlCrossRefPubMed
  13. ↵
    1. Dennis G Jr.,
    2. Sherman BT,
    3. Hosack DA,
    4. Yang J,
    5. Gao W,
    6. Lane HC,
    7. Lempicki RA
    : DAVID: Database for annotation, visualization, and integrated discovery. Genome Biol 4(5): P3, 2003.
    OpenUrlCrossRefPubMed
  14. ↵
    1. Takashima Y,
    2. Hamano M,
    3. Yoshii K,
    4. Hayano A,
    5. Fukai J,
    6. Iwadate Y,
    7. Kajiwara K,
    8. Hondoh H,
    9. Yamanaka R
    : Reciprocal expression of the immune response genes CXCR3 and IFI44L as module hubs are associated with patient survivals in primary central nervous system lymphoma. Int J Clin Oncol 28(3): 468-481, 2023. DOI: 10.1007/s10147-022-02285-8
    OpenUrlCrossRefPubMed
  15. ↵
    1. Takashima Y,
    2. Sasaki Y,
    3. Hayano A,
    4. Homma J,
    5. Fukai J,
    6. Iwadate Y,
    7. Kajiwara K,
    8. Ishizawa S,
    9. Hondoh H,
    10. Tokino T,
    11. Yamanaka R
    : Target amplicon exome-sequencing identifies promising diagnosis and prognostic markers involved in RTK-RAS and PI3K-AKT signaling as central oncopathways in primary central nervous system lymphoma. Oncotarget 9(44): 27471-27486, 2018. DOI: 10.18632/oncotarget.25463
    OpenUrlCrossRefPubMed
  16. ↵
    1. Takashima Y,
    2. Kawaguchi A,
    3. Sato R,
    4. Yoshida K,
    5. Hayano A,
    6. Homma J,
    7. Fukai J,
    8. Iwadate Y,
    9. Kajiwara K,
    10. Ishizawa S,
    11. Hondoh H,
    12. Nakano M,
    13. Ogawa S,
    14. Tashiro K,
    15. Yamanaka R
    : Differential expression of individual transcript variants of PD-1 and PD-L2 genes on Th-1/Th-2 status is guaranteed for prognosis prediction in PCNSL. Sci Rep 9(1): 10004, 2019. DOI: 10.1038/s41598-019-46473-5
    OpenUrlCrossRefPubMed
  17. ↵
    1. Takashima Y,
    2. Kawaguchi A,
    3. Hayano A,
    4. Yamanaka R
    : CD276 and the gene signature composed of GATA3 and LGALS3 enable prognosis prediction of glioblastoma multiforme. PLoS One 14(5): e0216825, 2019. DOI: 10.1371/journal.pone.0216825
    OpenUrlCrossRefPubMed
  18. ↵
    1. Takashima Y,
    2. Kawaguchi A,
    3. Fukai J,
    4. Iwadate Y,
    5. Kajiwara K,
    6. Hondoh H,
    7. Yamanaka R
    : Survival prediction based on the gene expression associated with cancer morphology and microenvironment in primary central nervous system lymphoma. PLoS One 16(6): e0251272, 2021. DOI: 10.1371/journal.pone.0251272
    OpenUrlCrossRefPubMed
  19. ↵
    1. Cermakova K,
    2. Hodges HC
    : Next-generation drugs and probes for chromatin biology: from targeted protein degradation to phase separation. Molecules 23(8): 1958, 2018. DOI: 10.3390/molecules23081958
    OpenUrlCrossRef
    1. Ciechanover A
    : The ubiquitin-proteasome pathway: on protein death and cell life. EMBO J 17(24): 7151-7160, 1998. DOI: 10.1093/emboj/17.24.7151
    OpenUrlFREE Full Text
    1. Nalepa G,
    2. Rolfe M,
    3. Harper JW
    : Drug discovery in the ubiquitin–proteasome system. Nat Rev Drug Discov 5(7): 596-613, 2006. DOI: 10.1038/nrd2056
    OpenUrlCrossRefPubMed
  22. ↵
    1. Li H,
    2. Dong J,
    3. Cai M,
    4. Xu Z,
    5. Cheng XD,
    6. Qin JJ
    : Protein degradation technology: a strategic paradigm shift in drug discovery. J Hematol Oncol 14(1): 138, 2021. DOI: 10.1186/s13045-021-01146-7
    OpenUrlCrossRef
    1. Lu X,
    2. Guo J,
    3. Hsieh TC
    : PC-SPES inhibits cell proliferation by modulating p21, cyclins D, E and B and multiple cell cycle-related genes in prostate cancer cells. Cell Cycle 2(1): 59-63, 2003.
    OpenUrlPubMed
    1. Wang J,
    2. Chen Y,
    3. Li S,
    4. Liu W,
    5. Zhou XA,
    6. Luo Y,
    7. Xu Z,
    8. Xiong Y,
    9. Cheng K,
    10. Ruan M,
    11. Yu W,
    12. Li X,
    13. Wang W,
    14. Wang J
    : PP2A inhibition causes synthetic lethality in BRCA2-mutated prostate cancer models via spindle assembly checkpoint reactivation. J Clin Invest 134(1): e172137, 2024. DOI: 10.1172/JCI172137
    OpenUrlCrossRefPubMed
    1. Feng R,
    2. Li Z,
    3. Ge G,
    4. Wang C,
    5. Jia Y,
    6. Ouyang J
    : NEDD4L represses prostate cancer cell proliferation via modulating PHF8 through the ubiquitin–proteasome pathway. Clin Transl Oncol 25(1): 243-255, 2022. DOI: 10.1007/s12094-022-02933-5
    OpenUrlCrossRefPubMed
    1. Soond SM,
    2. Smith PG,
    3. Wahl L,
    4. Swingler TE,
    5. Clark IM,
    6. Hemmings AM,
    7. Chantry A
    : Novel WWP2 ubiquitin ligase isoforms as potential prognostic markers and molecular targets in cancer. Biochim Biophys Acta 1832(12): 2127-2135, 2013. DOI: 10.1016/j.bbadis.2013.08.001
    OpenUrlCrossRef
    1. Feng Y,
    2. Sessions EH,
    3. Zhang F,
    4. Ban F,
    5. Placencio-Hickok V,
    6. Ma CT,
    7. Zeng FY,
    8. Pass I,
    9. Terry DB,
    10. Cadwell G,
    11. Bankston LA,
    12. Liddington RC,
    13. Chung TDY,
    14. Pinkerton AB,
    15. Sergienko E,
    16. Gleave M,
    17. Bhowmick NA,
    18. Jackson MR,
    19. Cherkasov A,
    20. Ronai ZA
    : Identification and characterization of small molecule inhibitors of the ubiquitin ligases Siah1/2 in melanoma and prostate cancer cells. Cancer Lett 449: 145-162, 2019. DOI: 10.1016/j.canlet.2019.02.012
    OpenUrlCrossRefPubMed
    1. Rougier JS,
    2. van Bemmelen MX,
    3. Bruce MC,
    4. Jespersen T,
    5. Gavillet B,
    6. Apothéloz F,
    7. Cordonier S,
    8. Staub O,
    9. Rotin D,
    10. Abriel H
    : Molecular determinants of voltage-gated sodium channel regulation by the Nedd4/Nedd4-like proteins. Am J Physiol Cell Physiol 288(3): C692-701, 2005. DOI: 10.1152/ajpcell.00460.2004
    OpenUrlCrossRefPubMed
  29. ↵
    1. Chen C,
    2. Matesic LE
    : The Nedd4-like family of E3 ubiquitin ligases and cancer. Cancer Metastasis Rev 26(3-4): 587-604, 2007. DOI: 10.1007/s10555-007-9091-x
    OpenUrlCrossRefPubMed
  30. ↵
    1. Park R,
    2. Kim H,
    3. Jang M,
    4. Jo D,
    5. Park YI,
    6. Namkoong S,
    7. Lee JI,
    8. Jang IS,
    9. Park J
    : LNX1 contributes to tumor growth by down-regulating p53 stability. FASEB J 33(12): 13216-13227, 2019. DOI: 10.1096/fj.201900366R
    OpenUrlCrossRefPubMed
    1. Karin M,
    2. Ben-Neriah Y
    : Phosphorylation meets ubiquitination: the control of NF-κB activity. Annu Rev Immunol 18(1): 621-663, 2000. DOI: 10.1146/annurev.immunol.18.1.621
    OpenUrlCrossRefPubMed
  32. ↵
    1. Sun J,
    2. Liu P,
    3. Mao J,
    4. Han W,
    5. Jiang X,
    6. Gu Y
    : TRIM52 positively mediates NF-κB to promote the growth of human benign prostatic hyperplasia cells through affecting TRAF2 ubiquitination. Life Sci 259: 118380, 2020. DOI: 10.1016/j.lfs.2020.118380
    OpenUrlCrossRefPubMed
  33. ↵
    1. Visintin R,
    2. Prinz S,
    3. Amon A
    : CDC20 and CDH1: a family of substrate-specific activators of APC-dependent proteolysis. Science 278(5337): 460-463, 1997. DOI: 10.1126/science.278.5337.460
    OpenUrlAbstract/FREE Full Text
    1. Kops GJ,
    2. van der Voet M,
    3. Manak MS,
    4. van Osch MH,
    5. Naini SM,
    6. Brear A,
    7. McLeod IX,
    8. Hentschel DM,
    9. Yates JR 3rd.,
    10. van den Heuvel S,
    11. Shah JV
    : APC16 is a conserved subunit of the anaphase-promoting complex/cyclosome. J Cell Sci 123(Pt 10): 1623-1633, 2010. DOI: 10.1242/jcs.061549
    OpenUrlAbstract/FREE Full Text
  35. ↵
    1. Wild T,
    2. Budzowska M,
    3. Hellmuth S,
    4. Eibes S,
    5. Karemore G,
    6. Barisic M,
    7. Stemmann O,
    8. Choudhary C
    : Deletion of APC7 or APC16 allows proliferation of human cells without the spindle assembly checkpoint. Cell Rep 25(9): 2317-2328.e5, 2018. DOI: 10.1016/j.celrep.2018.10.104
    OpenUrlCrossRefPubMed
  36. ↵
    1. Zeng Y,
    2. Abdallah A,
    3. Lu JP,
    4. Wang T,
    5. Chen YH,
    6. Terrian DM,
    7. Kim K,
    8. Lu Q
    : delta-Catenin promotes prostate cancer cell growth and progression by altering cell cycle and survival gene profiles. Mol Cancer 8: 19, 2009. DOI: 10.1186/1476-4598-8-19
    OpenUrlCrossRefPubMed
  37. ↵
    1. Semb H,
    2. Christofori G
    : The tumor-suppressor function of E-cadherin. Am J Hum Genet 63(6): 1588-1593, 1998. DOI: 10.1086/302173
    OpenUrlCrossRefPubMed
    1. Umbas R,
    2. Schalken JA,
    3. Aalders TW,
    4. Carter BS,
    5. Karthaus HF,
    6. Schaafsma HE,
    7. Debruyne FM,
    8. Isaacs WB
    : Expression of the cellular adhesion molecule E-cadherin is reduced or absent in high-grade prostate cancer. Cancer Res 52(18): 5104-5109, 1992.
    OpenUrlAbstract/FREE Full Text
  39. ↵
    1. Giroldi LA,
    2. Schalken JA
    : Decreased expression of the intercellular adhesion molecule E-cadherin in prostate cancer: Biological significance and clinical implications. Cancer Metastasis Rev 12(1): 29-37, 1993. DOI: 10.1007/BF00689788
    OpenUrlCrossRefPubMed
Back to top

In this issue

Print
Download PDF
Article Alerts
Email Article
Citation Tools
Reprints and Permissions
Ubiquitin-proteasome Pathway-linked Gene Signatures as Prognostic Indicators in Prostate Cancer
YASUO TAKASHIMA, KENGO YOSHII, MASAMI TANAKA, KEI TASHIRO
Anticancer Research May 2025, 45 (5) 1825-1841; DOI: 10.21873/anticanres.17562
Twitter logo Facebook logo Mendeley logo

Jump to section

Keywords