The LNCaP, 22RV1, DU145, and PC3 cell lines were generously provided by the Stem Cell Bank, Chinese Academy of Sciences (Shanghai, China). These cell lines were cultured in RPMI 1640 medium (L210KJ; BasalMedia, Shanghai, China) supplemented with 10% fetal bovine serum (S660JJ; BasalMedia), 1% penicillin/streptomycin (15070063; Gibco, Grand Island, NY, United States), and 1% HEPES (15630080; Gibco). The prostate normal cell lines BPH-1 and RWPE-1 were purchased from the Stem Cell Bank, Chinese Academy of Sciences and cultured in keratinocyte complete medium (C120JV; BasalMedia). The method used to establish the LNCaP_AI cell line was described in our previous study (Zhou et al., 2024). All cells were maintained at 37°C in a humidified incubator with 5% CO₂.

The short hairpin RNA (shRNA) sequences for each gene are detailed in Table 1. All shRNA sequences were cloned into the pLKO.1 vector. Lentiviruses targeting each gene were produced by transfecting HEK293T cells with the shRNA constructs, psPAX2 (#12260; Addgene, Watertown, MA, United States), and pCMV-VSVG (#8454; Addgene) using PEI 40K (G1802; Servicebio, Shanghai, China). For AR overexpression, full-length AR cDNA with a Flag tag was cloned into the pLVX-IRES-Puro vector between the XhoI and BamHI sites. Lentivirus-containing supernatants were collected at 48 and 72 h after transfection of 293T cells and used to infect 22RV1 and PC3 cells. Puromycin (5 μg/mL; Sigma-Aldrich, St. Louis, MO, United States) was used to select stable transformants.

Cell growth evaluation was conducted using the MTT assay system. Cells (2,000 per well) were seeded into 96-well plates at 100 μL/well and incubated at 37°C under 5% CO₂ until cell attachment (typically 24–48 h). Subsequently, 20 μL of MTT solution (final concentration of 0.5 mg/mL) was added to each well. The plate was gently mixed and incubated for 2–4 h to allow viable cells to reduce MTT to purple formazan crystals via mitochondrial dehydrogenases. After incubation, the medium was carefully aspirated without disturbing the cell layer or crystals, and 200 μL of solubilization solution (dimethyl sulfoxide) was added to dissolve the formazan crystals fully by shaking for 15 min. Finally, the optical density of each well was measured at 490 nm using a microplate reader. For the colony-formation assay, 22RV1 (1,000 cells/well) and PC3 cells (500–1,000 cells/well) were seeded in six-well plates with complete medium and cultured for 10–14 days, depending on colony size. Cells were then fixed in methanol for 10 min and stained with 0.1% crystal violet for 1 h.

Cells were gently washed three times with phosphate-buffered saline and lysed in lysis buffer. Protein samples were separated by sodium dodecyl sulfate–polyacrylamide gel electrophoresis and then transferred to a nitrocellulose membrane (10600001; Amersham, Marlborough, MA, United States). The membranes were blocked with 5% bovine serum albumin in Tris-buffered saline with Tween for three washes, then incubated with primary antibodies at 4°C overnight. Subsequently, the membranes were incubated with secondary antibodies corresponding to the primary antibodies at room temperature for 1 h and washed three times in Tris-buffered saline with Tween. Signal detection was performed using the Tanon Imaging System (Tanon-5200; Tanon, Shanghai, China). Antibodies used in the Western blotting assay were as follows: β-tubulin (SB-AB2002; ShareBio, Shanghai, China), AR (SC-56824; Santa Cruz Biotechnology, Dallas, TX, United States), USP42 (A15911; ABclonal, Wuhan, China), Vinculin (A2752; ABclonal), and γ-H2A.X (AP0099; ABclonal). All antibodies were diluted to the recommended concentrations according to the manufacturers’ instructions.

Total RNA was isolated using Total RNA Extractor (TRIzol) (B511311-0100; Sangon Biotech, Shanghai, China) in accordance with the manufacturer’s instructions, then converted to cDNA using the ABScript III RT Master Mix for qPCR with gDNA Remover (RK20429; ABclonal). Quantitative RT-PCR was performed using QuantStudio™ Real-Time PCR Software v1.7.1 (Applied Biosystems, Foster City, CA, United States) and the 2X Universal SYBR Green Fast qPCR Mix (RK21203; ABclonal). Relative gene expression levels were analyzed using the 2^−ΔΔCT method, with GAPDH serving as the internal control. The sequences of primers used to knock down the target genes are listed in Table 2.

Tissues were fixed in 10% buffered formalin for 24 h and embedded in paraffin. The paraffin-embedded tissues were sectioned and placed on charged glass slides, followed by hematoxylin–eosin or IHC staining using an IHC staining kit (HPA006752; Sigma-Aldrich), according to the manufacturer’s instructions. IHC scores were calculated using the formula: IHC score = intensity score × percentage score. The intensity score was based on staining intensity (0–4), and the percentage score was determined by the proportion of stained cells (0: 0%, 1: 1%–25%, 2: 26%–50%, 3: 51%–75%, and 4: 76%–100%). The antibodies used in the IHC assay were USP42 (HPA006752; Sigma-Aldrich) and AR (SC-56824; Santa Cruz Biotechnology). All antibodies were diluted to the recommended concentrations according to the manufacturers’ instructions.

Cells were seeded at a density of 2 × 10⁴ cells/well onto glass slides in a 24-well plate and cultured for 24 h. The cells were then fixed with paraformaldehyde, permeabilized with 0.1% Triton X-100 for 10 min, incubated with γ-H2A.X antibody (AP0099, 1:200; ABclonal) at 4°C overnight, and subsequently incubated with FITC-labeled secondary antibodies for 1 h. Nuclei were counterstained with DAPI. Cell images were captured using a laser scanning confocal microscope (FV3000; Olympus, Tokyo, Japan).

22RV1 cells (1 × 10⁶), infected with lentiviruses targeting USP42 or a control gene, were mixed with Matrigel (1:1, v/v) and subcutaneously injected into BALB/c nude mice (Si Pei Fu Biotech, Beijing, China). A tumor-free xenograft was defined as one that did not reach the flank. All mice were sacrificed after 30 days, and the xenografts were extracted, weighed, and photographed.

The experimental protocol was approved by the Experimental Animal Ethics Committee of the Department of Laboratory Animal Science, Fudan University. All animal experiment designs complied with the 3R principles (Replacement, Reduction, Refinement). The care and use of animals followed institutional guidelines. The BALB/c nude mice, which were specific pathogen-free, were purchased from Si Pei Fu Biotech. All mice were housed in a monitored environment at 23°C ± 1°C with 50%–60% relative humidity and a 12-h light/12-h dark cycle, with water and food provided ad libitum.

The mice were randomly divided into two groups, and different cells were subcutaneously injected into the flank. Tunnel handling was used to pick up the mice, and they were restrained using three fingers. Euthanasia was performed by CO₂ asphyxiation followed by cervical dislocation. CO₂ was dispensed from a commercial cylinder using a fixed-pressure regulator and inline restrictor, controlling gas flow within 30%–70% of the chamber volume per minute, in accordance with the 2020 American Veterinary Medical Association guidelines. CO₂ flow was maintained for more than 60 s after respiratory arrest (which may take up to 5 min), followed by cervical dislocation to ensure euthanasia.

Total RNA extracts from 22RV1 cells were subjected to RNA sequencing at Majorbio Biopharm Technology (Shanghai, China), and expression profiles were generated using the Majorbio Cloud Platform. Data-independent acquisition (DIA) proteomic analysis of 22RV1 cells was performed using an Orbitrap Astral mass spectrometer (Thermo Fisher Scientific, Waltham, MA, United States), and the resulting data were also analyzed on the Majorbio Cloud Platform.

Gene expression datasets of human PCa samples, including TCGA-PRAD (Hock et al., 2011; Hock et al., 2014) and GSE21034 (Hock et al., 2011; Hock et al., 2014), were obtained from The Cancer Genome Atlas (TCGA) database (http://portal.gdc.cancer.gov/) and the Gene Expression Omnibus database (http://www.ncbi.nlm.nih.gov/gds/), respectively. Gene set enrichment analysis (GSEA) was conducted using the GSEA software (version 4.2.2) provided by the Broad Institute (http://www.broadinstitute.org/gsea/index.jsp), with curated hallmark gene sets from the Molecular Signatures Database. The transcriptomic data have been deposited under BioProject ID: PRJNA1230828, and the mass spectrometry-based proteomic data are accessible via iProX ID: PXD061421.

All statistical analyses were performed using GraphPad Prism software Version 9.0 (GraphPad Software, San Diego, CA, United States). Quantitative data are presented as means ± standard deviation (SD). P-values of <0.05 were considered statistically significant.

To explore potential targets that promote the emergence and progression of PCa, we initially screened the expression levels of all DUB members in tumor and normal tissues using two large public PCa clinical datasets (TCGA (Cancer Genome Atlas Research, 2015) and MSKCC (Taylor et al., 2010)). As shown in Figure 1A, six DUBs were identified for further investigation. We designed shRNAs targeting each gene, and RT-PCR was performed to confirm the knockdown efficiency of each shRNA in PCa cells (Figure 1B). The proliferation rates of PCa cells were then measured using the MTT assay (Figure 1C), and their colony-forming ability was also assessed (Figure 1D). Based on its effect on PCa cell growth, we selected USP42—which remains functionally uncharacterized in PCa development and progression—for further study.

We examined USP42 expression in normal (n = 42) and PCa tissues (n = 43) via IHC. The scores were significantly higher in PCa tissues than in normal tissues (Mann–Whitney test, P < 0.05) (Figure 1E). Staining intensity was weaker in normal prostate tissues than in PCa tissues (Figure 1F). Furthermore, we assessed USP42 expression levels in both normal and tumor-derived prostate cell lines. Compared with normal cells, USP42 expression was elevated in prostate tumor cells (Figure 1G). In addition, a Kaplan–Meier plot revealed a significant association between higher USP42 expression and shorter PSA recurrence-free survival in PCa patient cohorts (Figure 1J). Consistent with the IHC results, the mRNA level of USP42 was also upregulated in PCa tissues (Figures 1H,I). These findings imply that abnormally elevated USP42 expression may play an important role in PCa development and progression.

ADT is the most common treatment for patients with PCa; therefore, we examined USP42 expression under low-androgen conditions. We found that in the androgen-sensitive PCa cell line LNCaP, USP42 was significantly downregulated after 3 days of low-androgen culture (Figure 2A). In the same experiment, we observed that USP42 expression in the androgen-independent cell line LNCaP_AI partially recovered (Figure 2A), implying that the presence of USP42 might contribute to PCa progression. Moreover, we knocked down AR using two different shRNAs or overexpressed AR via lentiviral transduction in PCa cells. The results show that USP42 expression decreased following AR knockdown and increased following AR overexpression (Figure 2B). This regulatory relationship between AR and USP42 was further confirmed at the transcriptional level (Figure 2C). To assess the correlation at the protein level, serial sections of human prostatectomy samples were analyzed using USP42 and AR IHC. USP42 and AR expression were significantly correlated in 68 human prostate tissue samples (Figures 2D,E). Additionally, to explore whether USP42 expression is linked to AR expression in PCa, we analyzed published human PCa datasets. Strikingly, a strong positive correlation was observed between USP42 and AR expression in two primary PCa cohorts (Figures 2E,F). Furthermore, USP42 expression levels showed a positive correlation with tumor stage Supplementary Figure S1A–C). These findings imply that USP42 may act downstream of AR in PCa.

To investigate the role of USP42 in PCa cell proliferation, we silenced USP42 using two shRNAs targeting distinct sequences. The knockdown efficiency was confirmed by Western blotting (Figure 3A). Depletion of USP42 in 22RV1 and PC3 cells significantly inhibited cell growth (Figure 3B). Similarly, in the cell viability assays, USP42 knockdown markedly reduced the numbers of cell colonies (Figures 3C,D). Next, we subcutaneously injected PCa cells—with or without USP42 knockdown—into nude mice. During the 1-month follow-up, mice injected with USP42-knockdown cells exhibited markedly delayed tumorigenesis (Figure 3G). At the end of the study, all mice were euthanized and the xenografts were examined. Tumors were significantly larger (Figure 3E) and heavier (Figure 3F) in the control group than in the USP42-knockdown group, indicating that suppression of USP42 significantly inhibited PCa tumor growth in vivo. These results demonstrate that USP42 is required for proliferation of PCa cells both in vitro and in vivo.

To investigate the possible mechanisms underlying the inhibitory effects of USP42 on PCa cell growth, we performed RNA sequencing and DIA-based proteomics following USP42 knockdown in 22RV1 cells. GSEA was used to analyze potential changes in biological states or processes. Gene sets related to the G2/M checkpoint and DNA repair were among the most highly enriched in control PCa cells (Figures 4A,B). Furthermore, KEGG analysis of downregulated genes in USP42-depleted PCa cells revealed significant enrichment in pathways related to DNA repair (Figure 4C). Similarly, relative normalized enrichment scores and P-values demonstrated that the G2/M checkpoint gene set was the most significantly enriched in patients with high USP42 expression (Figure 4D). Given these findings, we hypothesized that USP42 inhibition might be associated with increased DNA damage. To test this, we assessed γ-H2A.X, a marker of DNA damage, in PCa cells. USP42 knockdown resulted in a significant increase in γ-H2A.X foci (Figures 4E,F), and the protein levels of γ-H2A.X were also clearly elevated following USP42 depletion (Figure 4G). These data imply that USP42 deficiency induces DNA damage, contributing to its antitumor effects.

Olaparib, a PARP inhibitor, has been applied to treatment of metastatic PCa with DNA repair defects, and it has been confirmed that such patients can derive a survival benefit (Mateo et al., 2015). We investigated whether USP42 deficiency could enhance the efficacy of PARP inhibition in two olaparib-resistant PCa cell lines, 22RV1 and PC3. The combination of olaparib and USP42 knockdown led to more profound inhibition of cell growth than did either treatment alone (Figures 5A,B). Similarly, colony-formation assays showed that USP42 knockdown reduced cell survival under olaparib treatment (Figures 5C,D). Moreover, both the intensity of γ-H2A.X foci and its protein level were significantly increased in cells treated with both USP42 knockdown and olaparib (Figures 5D,F). These findings imply that the expression level of USP42 may influence the efficacy of olaparib in PCa.