The human PCa cell lines (LNCaP, C4-2B, VCaP, 22Rv1, PC3, DU145) and HEK293T cell (CRL-3216) were obtained from the American Type Culture Collection (Rockville, MD). Fetal bovine serum (FBS) and Charoal-stripped fetal bovine serum (CSS, depleted androgen and any other steroids) were purchased from Hyclone (South Logan, UT, USA). Dihydrotestosterone (DHT) was acquired from Meilunbio (Dalian, China). Thapsigargin and BAPTA-AM were purchased from APExBIO (Houston, USA). STO-609 was obtained from MedChemExpress (Monmouth Junction, NJ, USA).

The immunofluorescence protocol followed our established methodology (22, 23). Briefly, pretreated PCa cells were plated onto glass coverslips. After being fixed, cells were incubated with primary antibodies overnight at 4°C, Subsequently, the cells were incubated with secondary antibodies. To stain the nuclei, DAPI (Invitrogen, Carlsbad, CA, USA) was employed. The images of Immunofluorescence were captured and processed using Confocal Microscope (LSM880; ZEISS, Oberkochen, Germany). The primary antibody includes PCP4 (Proteintech, Wuhan, China) and CAMKK2 (Proteintech, Wuhan, China).

Western blotting assay was described in our previously study (24). Cells were harvested and lysed in RIPA buffer supplemented with protease inhibitors followed by protein quantification performed using BCA protein assay kit (Beyotime Biotechnology, Beijing, China). Then total protein samples along with protein ladder were loaded into the wells of SDS-PAGE gels. Wet transfer systems were used to transfer the proteins to the methanol activated PVDF membrane (Merck Millipore, Billerica, MA, USA). Then, we incubated the membranes in 10% milk for more than 1 hour to block of membranes. Primary antibody for PCP4 was obtained from Abcam (Cambridge, MA, USA). Antibodies for CAMKK2 and CaM were from Proteintech (Wuhan, China). Antibodies targeting p-AMPK, AMPK, AR, GAPDH and LaminA/C were form Cell Signaling Technology (Boston, MA, USA). Immunoprecipitation assay was conducted following the manufacturer’s instructions.

The total RNA was extracted from cultured cells by the TRIzol regents (Invitrogen, Carlsba d, CA, USA) to break open the cells and release the RNA. Then mRNA was reverse-transcribed into cDNA by the ReverTra Ace qPCR RT kit (TOYOBO, Japan). FastStart Universal SYBR Green Master (Roche, Base I, Switzerland) was used for qRT-PCR assay. The details of the primers of genes used in this study are listed in Supplementary Table S1 .

PCP4-specific siRNAs, PCP4 plasmids and corresponding negative controls for transient transfection were designed and synthesized by Geme-Pharma (Shanghai, China). The effective sequences of siRNA are presented in Supplementary Table 1 . To minimize off-target effects, co-transfection of two RNAs with better interference efficiency was performed. PCa cell lines were transiently transfected with siRNA or plasmids, along with their matched controls using Lipofectamine 3000 (Invitrogen, Carlsbad, CA, USA). Lentiviral vectors encoding PCP4 and the empty vector controls were procured from Genechem (Shanghai, China), with infections performed according to the manufacturer’s instruction. Then incubate the cells with puromycin for 2 weeks to obtain cell lines with stable expression.

3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfoohenyl)-2H-tetrazolium inner salt (MTS) (Promega, Madison, WI, USA) assay was applied to evaluate cell proliferation. Pretreated cells were seeded into 96-well cell plates at a density of 1.5 × 10 (3) cells per well, with 3 replicated wells per experimental group. 10 μL of MTS solution was introduced to each well and incubated at 37 °C for 2.5 hours. Then transfer the plate to a microplate reader. Measure the absorbance at a wavelength of 490 nm. The absorbance value was proportional to the number of viable cells, reflecting the cell proliferation rate. The transwell assay was used to measure the migration and invasion ability of PCa cells.

Male nude mice (4-6-week-old) were purchased from Weitonglihua Biotechnology (Beijing, China) and randomly divided into the experimental groups and control groups (n = 5/group). C4-2B cells stably expressing control or PCP4 (n = 1 × 10 (7)) were collected and resuspended in 100 μL PBS (50% Matrigel) and then injected subcutaneously into the mice. Tumor volume [length (mm)*width (2) (mm)*0.5] and tumor weight (g) were measured until 24 days after implantation. The experimental protocol was approved by the Shandong University Animal Care committee (Document No. ECSBMSSDU2021-2-126). All procedures were performed in compliance with the institution’s guidelines.

We performed RNA-seq analyses (ShenZhen BGI Genomics Co., Ltd. China) to compared the mRNA expression profiles between control (Vector) and PCP4 overexpression (PCP4) C4-2B cells. The raw data is available in NCBI website (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE293745). Gene Ontology/Kyoto Encyclopedia of Genes and Genomes (GO/KEGG) database was used to demonstrate the related pathways of PCP4 (25–27).

Datasets including GSE35988, GES68882, GES32269, GSE26367, GSE21034, GSE70770 and GSE70769 were downloaded from GEO database (http://www.ncbi.nlm.nih.gov/geo, modified: July 16, 2024). CBioPortal (http://www.cbioportal.org/, v6.0.20) tool was applied to assess PCP4 alteration frequency, mutation type and CNA (copy number alteration) of prostate cancer samples (28–30). The Cancer Genome Atlas (TCGA, prad_tcga_pan_can_atlas_2018), the Stand Up To Cancer (SU2C, prad_su2c_2019), MSKCC (MSKCC cancer cell 2010), and prostate_dkfz_2018 datasets were downloaded from CBioPortal website.

GSEA v4.0 program was used for gene set enrichment analysis according to the instructions. Androgen-induced and -repressed gene sets were obtained from Zhang et al. (31) and the data source and lists of these gene sets were showed in Supplementary Table 2 .

Cytosolic Ca²⁺ concentration was measured with the fluorescent using Fluo-acetoxymethylester probe (Fluo-4 AM) (Beyotime, Shanghai, China) according to manufacturer’s instruction. Briefly, after treatment, pretreated cells in 6-well plates were incubated with 5 μM Fluo-4 AM at 37°C for 1 hour in the dark and then washed 3 times with PBS to remove the extracellular Fluo-4 AM. Then, cells were incubated for another 20 min to ensure that Fluo-4 AM was completely convert into Fluo-4. Ca²⁺ imaging was acquired with inverted Fluorescence Microscope (EVOS M5000) and the fluorescent intensity was detected using flow cytometry.

Statistical analyses were performed using SPSS 20.0 (IBM Corp., Armonk, NY, USA) or GraphPad Prism 9.3.0. Each assay in vivo was performed in biological triplicate and the data are presented in the form of the mean ± standard deviation (SD). Statistical comparisons between groups were carried out using two-sided Student’s t test. Kaplan-Meier analysis was used to compare the survival information. The statistical significance of differences among groups was indicated by asterisks (*, p < 0.05; **, p < 0.01; ***, p < 0.001; ****, p < 0.0001). P < 0.05 was considered statistically significant.

Aiming to pinpoint the potential genes underlying the progression of PCa, particularly that of CRPC, we contrasted several datasets of primary localized PCa samples and CRPC samples, including GSE35988, GSE68882 and GSE32269. The transcriptomes between metastatic CRPC (mCRPC) and primary localized PCa samples were analyzed. By Cross-comparing all clustered genes, we obtained a common list of 5 genes that exhibited the most significantly changes in the CRPC group (> 3-fold change, p < 0.01; Figure 1A , Supplementary Table 3 ). Subsequently, the mRNA levels of these 5 genes were measured in LNCaP cells and C4-2B cells, which are a hormone sensitive prostate cancer (HSPC) model cell line and a CRPC model cell line, respectively. Among these 5 genes, PCP4 showed the most significant change, specifically a down-regulation ( Figure 1B ). A high frequency of PCP4 gene deletion was observed in different prostate adenocarcinoma studies based on cBioportal database. Moreover, the deletion of PCP4 DNA was found to be correlated with a lower transcriptome expression level of PCP4 ( Figure 1C ). Consistent with our findings, PCP4 expression was decreased in CRPC group than in primary prostate cancer tissues and normal prostate tissues in datasets such as GSE6919, GSE21034 and GES35988 ( Figure 1D ). Microarray analysis of the copy-number variation of PCP4 revealed a positive correlation between PCP4 copy-number variation and its expression (16).

PCP4 expression in different prostate cancer cell lines was performed in both transcriptome and protein level. Highest expression was observed in LNCaP cell line and lower expression lever was observed in C4-2B cell line, while extremely low expression was noted in other cell lines ( Figure 1E ).

To determine the frequency of PCP4 deletion and TMPRSS2-ERG genomic fusions in different prostate tissue types, we undertook an in-depth analysis of public datasets including TCGA and the SU2C dataset ( Figure 1F ). Regarding TCGA patients, TMPRSS2-ERG fusion was detected in 28.6% (140/489) of localized primary prostate cancer out of which 60% (84/140) exhibited both PCP4 deletion and TMPRSS2-ERG fusion. Meanwhile, concomitance of PCP4 deletion and TMPRSS2-ERG fusion was observed in CRPC patients in SU2C database.

TCGA, MSKCC, prostate_dkfz_2018 or GEO databases showed that reduced PCP4 mRNA expression was significantly associated with higher T stages ( Figure 2A ), higher Gleason scores ( Figure 2B ), poorer disease-free survival (DFS) ( Figure 2C ) or biochemical recurrence-free survival (BCRFS) ( Figure 2D ).

RNA silencing was applied to deplete PCP4 expression in androgen dependent cell line LNCaP, while plasmid to overexpress PCP4 in CRPC cell line C4-2B and 22Rv1 ( Figure 3A ). Downregulation of PCP4 led to increased cell growth in LNCaP cells, regardless of whether androgen was depleted (using CSS) or not. Conversely, PCP4 overexpression decreased cell growth in C4-2B and 22Rv1 cells ( Figure 3A ). Downregulation of PCP4 also resulted in increased migration and invasion capabilities in LNCaP cells both in the presence and absence of androgen depletion. Similarly, PCP4 overexpression reduced cell growth in C4-2B and 22Rv1 cells ( Figure 3B ). Intriguingly, PCP4 knockdown significantly augmented the proliferation and invasion of LNCaP cells under androgen depletion compared to androgen-repletion conditions (using FBS).

Furthermore, PCP4 overexpression in C4-2B xenografts in castrated nude mice resulted in delayed tumor progression ( Figure 3C ). The mean tumor volume at 1 351 ± 353.0 mm (3) in C4-2B-Vector xenografts and 299 ± 104.0 mm (3) in C4-2B-PCP4 xenografts with tumor weight 1.180 ± 0.084 g and 0.340 ± 0.114 g, respectively.

PCP4 protein is distributed mainly in the cell nucleus and a small amount in the cytoplasm in LNCaP cells, regardless of whether androgen was depleted (using CSS) or not ( Figure 4A ). The androgen-AR signaling plays a pivotal role in prostate tumorigenesis and progression, especially in the growth of CRPC. Given that PCP4 expression is most prominent in AR-positive LNCaP cells and AR serves as the primary target of DHT action in LNCaP cells (32), we utilized LNCaP cell line to identify whether PCP4 responds to androgen. Prostate specific antigen (PSA), an endogenous AR downstream target, was utilized as a positive control. LNCaP cells were first cultured in CSS medium for 3 days and then exposed to 0–10 nM DHT for 24 hours or 1 nM DHT for 0–48 hours. As depicted in Figure 4B , upon treatment with DHT, PCP4 decreased while PSA increased in a dose- and time-dependent manner at both mRNA and protein levels in LNCaP cells. The public dataset from GSE114267 also showed that PCP4 expression decreased when LNCaP was treated with DHT, even though the data was not significant which might be due to small sample size ( Figure 4C ). Conversely, down-regulation of AR in LNCaP and C4-2B cell lines led to PCP4 overexpression in both these two cell lines ( Figure 4D ). Based on these results, we infer that PCP4 is a novel DHT-repressed AR-target gene in PCa cells.

Multiple public datasets were utilized to elucidate how PCP4 regulates AR signaling. GSEA analysis based on TCGA dataset demonstrated that AR responded gene sets were significantly enriched in PCP4-depleted samples (NES = -1.86; p < 0.001; FDR q < 0.001) and AR-induced genes were also enriched in PCP4 depleted samples (NES = -1.50; p = 0.005; FDR q = 0.007) ( Figure 4E ). Meanwhile, PCP4 mRNA expression was negatively correlated with AR expression in several public datasets, such as SU2C and GSE35988 ( Figure 4F ).

To explore the mechanisms of how PCP4 regulates AR signaling, we detected key molecules in AR signaling pathways, which were reported as downstream effectors of AR, including kallikrein related peptidase 2 (KLK3), TMPRSS2, kallikrein related peptidase 2 (KLK2) ( Figure 4G ). Down regulation of PCP4 in LNCaP cells increased the mRNA levels of AR, KLK3, TMPRSS2, KLK2. While PCP4 overexpression in C4-2B reduced the mRNA levels of AR and TMPRSS2. These results suggested that AR signaling was negatively related with PCP4 expression in PCa. Interestingly, qRT-PCR results showed that PCP4 downregulation enhanced the expression of the AR-regulated gene KLK3 in a dose-dependent manner ( Figure 4H ).

Then we carried out RNA-sequencing to investigate the mechanisms through PCP4 suppresses the aggressive phenotype of PCa cells. C4-2B cells were transfected with either negative vectors or PCP4 plasmids for 48 hours. After that, total RNA was extracted for RNA sequencing analysis. KEGG analysis showed that the dysregulated genes induced by PCP4 overexpression were mostly enriched in Calcium signaling pathway, which was consistent with its molecular function reported before ( Figure 5A ). The relative expression levels of genes associated with Calcium signaling pathway were showed in Figure 5B .

Intracellular Ca²⁺ acts as one of the most widespread and crucial second messenger molecules. The Ca²⁺ fluorescence images and flow cytometry results showed that intracellular Ca²⁺ concentration decreased remarkable in C4-2B cells after PCP4 overexpression compared to the control while elevated in LNCaP cells after PCP4 downregulation ( Figures 5C, D ).

Previous study indicated that DHT could accelerate Ca²⁺ influx and increase the Ca²⁺ concentration in PCa cancer cells (33). In this study, we observed that DHT elevated Ca²⁺ concentration was partially reversed by PCP4 overexpression ( Figure 5E ). These results indicated that DHT induced PCP4 downregulation was involved in the process of DHT induced Ca²⁺ influx.

Among all the molecules involved in the Calcium signaling pathway, the pro-tumorigenic role of CAMKK2 in prostate cancer was widely established. Previous evidence has demonstrated that CAMKK2 expression was induced by AR and it in turn stabilized AR to promote its transcriptional activity and cell cycle progression. This interaction forms a novel positive feedback loop within the PCa (20, 21). Consequently, we hypothesized that the mechanism of PCP4 regulating PCa progression might be linked to Ca²⁺/CAMKK2/AR signaling.

CAMKK2, a crucial factor in the progression of PCa and transition to CRPC, has mostly been linked to phosphorylation of AMPK at threonine 172 (AMPK(Thr172)) in an AR-dependent manner (19). To examine whether PCP4 regulates CAMKK2 signaling pathway, AMPK and p-AMPK were tested as downstream effectors of CAMKK2. PCP4 did not induce remarkable changes of CAMKK2 in mRNA level. However, downregulation of PCP4 in LNCaP cells increased the protein levels of both CAMKK2 and p-AMPK. Conversely, the overexpression of PCP4 in C4-2B cells reduced CAMKK2 and p-AMPK protein expression levels ( Figure 5F ).

Additionally, treatment with Ca²⁺-chelator BAPTA-AM effectively reversed the PCP4-mediated inhibition of CAMKK2; Conversely, the CAMKK2-activating effect induced by PCP4 knockdown was significantly reversed by thapsigargin (a Ca²⁺-ATPase inhibitor) ( Figure 5G ). Critically, CAMKK2 depletion reversed PCP4 depletion-induced activation of p-AMPK and cell proliferation in LNCaP cells ( Figure 5H ). STO-609 is a regular inhibitor of CAMKK2 (34). The MTS results showed that whether it was siRNA or STO-609, both reversed PCP4 depletion-induced cell proliferation in LNCaP cells ( Figures 5I, J ). These findings collectively suggested that the downregulation of PCP4 promoted PCa progression by activating Ca²⁺/CAMKK2/AMPK/AR signaling axis.

Since PCP4 modulates CAMKK2 post-transcriptionally without altering its transcriptional expression, we next investigated whether PCP4 affects CAMKK2 or AR protein degradation in PCa cells. LNCaP and C4-2B cell lines were treated with cycloheximide (CHX) to block de novo protein synthesis. In LNCaP cells, downregulation of PCP4 significantly decelerated the degradation of CAMKK2 in LNCaP cells, while PCP4 overexpression remarkably accelerated the degradation of CAMKK2 in C4-2B cells ( Figures 6A, B ). These results indicated that PCP4 could promote the degradation of CAMKK2 protein.

Previous study suggested that CAMKK2 staining intensity increased with the Gleason score of tumors and the staining pattern shifted from predominantly cytoplasmic to perinuclear and nuclear (21). To explore whether PCP4 influences the nuclear translocation of CAMKK2, we treated with DHT or vehicle, then tested CAMKK2 localization by subcellular fractionation and immunofluorescence microscopy in PCP4 silenced LNCaP cells compared with control. The results showed that intracellular CAMKK2 was increased when PCP4 was downregulated whether LNCaP cells were treated with DHT or DMSO. Compared with DMSO group, CAMKK2 was increased after DHT treatment, and downregulation of PCP4 enhanced the CAMKK2 upregulation induced by DHT treatment. The increasement of CAMKK2 were both observed in both cytoplasmic and nuclear compartments ( Figures 6C, D ).

We noticed that downregulation of CAMKK2 increased the PCP4 protein expression level in LNCaP cells ( Figure 5H ). To confirm this finding, downregulation of CAMKK2 using siRNA was carried out in both LNCaP and C4-2B cell lines and the results verified this finding ( Figure 6E ). This suggests that CAMKK2 might form a feedback loop with PCP4.

Co-IP results showed that both PCP4 and CAMKK2 protein could bind with CaM, while PCP4 did not bind with CAMKK2 ( Figure 6F ). The binding between CAMKK2 and CaM increased when PCP4 was downregulated in LNCaP and decreased when PCP4 was overexpressed in C4-2B, which might a result of the difference in CAMKK2 protein level ( Figure 6G ).