OR Oncology Reports 1021-335X 1791-2431 D.A. Spandidos 10.3892/or.2021.8147 OR-0-0-8147 Articles Screening of hub genes and evaluation of the growth regulatory role of CD44 in metastatic prostate cancer LinJunhao ChenZhi LiZuan NongDeyong LiXiming HuangGuihai HaoNan LiangJianbo LiWei Department of Urology, The People's Hospital of Guangxi Zhuang Autonomous Region, Nanning, Guangxi Zhuang Autonomous Region 530021, P.R. China Correspondence to: Dr Wei Li, Department of Urology, The People's Hospital of Guangxi Zhuang Autonomous Region, 6 Taoyuan Road, Nanning, Guangxi Zhuang Autonomous Region 530021, P.R. China, E-mail: liwei95_2000@163.com 09 2021 19 07 2021 46 3 196 16102020 12052021 Copyright: © Lin et al. 2021 This is an open access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License, which permits use and distribution in any medium, provided the original work is properly cited, the use is non-commercial and no modifications or adaptations are made.

Prostate cancer (PCa) is the most common cancer type in men worldwide. Currently, the management of metastatic PCa (mPCa) remains a challenge to urologists. The analysis of hub genes and pathways may facilitate the understanding of the molecular mechanism of PCa. In the present study, to identify the hub genes in the mPCa, the three datasets GSE3325, GSE6919 and GSE38241 were downloaded from the platform of the Gene Expression Omnibus and function enrichment analysis of differentially expressed genes (DEGs) was performed. A total of 168 DEGs were obtained and the DEGs were significantly enriched in ‘cell junction’ and ‘cell adhesion’, among others. The Kyoto Encyclopedia of Genes and Genomes pathway enrichment analysis demonstrated that DEGs were enriched in three pathways including ‘focal adhesion’, ‘renal cell carcinoma’ and ‘Hippo signaling pathway’. The results of the protein-protein interaction network revealed that the hub genes in mPCa were separately PTEN, Rac GTPase-activating protein 1, protein regulator of cytokinesis 1, PDZ binding kinase, centromere-associated protein E, NUF2 component of NDC80 kinetochore complex, TPX2 microtubule nucleation factor, SOX2, CD44 and ubiquitin-like with PHD and ring finger domains 1. As a hub gene, CD44 was differentially expressed in PCa, as determined by Oncomine analysis. Further experiments in vivo demonstrated that SB-3CT, a selective matrix metalloproteinase inhibitor that has been reported to block CD44 cleavage and inhibit the downstream signaling pathway, suppressed the tumorigenicity of PCa cells by decreasing the expression levels of pyruvate dehydrogenase kinase 1 and 6-phosphofructo-2-kinase/fructose-2,6-biphosphatase 4. Moreover, the combination therapy with SB-3CT and docetaxel was more effective in inhibiting PCa compared with monotherapy. In conclusion, the identification of DEGs and the in vivo experimental results helped to elucidate the molecular mechanisms of PCa and provided a potential strategy for the treatment of PCa.

 bioinformatics analysis hub gene CD44 prostate cancer National Natural Science Foundation of Chinahttp://dx.doi.org/10.13039/501100001809 81460387 This work was funded by the National Natural Science Foundation of China (grant no. 81460387). Introduction

Prostate cancer (PCa) is the second most common cancer type worldwide and was ranked fifth with regards to cancer-related mortality rates in men in 2018 globally (1). Localized PCa can be treated using radical prostatectomy or radiation therapy (2). However, the disease control of metastatic PCa (mPCa) remains unsatisfactory (3). Although hormonal therapy has been widely used for mPCa, recurrence nearly always occurs after the initial period of treatment response and the cancer inevitably progresses to metastasis castrate-resistant PCa, which is extremely difficult to treat (4,5). Therefore, it is crucial to identify the exact molecular mechanism of the progression of mPCa, which may provide novel diagnostic and therapeutic targets.

Over the last decades, the technology of gene microarray and bioinformatic analysis has been applied for the examination of genetic alterations, which has enabled the identification of differentially expressed genes (DEGs) between mPCa and normal prostate tissues (6). In the present study, different microarray datasets were downloaded from the Gene Expression Omnibus (GEO) and multivariate statistical techniques were used for analysis. Analysis of the protein-protein interaction (PPI) network, enrichment analysis of Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway were performed to predict the hub genes and the molecular mechanism of mPCa among the DEGs, which could guide future experiments in vitro and in vivo.

CD44 is a cell-surface receptor that is expressed in the majority of normal and cancer tissues (7). CD44 is a cell-surface marker that is associated with the stemness, initiation and invasiveness of tumor cells (8). It has been reported that cell adhesion is primarily mediated by the CD44 signaling pathway, which is initiated by cleavages of CD44 ectodomain. Cleavages of CD44 ectodomain induced CD44 intracellular domain cleavage, and the subsequently generated intracellular domain fragment, can regulate signaling transcription (9). It has been revealed that MMPs are involved in the cleavage of CD44 ectodomain (10). Previous findings have observed that PC-3 cells expressed CD44, while LNCaP cells did not (11). It was also reported that CD44 could regulate cell proliferation, invasion and migration via pyruvate dehydrogenase kinase 1 (PDK1) and 6-phosphofructo-2-kinase/fructose-2,6-biphosphatase 4 (PFKB4) in PCa cells (12). Additionally, MMP inhibitor (SB-3CT) could decrease glycolytic activity via the inhibition of CD44 in PCa cells, and combination therapy with SB-3CT and docetaxel was more effective in inhibiting PCa compared with monotherapy (12).

Based on the results of previous studies in vitro, the present study performed additional experiments in vivo to further determine the role of CD44 in the progression of PCa and the role of SB-3CT in PCa.

 Materials and methods <sec> <title>Information of microarray data

GEO (http://www.ncbi.nlm.nih.gov/geo) is a public repository of high-throughput functional genomics data (13,14). In total, three datasets (GSE3325, GSE6919 and GSE38241) were downloaded from GEO to compare gene expression between metastasis prostate cancer tissues and normal prostate tissues. The organism of the selected datasets was homo sapiens, and the experiment type was expression profiling array. The probes of each dataset were annotated by the gene symbol, according to the information of the platform. The dataset of GSE3325 contained six mPCa tissue samples and six normal prostate tissue samples (15), while GSE6919 contained 25 mPCa tissue samples and 17 normal prostate tissue samples (16,17), and GSE38241 contained 18 mPCa tissue samples and 21 normal prostate tissue samples (18).

 Identification of differentially expressed genes

The DEGs between mPCa and normal prostate samples were screened using statistical software R (https://www.r-project.org; version 3.6.0). Data standardization and quality detection were performed prior to analysis. DEGs were identified using the Empirical Bayes method according to the ‘limma’ package of Bioconductor (https://bioconductor.org) (19). In addition, |log2FC|>1 was set as the cut-off value and adjusted P<0.05 was considered significant. A volcano plot of the DEGs was conducted based on the ‘ggplot2’ package of R (https://CRAN.R-project.org/package=ggplot2; version 3.2.1). A heatmap of the DEGs was constructed using the ‘pheatmap’ package of R (https://CRAN.R-project.org/package=pheatmap; version 1.0.12).

 Enrichment analysis of GO term and KEGG pathway

GO term enrichment analysis of DEGs including biological process (BP), cellular component (CC) and molecular function (MF) was conducted using the ‘clusterProfiler’ package of Bioconductor (20). The KEGG pathway enrichment analysis of DEGs was performed using the DAVID database (https://david.ncifcrf.gov; version 6.8), which provides functional annotation tools online for understanding biological processes. P<0.05 was considered to indicate a statistically significant difference (21,22).

 Analysis of PPI network and hub gene identification

PPI network analysis was performed to identify hub genes and to evaluate the interactions among DEGs using the online database of Search Tool for the Retrieval of Interacting Genes (STRING; https://string-db.org; version 11.0) and Cytoscape software (www.cytoscape.org; version 3.7.1) (23,24). Firstly, the network of DEGs was mapped using STRING database with an interaction score >0.4. Then, the network was visualized using Cytoscape software. The top 10 hub genes among the DEGs were identified using the cytoHubba plugin of Cytoscape (25).

 PCa cell line and cell culture

LNCap and PC-3 cells (obtained from the Cell Bank of Type Culture Collection of Chinese Academy of Sciences) were cultured in RPMI-1640 medium (Thermo Fisher Scientific, Inc.) supplemented with 10% FBS (Thermo Fisher Scientific, Inc.). The cells were cultured in an incubator at 37°C and 5% CO2. Negative control groups were designed and performed in the experiments, and experiments were repeated at least three times.

 Cell transfection

The sequences of short hairpin (sh)RNA targeting PDK1 or PFKFB4 and the negative control were inserted into the lentiviral vector. A non-targeting sequence (forward sequence 5′-CCGGCAACAAGATGAAGAGCACCAACTCGAGTTGGTGCTCTTCATCTTGTTGTTTTT-3′) was used as the negative control. Lentiviruses with the packaging plasmid (PG-P1-VSVG, PG-P2-REV, PG-P3-RRE and pGLV3/H1/GFP) and shRNA plasmid were produced via the transfection of 293T cells. Supernatants with lentiviral were collected after transfection and filtered using a 0.45-µm strainer. The lentiviral-expressing CD44 was harvested by inserting the sequences of CD44 into a pLVX-EF1α vector. An empty vector was used as the negative control. The lentiviral vector was used to infect PC-3 cells for 24 h at 37°C. Using the same method, LNCaP cells were transfected with CD44 overexpression lentiviral vector. PC-3 and LNCaP cells infected by lentiviral vector were cultured for 72 h before subsequent experiments. Plasmid was purchased from BioVector NTCC Inc.

 Western blot analysis

Cell extraction was performed using lysis buffer (Thermo Fisher Scientific, Inc.) and protein was separated via 10% SDS-PAGE. The concentration of protein was quantified by the bicinchoninic acid method. The protein extracted from the NC, SB-3CT, SB-3CT + Docetaxel (5 mg/kg), and SB-3CT + Docetaxel (10 mg/kg) groups were loaded in different western blot lanes, respectively. Then, the separated protein was transferred onto a polyvinylidene fluoride membrane. After blocking with 5% skimmed milk, the membrane was incubated with primary antibodies specific for PDK1 (cat. no. ab110025; 1:500 dilution, Abcam), PFKFB4 (cat. no. ab137785; 1:500 dilution, Abcam) or GAPDH (cat. no. ab181602; 1:1,000 dilution, Abcam), followed by incubation with secondary antibody IgG (cat. no. R4880; 1:1,000 dilution, Sigma-Aldrich; Merck KGaA). The protein was visualized using enhanced chemiluminescence and analysed using software by Labworks Analysis Software.

 Immunohistochemical staining

Immunohistochemical staining was performed on paraffin-embedded tumor tissue sections using anti-PDK1 (cat. no. ab110025; 1:1,000 dilution, Abcam) or anti-PFKFB4 (cat. no. ab137785; 1:1,000 dilution, Abcam) antibodies according to the manufacturer's protocol. The tissue sample was fixed with 4% paraformaldehyde at 4°C for 12 h. After staining, the sections (5 µm) were observed at ×100 and ×400 magnification using a light microscope. Positive cells were distinguished by strong staining of the membrane.

 Tumor xenograft model in vivo

Animal experiments were permitted by the Ethics Committee of the People's Hospital of Guangxi Zhuang Autonomous Region (approval no. 2014-010). A total of 40 BALB/c nude mice (age, 4–6 weeks; male; weight, 20–25 g) were obtained from Guangdong Medical Laboratory Animal Center and maintained in a specific pathogen-free environment which consisted of individually ventilated cages and isolator modules. The mice were injected with treated PCa cells in the armpit. PC-3 cells, PC-3 cells infected with shRNA-PDK1 or shRNA-PFKFB4, and LNCaP cells infected with vector-expressing CD44 or negative control were subcutaneously injected into BALB/c nude mice. The BALB/c nude mice injected with PC-3 cells were considered as the negative control. The tumor volume was observed and measured up to 33 days. Tumor weight was measured after mice were euthanized.

For the evaluation of CD44 in the treatment of PCa in vivo, PC-3 cells were subcutaneously injected into BALB/c nude mice and SB-3CT or SB-3CT combined with docetaxel (5 or 10 mg/kg) was injected into the mice via the tail vein. The tumor volume and weight were measured, and the tumor tissues were dissected for western blotting and immunohistochemical staining. BALB/c nude mice injected with PC-3 cells were considered the negative control.

During the experiment, 40 BALB/c nude mice were used. Mice were kept on a 12 h light-dark cycle at 50-60% humidity and 23–25°C and fed chow and water ad libitum. The health and behavior of the mice were monitored twice daily, which included weight, water and food intake, and animal posture. Sign of weight loss, rapid breathing, bloating, reduced food intake, and visible tumor under the skin was regarded as illness, which led to the euthanasia of mice. All 40 mice were euthanized as tumors were observed under the skin, using pentobarbital sodium (100 mg/kg) in the study. The pentobarbital sodium was injected via the tail vein. The maximum tumor size in the mice allowed to grow was 2,000 mm3 (not exceed 20 mm in any direction) before euthanasia. In the research, ulceration of tumors was not observed in any of the mice, and metastatic tumors to the lung were evident in 8 of the 40 mice.

 Statistical analysis

The statistical analysis was performed using SPSS software (version 19.0; IBM Corp.) and the graphs were created using GraphPad Prism software (version 6.0; GraphPad Software, Inc.). Data were analysed using the Student's t-test (independent t-test) or a one-way ANOVA with Tukey's post hoc test. P<0.05 was considered to indicate a statistically significant difference.

 Results <sec> <title>Identification of DEGs in mPCa

The datasets of GSE3325, GSE6919 and GSE38241 were downloaded from the GEO platform. A total of 4,790, 1,144 and 1,920 DEGs were screened from GSE3325, GSE6919 and GSE38241, respectively (Fig. 1A-C). In addition, 168 common DEGs were identified among the three datasets (Fig. 2A) and the expression levels of the common DEGs in the three datasets are presented (Fig. 2B-D).

 Enrichment analysis of GO term

The results of GO enrichment analysis varied with regards to the GO term and the different expression of common DEGs (Fig. 3A). The result of GO enrichment analysis in BP showed that the common DEGs were significantly enriched in ‘cell junction’, ‘cell adhesion’, ‘epithelial to mesenchymal transition’ and ‘epithelial cell proliferation’, among others (Fig. 3B). With regards to CC, the common DEGs were significantly enriched in ‘adherens junction’, ‘cell junction’, ‘focal adhesion’ and ‘myofibril’, among others (Fig. 3C). For MF, the common DEGs were significantly enriched in ‘actin binding’, ‘collagen binding’, ‘proximal promoter sequence-specific DNA binding’, ‘guanyl nucleotide binding’ and ‘guanyl ribonucleotide binding’ (Fig. 3D).

 Enrichment analysis of KEGG pathway

The results of KEGG pathway enrichment analysis demonstrated that the common DEGs were significantly enriched in the ‘Focal adhesion’, ‘Hippo’ and ‘Renal cell carcinoma’ signaling pathways (Fig. 4).

 Analysis of PPI network and hub gene

The PPI network consisted of 167 nodes and 168 edges based on STRING database, and the network was visualized using Cytoscape software (Fig. 5). The top 10 common DEGs with the highest degree were screened as the hub genes of mPCa and their names and functions are presented in Table I.

 Knockdown of PDK1 or PFKFB4 inhibits tumorigenicity of PCa cells in vivo

The BALB/c nude mice were injected subcutaneously with PC-3 cells infected with shRNA-PDK1 or shRNA-PFKFB4. Tumors dissected from mice were imaged and measured (Fig. 6A). The maximum tumor size was 1,775.74 mm3. It was found that knockdown of PDK1 or PFKFB4 inhibited the tumor growth of PCa cells in vivo (Fig. 6B). Similar results were observed for tumor weight (Fig. 6C).

 Overexpression of CD44 promotes tumorigenicity of PCa cells in vivo

The BALB/c nude mice were subcutaneously injected with LNCaP cells transfected with CD44 overexpression vector or NC. Tumors dissected from mice were imaged and measured (Fig. 7A). The maximum tumor size was 1,932.95 mm3. The results indicated that overexpression of CD44 promoted the tumor growth and tumor weight of PCa xenografts in vivo (Fig. 7B and C).

 Inhibition of CD44 suppresses tumorigenicity of PCa cells in vivo and the CD44 inhibitor (SB-3CT) combined with docetaxel inhibits the tumorigenicity of PCa

The BALB/c nude mice were subcutaneously injected with PC-3 cells. After inoculation, SB-3CT or SB-3CT combined with docetaxel (5 or 10 mg/kg) was injected into mice via the tail vein. Tumors dissected from mice were imaged and measured (Fig. 8A). The maximum tumor size was 1,775.74 mm3. It was identified that combined therapy with CD44 inhibitor (SB-3CT) and docetaxel could significantly inhibit tumor growth compared with treatment with the CD44 inhibitor (SB-3CT) alone. Moreover, it was found that a high concentration of docetaxel (10 mg/kg) could achieve higher inhibitory effects compared with the low concentration (5 mg/kg) (Fig. 8B and C). The expression levels of PDK1 and PFKFB4 in tumor tissue were examined, and were found to be significantly downregulated both in the monotherapy and combined therapy groups. Similar results were also observed in the results of immunohistochemical staining (Fig. 8D-G). Based on these aforementioned results, it was suggested that SB-3CT combined with a high dose of docetaxel could inhibit tumor growth more effectively than SB-3CT alone.

 Discussion

Although significant progress has been achieved in the management of mPCa, the pathogenesis of mPCa has not been fully elucidated due to the potentially complex biological traits of cancer. Microarray technology enables researchers to screen hub genes and primary pathways that are associated with mPCa, and has proven to be a helpful technology (26). Therefore, microarray technology has been used to identify genetic alterations involved in the pathogenesis and progression of diseases.

In the present study, three microarray databases were accessed to identify DEGs between mPCa tissues and normal prostate tissues. A total of 168 common DEGs were obtained for further analysis. To reveal interactions among the common DEGs, GO and KEGG pathways, enrichment analysis was performed. GO enrichment analysis indicated that the DEGs were mostly enriched in ‘cell junction’ and ‘cell adhesion’. The results of GO analysis are consistent with previous studies, which reported that ‘cell junction’ is associated with paracellular diffusion regulation and that ‘cell adhesion’ serves a crucial role in the transformation and progression of cancer (27–29). The KEGG pathway enrichment analysis indicated that the DEGs were mostly enriched in ‘Hippo signaling pathway’, ‘focal adhesion’ and ‘renal cell carcinoma’. Previous studies have reported that the hippo signaling pathway is involved in the regulation of cell proliferation and cell apoptosis, and is upregulated in tumors (30,31). In the present study, the top 10 common DEGs with highest degree were screened as the hub genes, including PTEN, Rac GTPase-activating protein 1, Protein regulator of cytokinesis 1, PDZ binding kinase, Centromere-associated protein E, NUF2 component of NDC80 kinetochore complex, TPX2 microtubule nucleation factor, SOX2, CD44 and ubiquitin-like with PHD and ring finger domains 1. Previous findings have revealed that CD44 is an adhesion molecule and is involved in the processes of invasion and metastasis in tumor cells (10). The present bioinformatics analysis results were consistent with these aforementioned findings. Based on the current results of the bioinformatics analysis, it was suggested that CD44 could regulate cell proliferation, migration and invasion in PCa.

The results of our previous study revealed that inhibition of CD44 using SB-3CT could suppress proliferation, invasion and migration in PCa cells by regulating PDK1 and PFKFB4 expression levels (12). Based on the present bioinformatics analysis results and our previous study, additional experiments in vivo were performed, including tumor formation assay and tumor metastasis experiments. In the present study, the results of tumor xenograft implantation demonstrated that knockdown of PDK1 or PFKFB4 in PC-3 cells inhibited the tumorigenicity of PCa in vivo. Further experiments indicated that inhibition of CD44 using an MMP inhibitor (SB-3CT) in PC-3 cells suppressed the tumorigenicity of PCa and inhibited the expression levels of PDK1 and PFKFB4 in PC-3 cells. SB-3CT, a selective MMP inhibitor, has been reported to block CD44 cleavage and inhibit downstream signaling pathway (32). Taken together, the present results indicated that CD44 suppressed the tumorigenicity of PCa via PDK1 and PFKFB4 in vivo, which was consistent with the results of previous study in vitro (12). As aforementioned, previous studies have reported that PC-3 cells expressed CD44, while LNCaP cells did not (11). In the present study, CD44 was overexpressed in LNCaP cells and this overexpression promoted the tumorigenicity of PCa. Tumor metastasis experiments were also performed. In the present study, metastatic tumors were found in the lung of the mice. However, the difference between two groups in the number of metastatic tumors in the lung was not significant.

Currently, hormonal therapy and chemotherapy are the first line choice for the treatment of mPCa, and adverse events were frequent in the two protocols (33). The combination of hormonal therapy and docetaxel became a novel therapy for mPCa. The results of the present study demonstrated that CD44 regulated the tumorigenicity of PCa in vivo, which suggested that inhibition of CD44 using SB-3CT is a novel potential treatment of PCa. According to current combined therapy of mPCa, it was suggested that the combination of CD44 inhibitor and docetaxel may be a beneficial strategy. Our previous study revealed that the combination of docetaxel and SB-3CT could significantly decrease the viability of PC-3 cells compared with single treatment of docetaxel at the concentration of 5 or 10 mg/kg (12). The present study evaluated the effect of combined therapy with CD44 inhibitor (SB-3CT) and docetaxel. The results indicated that treatment with CD44 inhibitor and docetaxel inhibited tumor growth and decreased expression levels of PDK1 and PFKFB4. Moreover, it was identified that treatment with high concentration of docetaxel induced a more positive response compared with the low concentration. Although the combined therapy was effective, sequential therapy is another potential therapy, but requires further investigation (34).

CD44 is a cell-surface receptor for hyaluronic acid and extracellular matrix components, and it serves a critical role in connecting the microenvironments in cancer. CD44 enables cancer cells to perceive the changes of microenvironments and can mediate the transduction of growth factor and cytokine signaling which can promote cell invasion and metastasis. Growth factors from microenvironments mediated by CD44, including EDF, FGF, HGF, VEGF, TGF-β, can also regulate tumorigenicity (35). It has been reported that CD44 could regulate the activation of macrophages in tumors, which was associated with tumorigenicity (36,37). As aforementioned, the regulation of microenvironments or macrophages by CD44 could affect tumorigenicity, but this required further examination. CD44 can also regulate EMT and reactive oxygen species (ROS) metabolism. Moreover, CD44 may regulate glucose metabolism in PCa (38). With increased glycolytic activity, reduced mitochondrial respiration leads to decreased ROS levels (39). Previous in vitro studies have reported that inhibition of CD44 expression could decrease glucose consumption and increase ROS level. Based on these results, it was suggested that CD44 could regulate the tumorigenicity of PCa cells via the regulation of ROS via PDK1 or PFKFB4.

The 70-kDa ribosomal protein S6 kinase, known as p70S6K, is a dual pathway kinase which acts downstream of PI3K pathway and mTOR pathway in response to growth factors and cytokines to regulate cell growth and inhibit cell apoptosis. Previous findings showed that p70S6K was regulated by PDK1 in PI3K pathway and could suppress BAD-induced cell apoptosis by the phosphorylation of Ser-136 on BAD (40). In our study, it was suggested that CD44 could suppress the tumorigenicity of prostate cancer by decreasing PDK1 and PFKFB4. Consequently, we hypothesized that CD44 could regulate the apoptosis of prostate cancer cells through p70S6K. In future, the effect of CD44 on the apoptosis of prostate cancer and its mechanism may be the focus of future research.

However, the present study has some limitations. CD44 has two variable regions and various CD44 isoforms produced by alternative splicing, which may have diverse effects on cancer progression. Thus, the CD44 isoforms should be analyzed and examined further in PCa tissues. Our previous findings demonstrated the difference between combined treatment and docetaxel alone in vitro. In addition, docetaxel was the first-line treatment for castration-resistant prostate cancer, which had been confirmed by lots of experiments in vivo and clinical practices. The main purpose of the present study was confirming the benefits of combined treatment with docetaxel and SB-3CT, a novel compound for prostate cancer treatment. Based on the aforementioned findings, SB-3CT alone was designed as the control group; however, it would be more reasonable to add docetaxel as monotherapy as a further control group.

In conclusion, a total of 168 common DEGs were identified and 10 hub genes were considered as biomarkers for mPCa. Further experimental results indicated that CD44 regulated the tumorigenicity of PCa via PDK1 and PFKFB4 in vivo. The present results demonstrated that the combination of SB-3CT and docetaxel was more effective in the inhibition of tumor growth, which suggested that combination therapy is a potential therapeutic strategy for mPCa.

 Acknowledgements

Not applicable.

 Funding

This work was funded by the National Natural Science Foundation of China (grant no. 81460387).

 Availability of data and materials

The data are available on reasonable request from the corresponding author.

 Authors' contributions

WL contributed to the experiment conception and design, data analysis, and manuscript draft. JL, ZC, and XL conducted the experiments. WL, JL, DN, and GH contributed to manuscript draft and data analysis. NH, ZL, and JL contributed to interpretation of data, manuscript draft and manuscript revision. NH, ZL, and JL are responsible for confirming the authenticity of all the raw data. All authors read and approved the final manuscript.

 Ethics approval and consent to participate

The Ethics Committee of the People's Hospital of Guangxi Zhuang Autonomous Region approved the study (approval no. 2014-010). Informed consent was provided by all the participants.

 Patient consent for publication

Not applicable.

 Competing interests

The authors declare that they have no competing interests.

 References BrayFFerlayJSoerjomataramISiegelRLTorreLAJemalAGlobal cancer statistics 2018: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countriesCA Cancer J Clin68394424201810.3322/caac.2149230207593 CocciACitoGRomanoALarganàGVignoliniGMinerviniADi MaidaFCampiRCariniMMondainiNRussoGIRadical prostatectomy and simultaneous penile prosthesis implantation: A narrative reviewInt J Impot Res32274280202010.1038/s41443-019-0176-131395986 WeinerABNetteyOSMorgansAKManagement of metastatic hormone-sensitive prostate cancer (mHSPC): An evolving treatment paradigmCurr Treat Options Oncol206201910.1007/s11864-019-0668-831286275 HeidenreichABastianPJBellmuntJBollaMJoniauSvan der KwastTMasonMMatveevVEAU guidelines on prostate cancer. Part II: Treatment of advanced, relapsing, and castration-resistant prostate cancerEur Urol65467479201410.1016/j.eururo.2013.11.00224321502 WülfingCBögemannMGoebellPJHammererPMachtensSPfisterDSchwentnerCSteuberTvon AmsbergGSchostakMTreatment situation in metastastic Castration Naive Prostate Cancer (mCRPC) and the implications on clinical routineUrologe A58106610722019(In German)10.1007/s00120-019-0925-2 DaileyALMetabolomic bioinformatic analysisMethods Mol Biol1606341352201710.1007/978-1-4939-6990-6_2228502011 SpringFADalchauRDanielsGLMallinsonGJudsonPAParsonsSFFabreJWAnsteeDJThe Ina and Inb blood group antigens are located on a glycoprotein of 80,000 MW (the CDw44 glycoprotein) whose expression is influenced by the In(Lu) geneImmunology64374319882454887 ProchazkaLTesarikRTuranekJRegulation of alternative splicing of CD44 in cancerCell Signal2622342239201410.1016/j.cellsig.2014.07.01125025570 Miletti-GonzálezKEMurphyKKumaranMNRavindranathAKWernyjRPKaurSMilesGDLimEChanRChekmarevaMIdentification of function for CD44 intracytoplasmic domain (CD44-ICD)J Biol Chem2871899519007201210.1074/jbc.M111.318774 NaganoOSayaHMechanism and biological significance of CD44 cleavageCancer Sci95930935200410.1111/j.1349-7006.2004.tb03179.x15596040 TaiSSunYSquiresJMZhangHOhWKLiangCZHuangJPC3 is a cell line characteristic of prostatic small cell carcinomaProstate7116681679201110.1002/pros.2138321432867 LiWQianLLinJHuangGHaoNWeiXWangWLiangJCD44 regulates prostate cancer proliferation, invasion and migration via PDK1 and PFKFB4Oncotarget86514365151201710.18632/oncotarget.1782129029419 BarrettTWilhiteSELedouxPEvangelistaCKimIFTomashevskyMMarshallKAPhillippyKHShermanPMHolkoMNCBI GEO: Archive for functional genomics data sets-updateNucleic Acids Res41(Database Issue)D991D995201323193258 EdgarRDomrachevMLashAEGene Expression Omnibus: NCBI gene expression and hybridization array data repositoryNucleic Acids Res30207210200210.1093/nar/30.1.20711752295 VaramballySYuJLaxmanBRhodesDRMehraRTomlinsSAShahRBChandranUMonzonFABecichMJIntegrative genomic and proteomic analysis of prostate cancer reveals signatures of metastatic progressionCancer Cell8393406200510.1016/j.ccr.2005.10.00116286247 ChandranURMaCDhirRBiscegliaMLyons-WeilerMLiangWMichalopoulosGBecichMMonzonFAGene expression profiles of prostate cancer reveal involvement of multiple molecular pathways in the metastatic processBMC Cancer764200710.1186/1471-2407-7-6417430594 YuYPLandsittelDJingLNelsonJRenBLiuLMcDonaldCThomasRDhirRFinkelsteinSGene expression alterations in prostate cancer predicting tumor aggression and preceding development of malignancyJ Clin Oncol2227902799200410.1200/JCO.2004.05.15815254046 AryeeMJLiuWEngelmannJCNuhnPGurelMHaffnerMCEsopiDIrizarryRAGetzenbergRHNelsonWGDNA methylation alterations exhibit intraindividual stability and interindividual heterogeneity in prostate cancer metastasesSci Transl Med5169ra10201310.1126/scitranslmed.300521123345608 SmythGKLinear models and empirical bayes methods for assessing differential expression in microarray experimentsStat Appl Genet Mol Biol3Article3200410.2202/1544-6115.102716646809 YuGWangLGHanYHeQYclusterProfiler: An R package for comparing biological themes among gene clustersOMICS16284287201210.1089/omi.2011.011822455463 Huang daWShermanBTLempickiRASystematic and integrative analysis of large gene lists using DAVID bioinformatics resourcesNat Protoc44457200910.1038/nprot.2008.211 Huang daWShermanBTLempickiRABioinformatics enrichment tools: Paths toward the comprehensive functional analysis of large gene listsNucleic Acids Res37113200910.1093/nar/gkn923 ShannonPMarkielAOzierOBaligaNSWangJTRamageDAminNSchwikowskiBIdekerTCytoscape: A software environment for integrated models of biomolecular interaction networksGenome Res1324982504200310.1101/gr.123930314597658 SzklarczykDGableALLyonDJungeAWyderSHuerta-CepasJSimonovicMDonchevaNTMorrisJHBorkPSTRING v11: Protein-protein association networks with increased coverage, supporting functional discovery in genome-wide experimental datasetsNucleic Acids Res47D607D613201910.1093/nar/gky113130476243 ChinCHChenSHWuHHHoCWKoMTLinCYcytoHubba: Identifying hub objects and sub-networks from complex interactomeBMC Syst Biol8(Suppl 4)S11201410.1186/1752-0509-8-S4-S1125521941 AlbertsonDGPinkelDGenomic microarrays in human genetic disease and cancerHum Mol Genet 12 Spec No2R145R152200310.1093/hmg/ddg26112915456 HejmejABilinskaBA role of junction-mediated interactions in cells of the male reproductive tract: Impact of prenatal, neonatal, and prepubertal exposure to anti-androgens on adult reproductionHistol Histopathol29815830201424481863 HejmejABilinskaBThe effects of flutamide on cell-cell junctions in the testis, epididymis, and prostateReprod Toxicol81116201810.1016/j.reprotox.2018.06.01429958919 YuCCChenLCLinVCHuangCYChengWCHsiehARChangTYLuTLLeeCHHuangSPBaoBYEffect of genetic variants in cell adhesion pathways on the biochemical recurrence in prostate cancer patients with radical prostatectomyCancer Med827772783201910.1002/cam4.216330993852 PanDThe hippo signaling pathway in development and cancerDev Cell19491505201010.1016/j.devcel.2010.09.01120951342 SaucedoLJEdgarBAFilling out the Hippo pathwayNat Rev Mol Cell Biol8613621200710.1038/nrm222117622252 ThorneRFLeggJWIsackeCMThe role of the CD44 transmembrane and cytoplasmic domains in co-ordinating adhesive and signalling eventsJ Cell Sci117373380200410.1242/jcs.0095414702383 TannockIFde WitRBerryWRHortiJPluzanskaAChiKNOudardSThéodoreCJamesNDTuressonIDocetaxel plus prednisone or mitoxantrone plus prednisone for advanced prostate cancerN Engl J Med35115021512200410.1056/NEJMoa04072015470213 TakeuchiHMmejeCOJineshGGTaokaRKamatAMSequential gemcitabine and tamoxifen treatment enhances apoptosis and blocks transformation in bladder cancer cellsOncol Rep3427382744201510.3892/or.2015.422026323344 YanYZuoXWeiDConcise review: Emerging role of CD44 in cancer stem cells: A promising biomarker and therapeutic targetStem Cells Transl Med410331043201510.5966/sctm.2015-004826136504 RaoGWangHLiBHuangLXueDWangXJinHWangJZhuYLuYReciprocal interactions between tumor-associated macrophages and CD44-positive cancer cells via Osteopontin/CD44 promote tumorigenicity in colorectal cancerClin Cancer Res19785797201210.1158/1078-0432.CCR-12-278823251004 TakeuchiHTanakaMTanakaATsunemiAYamamotoHPredominance of M2-polarized macrophages in bladder cancer affects angiogenesis, tumor grade and invasivenessOncol Lett1134033408201610.3892/ol.2016.439227123124 LiWCohenASunYSquiresJBraasDGraeberTGDuLLiGLiZXuXThe role of CD44 in glucose metabolism in prostatic small cell neuroendocrine carcinomaMol Cancer Res14344353201610.1158/1541-7786.MCR-15-046626832214 TakeuchiHTaokaRMmejeCOJineshGGSafeSKamatAMCDODA-Me decreases specificity protein transcription factors and induces apoptosis in bladder cancer cells through induction of reactive oxygen speciesUrol Oncol34337.e11e18201610.1016/j.urolonc.2016.02.02527038699 HaradaHAndersenJSMannMTeradaNKorsmeyerSJp70S6 kinase signals cell survival as well as growth, inactivating the pro-apoptotic molecule BADProc Natl Acad Sci USA9896669670200110.1073/pnas.17130199811493700 ChuECTarnawskiASPTEN regulatory functions in tumor suppression and cell biologyMed Sci Monit10RA235RA241200415448614 GlotzerMCytokinesis: Centralspindlin moonlights as a membrane anchorCurr Biol23R145R147201310.1016/j.cub.2013.01.00623428321 JiangWJimenezGWellsNJHopeTJWahlGMHunterTFukunagaRPRC1: A human mitotic spindle-associated CDK substrate protein required for cytokinesisMol Cell2877885199810.1016/S1097-2765(00)80302-09885575 GaudetSBrantonDLueRACharacterization of PDZ-binding kinase, a mitotic kinaseProc Natl Acad Sci USA9751675172200010.1073/pnas.09010239710779557 TestaJRZhouJYBellDWYenTJChromosomal localization of the genes encoding the kinetochore proteins CENPE and CENPF to human chromosomes 4q24->q25 and 1q32->q41, respectively, by fluorescence in situ hybridizationGenomics23691693199410.1006/geno.1994.15587851898 NabetaniAKoujinTTsutsumiCHaraguchiTHiraokaYA conserved protein, Nuf2, is implicated in connecting the centromere to the spindle during chromosome segregation: A link between the kinetochore function and the spindle checkpointChromosoma110322334200110.1007/s00412010015311685532 HollandAJClevelandDWLosing balance: The origin and impact of aneuploidy in cancerEMBO Rep13501514201210.1038/embor.2012.5522565320 JohanssonHSimonssonSCore transcription factors, Oct4, Sox2 and Nanog, individually form complexes with nucleophosmin (Npm1) to control embryonic stem (ES) cell fate determinationAging (Albany NY)2815822201010.18632/aging.10022221076177 MudbharyRHoshidaYChernyavskayaYJacobVVillanuevaAFielMIChenXKojimaKThungSBronsonRTUHRF1 overexpression drives DNA hypomethylation and hepatocellular carcinomaCancer Cell25196209201410.1016/j.ccr.2014.01.00324486181

Volcano plots of DEGs. Volcano plots of DEGs in metastasis prostate cancer and normal prostate samples in the datasets of (A) GSE3325, (B) GSE6919 and (C) GSE38241. DEGs were selected using |log2FC|>1 and an adjusted P<0.05. Colors refer to the different genes: Grey represents genes without significantly different expression. Red represents genes that are upregulated in metastasis prostate cancer tissues compared with normal prostate tissues and blue indicates genes that are downregulated. DEG, differentially expressed genes.

Venn diagram and heatmap of common DEGs. (A) Venn diagram indicates the number of common DEGs in the three datasets. Heatmap of common DEGs in the datasets of (B) GSE38241, (C) GSE3325 and (D) GSE6919. In the heatmap of common DEGs, blue indicates downregulated and red represents upregulated genes. DEG, differentially expressed genes.

GO enrichment analysis of common DEGs. (A) Results of GO enrichment analysis. The x-axis indicates the number of genes enriched in the marked category. The y-axis represents the functional categories. Only categories of GO with P<0.05 are shown. Bubble charts for (B) BPs, (C) CCs and (D) MFs. The x-axis represents the ratio of common DEGs enriched in the marked category. The size of the bubble indicates the number of common DEGs enriched in the marked category and color refers to the value of difference. GO, Gene Ontology; BP, Biological Processes; CC, Cellular Components; MF, Molecular Functions; DEG, differentially expressed genes.

KEGG pathway enrichment analysis of common DEGs. Bubble chart for KEGG pathway enrichment analysis is presented. The x-axis represents the ratio of common DEGs enriched in the marked category. The size of the bubble indicates the number of common DEGs enriched in the marked category and color refers to the value of differences. The categories of ‘Focal adhesion’, ‘Hippo signaling pathway’ and ‘Renal cell carcinoma’ were considered significant, while the categories of ‘Regulation of actin cytoskeleton’, ‘Melanoma’ and ‘Inflammatory bowel disease’ were considered not significant. DEG, differentially expressed genes; KEGG, Kyoto Encyclopedia of Genes and Genomes.

PPI network and module analysis of common DEGs. A total of 167 DEGs were screened using the STRING database and the PPI network was constructed using Cytoscape. The color is indicative of the different genes: Blue represents downregulated genes, while red indicates upregulated genes. Yellow represents genes that are upregulated in some probes but downregulated in other probes. DEG, differentially expressed genes; PPI, protein-protein interaction.

Knockdown of PDK1 or PFKFB4 inhibits tumorigenicity of prostate cancer cells in vivo. (A) Tumors dissected from 15 BALB/c nude mice are presented (n=5 for each group). (B) Tumor volume curve of NC, PDK1 and PFKFB4 treatment groups. Asterisk in the red tumor volume curve indicates that the shRNA-PDK1 group is significantly different compared to the NC group. Asterisk in the blue tumor volume curve indicates that the shRNA-PFKFB4 group is significantly different compared to the NC group. (C) Tumor weight of NC, PDK1 and PFKFB4 treatment groups. Knockdown of PDK1 or PFKFB4 inhibited the tumor growth of prostate cancer cells in vivo. Data are analysed using an ANOVA with Tukey's post hoc test. Data are presented as the mean ± SD. *P<0.05. NC, negative control; PDK1, pyruvate dehydrogenase kinase 1; PFKB4, 6-phosphofructo-2-kinase/fructose-2,6-biphosphatase 4.

Overexpression of CD44 promotes tumorigenicity of prostate cancer cells in vivo. (A) Tumors dissected from 10 BALB/c nude mice are presented (n=5 for each group). (B) Tumor volume curve of NC and CD44 treatment groups. (C) Tumor weight of NC and CD44 treatment groups. Overexpression of CD44 promoted the tumor growth of prostate cancer cells in vivo. Data are analysed using a Student's t-test (independent t-test) or an ANOVA with Tukey's post hoc test. Data are presented as the mean ± SD. *P<0.05. NC, negative control; PDK1, pyruvate dehydrogenase kinase 1; PFKB4, 6-phosphofructo-2-kinase/fructose-2,6-biphosphatase 4.

Inhibition of CD44 suppresses tumorigenicity of prostate cancer cells in vivo and the CD44 inhibitor (SB-3CT) combined with docetaxel inhibits tumorigenicity of prostate cancer. (A) Tumors dissected from BALB/c nude mice are presented (n=5 for each group). (B) Tumor volume curve of NC, SB-3CT, SB-3CT + Docetaxel (5 mg/kg) and SB-3CT + Docetaxel (10 mg/kg) treatment groups. (C) Tumor weight of NC, SB-3CT, SB-3CT + Docetaxel (5 mg/kg) and SB-3CT + Docetaxel (10 mg/kg) treatment groups. (D and E) PDK1 and PFKFB4 expression levels were downregulated in the PC-3 + SB-3T group, PC-3 + SB-3T + Docetaxel (5 mg/kg) group and PC-3 + SB-3T + Docetaxel (10 mg/kg) group compared with the PC-3 + NC group. (F) PDK1 expression levels were downregulated in the PC-3 + SB-3T group, PC-3 + SB-3T + Docetaxel (5 mg/kg) group and PC-3 + SB-3T + Docetaxel (10 mg/kg) group compared with the PC-3 + NC group. (G) PFKFB4 expression levels were downregulated in the PC-3 + SB-3T group, PC-3 + SB-3T + Docetaxel (5 mg/kg) group and PC-3 + SB-3T + Docetaxel (10 mg/kg) group compared with the PC-3 + NC group. Data are analysed using an ANOVA with Tukey's post hoc test. Data are presented as the mean ± SD. *P<0.05. NC, negative control; PDK1, pyruvate dehydrogenase kinase 1; PFKB4, 6-phosphofructo-2-kinase/fructose-2,6-biphosphatase 4.

Functional roles of top 10 hub genes with a degree ≥8.

No. Gene symbol Full name Function (Ref.)
1 PTEN Phosphatase and tensin homolog PTEN acts as a tumor suppressor through regulating AKT/PKB signaling pathway negatively (41)
2 RACGAP1 Rac GTPase-activating protein 1 RACGAP1 can regulate the progression of cytokinesis, cell growth and differentiation (42)
3 PRC1 Protein regulator of cytokinesis 1 PRC1 is associated with cytokinesis (43)
4 PBK PDZ binding kinase High expression of PBK is associated with tumorigenesis (44)
5 CENPE Centromere-associated protein E CENPE is necessary for stable spindle microtubule capture (45)
6 NUF2 NUF2 component of NDC80 kinetochore complex NUF2 is associated with centromeres of mitotic (46)
7 TPX2 TPX2 microtubule nucleation factor TPX2 is implicated as a regulator of cell apoptosis (47)
8 SOX2 SRY-box transcription factor 2 Transcription factors encoded by SOX2 play a regulatory role in embryonic development (48)
9 CD44 CD44 molecule CD44 is associated with cell-cell interaction, cell migration and adhesion (10)
10 UHRF1 Ubiquitin-like with PHD and ring finger domains 1 UHRF1 is overexpressed in various types of cancer (49)