Cell culture
The cell lines LNCaP (SCSP-5021, an androgen-dependent PCa cell line), PC-3 (SCSP-532, an androgen-independent cell line lacking AR expression), VCaP (SCSP-5034, a PCa cell line with wild-type AR expression), and 22RV1 (SCSP-5022, CRPC cell line with the expression of AR splice variants) were acquired from the Chinese Academy of Sciences, China. The LNCaP cells were cultured in 90% RPMI-1640 medium (61870127, Gibco, USA) with 10% fetal bovine serum (FBS) (10099141, Gibco, USA) and the VCaP cells were cultured in 90% DMEM (C12571500BT, Gibco, USA) with 10% FBS, 1 mM sodium pyruvate. The 22RV1 cells were cultured in 90% RPMI-1640 medium with 10% FBS, and the PC-3 cells were cultured in 90% F12K nutrient mixture medium (21127-022, Gibco, USA) with 10% FBS. All cell lines were cultured at 37°C with 5% of CO2 in a humidified atmosphere.
The cell proliferation assay
To assess the effect of Abiraterone on survival, cells were separately treated with 0, 10, 20, 40, 80, 150, 300, or 500 µM Abiraterone (A125745, Aladdin, Shanghai Aladdin Biochemical Technology Co., Ltd). In addition, cell survival was measured using a cell counting kit-8 (CCK-8, Sigma, USA) at 24, 48, or 72 h. The results were detected utilizing a microplate reader (Infinite M100 PRO, TECAN, Switzerland) at a wavelength of 450 nm.
The cell apoptosis assay
The 22RV1 cells (7 ⅹ105) and LNCaP cells (2 ⅹ105) were treated with 175 µM Abiraterone, and cell proliferation studies were performed at 24, 48 and 72 h using an annexin V-fluorescein isothiocyanate (FITC)-propidium iodide (PI) cell apoptosis kit (Invitrogen, Thermo Fisher Scientific, Inc.). The cells were washed twice with PBS buffer (pH 7.4) then resuspended in the binding buffer. Moreover, 5 µl PI and 5 µl Annexin V-FITC were mixed with the cells. The mixtures were incubated at 25°C for 15 min in the dark, and then the combinations were measured utilizing FACScan flow cytometry (FACSCalibur, BD).
Western blot analysis
Cell samples [untreated cell lines (LNCaP, PC-3, VCaP, and 22RV1), and cell lines (LNCaP and 22RV1) treated for 48 or 72 h with Abiraterone were homogenized in RIPA buffer (P0013B, Biyuntian Biotechnology Co. LTD, China), and a BCA protein kit (PL212989, Thermo, USA) was utilized to detect the protein concentration. Then, 4-12% Sodium dodecylsulphate polyacrylamide gel electrophoresis (SDS-PAGE) was used to separate cell lysates. Subsequently, the extracted proteins were transferred to polyvinylidene difluoride (PVDF) membranes (IPVH00010, Millipore, United States). The PVDF membrane was blocked with 5% skim milk in TBST at 37°C for 2 h, and then incubated with diluted primary rabbit and mouse monoclonal antibodies (1:1,000) overnight at 4°C, containing AR-V7 (ab198394, Abcam) and β-actin (Ab8226, Abcam), and then incubated with the secondary antibody [peroxidase AffiniPure Goat Anti-Rabbit IgG (H+L) (1:10 000, 111-035-045, Jackson] and Peroxidase AffiniPure Donkey Anti-Mouse IgG (H+L) (1: 5,000; 715-035-151, Jackson)] for 2 h at 37°C. The ECL (electro-chemiluminescence) substrate (SB-WB012, Share-bio, China) was used to detect protein bands via a chemiluminescence imaging system (4600, Tanon, China). During this experiment, β-actin was utilized as an internal control.
Real-time quantitative reverse transcription PCR (qRT-PCR)
The total RNA of LNCaP and 22RV1 cells (Abiraterone-treated at 48 or 72 h) was isolated from cell samples utilizing TRIzol (9109, TaKaRa, Japan). After that, 1 µg of the RNA was reverse-transcribed to cDNA. Based on provided protocols (7900HT FAST, ABI, USA), SYBR green-mediated RT-PCR amplification and real-time fluorescence detection were performed. The relative mRNA expression of each gene was normalized to GAPDH. Primers used included: AR-V7 (F):5′-CGGAAATGTTA TGAAGCAGGGATGA-3′, AR-V7 (R): 5′-CTGGTCATTTTGAGATGCTTGCAAT-3′; GAPDH (F): 5′-TGACAACTTTGGTATCGTGGAAGG-3′, GAPDH (R): 5′-AGGCAGGGATGATGTTCTGGAGAG-3′.
RNA isolation, sequencing, and data preprocessing
Total RNA from 22RV1 cells [a control group (Abiraterone-treated 0 h), the JY-48 h group (Abiraterone-treated for 48 h), and the JY-72 h group (Abiraterone-treated for 72 h), three samples in each group] was extracted as mentioned above. The RNA concentration and purity were performed utilizing a Nanodrop 2000 system (Thermo, USA), and the RNA integrity was electrophoresed onto a 1% agarose gel. Sequencing was performed using an Illumina Hiseq X-ten platform (Illumina, Inc., San Diego, CA, USA). The FASTQ sequence data were acquired from the RNA-sequencing data using Base Calling version 0.11.4. In addition, the FASTQ data quality control was performed utilizing the READ QC tool in FastQC version 0.11.4. Cutadapt version 1.16 was used for raw data trimming. Also, to acquire clean reads, the low-quality reads, including sequences with a quality score < 15, adaptor sequences, and sequences with an N number of raw reads > 1 were removed.
The clean reads were localized to the human reference genome (GENCODE, Homo_GRCh38) using HISAT2 version 2.1.0 [16]. FeatureCounts software version 1.6.0 [17] was used to acquire the read count information mapped on each gene, according to GENCODE [18], which provided human gene annotation information. The “protein-coding” annotation information was used as mRNA, and the annotation information of “antisense,” “sense_intronic,” “sense_overlapping,” “lincRNA,” “processed_transcript,” “bidirectional_promoter_lncRNA,” “non_coding” were used as lncRNA.
Expression level and Pearson correlation coefficient analysis of samples
The cor function in R language was employed to calculate the Pearson correlation coefficient (P) between each pair of samples. The prcomp algorithm in R language was utilized to conduct PCa of samples. Finally, the PCa diagram was performed using the Ggfortify package version 0.4.5.
Differential expression analysis
The voom function in the limma package [19] was employed to standardize the raw counts. The differentially expressed (DE)mRNAs and DE Long non-coding RNAs (lncRNAs) between the JY-48 h group and the control group, or the JY-72 h group and the control group were screened utilizing the limma package with the threshold of adj.P < 0.05 and a |log fold change (FC)| > 1. The Benjamini and Hochberg (BH) method was used to adjust the P-value for multiple comparisons. In addition, to obtain the AR-V7-related RNAs affecting Abiraterone treatment of CRPC, the DEmRNAs and DElncRNAs screened from the JY-48 h group, and the control group were intersected with DEmRNAs and DElncRNAs screened from the JY72 h group and the control group, and the overlapping DEmRNAs and DElncRNAs were obtained. Then, the overlapping DEmRNAs and DElncRNAs were removed from the DEmRNAs, and DElncRNAs screened from the JY-48 h group and the control group, and the remaining DEmRNAs and DElncRNAs were used for subsequent analyses. Moreover, the featured lncRNAs recorded in the LncBook database [20] were intersected with the other DElncRNAs, and the overlapping lncRNAs were obtained for subsequent analyses.
DEmRNA-DElncRNA co-expression network construction
The P of DEmRNAs and DElncRNAs was calculated based on mRNA and lncRNA data-matched samples. In addition, the BH method was performed to adjust the P values, and the DEmRNAs-DElncRNAs pairs were screened with the cutoff value of r > 0.95 and adj.P < 0.01. Then, the DEmRNA-DElncRNA co-expression networks were constructed utilizing Cytoscape software [21]. Moreover, “prostate cancer” was used as a keyword to search disease-related lncRNAs based on the LncBook database, and the disease-related lncRNAs were marked in the DEmRNA-DElncRNA co-expression networks. Also, “prostate neoplasms, castration-resistant” were used as keywords to search disease-related RNAs based on the CTD database [22] (http://ctdbase.org/) and the GeneCards [23] (https://www.genecards.org/) database, and the disease-related RNAs were marked in the DEmRNA-DElncRNA co-expression networks.
Functional enrichment analysis
Both Gene Ontology (GO), and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analyses were conducted on the target genes of lncRNAs utilizing the Metascape online tool [24] based on the DEmRNAs-DElncRNAs pairs. The parameters were set to min overlap = 3, P = 0.05, min enrichment = 1.5. Based on the mRNAs of the DEmRNAs-DElncRNAs pairs, the DAVID tool [25] was used to perform GO and KEGG pathway enrichment analyses with a cutoff value of P < 0.05 and count ≥ 2.
Protein-protein interaction network building
The Search Tool for the Retrieval of Interacting Genes (STRING) database [26] was performed to analyze the interactions between protein and protein encoded by mRNAs in the DEmRNA-DElncRNA co-expression network. The PPI score was set as 0.4 (referred to as the median confidence). Afterward, the PPI network was built using Cytoscape software. The CytoNCA plugin [27] was used to analyze the degree of nodes in the network, and the parameter was set to “without weight.”
lncRNA-miRNA-mRNA network construction
According to the mRNAs in the DEmRNA-DElncRNA co-expression network, the miRTarBase database [28] was utilized to predict microRNAs (miRNAs), and the miRNA-mRNA pairs were obtained. The DIANA-LncBase database [29] was employed to predict miRNAs based on lncRNAs in the DEmRNA-DElncRNA co-expression network, and the lncRNA-miRNA pairs were obtained. The lncRNA-miRNA-mRNA pairs were further identified according to the positive co-expression relationships between mRNA and lncRNA (r > 0.95), and the competing endogenous RNAs (ceRNA) network was built.
Survival analysis
Based on the lncRNAs and mRNAs in the ceRNA network, the survival analysis was performed utilizing the GEPIA2 online tool [30] (http://gepia2.cancer-pku.cn/#survival), and the statistical significance level was defined as P < 0.05.
Statistical Analysis
The data from each assay were shown as the mean ± standard deviation. One-way analysis of variance (ANOVA), t-tests, and Newman-Keuls multiple comparison tests were performed to assess the statistical significance between groups. All data were analyzed using SPSS 17.0 software and GraphPad Prism. A P < 0.05 was regarded as significant.