Cell lines and clinical samples

The prostate cancer cell line LNCaP was obtained from American Type Cell Collection (ATCC, Manassas, VA, USA) and DuCaP was kindly provided by Dr Jack Schalken (Radboud University Nijmegen Medical Center, Nijmegen, the Netherlands). LNCaP cells with deletion of the promoter and 1st and 2nd exon of EPCART (EPCART-del; clones del-4 and del-56) and their wild-type control (WT) were previously established by our group [11]. To construct the EPCART overexpression plasmid, the sequence for the EPCART transcript (exons 2–5) was synthesized with additional restriction sites (NheI and XhoI) and then added to the pcDNA3.1(+) plasmid (Invitrogen) by GenScript (Piscataway, New Jersey, USA). Either the pcDNA3.1(+) empty expression vector (Invitrogen) or pcDNA3.1(+) containing EPCART was transfected into LNCaP cells with Lipofectamine 3000 transfection reagent (Invitrogen) according to the manufacturer’s instructions. A stable cell pool was selected with 400 μg/ml geneticin (G418; Invitrogen) for several weeks, after which a lower geneticin concentration (200 μg/ml) was used for stable cell line maintenance. EPCART expression was determined by quantitative reverse transcription PCR (qRT-PCR). All cell lines were cultured as recommended by the suppliers and tested for mycoplasma contamination regularly.

A formalin-fixed paraffin-embedded (FFPE) tissue specimen of an untreated primary PCa (n = 1) for RNA in situ hybridization studies, fresh-frozen tissue samples of untreated primary PCas (n = 2) for RACE, and 171 prostate tissue microarray (TMA) samples of untreated primary PCas (n = 111) and locally recurrent CRPCs (n = 60) for IHC analysis were obtained from Tampere University Hospital (Tampere, Finland).

Data acquisition and analysis

RNA-seq data curated from different cancer and tissue types were retrieved from MiTranscriptome catalog [8]. RNA-seq data from TCGA-PRAD samples [15] were retrieved from the Genomic Data Commons Data Portal (https://portal.gdc.cancer.gov/) and analyzed as previously described [11]. Clinical data and protein array data from TCGA-PRAD samples were retrieved from cBioPortal (https://www.cbioportal.org/) [16,17,18]. For PDCD4 expression analyses, Taylor et al. [19] whole transcript expression data for human primary and metastatic PCa samples were retrieved from GSE21034, proteome quantification data for primary PCa and localized CRPC (called Tampere PC cohort) from Latonen et al. [20]. RNA-seq normalized expression data for the same samples from Annala et al. [21], and whole proteome quantification data for primary PCa and metastatic CRPC from Iglesias-Gato et al. [22].

qRT-PCR analysis

RNA was extracted from EPCART-del and WT cells with TRIzol Reagent (Invitrogen) according to the manufacturer’s instructions. RNA was converted to cDNA by random hexamer primers (Thermo Scientific) and Maxima reverse transcriptase (Thermo Scientific) following the manufacturers’ instructions. Quantitative PCR was performed by either CFX Opus 96, CFX96, or CFX384 real-time PCR detection system (Bio-Rad). Primer sequences are listed in Supplementary Table 1.

Relative expression values were calculated from quantification cycle (Cq) values, and the target gene measurements were normalized to reference gene (e.g., TBP) values and averaged. Relative gene expression changes were calculated using the 2−ΔΔCq method. 2ΔCq values were used to calculate the significance between each pair (e.g., deletion clone vs. WT).

5′ and 3′ end determination

RACE was performed using the SMARTer RACE 5′/3′ Kit (Takara Bio) according to the manufacturer’s instructions. RNA from two fresh-frozen primary PC tissue samples was extracted as previously described [9]. RACE PCR products were obtained using the supplied primers and the appropriate gene-specific primers listed in Supplementary Table 1 and separated on a 1.2% agarose gel. Different sized gel products were extracted with NucleoSpin Gel and PCR Clean-Up Kit (Macherey-Nagel). 5′ RACE products were cloned into pRACE vectors by In-Fhusion HD cloning kit (Takara Bio) and purified by NucleoSpin Plasmid Mini kit (Macherey-Nagel) following manufacturers’ instructions. Purified 5′ RACE-vectors and 3′ RACE PCR-products were sequenced bidirectionally by Sanger sequencing using kit’s universal, gene-specific, or M13 primers (Supplementary Table 1). The Sanger sequencing was performed using BigDye® Terminator v3.1 Cycle Sequencing Kit (Applied Biosystems) and 3500xL Genetic Analyzer (Applied Biosystems) according to the manufacturer’s instructions.

RNA in situ hybridization

EPCART localization in PCa tissues was studied by RNA in situ hybridization. FFPE tissue sections were treated using ViewRNA™ ISH Tissue 2-Plex Assay (Affymetrix) according to the manufacturer’s instructions for a 1-plex assay, as only Fast Red was used for detection. First, slides were briefly deparaffinized in xylene and dehydrated in 100% ethanol. Sections were then pretreated and boiled, and a target probe for EPCART (VA1-19503, Affymetrix) and signal amplifiers were hybridized using the ThermoBrite System (Leica Biosystems). A probe for human housekeeping genes (GAPDH (glyceraldehyde 3-phosphate dehydrogenase), ACTB (actin beta), and PPIB (peptidyl-prolyl cis-trans isomerase B); VA1-15726, Affymetrix) was used as a positive control, and a probe for dihydrodipicolinate reductase (dapB; VF1-11712, Affymetrix) was used as a negative control in every assay. Signal detection was performed using Fast Red substrate. Slides were counterstained with Gill’s hematoxylin (Sigma-Aldrich). Finally, slides were mounted, first with ImmunoHistoMount (Sigma-Aldrich), and secondly with organic mounting medium. Slides were scanned with Aperio ScanScope XT scanner (Aperio Technologies, Inc.), and imaged at a higher resolution under LSM780 Laser Scanning Confocal Microscope (Zeiss).

Cellular fractionation

EPCART localization was studied in subcellular fractions in LNCaP and DuCaP cells. Nuclear and cytoplasmic RNA was extracted with SurePrep Nuclear or Cytoplasmic RNA Purification Kit (Fisher BioReagents) following the manufacturer’s instructions. Expression of EPCART, cytoplasmic control (GAPDH), and nuclear control (U1) were analyzed by qRT-PCR. Primer sequences are listed in Supplementary Table 1. RNA content in subcellular fractions was calculated as % of transcript abundance = \(^}(})+}(})]}\times 100\), where total RNA abundance is a sum of nuclear and cytoplasmic fractions.

RNA-sequencing

For RNA-seq of EPCART-deleted and WT clones, three biological replicates were used. Cells were grown for 48 h in a normal medium. RNA was isolated using Trizol (Invitrogen, Thermo Fisher Scientific), treated with RNase-free DNase set (Qiagen), and purified by Monarch RNA Cleanup Kit (New England Biolabs) according to manufacturers’ protocols. The purified RNA was quantified by Qubit 4 Fluorometer (Invitrogen, Thermo Fisher Scientific) and Qubit RNA Broad Range Assay Kit (Invitrogen, Thermo Fisher Scientific), and its purity was assessed by the 260 nm/280 nm ratio. RNA integrity was checked using the 5300 Fragment Analyzer System (Agilent Technologies).

Library preparation was performed using standard polyA enrichment and strand-specific library protocol. Sequencing was performed with a Novaseq6000 (Illumina) in Novogene (Hong Kong, China) for 150 bp paired-­end reads. On average, 105 million cleaned reads per sample were obtained in strand-specific RNA-seq.

RNA-seq alignment, expression quantification, and differential expression analysis

Read quality of strand-specific RNA-seq data from EPCART-del and WT samples were assessed with FastQC v0.11.8 (https://www.bioinformatics.babraham.ac.uk/projects/fastqc/) and reads were aligned to GRCh38 using STAR v2.71a [23] followed by indexing with Samtools v1.8 [24]. Read counts of protein-coding transcripts were calculated using BEDTools v. 2.27.1 sub-command multicov [25] and GENCODE v.38 annotation was used for gene calls. For visualization and clustering of the RNA-seq data, we used variance stabilizing transformation counts calculated by DESeq2 (version 1.22.2) [26], and PCA plots were generated. Differential expression analysis between EPCART-del and WT clones was performed using the DESeq2 R package (version 1.22.2), where p-values were attained by the Wald test and corrected for multiple testing using the Benjamini and Hochberg method [26].

Pathway analysis for RNA-seq data

Differentially expressed protein-coding genes (p < 0.05) from RNA-seq data of EPCART-del and WT clones were analyzed by Ingenuity Pathway Analysis (Qiagen). Canonical Pathways were filtered only to show Signaling Pathways for further analysis for each sample pair (del-4 vs. WT or del-56 vs. WT). Comparison Analysis of del-4 vs. WT and del-56 vs. WT was performed for functions and diseases; only molecular and cellular functions were filtered to be shown for further analysis. IPA uses the p-value of overlap, calculated using the right-tailed Fisher’s exact test, to identify significant pathways. The overall activation/inhibition states of Canonical Pathways are predicted based on a z-score algorithm. Z-scores that are greater than or equal to 2 represent predictions of activation, while predictions of inhibition are made for z-scores less than or equal to −2. Log p-values > 1.3 (=p < 0.05) are considered as significant.

RNA stability assay

Experiments were performed in biological triplicates. Cells from WT clones were pretreated for 2 h with CHX (at 100 µg/ml) prior to the addition of ActD (at 5 µg/ml) to block transcription. Control cells were treated identically, except that no CHX was added. Samples were taken at 0 and 6 hours of ActD treatment, and the latter was normalized to the former. As there is no transcription in the presence of ActD, the decrease in RNA level between 0 and 6 h is indicative of the degradation rate of that RNA. RNA extraction, cDNA synthesis, and qRT-PCR were carried out as above. Relative gene expression changes were calculated using the 2−ΔΔCq method and GAPDH as a reference gene.

Western blot

Protein samples from cell lysates were prepared as previously described [11]. Proteins were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) on Mini-PROTEAN TGX Precast Protein Gels (Bio-Rad) and transferred to PVDF membrane (Immobilon-P; Milli-pore). Primary antibodies against target proteins (Supplementary Table 2) were used and detected by anti-mouse HRP-conjugated antibodies produced in rabbit (dilution 1:3000; DAKO) or by anti-rabbit HRP-conjugated antibodies produced in swine (dilution 1:5000; DAKO) and Clarity Western ECL Substrate (Bio-Rad) or SuperSignal West Femto Maximum Sensitivity Substrate (Thermo Scientific) with ChemiDoc MP Imaging System (Bio-Rad).

Protein bands were quantified with ImageJ by calculating the relative amounts as a ratio of each protein band relative to the lane’s loading control. Relative protein levels were calculated as fold changes for each pair (e.g., deletion clone/WT) and used for graphs. Quantified ratios were used to calculate the significance between each pair (e.g., deletion clone vs. WT).

Polysome profiling

Three biological replicates of EPCART-del and WT clones were lysed separately for polysome profiling. Each cell lysate was prepared from two 90% confluent 150 mm plates, according to McGlincy et al. [27] with minor changes. Briefly, the complete cell growth medium was changed 2 h prior to harvesting and placed back in a +37 °C/CO2 incubator. Then, plates with cells were placed on ice and washed with 8 mL of ice-cold PBS supplemented with 100 µg/mL CHX. PBS was removed and plates were floated in liquid N2 to snap-freeze cells. While still frozen, 400 µL of freshly prepared lysis buffer (20 mM Tris, pH 7.4; 150 mM NaCl, 10 mM MgCl2, 1% Triton X-100, 1 mM dithiothreitol (DTT), 10 U/mL DNaseI, 100 µg/mL CHX) was dripped onto each plate. Cells were scraped from plates and let slowly melt on ice. Lysates were triturated 10 times through a 26 G needle and clarified by centrifugation for 10 min at 10,000×g, +4 °C. The RNA concentration in lysates was measured by a Qubit Broad Range kit (Thermo Fisher Scientific).

In total, 150 µg of lysate was layered onto 10–50% sucrose gradient prepared in polysome buffer (20 mM Tris, pH 7.4; 150 mM NaCl, 10 mM MgCl2, 1 mM DTT, 100 µg/mL CHX) and centrifuged at 35 000 rpm (209627.4 × g) for 3 h, +4 °C in TH641 rotor (Sorvall). Gradients were fractionated into 15 × 750 µL fractions using an automated piston fractionator (Biocomp) with dual-wavelength A260/A280 detection flow cell.

RNA was extracted from polysome fractions with TRIzol LS Reagent (Invitrogen) according to the manufacturer’s instructions with the following changes: 150 µL of chloroform was used; phase separation was performed at 14,000 × g for 10 min; 15 µg of GlycoBlue Coprecipitant (Invitrogen) was added to aqueous phase; RNA precipitation was performed at 18,000 × g for 30 min. The same volume of RNA solution from each fraction was used for cDNA synthesis. cDNA synthesis and qRT-PCR were carried out as described above. The arithmetic mean of Cq values was calculated for the three technical replicates of each sample, and the RNA percentage for each fraction was calculated as \(}=2}^}}_}/^}}_}+^}}_}+\ldots +^}}_})\times 100\), where x = the number for the fraction that is calculated and y = the number for the total number of fractions.

Sample preparation for mass spectrometry

Five replicate samples from each clone (WT, del-4, del-56) were prepared for MS analysis. Cell pellets (~1 × 106 cells/sample) brought up in cold RIPA lysis buffer with 1% Halt protease inhibitor cocktail (Thermo Scientific) were lysed using ultrasonication for 5 min and incubated for 25 min on ice. The clear supernatant of the cell lysate was collected by centrifugation, avoiding the cell debris, and the total protein concentrations were measured with Bio-Rad DC Protein Assay (Bio-Rad). In total, 50 µg of protein were precipitated with cold acetone, and the precipitate was collected by centrifugation and dissolved in 2% SDS (Sigma-Aldrich) in 50 mM triethylammonium bicarbonate (TEAB) (Honeywell). Protein cysteine disulfide bond reduction was performed with a reducing agent to a final concentration of 3 mM tris-(2-carboxyethyl)-phosphine (Sigma-Aldrich), incubating for 1 h at +60 °C. Samples were transferred to 30 kDa molecular weight cut-off filters (Pall Laboratory), flushed with 8 M urea in 50 mM Tris-HCl (Sigma-Aldrich), and subsequent alkylation of the free reduced cysteine thiols was performed by incubation in dark for 20 min to a final concentration of 50 mM iodoacetamide (Sigma-Aldrich). The protein samples were rinsed multiple times with aliquots of 8 M urea buffer and 50 mM TEAB after which, TPCK-treated trypsin (Sciex; trypsin to protein ratio 1:25) was used to digest the proteins for 16 h at +37 °C. After multiple rinses with aliquots of 50 mM TEAB, peptides were eluted from the filter with 0.5 M sodium chloride (Sigma-Aldrich) and dried in a vacuum centrifuge. The peptide samples were reconstituted in 0.1% trifluoroacetic acid (TFA) and cleaned and desalted with C18 tips (Thermo Scientific). Tips were washed with 2.5% acetonitrile (ACN), 0.1% trifluoroacetic acid and the peptides were eluted from the tips with 80% ACN, 0.1% formic acid (FA) and dried in vacuum centrifuge to be stored for future use. For the MS analysis, the peptide samples were resuspended in 2% ACN, 0.1% FA to a 1.5 µg/µL concentration.

Mass spectrometry analysis, protein identification, and quantification

Quadrupole time-of-flight mass spectrometer TripleTOF5600 (AB Sciex) coupled to an Eksigent 425 Nano LC system and an Eksigent nano flex cHiPLC system, with Nanospray III electrospray interface (AB Sciex) was used for analysis. 3 µg of the peptide sample was loaded onto a trap column (cHiPLC® ChromXP C18-CL, 3 µm particle size, 120 Å, 75 µm i.d × 5 mm) and loading and desalting were carried out with loading solvent: 2% ACN and 0.1% FA at a 2 µl/min flow rate for 10 min. Consecutively, the trap column was switched to be in line with the reversed phased analytical nano cHiPLC column (cHiPLC® ChromXP C18-CL, 3 µm particle size, 120 Å, 75 µm i.d × 15 cm). The peptide separation was performed using a 120-min gradient of mobile phases A and B, where A is 0.1% FA, 1% ACN in water and B is 0.1% FA in ACN at a 300 nl/min flow rate. The eluted peptides were electro-sprayed into the mass spectrometer via a fused silica emitter (New Objective).

Data dependent acquisition (DDA) method was implemented to generate MS data used to create a spectral library. All 15 samples were used to generate this spectral library containing 2,47,249 spectra, 25,144 peptides from 2519 proteins (at FDR 1%) by searching against the Swiss-Prot human database (canonical 20,370 genes) in the Protein Pilot software® 4.5 (AB Sciex). All 15 samples (two runs/sample) were then rerun again on the same instrument using the same LC conditions, with a data-independent (SWATH) acquisition mode to acquire protein quantification data. Retention time normalization was carried out using 6 peptides, each of the two highest-score proteins (HSPD1, HSPA8). Altogether 2083 proteins were quantified (at FDR 1%) for each sample processing against the spectral library using the PeakView® (AB Sciex) and MarkerView® software.

Immunohistochemical analysis

PDCD4 protein expression levels in prostate carcinomas were validated by IHC analysis from FFPE TMA samples. IHC staining of PDCD4 was performed by Ventana BenchMark GX IHC/ISH system (Ventana Medical Systems, Roche), ultraView Universal Dab Detection Kit (Roche), and anti-PDCD4 antibody (EPR3431, Abcam) in 1:4000 dilution. Slides were scanned with NanoZoomer S60 Digital slide scanner (Hamamatsu Photonics) with a 20× objective.

Nuclear and cytoplasmic staining intensities of PDCD4 were classified on a scale from 0 to 3 with negative (0), weak (1), moderate (2), or strong (3) staining within cancerous areas. If possible, a minimum of 200 cells were calculated for each sample. The Histoscore (H-score) was calculated as H-score = (0 × percentage of cells with absent cytoplasmic staining) + (1 × percentage of “1+” cells) + (2 × percentage of “2+” cells) + (3 × percentage of “3+” cells).

Statistical analyses

Mann–Whitney U tests were used to analyze the association between PTEN-sample groups in TCGA-PRAD data. Unpaired two-tailed Student’s t tests were used to calculate the significance between control and experimental conditions in PCR and immunoblot experiments. P-values < 0.05 were considered statistically significant.

Kaplan–Meier survival analysis and log-rank tests for TCGA-PRAD and Taylor et al. data were used to determine the progression-free survival between samples divided by their first quartile expression. For correlation analysis in Tampere PCa and TCGA-PRAD cohorts, Spearman’s rank correlation coefficient was calculated for EPCART and PDCD4 expressions in a pairwise manner.

The protein quantification data from EPCART-del and WT clones were log2-transformed, and the replicate MS analyses were combined by taking means. The coefficient of variation (CV) was calculated for samples originating from the same sample type and passage in order to identify and exclude quantified proteins with poor repeatability (CV ≥ 30%). Due to the small number of samples processed, only descriptive analysis and results are reported for individual proteins. These include means by sample types and the associated log2 fold changes (log2FC) between different sample types. Proteins with log2FC > log2(1.5) and <log2(0.67) were included in the pathway analyses. R software (v4.1.2, R Core Team) was used to process the data and perform the descriptive analyses.