S-Adenosylmethionine Treatment Diminishes the Proliferation of Castration-Resistant Prostate Cancer Cells by Modulating the Expression of miRNAs

Cell culture and treatment with SAM

2.1.

PC-3 cells were grown in a RPMI 1640 medium supplemented with 10% fetal calf serum and 1.2% penicillin/streptomycin as described previously (Schmidt et al. 2016) (PAN-Biotech GmbH, Aidenbach, Germany). For RNA analysis, cells were grown to a subconfluence of 70–80%. The cells were treated with vehicle (0.005 M H₂SO₄ and 10% Ethanol) or 200 μM of SAM (32 mM prepared in 5 mM sulfuric acid and 10% ethanol; New England Biolabs, Ipswich, MA, USA) for 120 h. Under both conditions, the medium was changed once per 24 h. Cells were trypsinized and counted. A number of 1 Mio cells was used for the isolation of RNA.

Isolation of total RNA, library preparation, and miRNA-seq

2.2.

The total RNA was isolated from PC-3 cells using a NucleoSpin miRNA Kit Macherey Nagel (Düren, Germany) according to the manufacturer’s instructions. By following ligation of the RNA to modified 3′ and 5′ adaptors, the products were reverse transcribed (Super Script III; Thermo Fisher Scientific, Waltham, MA, USA), purified (SPRIselect reagent kit, Beckman Coulter, Brea CA, USA), and PCR amplified for 14 cycles (KAPA HiFi HotStart PCR Kit, KAPA Biosystems, Wilmington, MA, USA). After size selection by polyacrylamide gel electrophoresis, miRNAs containing flanking p5 and p7 adapters were sequenced on a HiSeq 2000 instrument (Illumina; library preparation and sequencing were performed by GenXPro GmbH, Frankfurt, Germany). The results were uploaded to EMBL Array Express. The accession number is E-MTAB-14388.

Reverse transcription-quantitative PCR (RT–qPCR)

2.3.

The total RNA was reverse transcribed into cDNA using Super Script III (Thermo Fisher Scientific) according to the manufacturer’s instructions. qPCR was subsequently performed using a TaqMan miRNA assay kit Applied Biosystems (Waltham MA, USA); (Thermo Fisher Scientific) according to the manufacturer’s protocol to amplify miRNAs on an ABI 7500 Real-Time PCR system (Applied Biosystems; Thermo Fisher Scientific). The thermocycling reactions were performed in the following three steps: initial denaturation at 95°C for 5 min, 40 cycles of 95°C for 10 s, and 60°C for 30 s; and the following primer pairs for qPCR are as follows: miR-192-5p forward, 5′-GGACTTTCTTCATTCACA CCG-3′; reverse, 5′-GACCACTGAGGTTAGAGCCA-3′; and U6 forward, 5′-TCGCTTCGGCAGCACATATACT-3′ and reverse, 5′-ACGCTTCGGGAATTTGCGTGTC-3′. StarD13 forward, 5′-ACTGTCTGTGGTGGGAAACA-3′; StarD13 reverse, TGAGGCACACTTGTTCAACG-3′; GAPDH forward, 5′-TCAAGAAGGTGGTGAAGCAGG-3′; and GAPDH reverse, 5′-TCAAAGGTGGAGGAGTGGGT-3′. The expression levels were quantified using the 2^−ΔΔCt method, miRNA expression was normalized to that of U6, and mRNA expression was normalized to that of GAPDH. The reactions were performed in triplicate for each sample, with at least three independent runs.

miRNA inhibitor transfection

2.4.

PC-3 cells were seeded in 6-well plates. Transfection of the miR-192-5p antagomir (CUGACCUAUGAAUUGACAGCC) and the corresponding negative control miRNA (miR-NC, UUUGUACUACACAAAAGUACUG) were purchased from Thermo Fisher Scientific Company. Transfection into PC-3 cells was performed using Lipofectamine 3000 in 250 mL of Opti-MeM (Reduced Serum Medium) (Invitrogen, Waltham MA, USA) according to the manufacturer’s instructions. Then, the cells were trypsinized and counted for subsequent analyses of hsa-microRNA-192-5p expression or for use in cell proliferation assays.

Cell proliferation assay

2.5.

To assess the proliferation of transfected or treated PC-3 cells, the CellTiter 96^® AQ_ueous Nonradioactive Cell Proliferation Assay (MTS) was used (Promega, Mannheim, Germany). Transfected or treated PC-3 cells were seeded in 96-well plates (2000 cells per well), after which MTS was added, and the absorbance was measured after 24 h, 48 h, and 120 h. At least three independent experiments were performed in triplicate. The data are expressed as percentages of proliferation, where the proliferation rates for the control were arbitrarily set to 100%.

Fluorogenic caspase activity assay

2.6.

PC-3 cells were washed with phosphate buffered saline. Cell lysis buffer was added (#70108, Cell Signaling Technology, Danvers, MA, USA), and cell lysates were collected in Eppendorf cups. Subsequently, the caspase activity assay (Caspase-3 Activity Assay Kit #5723, Cell Signaling Technology) was performed according to the manufacturer’s instructions. Fluorescence was measured using a plate reader (excitation wavelength: 490 nm; emission wavelength: 525 nm) and expressed in relative fluorescence units.

Data analysis – quantification of miRNA expression by omiRas

2.7.

The analysis of the miRNA-seq libraries was conducted using omiRas (Müller et al. 2013). After removing adapters and low-quality reads, the remaining reads were summarized as tags, and singletons were excluded from the dataset. Subsequently, the tags were mapped to the human genome (hg19) with bowtie (Li and Durbin 2009). The annotation of tags and mapping to the hg19 were performed using various databases of coding and ncRNAs retrieved from the UCSC Table Browser. ncRNAs that are not mapped to exonic regions of coding genes were quantified in each library. By following normalization of the read number for each tag to the number of mapping loci, the differential expression analysis was performed between treated samples and controls for each miRNA using the DEGseq bioconductor package (Wang et al. 2010). A clustered heatmap was generated using the heatmap package of R software (version 3.5.0). A volcano plot was created using the R package ggplot.

Target prediction and identification of miRNA–mRNA pairs

2.8.

To identify miRNA-correlated genes, a two-step approach was used. First, three miRNA databases were used to predict potential mRNA targets of miRNAs (miRDB, TargetScan, and miRanda). Second, Pearson correlation coefficients of the intersecting miRNA–mRNA pairs were calculated. (R < ‒0.7 and p < 0.05) for the downstream analysis. The results intersected with previously published RNA-seq data (GSE71070).

Gene ontology (GO) term and KEGG pathway enrichment analyses

2.9.

GO and KEGG pathway enrichment analyses of the DEGs were performed using the DAVID bioinformatics resource (Dennis et al. 2003). We used a false discovery rate (FDR) <0.01 as screening threshold to determine the significance of the functions and pathways, and the “GO KEGG bar dot” function of the bioinformatics analysis software “SR plot” (Tang et al. 2023) was used to visualize the GO terms and KEGG enrichment.

miRNA prediction and expression

3.1.

We analyzed the miRNA-seq dataset of PC-3 cell samples treated with SAM and untreated controls by means of omiRas. We detected 17 differentially expressed mature miRNAs that satisfied the criterion of an FDR-corrected P < 0.005. The set of miRNAs included a total of seven upregulated miRNAs in the treated samples (Figure 1a), many of them with tumor suppressor function, for example, hsa-miR-4454 and hsa-miR-503. In contrast, we detected 10 downregulated miRNAs, displaying mainly oncogenic functions (Figure 1a), such as hsa-miR-429 and also hsa-miR-9-5p, which is highly expressed in prostate cancer cells under normal conditions. The volcano plot in Figure 1b displays the differential expression of upregulated and downregulated miRNAs between SAM-treated samples and controls, whereas the heatmap in Figure 1a clearly shows that the hierarchical clustering analysis performed with omiRas completely discriminated the two groups (SAM-treated and control samples) as well as those upregulated from downregulated samples. Target identification of the respective miRNAs was performed using three different miRNA databases (miRDB, TargetScan, and miRanda) to predict potential mRNA targets of miRNAs. We selected only those miRNAs that were found to be predicted in at least two of the three databases. In total, we found 2775 predicted mRNA targets for upregulated miRNAs (Figure 2a) and 5384 predicted target genes for downregulated miRNAs (Figure 2b).

Fig 1.

Fig 2.

Transcripts regulated by miRNAs

3.2.

Based on the results of the target identification, we concentrated especially on those potential transcript-miRNA interactions in which the differential expression of transcripts and miRNAs among the treated samples and controls exhibited the opposite trend, i.e., when transcripts were upregulated in a particular comparison, the miRNA was downregulated, and this was the opposite. Hence, 92 target transcripts were downregulated (Figure 2a), 32 of which exhibited oncogenic functions, while the expression of the mature miRNA was upregulated in the treated samples compared with the control samples. Among those downregulated transcripts, we identified, for example, transforming growth factor beta 2 (TGFB2) (Figure 2a), which is known to promote cancer progression in later stages of prostate cancer. In contrast, 41 targets were upregulated (Figure 2b) in the treated samples compared with the control samples, while the mature miRNAs were downregulated. Eleven of those upregulated genes are known to act as tumor suppressor genes. Among the upregulated genes, SERPINE1 (Figure 2b), which encodes plasminogen activator inhibitor-1 (PAI-1) and is known to be increased in prostate cancer cell lines as well as in tumor samples of patients (Kubala and De Clerck 2019), was detected. PAI-1 usually inhibits the urokinase plasminogen activator (uPA), which has been implicated in tumor cell invasion, survival, and metastasis in a variety of cancer entities, including prostate cancer (Kubala and De Clerck 2019). Since an increase in StarD13 gene expression in the transcriptome could not be detected despite the significant upregulation of hsa-miR-9-5p, we performed real-time PCR for StarD13 in SAM-treated PC-3 cells and ultimately found significant upregulation of its expression (211% of the untreated control) (Figure 2c).

Downregulation of hsa-miR-192-5p diminishes the proliferation of PC-3 cells

3.3.

To confirm the impact of SAM on the differential expression of miRNAs in PC-3 cells, we carried out quantitative PCR on PC-3 cells treated with SAM. We used primers against hsa-miR-192-5p since it is known to be overexpressed and promote cell proliferation in prostate cancer cells (Chen et al. 2019b). SAM treatment caused significant downregulation of hsa-miR-192-5p expression after 48 h and 120 h (Figure 3a). After transfection of cells with an inhibitor against hsa-miR-192-5p, the expression significantly decreased after 24 h, 48 h, and 120 h (Figure 3a). A combination of both treatments even increased the downregulation of hsa-miR-192-5p expression (Figure 3a). We also examined if the downregulation of hsa-miR-192-5p causes downregulation of proliferation. After 48 h and 120 h, the proliferation of the samples transfected with the hsa-miR-192-5p agomir significantly decreased (Figure 3b). A combination of SAM and the hsa-miR-192-5p agomir had the most significant downregulatory effect on PC-3 cell proliferation after 120 h (decrease of 77% compared with that of the untreated controls) (Figure 3b). To clarify if the results of the MTS assay are linked to a loss in cell proliferation or to cell death, a fluorogenic caspase-3 activity assay was performed, wherein we could detect a significantly induced caspase-3 activity in PC-3 cells transfected with the hsa-miR-192-5p agomir after 120 h of cell culture (Figure 3c). Again, a combination of SAM-treatment and transfection with the hsa-miR-192-5p agomir had the strongest effect after 120 h of cell culture (Figure 3c). We conclude that downregulation of hsa-miR-192-5p may induce apoptosis in PC-3 cells.

Fig 3.

GO and KEGG pathway enrichment analysis

3.4.

GO term enrichment and KEGG pathway analysis of the differentially expressed genes were performed by comparing the treatment group (treated with SAM) and the control group. Among the significantly enriched GO terms for the downregulated transcripts, “cellular response for bone morphogenic protein (BMP) stimulus” (GO:0071773), “response to BMP” (GO:0071772), “regulation of BMP signaling pathway” (GO:0030510), “BMP signaling pathway” (GO:0030509), “positive regulation of epithelial cell migration” (GO:0010634), “regulation of epithelial cell migration” (GO:0010632), “regulation of cellular response to growth factor stimulus” (GO:0090287), and “negative regulation of cytoskeleton organization” (GO:0051494) were found in the BP category (Figure 4a and Table S2). In terms of molecular functions, “insulin-like growth factor I binding” (GO:0031994), “transcription coregulator activity” (GO:0003712), “fibroblast growth factor binding” (GO:0017134), and “transforming growth factor beta receptor binding” (GO:0005160) were found to be enriched significantly in the set of downregulated mRNAs (Figure 4a and Table S2); whereas, for the cellular component (CC) category, we detected enrichment in the set of downregulated transcripts for “actin-based cell projection” (GO:0098858), “actomyosin” (GO:0042641), and “cell–cell junction” (GO:0005911) (Figure 4a and Table S3). The KEGG pathway analysis revealed particular enrichment of miRNAs in cancer and the TGF-β signaling pathway (Figure 4a and Table S4). For the TGF-β signaling pathway, five downregulated transcripts, i.e., FST, SMAD6, TGFB2, INHBB, and BMPR2, were detected (Figure 4c).

Fig 4.

Considering upregulated gene expression, we detected, for example, significant enrichment for “cell–cell adhesion via plasma-membrane adhesion molecules” (GO:0098742), “cell killing” (GO:0001906), “negative regulation of cytokine production” (GO:0001818), and “positive regulation of cell killing” (GO:0031343) in the BP category (Figure 4b and Table S5); “growth factor binding” (GO:0031343), “extracellular matrix structural constituent” (GO:0005201), and “cytokine activity” (GO:0005201) in the molecular functions category (Figure 4b and Table S6); and “plasma membrane bounded cell projection cytoplasm” (GO:0032838), “cytoplasmic region” (GO:0099568), and “collagen-containing extracellular matrix” (GO:0062023) in the CC category (Figure 4b and Table S7). The KEGG pathway analysis of the upregulated transcripts revealed enrichment of pathways involved in “cell adhesion molecules” and “cytokine–cytokine receptor interaction” (Figure 4b and Table S8).