A series of 111 formalin-fixed paraffin-embedded (FFPE) prostatectomy specimens with primary prostate adenocarcinoma (PCa) from the Department of Pathology, Ruhr University Bochum, diagnosed between 2009 and 2011 were collected for the study. The composition of the cohort concerning Gleason grade, age, PSA levels and stage is shown in Table I. The PCa specimens were subjected to histological examination by an expert pathologist for confirmation of the Gleason grading, WHO classification of the tumour and staging according to the tumour-node-metastasis system. No patients had distant metastases at the time of diagnosis. The study was approved by the local Ethics Committee (Votum no. 3991-11).

The histopathological tumour regions of interest were micro-dissected with the corresponding HE slide as a guide (tumour was marked with a pen on the HE slide) and used for RNA extraction using the miRNA FFPE kit (Qiagen, Hilden, Germany), following the manufacturer’s instructions.

For the detection of miRNA levels, 500 ng total RNA was transcribed using the miScript reverse transcription kit (Qiagen); qRT-PCR was carried out with 25 ng cDNA per reaction using the miScript SYBR green PCR system (Qiagen) with the miR-205 primer (#MS00003780) on an Eppendorf Realplex4 cycler (Eppendorf, Hamburg, Germany). RNU6B (U6 small nuclear RNA 2, #MS00014000) was used for the normalization to total RNA (23). For the detection of BCL2 in FFPE samples and cell culture, RNA was transcribed to cDNA using the High Capacity cDNA system (Life Technologies, Darmstadt, Germany). qRT-PCR was carried out with the TaqMan system (primer set #Hs99999018_m1) with 2X Universal Master Mix (Life Technologies). Template cDNA (25 ng) was used per PCR. 18S rRNA (#Hs99999901_s1) was used as internal control.

The prostate cancer cell lines PC3 and LnCap and the osteosarcoma cell line U2-OS were obtained from ATCC (Manassas, VA, USA). PC3 and LnCap cells were cultivated in RPMI-1640 (Pan Biotech, Aidenach, Germany) and U2-OS was cultivated in DMEM high glucose (Pan Biotech), both supplemented with 10% fetal calf serum (Hyclone-Thermo Scientific, Bonn, Germany), 100 U/ml penicillin and 100 μg/ml streptomycin (PAN Biotech); cisplatin and doxorubicin were obtained from Sigma-Aldrich (Taufkirchen, Germany).

PC3 cells were seeded at 5×10⁵ per well and LnCap were seeded at 8×10⁵ cells per well in 6-well plates in their normal medium and transfected in suspension with 10 nM miRNA mimics (Qiagen) or AllStars negative control siRNA, labelled with AlexaFluor-647 (Qiagen), using 6 μl HiPerFect (Qiagen). The transfection efficiency was determined by measuring the fluorescence of the AllStars oligos by flow cytometry and was between 90 and 96%. Whenever the assay required, the cells were treated with cisplatin or doxorubicin 72 h after transfection.

Fragments of the 3′ untranslated region (UTR) of E2F1 (262 bp), E2F5 (389 bp) and BCL2 (438 bp) containing the potential miR-205 binding sites were cloned into pGL3 as 3′UTR of the firefly luciferase gene by PCR. U2-OS cells were transiently transfected in triplicates with 40 ng of the pGL3 construct, 10 ng of the expression vector for renilla luciferase pRL-CMV (Promega, Mannheim, Germany) and 10 nmol of miRNA mimics (Qiagen), using the transfection reagent Attractene according to the manufacturer’s instructions (Qiagen). Cells were harvested 30 h after transfection and activities of firefly and renilla luciferase were measured using the Dual-Luciferase reporter assay system (Promega). The resulting luminescence was measured with a Tecan M200 microplate reader (Tecan, Crailsheim, Germany).

Cells were plated at 10,000 (PC3) or 5,000 (LnCap) per well in 24-well plates. One, three, five and seven days after plating, respectively, 100 μl of 5 mg/ml MTT (thiazolyl blue tetrazolium bromide, Carl Roth, Karlsruhe, Germany) in PBS was added per well; cells were lysed after 4 h by addition of 250 μl triplex solution (10% SDS; 5% isobutanol, 0.012 M HCl). Absorbance was measured at 562 nm.

Caspase activities were measured using the Caspase Glo 3/7 assay (Promega), following the manufacturer’s instructions. Values were corrected for differences in cell numbers by doing an MTT assay in parallel.

The mitochondrial membrane potential was determined in LnCap by JC-1 (Axxora, Lörrach, Germany) staining. JC-1 was added to the cells at a final concentration of 5 μg/ml, incubated for 10 min at 37°C, after which the cells were trypsinized and washed extensively with PBS. In PC3, the mitochondrial membrane potential was instead measured using the FlowCellect™ MitoPotential Red kit (Millipore, Schwalbach, Germany), following the manufacturer’s instructions. In both cases, treatments were performed in triplicates; fluorescence was measured on a Guava easyCyte 8HT flow cytometer (Millipore).

For the determination of the sub-G1 fraction, prostate carcinoma cells were transfected with miR-205 mimics or AllStars control oligos, trypsinized, plated at 5×10⁴ cells per well in 6-well plates in triplicates the next day and treated with doxorubicin or cisplatin 48 h after transfection. Forty-eight hours after treatment with doxorubicin and cisplatin, respectively, cells were harvested by trypsinisation and combined with the floating cells. Cells were fixed in 70% ice-cold ethanol and stained in 0.1% Triton X-100 (Carl Roth), 0.5 mg/ml RNaseA (Sigma-Aldrich) and 60 μg/ml propidium iodide (MP Biochemicals, Illkirch, France) in PBS. The sub-G1 fraction of propidium iodide stained cells were measured on a Guava easyCyte 8HT flow cytometer (Millipore).

Cells were harvested by scraping into ice-cold RIPA buffer (1% NP40, 0.5% sodium deoxycholate, 0.1% SDS, 150 mM NaCl, 50 mM Tris-HCl pH 8.0) containing EDTA-free complete protease inhibitors (Roche, Mannheim, Germany). Protein concentration was measured using the BCA assay (Thermo Scientific) according to the manufacturer’s instructions. Incubation with primary antibodies anti-BCL2 (Cell Signaling, Frankfurt, Germany) or anti-α-tubulin (clone DM1A, Thermo Fisher, Bonn, Germany) was done overnight at 4°C. Blots were incubated with horseradish peroxidase labelled secondary antibodies (Cell Signaling) and the signal was detected using ECL (Thermo Scientific). Changes in protein expression were quantified using ImageJ Software (V.1.43u; NIH; USA). By normalizing integrated density values (IDV) of the protein bands of interest were normalized to the corresponding tubulin IDVs.

Comparisons between pairs of values were done using the t-test; the Welch t-test was applied if the F-test indicated a significant difference between the variances. Statistical inferences were derived from single experiments. For statistical testing on expression levels of miR-205 and BCL2, data were log2 transformed, to allow the use of normal distribution-based tests and Pearson’s correlation coefficient. For post hoc testing between individual pairs of conditions, the Scheffé test was applied. Statistical analysis was carried out by the Statistica 10 program (StatSoft, Tulsa, OK, USA).

We determined the expression of miR-205 by qRT-PCR in 111 FFPE specimens of prostate carcinoma. The miR-205 expression was found to be strongly reduced in 102 of 111 analyzed prostate cancer samples, with a median relative expression level of 0.160 in comparison to benign tissue from the same patient (interquartile range, IQR, 0.085–0.350). MiR-205 expression levels decreased with increasing Gleason score from 7a=3+4 to 8=4+4, but Gleason score 6=3+3 samples showed lower expression than Gleason score 7a samples (median expression levels: Gleason 3+6, 0.168; Gleason 7a, 0.303; Gleason 7b, 0.150; Gleason 8, 0.087; p=0.050 by ANOVA on log transformed data) (Fig. 1A). Its expression was also higher in tumours confined to the prostate (pT2) than in those that extend beyond the prostate (pT3) (p=0.0425). However, differences in miR-205 expression between perineural invasion categories were not significant (Fig. 1A). Only 9 of 111 samples showed an increased expression of miRNA 205 in tumour tissue in comparison to the benign controls; these were all relatively small tumours (<4 cm³). The inverse correlation between miR-205 expression and tumour size was significant (r=−0.234, p=0.0087) (Fig. 1B).

The reduced expression of miR-205 in prostate cancer indicates that its targets may aid tumour progression. From the targets suggested by the miRGen program (www.diana.pcbi.upenn.edu/miRGen/html), we selected BCL2 as a novel candidate. BCL2 appears interesting in this context, as its overexpression in prostate cancer is a known marker for poor prognosis (24).

In the ∼5-kb long 3′UTR, 3 potential miR-205 binding sites are found, one weak (6-mer + a) centrally located at position 2380 and two potentially stronger sites near the 3′ end of the 3′UTR at positions 3906 (7-mer-m8) and 4094 (7-mer-1a) (Fig. 2A), both with a ΔG above −13.40 kCal/mol (25). For a random sequence of this length, a ΔG of −13.40 kCal/mol would be expected (26). Additionally, these sites are located relatively close together and near the 3′ end of a long 3′UTR, which are also typical characteristics of functional miRNA binding sites (27).

The fragment containing these two most 3′ terminal sites was cloned into the pGL3 vector as 3′UTR for the luciferase gene (28). A reporter assay was done in U2-OS cells with miR-205 or control oligos to test for reduced luciferase activity after binding the microRNA. The osteosarcoma cell line U2-OS was used for this assay due to its ease of transfection and the stability of the assay in this cell line. As positive controls, equivalent constructs containing miR-205 binding sites in E2F1 and E2F5 were used. In comparison to the control oligos, miR-205 caused a 34% reduction in the luciferase signal when the BCL2 UTR was tested. In comparison, the E2F1 UTR caused a 51% percent reduction in signal intensity and the E2F5 UTR only 6%. The decrease in luminescence was significant for both BCL2 and E2F1 (p<0.01, t-test). The control vector did not show significantly reduced luciferase activity in the presence of miR-205 (Fig. 2B).

Next, we analyzed if overexpression of miR-205 leads to a decrease in the BCL2 protein levels of the prostate cancer cell line PC3. We observed a decrease in the BCL2 level by western blotting at 72 and 96 h after transfection with miR-205 mimics, in comparison those transfected with control oligos (Fig. 2C).

Having confirmed BCL2 as a miR-205 target, we determined the expression of BCL2 at the mRNA level by TaqMan assay in the series of archival samples for which miR-205 levels were measured by qRT-PCR. The quality of the RNA was sufficient for reliable PCR results in 84 of 111 samples. The mRNA levels of BCL2 have been found to correlate well with its protein level (29), allowing for the use of qRT-PCR to measure BCL2 expression. BCL2 is strongly expressed in the basal cells of benign glands (24,30), making the comparison to benign tissue difficult for this gene. Therefore, we calculated BCL2 expression levels as ΔCt relative to 18S levels in tumour tissue. BCL2 mRNA expression increased significantly with increasing Gleason grade (median expression relative to 18S: Gleason 3+3, 0.0074% (n=12); Gleason 3+4, 0.0148% (n=33); Gleason 4+3, 0.022% (n=31); Gleason 4+4, 0.100% (n=8); p=0.173. Expression was also stronger in tumours with extension beyond the prostate (pT3, median, 0.0488%) in comparison to those confined to the prostate (pT2, median, 0.0148%) (p=0.0399). There was no significant difference between those with (Pn1, median, 0.0172%) and without perineural invasion (Pn0, median, 0.0118%) (Fig. 3A). BCL2 levels did not correlate to tumour size either.

Although there was no linear correlation between miR-205 and BCL2 expression, all the samples except one with high BCL2 expression, had low (<30% of benign) miR-205 expression (Fig. 3B).

Having confirmed anti-apoptotic gene BCL2 as a miR-205 target, we investigated the effect of miR-205 overexpression on the induction of cell death in prostate cancer cells. We measured caspase 3/7 activity in cells transfected with miR-205 mimics after treatment for 24 h with 0.1 μg/ml doxorubicin (0.1724 μM), 4 and 10 μM cisplatin (PC3) or 10 and 20 μM (LnCap) cisplatin. A consistent increase in caspase activity was observed in cells transfected with miR-205 in comparison to control oligos in both cell lines in combination with cisplatin treatment (Fig. 4A). Cisplatin (4 μM) does not induce significant cell death in LnCap cells, therefore higher doses were used in this cell line (Fig. 4B and data not shown). Doxorubicin only caused a significant difference in PC3 cells in this assay.

As a marker for apoptosis, we determined the sub-G1 fraction by flow cytometry 48 h after genotoxic treatment. We observed a significant increase in the fraction of apoptotic cells both in control cells and cells treated with doxorubicin or 4 μM cisplatin (Fig. 4C) in PC3. LnCap shows the strongest increase in apoptosis at 10 μM cisplatin concentration in combination with miRNA205 mimics. In these cells there was no increase in the spontaneous apoptosis rate after miR-205 transfection, in contrast to PC3. Apoptosis is also characterized by loss of mitochondrial membrane potential (ΔΨM), which can be measured through a change in the fluorescence spectrum of the JC-1 dye. Loss of ΔΨM was significantly increased in LnCap cells transfected with miR-205 after treatment with 10 μM cisplatin, as compared to control transfected cells, confirming the results for caspase 3/7 activation (Fig. 4B).

We next determined the effect of miR-205 overexpression on cell growth. For this, PC3 and LnCap cells were transfected with either AllStars negative control siRNA or miR-205 mimics and then plated out at low densities. An MTT assay was performed at the starting point and at 3, 5 and 7 days thereafter. We observed a significantly lower cell number in the miR-205 transfected cells (p<0.05), in comparison to control transfected cells at 7 days (Fig. 5) in both cell lines.