Fludarabine treatment favors the retention of miR-485-3p by prostate cancer cells: implications for survival

To build a reliable in vitro model to study miRNA release, we verified whether the DU-145 prostate cancer cell line spontaneously releases miRNAs into the growth medium. To address this question we considered 5 prostate cancer secretory miRNAs (PCS-miRNAs) and 5 secretory miRNAs (S-miRNAs) representative respectively of the miRNAs overrepresented or not overrepresented in the plasma/serum of patients with prostate cancer [3, 14, 16]. We found that in comparison to the immortalized prostate epithelial cell line PNT1-A, all PCS-miRNAs (with the exceptions of miR-21) were more abundant in the growth medium of DU-145 cells, unlike S-miRNAs (Figure 1A). These data indicate that DU-145 cells were able to release the same miRNAs that are overrepresented in the plasma of prostate tumor patients.

We then asked if the release of miRNAs was influenced by their endogenous expression levels. We quantified the intracellular levels of our panel of miRNAs and found elevated intracellular levels of almost all PCS-miRNAs in comparison to PNT1A cells (Figure 1B). The dot plot of the intracellular versus the extracellular levels showed a positive correlation for many PCS-miRNAs whereas this correlation was not observed for S-miRNAs (Figure 1C) suggesting a positive relationship between the intracellular and the extracellular levels for PCS-miRNAs.

We investigated the release of miRNAs after the exposure of DU-145 cells to a cytotoxic drug. We selected fludarabine, a drug which we have already used to induce cytotoxicity in tumor cells, as representative of cytotoxic drugs [19]. DU-145 cells were exposed to increasing concentrations of fludarabine for 48 h and from the dose-response curve we selected 10 μg/ml fludarabine as at this concentration cell proliferation was inhibited (Figure 2A). The cell-cycle analysis of DU-145 treated cells showed a blockage of cells in S and a strong depletion of G2 cells (Figure 2B). A slight but significant increase of apoptotic cells was also found (Figure 2C). The finding that after fludarabine treatment a substantial fraction (25%) of DU-145 cells was able to form colonies (Figure 2D) suggested that there were many surviving cells. Therefore, at the end of 48 h treatment we collected both attached cells and growth medium and measured the levels of our sets of miRNAs and calculated fold change of expression in treated versus untreated samples. The dot plot of the intracellular versus the extracellular fold changes showed that, while S-miRNAs had little variations after fludarabine treatment, almost all PCS-miRNAs were up regulated and released into the growth medium with the exception of miR-485-3p, which was less released by tumor cells despite its intracellular up regulation (Figure 2E). Interestingly, decreasing the concentration of fludarabine to 1 μg/ml, a concentration which reduced cell proliferation by 50% (Figure 2A), miR-485-3p was still retained (Figure 2F).

In order to define if miR-485-3p retention was common in prostate cancer cell lines, we extended the analysis of retention/release of miR-485-3p to PC-3, a prostate cancer androgen resistant cell line, and 22Rv1 and LNCaP, two androgen sensitive prostate cancer cell lines. Following the treatment with a concentration of fludarabine able to inhibit cell proliferation, we found that miR-485 was released in the growth medium of PC-3, 22Rv1 and LNCaP (Additional file 1: Figure S1A, B, C respectively), suggesting that the induction of release/retention is a process dependent on the genetic background. Then we asked whether the process was drug specific. We treated DU-145 cells with taxotere, a drug used for the treatment of prostate cancer. The dot plot of intracellular versus extracellular fold change of miRNAs levels showed that taxotere did not affected miR-485-3p retention/release (Additional file 1: Figure S1D) introducing the concept that a signature interaction between drug and genetic background exists.

To address if the release of PCS-miRNAs enhanced by fludarabine occurs through an active mechanism, we detected their loading on exosomes. We isolated exosomes from the growth medium of DU-145 cells treated (EXO-flu⁺) or untreated (EXO-flu^-) with fludarabine. Once we had verified that the amounts of EXO-flu⁺ and EXO-flu^- were similar and that exosomes were functionally active (Additional file 2: Figure S2), we determined the fold change of PCS-miRNAs and S-miRNAs in EXO-flu⁺ compared to EXO-flu^-. We found that fludarabine favored the association of almost all PCS-miRNAs and S-miRNAs to EXO-flu⁺ compared to EXO-flu^- with the exception of miR-485-3p (Figure 3A). MiR-485-3p was released less by treated cells and less loaded on EXO-flu⁺ than on EXO-flu^- suggesting that fludarabine affects the association of miR-485-3p with exosomes.

To support this hypothesis, DU-145 cells were transfected with either miR-485-3p or miR-100 and the expression of miRNAs was measured in the exosomes released by transfected cells treated (miR-485-3p/EXO-flu⁺ or miR-100/EXO-flu⁺) or untreated (miR-485-3p/EXO-flu^- or miR-100/EXO-flu^-) with fludarabine. Firstly, we observed that the two miRNAs were highly loaded on exosomes isolated from transfected cells (Figure 3B). When transfected cells were treated with fludarabine, we found that miR-485-3p/EXO-flu⁺ were less loaded with miR-485-3p than miR-485-3p/EXO-flu^- whereas miR-100/EXO-flu⁺ were loaded with miR-100 to a similar extent to miR-100/EXO-flu^- (Figure 3C). Since the association of miR-485-3p with exosomes was efficient in untreated cells but inhibited in fludarabine surviving cells, the possible explanations are that fludarabine alter the availability of miR-485-3p for the export or, alternatively, that fludarabine selected cells with a reduced ability to associate miR-485-3p with exosomes. In both cases the final result is that surviving cells were those enriched for miR-485-3p.

To investigate whether cells harboring high levels of miR-485-3p were induced or selected by fludarabine, we monitored the expression of miR-485-3p and its indirect target multidrug resistance gene 1 (MDR1) for 144 h (6 days) in fludarabine surviving cells. At the beginning of the treatment (t=0) the expression levels of miR-485-3p and MDR1 were respectively 2.5 and 5 fold higher in treated than in untreated cells and the high expression levels were stably maintained to late time points (Figure 3D). Due to the stable up regulation, it seems that fludarabine selected cells with a high intracellular level of miR-485-3p. To clarify this apparent relationship DU-145 cells were first transfected with an antagomiRNA specific for miR-485-3p and then treated with fludarabine. We found that cells depleted of miR-485-3p were more sensitive to fludarabine (Figure 3E) thus establishing that miR-485-3p was a modulator of the sensitivity to fludarabine.

To investigate the mechanism(s) that mediate(s) the reduction of sensitivity to fludarabine, we looked at the molecular targets of miR-485-3p. Among them we focused on the β subunit of nuclear factor-Y (NF-YB), which is post-transcriptionally inhibited by miR-485-3p and is a transcriptional repressor of topoisomerase IIα (Top2α) [20] (Figure 4A). The immunoblot analysis showed that NF-YB was down regulated in fludarabine-treated cells (Figure 4B) and accordingly Top2α was up regulated at both the RNA (Figure 4C) and protein levels (Figure 4D).

Interestingly two other genes, cyclin B2 (CCNB2) [21] and MDR1 [22], have been reported to be repressed by NF-YB when associated with p53 (Figure 4E) [21]. Co-immunoprecipitation experiments demonstrated that NF-YB and p53 interact in DU-145 cells (Figure 4F). In accordance with this, we observed the down regulation of NF-YB and the up regulation of CCNB2/MDR1 (Figure 4G) in fludarabine-treated cells. A possible scenario is that cells harboring high levels of miR-485-3p were able to proliferate as the down regulation of the transcription factor NF-YB triggered the transcription of multiple pro-survival genes.

miRNeasy mini kit, QuantiTect Reverse Trascription Kit, QuantiTect SYBER Green PCR kit, miScript Reverse Transcription Kit, miScript SYBR Green PCR Kit, Polyfect Transfection Reagent (QIAGEN, Hilden, Germany); RPMI 1640 medium, Ham’s nutrient mixture F-12 (EuroClone); crystal violet, anti α-tubulin, anti FLAG M2, Hoechst 33258 (Sigma-Aldrich Corporation, Seelze, Germany); ECL, Hybond-C extra membranes (GE Healthcare, Cleveland, Ohio); anti Top2α (BD Biosciences, San Jose, California); anti NF-YB (GeneSpin, Milan, Italy); anti-CD63, HRP-conjugated secondary antibodies (Santa Cruz Biotechnology, Dallas, Texas); anti p53 (DakoCytomation, Glostrup, Denmark); TMRE, DiIC18 (Invitrogen, Life Technologies, Carlsbad, California); Dynabeads Protein G (Novex, Life Technologies, Carlsbad, California); fludarabine (Teva, Petah Tikva, Israel); taxotere (Taxotere, Sanofi Aventis, Milan, Italy); single stranded cel-miR-39, miR-485-3p, miR-100, miR-NC, d-485-3p, d-NC (GenePharma, Shanghai, China).

DU-145, PNT1A, 22Rv1 and LNCaP cell lines were grown in RPMI 1640 medium. PC-3 were grown in Ham’s F-12. In all media 10% FBS, 1% penicillin/streptomycin (2 mM) and 1% L-glutammine (2 mM) were added to all media. Cells were incubated at 37°C in a humidified atmosphere containing 6% CO₂. For fludarabine treatment, cells were grown for 24 h and treated with 10 μg/ml for 48 h (DU-145 and PC-3 cells) or 30 μg/ml for 48 h (22Rv1 and LNCaP). For taxotere treatment, cells were grown for 24 h and treated with 10nM taxotere for 48 h.

Approximately 2×10⁵ cells were seeded in each well of a 6-well plate. After 24 h, cells were transfected with Polyfect Transfection Reagent according to the manufacturer instructions. We transfected the following oligonucleotides: either miR-485-3p (5′-GUCAUACACGGCUCUCCUCUCU-3′; 5′-AGAGAGGAGAGCCGUGUA UUUC-3′) or miR-100 (5′-AACCCGUAGAUCCGAACUUGU-3′; 5′-CAAGUUCGGAUCUACG GAUUAU-3′); antagomiRNA-485-3p (d-485-3p) (5′-AGAGAGGAGAGCCGUGUAUGACT-3′). As control we used either miR-NC (5′-UUCUCCGAACGUGUCACGUTT-3′; 5′-ACGUGAC ACGUUCGGAGAATT-3′) or antagomiRNA-NC (d-NC) (5 –CAGUACUUUUGUGUAGUAC AA-3′). After 6H hours from transfection, we added fresh growth medium and treated cells with fludarabine at 10 μg/ml.

Total RNA was extracted from 200 μl growth medium using the miRNEasy mini kit following the manufacturer recommendations (Purification of RNA from serum or plasma, Qiagen supplementary Protocol). According to the protocol, 5 μM cel-miR-39 (5′ -UCACCGGGUGUAAAUCAGCUUG- 3′) was added after the Qiazol step of the extraction protocol.

Cell proliferation was measured as follows: 2×10⁵ cells per 6 well plate were seeded and after the treatment with fludarabine or treatment plus transfection, cells were fixed in 2% paraformaldehyde in PBS and subsequently stained with 0.1 % crystal violet dissolved in 20% methanol and let dry at room temperature. Cells were then lysed with 10% acetic acid and the optical density (OD 590 nm) of the solution, detected with Plate Reader apparatus (SpectraCount, Packard), was used to measure cell proliferation.

At specified time points cells were stained with propidium iodide and the cell cycle analyzed using a FACScalibur cytofluorimeter (BD Biosciences, San Jose, California). The percentage of cells in each phase was determined and the fold change of treated in comparison to untreated cells was calculated.

Apoptosis was detected using the cationic dye TMRE, a fluorescent probe that measures the loss of mitochondrial membrane potential associated with apoptosis [36]. Briefly, DU-145 treated or untreated cells were harvested, incubated with 100 nM TMRE for 40 min at 37°C in the dark and then analyzed by flow cytometry. The percentage of apoptotic cells was measured from the SSC density plots obtained.

Cells exposed to fludarabine for 48 h were collected, counted and seeded at cell density of 200 cells/100 mm diameter culture dish to allow colony formation. After 10-12 days colonies were detected by staining dishes with 0.1 % crystal violet. Plating efficiency (no colonies/no seeded cells) of treated cells was normalized to that of the untreated cells.

Total RNA was extracted from 1×10⁶ cells using the miRNeasy mini kit following the manufacturer’s recommendations. To quantify multidrug resistance gene 1 (MDR1), Top2α and cyclin B2 (CCNB2), 1μg of total RNA was reverse transcribed using the QuantiTect Reverse Trascription Kit. Real-time PCR (qRT-PCR) was carried out with Rotor-Gene Q 2-Plex (Qiagen, Hilden, Germany) using the QuantiTect SYBR Green PCR kit. To quantify miRNAs, 1 μg of total RNA or 5-10 μl of RNA extracted from 200 μl growth medium were retrotranscribed with miScript Reverse Transcription Kit and qRT-PCR was carried out using the miScript SYBR Green PCR Kit. All reactions were performed in triplicate. Relative quantification of mRNAs and miRNAs expression was calculated with the fit point method. Transcript values were normalized to those obtained from the amplification of the internal control (glyceraldehyde 3-phosphate dehydrogenase (GAPDH) for mRNAs, U6 for intracellular miRNA and cel-miR-39 for extracellular miRNAs). The oligonucleotide sequences are reported in (Additional file 3: Figure S3).

Equivalent amounts of proteins were resolved on SDS-PAGE gels at different acrylamide percentages and transferred to Hybond-C extra membranes by electroblotting. The resulting blots were blocked with 5% nonfat dry milk or BSA solution in TBST. Anti α-tubulin (1:20000), anti CD63 (1:500), anti Top2α (1:1000), anti NF-YB (1:1000) and anti p53 (1:5000) primary antibodies were used. Incubation was performed overnight at 4°C and bands were revealed after incubation with the recommended secondary antibody coupled to peroxidase using ECL. Scanned images were quantified using OptiQuant software and normalized to α-tubulin.

1 mg of protein extracted from either DU-145 or DU-145 fludarabine treated cells were used in a coimmunoprecipitation assay. Sample were immunoprecipitated with 4 μg of either anti NF-YB or anti FLAG (as a control for aspecific immunoprecipitation) using Dynabeads Protein G following the manufacturers recommendations and separated on SDS-PAGE gels as described in the section Western Blot analysis.

Exosomes were isolated according to Wang [37], with some modifications. Briefly, growth media were centrifuged at 2,000 × g for 30 min to remove cells and debris. The supernatant was transferred to a new tube and centrifuged at 10,000 × g for 45 min to further remove cells and debris. The supernatant was filtered through a 0.2 μm filter to remove particles larger than 200 nm (mainly microvesicles). Exosomes were then pelleted from microvesicles-depleted supernatant by centrifuging at 100,000 × g for 2 h. Exosomes were washed with PBS and resuspended in 50 μl PBS.