miR-380-5p-mediated repression of TEP1 and TSPYL5 interferes with telomerase activity and favours the emergence of an “ALT-like” phenotype in diffuse malignant peritoneal mesothelioma cells

miR-380-5p synthetic precursor (pre-miR-380-5p) and the negative control oligomer (preNeg) were purchased as Pre-miR™ miRNA precursor molecules (Thermo Fisher Scientific). miR-380-5p inhibitor (LNA380) and the relative control oligomer (LNANeg) were purchased as miRCURY LNA™ microRNA Inhibitors (Exiqon A/S, Vedbaek, Denmark). Commercially available small interfering RNAs (siRNA) targeting TEP1 (siTEP1), TSPYL5 (siTSPYL5), p53 (sip53) and control siRNAs (siCTR) were obtained from Santa Cruz Biotechnology Inc. (Dallas, TX, USA). Target protector oligomers were purchased as miScript Target Protectors (Qiagen, Hilden, Germany). Doxorubicin hydrochloride was from Sigma-Aldrich (Milano, Italy). The reagents were dissolved in nuclease-free, sterile water, stored at −20 °C and diluted at the appropriate working concentration just before use.

DMPM (STO, MP8 and MP4) cells [10], A549 lung cancer cells (ATCC®-CCL-185™), U-2 Os osteosarcoma cells (ATCC® HTB-96™) and human adult mesothelial MES-F cells (Zen-Bio Inc., Research Triangle Park, NC) were maintained in the logarithmic growth phase in Dulbecco’s modified Eagle’s (DMEM), DMEM-F12 (Lonza Milano S.r.l, Treviglio, Italy) and MSO-1 (Zen-Bio Inc.) media, respectively, supplemented with foetal bovine serum, at 37 °C in a humidified incubator at 5% CO₂. Cells were periodically monitored for DNA profile by short tandem repeats analysis using the AmpFISTR Identifiler PCR amplification kit (Thermo Fisher Scientific, Monza, Italy).

For the transfection procedures, cells seeded at the appropriate density and allowed to attach for 72 h (DMPM cells) or 24 h (A549, U-2 Os, MES-F) were incubated with Lipofectamine-2000^TM (Thermo Fisher) in Opti-MEM I (Thermo Fisher). After 15 min at r.t., lipid-treated cells were added with Opti-MEM I medium containing the oligomers at the final concentrations of 20 nM (pre-miR-380-5p; preNeg), 100 nM (LNA380; LNA-Neg) or 50 nM (siTEP1, siTSPYL5 and sip53), and incubated for 4 h at 37 °C. For target protection experiments, 100 nM of custom designed TEP1 or TSPYL5 target protectors was transfected in combination with miR-380-5p precursor or preNeg, under the same transfection conditions described above. Cells were then harvested according to the timeline of each experiment and subsequently analysed. For the long-term assessment of miR-380-5p effects, the transfection procedure was reiterated every 7 days, up to 3 months. For the assessment of cell growth, cells were collected at different time intervals and their number and viability were evaluated by a dye exclusion test (0.4% Trypan Blue Solution, Sigma-Aldrich) using a TC20^TM automated cell counter (Bio-Rad Laboratories S.r.l., Segrate, Italy). For doxorubicin-based experiments, cells were treated with 3 μg/ml of the drug, preliminary defined as a sub-toxic concentration, 24 h after the transfection with sip53 or siCTR and cultured for additional 48 h. Non-transfected (NT) cells were run in parallel.

Total RNA was isolated using Qiagen RNeasy Mini kit (Qiagen, Hilden, Germany) and digested with 20 U RNase-free DNase. miR-380-5p expression levels were analysed using TaqMan® microRNA Assay (P/N 4427975, Applied Biosystems, Foster City, CA, USA), according to the manufacturer’s instruction. Gene expression analysis was carried on randomly primed total RNA using the High Capacity cDNA Reverse Transcription kit (Applied Biosystems). Each gene was amplified using the following TaqMan® Gene expression assays (Applied Biosystems): ATRX (Hs00230877); TEP1 (Hs00200091); and TSPYL5 (Hs00603217). Amplifications were run on the 7900HT Fast Real-Time PCR System (Applied Biosystems). Data were analysed by SDS 2.2.2 software (Applied Biosystems) and, if not otherwise specified, reported as relative quantity (RQ or Log₁₀(RQ)) with respect to a calibrator sample using the 2^−ΔΔCt method, where Ct represents the threshold cycle. RNU48 snRNA (P/N 4427975) and RNaseP (P/N 4316844) were used as normalisers during the analysis of miR-380-5p and mRNAs expression levels, respectively.

Forty micrograms of total protein extracts prepared according to standard methods was fractioned by SDS-PAGE, transferred onto nitrocellulose filters and incubated o.n. with primary antibodies: mouse monoclonal antibodies raised against β–actin, PAX8, TNKS, WRAP53; rabbit polyclonal antibodies raised against H2AX, γ- H2AX (S139), p21^Waf1, POT1, PRKCA (Y124), TEP1 and a goat polyclonal antibody raised against TNKS2 (Abcam, Cambridge, UK); rabbit polyclonal antibodies raised against TERF2, TSPYL5 and the mouse monoclonal raised against p53 (Santa Cruz Biotechnology); rabbit polyclonal antibodies raised against PARP1 and CPP32 (Cell Signaling Technology, Danvers, MA, USA). The filters were then probed with secondary peroxidase-linked whole antibodies (GE Healthcare, Milano, Italy) and detected by Novex® ECL HRP Chemiluminescent detection system (Thermo Fisher). Filters were then subjected to autoradiography. β-Actin (ACTB) was used on each blot to ensure equal protein loading.

To test the effect of miR-380-5p on its predicted target, the pEZX-MT05 vector (GeneCopoeia Inc., Rockville, MD, USA), hosting the whole 3′-UTR of TSPYL5 (GeneBank accession NM_033512.2) downstream the reporter Gaussia luciferase (GLuc) gene and containing a second reporter gene (Secreted Alkaline Phosphatase, SEAP) for transfection-normalization, was used. A mutated reporter construct (del43-54) characterized by the deletion of the seed region within the predicted miR-380-5p binding site was generated using the QuikChange II site-directed mutagenesis kit (Agilent Technologies Italia S.p.A., Cernusco S/N, Italy), according to the manufacturer’s instructions, in the presence of the following primer pair: F_5′-ctagctgcatcacttatcagtggtttcatctgtatctctc-3′ and R_5′-gagagatacagatgaaaccactgataagtgatgcagctag-3′. To assess the luciferase activity, STO cells were transfected with gene reporter constructs, in the presence or absence of pre-miR-380-5p or preNeg using Lipofectamine3000^TM (Thermo Fisher), according to the manufacturer’s protocol and incubated at 37 °C. GLuc and SEAP enzymatic activities were then measured in the culture media collected at 24, 48 and 72 h after transfection using the Secrete-Pair^TM Dual Luminescence Assay Kit (GeneCopoeia Inc.), according to manufacturer’s instructions. For the co-transfection experiments, 24 h after the transfection with pEZX-MT05 and pre-miR-380-5p, cells were transfected with miRNA inhibitors (LNA380 or LNA-Neg) and assessed for luciferase activity 48 h later. PEZX-MT05-expressing cells transfected with single oligomers (preNeg, pre-miR-380-5p, LNA-Neg or LNA380) were run in parallel.

Cells (2 × 10⁷) were suspended in RIP buffer (10 mM Tris, pH 7.5; 10 mM KCl; 2 mM MgCl₂; 1 mM DTT, 100 U/ml RNase inhibitor and a protease inhibitor mixture) for 15 min, homogenized with a Dounce homogenizer and centrifuged at 13,000 rpm at 4 °C for 10 min. Five percent of each solution was set apart to be used as control (Input) for the RT-PCR amplification. The remaining was incubated with IgGA/IgG-coated beads for 1 h at r.t. and centrifuged at 13,000 rpm for 1 min. The supernatants were then mixed with 5 μg of a rat monoclonal anti-human Ago2 antibody (Sigma-Aldrich) and a rabbit polyclonal human IgG (Abcam). After an overnight incubation at 4 °C, samples were re-incubated with IgGA/IgG-coated beads for 1 h at 4 °C, centrifugated at 2500 rpm for 1 min and washed twice with RIP buffer containing 150 mM NaCl and 0.5% Nonidet P-40 and once with buffer containing 400 mM NaCl and 0.5% Nonidet P-40. Finally, the RNA was extracted from the AGO2- and IgG-IP complexes using TRIzol (Thermo Fisher), according to the manufacturer’s instruction. The enrichment for miRNA binding sites was evaluated by RT-PCR using the High Capacity cDNA Reverse Transcription kit (Applied Biosystems). Site #1 was amplified by one-step PCR at 95 °C for 5 min followed by 33 cycles [95 °C 45 s; 57 °C 30 s; 72 °C 30 s] and 72 °C for 5 min using the following forward (5′-cagtccccaaagaggaacaa-3′) and reverse (5′-tcatggtccggggtttatta-3′) primer pair. The fragment corresponding to site #2 was first amplified at 95 °C for 5 min followed by 20 cycles [95 °C 45 s; 57 °C 30 s; 72 °C 30 s] using the following forward (5′-cgtaccttggccttcaatgt-3′) and reverse (5′-cagagaccctgaccacacct-3′) primer pair. One tenth of the amplification reaction was then re-amplified for 22 additional cycles under the same conditions followed by an elongation step at 72 °C for 5 min. The reaction products were resolved by agarose gel electrophoresis. Images were captured by UVI_DOC HD5 (UVITec Ltd., Cambridge, UK) and subjected to densitometric analysis using ImageJ 1.46r. The fold-enrichment was calculated after normalization towards Input RNA as the ratio between the mRNA levels detected in AGO2-IP or IgG-IP samples and the mRNA levels in IgG-IP sample from non transfected cells.

Telomerase activity was measured on 1 μg of protein by the Telomeric Repeat Amplification protocol (TRAP) using the (TRAPeze® kit; Merck S.p.A., Vimodrone, Italy) and quantified according to the manufacturer’s instructions [6]. Telomere length analysis was carried out on Rsa I/Hinf I-digested total DNA upon pulsed-field gel electrophoresis and southern blot, as previously described [6]. DNA-Hind III digest (Bio-Rad Laboratories) was used for size determination of telomere restriction fragments (TRF). Mean telomere length (mTL) was calculated by densitometric analysis (ImageJ 1.46r) as reported by Kimura et al. [11]. C-circle DNA levels were assessed by the CC assay (CCA) [12]. Briefly, genomic DNA was incubated at 30 °C for 8 h in the presence or absence of 7.5 U of φ29 DNA polymerase (New England BioLabs), heat-inactivated at 65 °C for 20 min and dot-blotted. After UV-cross-linking, filters were hybridized with a ³²P-(CCCTAA)₃ probe using PerfectHyb Plus hybridization buffer (Sigma-Aldrich), washed, exposed to a PhosphoImager screen and scanned on a Typhoon 8600 Variable Mode Imager (GE Healthcare). Signal intensity was quantified using the VisionWorks LS software (UVP Ltd.). Genomic DNA from ALT-positive U-2 Os and telomerase-positive HeLa cells was included as positive and negative control, respectively.

If not otherwise specified, data have been reported as mean values ± s.d. from at least three independent experiments. Two-sided Student’s t test was used to analyse differences between samples. P values <0.05 were considered statistically significant.

To unveil possible regulators of TA in human cancer, we focused on miR-380-5p, which was found to be significantly under-expressed in TA-positive compared to TA-negative DMPM tissue specimens or to normal peritoneal specimens [5, 13] by miRNA expression profiling (GSE99362) and subsequent validation by real-time RT-PCR (Additional file 1: Supplementary methods and Figure S1).

Functional studies were carried out in TA-positive DMPM cells (STO, MP8 and MP4), which show negligible endogenous levels of the miRNA compared to human normal mesothelial (MES-F) and to TA-negative, ALT-positive cancer cells (Fig. 1b).

The ectopic expression of mature miR-380-5p (Fig. 1c; Additional file 2: Figure S2A) induced a mild, though significant, reduction of DMPM cell growth over time (Fig. 1d; Additional file 2: Figure S2B), whereas it did not impair the growth of either normal mesothelial or telomerase-negative cancer cells (Additional file 2: Figure S2C). miR-380-5p-mediated DMPM cell growth impairment was associated to a remarkable inhibition of TA (61 ± 1.7%, P < 0.01; Fig. 1e) as well as to the accumulation of p53 and p21^waf1 and the occurrence of apoptosis, as revealed by the accumulation of PARP1 and Caspase-3 cleavage products and of γ-H2AX (Fig. 1f). Moreover, miR-380-5p-dependent inhibition of TA was paralleled by a marked decrease in the expression levels of TEP1, which is responsible for the proper functioning of the holoenzyme (Fig. 1a) [14]—both at mRNA and protein level, in all three TA-positive DMPM cell lines but not in the TA-negative U-2 Os cells (Fig. 1g, h).

To test whether the specific downregulation of TEP1 could phenocopy the effects of miR-380-5p, RNAi-mediated depletion of TEP1 was performed in STO cells. Specifically, cells transfected with a siRNA targeting TEP1 (siTEP1) were characterized by a remarkable decrease in TEP1 mRNA and protein amounts (Fig. 2a), which was paralleled by a pronounced inhibition of TA (Fig. 2b) and a significant impairment of cell growth (Fig. 2c), similarly to what was observed at the same time point upon the reconstitution of miR-380-5p. The evidence that the selective depletion of TEP1 reproduced the biological effects observed in miR-380-5p-transfected cells hinted at the possibility that the protein could be a target of the miRNA.

By an in silico target prediction analysis using miRWalk 2.0 [15], we found a set of genes within the list of putative miR-380-5p targets (Additional file 3: Figure S3A) known for being involved in the maintenance of telomere structure/function (e.g. POT1, TERF2, TNKS, TNKS2) as well as in the regulation of the catalytic activity (TERT, WRAP53, PRKCA) or the expression levels (TP53, PAX8) of telomerase. However, no miR-380-5p binding sites were predicted within the 3′-untranslated region (UTR) of TEP1 mRNA (Additional file 3: Figure S3A). Moreover, western immunoblotting analyses revealed that, except for a mild reduction in the amounts of PAX8—a transcription factor that transactivates the promoter of both TERT and TERC—none of the predicted telomere/telomerase-associated targets was found to be downregulated upon the ectopic expression of miR-380-5p (Additional file 3: Figure S3B).

Nonetheless, two out of five prediction tools within miRWalk database returned the presence of putative miRNA binding sites located within the open reading frame (ORF) of TEP1 and two of them (i.e. site #1 and #2 at positions 4875-4896 and 5773-5794, respectively, GeneBank accession number NM_007110.4, Fig. 3a) were specifically identified by RNA22 v1.0 [16]. To check whether the two sites were recognized by miR-380-5p, their interaction with Argonaute 2 (AGO2), which plays a central role in miRNA-mediated RNA silencing process [17], was examined. Such an analysis revealed a marked enrichment for both target sequences in miR-380-5p- with respect to preNeg-transfected cells (Fig. 3b), with a markedly higher fold-increase for target site #1 vs. #2. Target protection experiments were carried out to further verify whether miR-380-5p-mediated downregulation of TEP1 was consequent to the direct miRNA-mRNA interaction at the predicted binding sites. In particular, miR-380-5p precursor was co-transfected with a mix of target protectors (TEP1 TPs), namely single-stranded RNAs designed to specifically bind the miRNA binding sites within TEP1 ORF, thus protecting them from being targeted by miR-380-5p (Fig. 3a). Results revealed a partial, though significant, rescue of TEP1 protein amounts in miR-380-5p-expressing cells in the presence with respect to the absence of TEP1 TPs (Fig. 3c). These data revealed that protection of TEP1 ORF is sufficient to counteract, at least in part, the inhibitory effect of miR-380-5p and, in keeping with target enrichment analyses, indicate that TEP1 is indeed a direct target of the miRNA.

Furthermore, a prominent increase in p53 protein levels was observed upon miR-380-5p reconstitution (Fig. 1f), despite the protein has been reported to be a direct target of the miRNA [18]. This evidence, led us hypothesizing that other than directly targeting the ORF of TEP1, miR-380-5p could also affect the protein function by inhibiting the expression of one or more intermediary factors, the downregulation of which may favour the accumulation of p53, which in turn could impinge on TEP1 [14].

To address this issue, we paid specific attention to testis-specific protein, Y-encoded-like 5 (TSPYL5) gene, a member of TSPY-L family of genes [19]. The gene was predicted by 5/5 algorithms (Additional file 3: Figure S3A) and, according to miRWalk 2.0, bears a highly probable (P < 0.0002) miR-380-5p binding site with a 12-nucleotide long seed within the 3′-UTR (Fig. 3d). In keeping with this observation, a reduction of TSPYL5 protein amounts was appreciable in miR-380-5p expressing cells (Fig. 3e). Consequently, reporter gene assays were carried out to check if the gene could be a direct target of the miRNA. Specifically, a significant and time-dependent inhibition (P < 0.01) of reporter gene activity was observed upon the ectopic expression of miR-380-5p in cells transfected with the reporter construct bearing the wild-type compared to the deleted form of TSPYL5 3′UTR (Fig. 3d, f). Though apparently modest, the extent of miR-380-5p-mediated inhibition of luciferase activity is consistent with the presence of a single predicted miRNA binding site within the 3′-UTR of the target gene [20] and suggests that TSPYL5 is amenable of targeting by the miRNA. This evidence gained further support also by data showing that miRNA-mediated inhibition of reporter gene activity was significantly counteracted by the co-transfection of miR-380-5p with LNA-380, an inhibitor of the mature form of the miRNA (Fig. 3g).

It has been recently documented that, through its physical interaction with the ubiquitin specific protease 7, TSPYL5 leads to p53 protein level reduction and, consequently, to the impairment of p53-target gene regulation [19]. As such an event may have a role in regulating TA [14, 21], we envisaged a molecular circuitry by which miR-380-5p-mediated downregulation of TSPYL5 triggers p53 accumulation, which in turn may impinge on TEP1 function and may result in TA inhibition (Fig. 4a). The existence of such miR-380-5p-TSPYL5-TEP1 axis was confirmed in the TA-positive A549 cells, where the functional interplay between TSPYL5 and p53 was firstly documented [19].

Specifically, A549 cells are characterized by barely detectable levels of endogenous miR-380-5p (Fig. 1b) and, similarly to what observed in DMPM cells, the expression of miR-380-5p induced a significant and time-dependent slow down of cell growth (Additional file 4: Figure S4 A-B), which was accompanied by a remarkable reduction of TEP1 and TSPYL5 mRNA and proteins as well as by the accumulation of p53 (Additional file 4: Figure S4C-D). Of note, miR-380-5p failed to affect TSPYL5 RNA and protein levels and, consequently, of p53 in the TA-negative U-2 Os cell line (Additional file 4: Figure S4E-F), which showed barely detectable levels of endogenous TSPYL5 protein (Additional file 4: Figure S4E-F) compared to TA-positive DMPM and A549 cells.

The involvement of TSPYL5 in mediating miR-380-5p biological effects was hence explored by phenocopy experiments. Specifically, a remarkable reduction of TSPYL5 mRNA and protein amounts was observed in both STO and A549 cells after the transfection with a siRNA targeting TSPYL5 (siTSPYL5) (Fig. 4b, c). In addition, the selective depletion of TSPYL5 led to p53 accumulation and, similarly to what observed upon miR-380-5p reconstitution, to TEP1 protein reduction and TA inhibition (Fig. 4c, d). This evidence indicates that the depletion of TSPYL5 and the consequent accumulation of p53 may contribute, at least in part, to the direct effect of the miRNA on TEP1. In fact, we found that the inhibitory effect on TEP1 expression levels exerted by the miRNA was partly counteracted by the selective depletion of p53 (Additional file 4: Figure S4C and G). In addition, TSPYL5 target protection counteracted the inhibitory effect of miR-380-5p on the target and resulted in a reduced accumulation of p53 and in a rescue, though partial, of TEP1 protein amounts (Additional file 4: Figure S4 H and I).

The causative role of p53 in the regulation of TEP1 levels gained further support by the evidence that the remarkable accumulation of p53 caused by the exposure of STO and A549 cells to doxorubicin was accompanied by a dramatic reduction of TEP1 protein amounts (Fig. 4e), which instead were not remarkably affected upon silencing of p53 (Fig. 4e). These observations indicate that p53 may be actually involved in the control of TEP1 expression, as also suggested by the finding that doxorubicin induced a significant down-modulation of TEP1 mRNA levels in p53-proficient STO and A549 cells (Fig. 4f).

In accordance with the paradigm that a lag period is required to induce sufficient telomere shortening to impair cancer cell growth in the presence of inhibited TA [7], long-term reconstitution of miR-380-5p was pursued. Similarly to what observed in the short-term setting, after 3 months of weakly reiterated transfections with miR-380-5p, STO cells showed high levels of mature miR-380-5p (Fig. 5a) as well as a remarkable decrease in TEP1 mRNA expression levels and TA inhibition (Fig. 5b). Nevertheless, no evident telomere exhaustion was appreciable over this period of time (Table 1). Contrarily, upon 3 months of persistent reconstitution of miR-380-5p, a slight but significant increase in mean telomere length (about +1 Kb) was observed (Table 1). In addition, long-term miR-380-5p transfectants showed a significant reduction in the expression levels of ATRX (Fig. 5c) as well as detectable levels of C-circle DNA (Fig. 5d), two molecular markers associated with ALT activity [2, 22]. In addition, long-term STO cells expressing miR-380-5p were characterized by altered growing properties, in that compared to preNeg-transfected cells, they prevalently formed clusters with a remarkable component of vertical growth (Fig. 5e). However, despite these alterations in the growing properties, no growth arrest or massive culture loss was observed over the 3-month period of miR-380-5p reconstitution. Instead, miR-380-5p-expressing cells continued to grow in culture, though at significantly reduced rate compared to control cells (Fig. 5f).