MiR-4270 acts as a tumor suppressor by directly targeting Bcl-xL in human osteosarcoma cells

Chondrosarcomas and osteosarcomas are malignant bone tumors with a poor prognosis when unresectable or metastasized. Moreover, radiotherapy and chemotherapy could be ineffective. MiRNAs represent an alternative therapeutic approach. Based on high-throughput functional screening, we identified four miRNAs with a potential antiproliferative effect on SW1353 chondrosarcoma cells. Individual functional validations were then performed in SW1353 cells, as well as in three osteosarcoma cell lines. The antiproliferative and cytotoxic effects of miRNAs were evaluated in comparison with a positive control, miR-342-5p. The cytotoxic effect of four selected miRNAs was not confirmed on SW1353 cells, but we unambiguously revealed that miR-4270 had a potent cytotoxic effect on HOS and MG-63 osteosarcoma cell lines, but not on SaOS-2 cell line. Furthermore, like miR-342-5p, miR-4270 induced apoptosis in these two cell lines. In addition, we provided the first report of Bcl-xL as a direct target of miR-4270. MiR-4270 also decreased the expression of the anti-apoptotic protein Mcl-1, and increased the expression of the pro-apoptotic protein Bak. Our findings demonstrated that miR-4270 has tumor suppressive activity in osteosarcoma cells, particularly through Bcl-xL downregulation.


Introduction
Bone sarcomas are relatively rare tumors.Among the various bone sarcomas listed by the World Health Organization (WHO), the two main ones are osteosarcomas and chondrosarcomas.Osteosarcoma is the most common type of bone sarcoma (1,2).This aggressive cancer mainly affects children and young adults, and, in the great majority of cases, it develops on the long bones of the limbs: the lower end of the femur, the upper end of the tibia or the upper end of the humerus (3).The tumor cells characteristically synthesize an osteoid matrix.Chondrosarcoma is the second most common bone sarcoma, affecting adults after the age of 40 years.It preferentially develops on flat bones (pelvis, scapula, rib, spine) and long bones (femur, tibia, humerus) (4,5).This tumor is characterized by the formation of an extracellular cartilage matrix in a hypoxic environment (6).
In current treatments, osteosarcomas are the most curable sarcomas with a good prognosis for long-term survival.Neoadjuvant chemotherapy, including cisplatin, doxorubicin and methotrexate, coupled with surgery is the preferred therapy for this type of tumor (7,8).However, for people with unresectable metastases or older people, prognosis is poor (9).For chondrosarcomas, only surgery is efficient, because chondrosarcomas are resistant to chemotherapy and radiotherapy (10).Overall, new therapeutic strategies are absolutely needed for the treatment of both types of bone tumors.
MicroRNAs (miRNAs) are small non-coding RNAs of about 20-25 nucleotides.They are post-transcriptional regulators of several hundred different mRNAs.They function as inhibitors of translation or they favor the degradation of the mRNAs when they anneal to mRNAs through complementary base pairing (11).In recent years, miRNAs have emerged as key regulators within complex networks of targets that have not yet been fully elucidated (12,13).MiRNAs play an essential role in many physiological processes, but also in many pathological conditions, particularly in the progression of cancers, including bone sarcomas (14)(15)(16).Various miRNAs hold promise for use as diagnostic or prognostic biomarkers, and may also lead to the discovery of new therapeutic targets in bone sarcomas (14,(17)(18)(19).Many metabolic pathways regulated by miRNAs, such as survival, growth, angiogenesis or chemosensitivity, have been identified in bone sarcomas.However, few studies have focused on the effects of miRNAs directly targeting anti-apoptotic members of the Bcl-2 family in bone sarcomas.Several studies have reported the promising therapeutic potential of targeting these proteins for the treatment of bone sarcomas.The overexpression of Bcl-2 and Bcl-xL is indeed often associated with a poor prognosis and are partly responsible for the chemoresistance of tumor cells (20)(21)(22)(23)(24).One study has shown that pharmacological inhibition of Bcl-xL improves the sensitivity of osteosarcoma to doxorubicin (25).Furthermore, we previously showed that the direct inhibition of Bcl-xL by miR-491-5p and miR-342-5p induces apoptosis on osteosarcoma and chondrosarcoma cell lines (26,27).MiR-342-5p also directly inhibited the expression of Bcl-2 in some tested cell lines.
In this study, we continued our previous high-throughput miRNA screening of 1200 miRNAs mimics on the SW1353 chondrosarcoma cell line (26).Using miRNA target prediction, we selected four miRNAs inducing a potential cytotoxic effect and which may potentially target BCL2, BCL2L1 or MCL1 mRNAs.We subsequently performed an individual and functional study with miR-16-1-3p, miR-646, miR-3667-3p and miR-4270.MiR-342-5p was used as a positive control because its tumor suppressive effects have already been validated on chondrosarcoma and osteosarcoma cells (26,27).For the SW1353 chondrosarcoma cell line, the potential cytotoxic and chemosensitizing effects of miRNAs were studied in normoxia and hypoxia, because hypoxia is representative of the in situ physiopathological micro-environment of the tumor (28).Unfortunately, none of the newly selected miRNAs caused cytotoxic effects on SW1353 cells in the functional study.Because of their cytotoxic potential, we decided to test them on osteosarcoma cell lines.By contrast, miR-4270 had cytotoxic effects and induced cell death by apoptosis on two of the three osteosarcoma cell lines, suggesting a specific effect depending on cell type.Using westernblots and luciferase reporter assays, we demonstrated that miR-4270 directly targets BCL2L1 mRNA (Bcl-xL gene).Moreover, miR-4270 decreased the expression of Mcl-1 at the protein level, whereas it increased that of Bak.These results suggest that miR-4270 can act as a tumor suppressor in osteosarcoma cells.
The human articular chondrocytes (HACs) were obtained with appropriate ethical approval, they were prepared from macroscopically healthy zones of femoral heads obtained from patients undergoing joint arthroplasty, as previously described (26).The study was performed in full accordance with local ethics committee guidelines and all the cartilage samples were collected after written and informed consent of the donors according to French legislation.All the experimental protocols were approved by the Agence de la Biomedecine (research protocol n°PFS21-025) and French Ministry of Higher Education and Research (Ethics Committee for Research on Human Samples CODECOH: DC 2014-2325).

miRNA mimic screening
SW1353 cells were plated in 96-well E-plates onto a xCELLigence real-Time Cell Analysis (RTCA, ACEA, Ozyme, Saint Quentin-en-Yvelines, France) under normoxia as previously described (26).Briefly, cells were transfected in triplicate 24 h after seeding with 20 nM miRNAs using Interferin ™ (Polyplus-Transfection, Strasbourg, France).They were treated with or without CDDP 24 h post-transfection.Impedance of each well was continuously measured and expressed as a Cell index (CI) value ± SD (with the RTCA 2.1.0software.Morphological analyses were done with the Cellavista Imaging System (Roche, Basel) at the end of the experiment (120 h after seeding) as previously described (26).

Transfection and CDDP treatment
Growing chondrosarcoma and osteosarcoma cells were seeded at 1x10 4 cells/cm 2 per 25 cm 2 flask.Cell lines were transfected 24 h after seeding with 20 nM of miRNA mimics using INTERFERin ™ (Polyplus-Transfection) as described previously (26, 27).An additional CDDP treatment was carried out after 48 h with each sublethal dose of CDDP as described previously (26, 27).

XTT assay
SW1353 cells and osteosarcoma cell lines were seeded in a 96well plate at a density of 3.5x10 3 cells/well and 3.3x10 3 cells/well respectively in triplicate.Cells were transfected with 20 nM miRNA 24 h later with or without an additional CDDP treatment 48 h posttransfection.Cell metabolism activity was assessed 72 h posttransfection using the XTT assay (Roche, Basel, Switzerland) as previously described (26,27).Measurements of optical density (OD) were made with a microplate reader (Spark 10M, Tecan Lyon, France).

Bioluminescence cytotoxicity assay
The toxilight cytotoxicity assay kit (Interchim, Montlucon, France) was also used to evaluate miRNA-induced cytotoxicity 72 h post-transfection.Adenylate Kinase released was measured in the supernatant, as previously described (26).The bioluminescence was measured with a microplate reader (Spark 10M, Tecan Lyon, France) and expressed as relative luciferase units (RLU).

Cell cycle analysis by DNA content
DNA content after staining of trypsinized cells was performed by nuclear staining with propidium iodide (Sigma-Aldrich, Saint-Louis, MO, USA) using a Cytoflex S Flow Cytometer (Beckman Coulter France SAS, Paris, France) as previously described (26).

Statistical analysis
All experiments were repeated at least three times.Values are presented as means ± SD.Statistical analyses were performed using one-way ANOVA corrected for multiple comparisons using Dunnett's test, or two-way ANOVA corrected for multiple comparisons using Sidak's test.Alternatively, for representative experiments performed in triplicate, statistical analyses were done using a two-tailed unpaired Student's t-test with Welch's correction.Analyzes were done using GraphPad Prism 7 software (San Diego, CA, USA): NS, P > 0.05; *, p <0.05; **, p < 0.01; ***, p < 0.001).

Identification of antiproliferative miRNAs using high-throughput screening on the SW1353 chondrosarcoma cell line
We performed a real-time high-throughput screening of 1200 miRNA mimics on the SW1353 cells, as described previously (26).An additional cisplatin (CDDP) treatment was also realized 24 h after transfection of miRNAs to study their chemosensitizing effects (26).Biologicals effects of miRNAs were analyzed by following Cell Index (CI) in real-time using the xCELLigence System and by endpoint morphological observations of the cells using the Cellavista Imaging System.Finally, among all these miRNAs, we selected four novel miRNAs (miR-16-1-3p, miR-646, miR-3667-3p and miR-4270) with potential effectiveness on cell proliferation/attachment (Figure 1).
MiR-16-1-3p decreased the cell index from approximately 60 h post-transfection until the end of the experiment (5.7-fold decrease compared with miR-Ctrl, 120 h after seeding, i.e. 96 h after transfection, p<0.001, Figure 1A).In the presence of a sublethal dose of CDDP, the decrease in the cell index with miR-16-1-3p was delayed to the same degree as without CDDP 120 h after plating the cells (p<0.001, Figure 1A).Microscopic observations of the cells 120 h after seeding confirmed a decrease in proliferation with miR-16-1-3p, with or without CDDP (Figure 1).MiR-16-1-3p therefore showed antiproliferative effects on SW1353 cells without chemosensitizing effects.
MiR-646 also decreased the cell index from 48 h posttransfection until the end of the experiment (3-fold decrease relative to miR-Ctrl, 120 h after seeding, p<0.05, Figure 1B).When miR-646 was combined with CDDP, the decrease was 3.2fold (relative to miR-Ctrl + CDDP, 120 h after seeding, p<0.05, Figure 1B), suggesting no significant sensitization to CDDP.Moreover, miR-646-and miR-646/CDDP-treated cells showed a similar large reduction in confluence compared with the respective miR-Ctrl-treated cells (Figure 1).Therefore, miR-646 had only antiproliferative effects on SW1353 cells.
Similarly, the transfection of miR-3667-3p resulted in a decrease in the cell index by 4.7-fold at the end of the experiment with or without CDDP (p<0.001,relative to the respective miR-Ctrl treatments, Figure 1C).Cell attachment and cell density also decreased compared with respective miR-Ctrl (Figure 1).MiR-3667-3p appeared to have only antiproliferative effects on SW1353 cells.
Finally, miR-4270 also reduced the cell index, but its effects were delayed compared with the other three miRNAs, with an effect observed only from approximately 72 h post-transfection (Figure 1).At the end of the experiment, miR-4270 decreased the cell index by 2.5-fold with or without CDDP (p<0.01 compared with miR-Ctrl and p<0.05 compared with miR-Ctrl/CDDP, at 120 h after seeding).Cell confluency also decreased by at least 50% with or without CDDP, suggesting no chemosensitizing effect of this miRNA (Figure 1).Therefore, miR-4270 had only antiproliferative effects on SW1353 chondrosarcoma cells.

Individual functional analysis of miRNAs fails to confirm induction of cell death in the SW1353 chondrosarcoma cell line
Next, we carried out a functional analysis with these four miRNAs and with miR-342-5p, used as a positive control, at 72 h post-transfection on SW1353 cells cultured under normoxia and hypoxia.We also investigated the potential chemosensitizing effects of the miRNAs with a sublethal exposure to CDDP 48 h after their transfection.
The studies of SW1353 viability and proliferation and of the cytotoxicity of miRNAs showed the best antiproliferative and cytotoxic effects for miR-342-5p and no relevant effect of other miRNAs, except with miR-16-1-3p and miR-646 under hypoxia with CDDP, thereby revealing a possible chemosensitizing action (Figure S1A).Nevertheless, this effect was not further confirmed, because both of these latter miRNAs failed to induce cytotoxicity on SW1353 cells in hypoxia, with or without CDDP (Figure S1B), unlike miR-342-5p as previously described (26).
Regarding cell cycle progression, no cell cycle blockade was observed with miRNAs using flow cytometry (Figure S2).None of miRNAs induced a cell cycle blockade in SW1353 cells.On the contrary, the sublethal dose of CDDP induced a cell cycle blockade in the S and G2/M phases.Only miR-342-5p increased the number of cells in the sub-G1 phase in all four conditions of culture (approximately of 3-fold, p<0.01 or p<0.05 relative to the respective miR-Ctrl treatments).The proportions of the cells in the sub-G1 phase for miR-Ctrl and miR-342-5p respectively were 6.8% and 21.9% without CCDP under normoxia; 5.5% and 14% with CDDP under normoxia; 6.6% and 20.5% without CDDP under hypoxia; 4.4% and 14.4% with CDDP under hypoxia.
Microscopic observations at 72 h post-transfection further corroborated that only miR-342-5p induced cell death.Indeed, miR342-5p decreased the confluency and increased the number of apoptotic debris or cellular debris (Figure S3).MiR-646 significantly increased the percentage of sub-G1 events by 2-fold (p<0.05relative to miR-Ctrl/CDDP treated cells) only under normoxia in the presence of CDDP (Figure S2).MiRNAs did not increase the sub-G1 peaks in the presence of CDDP, suggesting no chemosensitizing effect for all miRNAs tested.
Overall, we concluded that the antiproliferative effects of these four selected miRNAs, observed in the high-throughput screening were not due to an induction of cell death in SW1353 cells.
High-throughput screening of miRNAs on SW1353 chondrosarcoma cells.SW1353 cells were grown for 24 h before transfection (arrow labeled with a T) with miR-Ctrl, miR-16-1-3p, miR-646, miR-3667-3p or miR-4270.The cells were additionally treated with CDDP (1 µg/ml) 24 h post-transfection (arrow labeled with a C).(A-D) Real-time growth curves monitoring was performed with the xCELLigence System.The measurement of Cell index was monitored every 3 h for 120 h.Analysis of one experiment per miRNA, with mean Cell index ± SD, is shown for 120 h (left panels) and at 120 h after seeding (right panels).The significance of the results was evaluated using an unpaired Student's t-test.(E) Endpoint morphological analysis of the cells was done 120 h after seeding with the Cellavista Imaging system.Ctrl, control; CDDP, cisplatin.*: p<0.05, **: p<0.01, ***: p<0.001.

Functional analysis of miRNAs reveals antiproliferative effects of miR-342-5p, miR-16-1-3, miR-646 and miR-4270 on osteosarcoma cell lines
In parallel, we performed a functional analysis on three osteosarcoma cell lines at 72 h post-transfection.Cells were treated under normoxia in the presence or absence of a sublethal dose of CDDP 48 h after the transfection of miRNAs.First, cell viability and proliferation were evaluated using an XTT assay (Figure 2).
In HOS cells (Figure 2), all miRNAs (except miR-3667) significantly decreased the cell viability and proliferation compared with miR-Ctrl (by about 1.8-fold for miR-342-5p and miR-646, p<0.001; by about 1.5-fold for miR-16-1-3p and miR-4270, p<0.01).In the presence of CDDP, all miRNAs decreased the cell viability and proliferation.MiR-342-5p and miR-646 were the most effective miRNAs with the same decrease with or without CDDP (by about 1.8-fold (p<0.001)).MiR-3667-3p was the least effective with a weak decrease of only 1.3-fold in the presence of CDDP (p<0.05).
Effects of miRNAs on cell viability and proliferation on osteosarcoma cell lines.(A) HOS, (B) MG-63 and (C) SaOS-2 cells were transfected with miRNAs 24 h after seeding and treated with CDDP (0.5 µg/ml) 48 h post-transfection for 24 (h) Analyses were then carried out 72 h posttransfection.Cell viability and proliferation were evaluated and expressed as the mean OD ± SD of four independent experiments for each cell line.The significance of the results between miR-Ctrl and miRNA-treated cells was evaluated using one-way ANOVA (*: p<0.05, **: p<0.01, ***: p<0.001).Regarding MG-63 cells, miR-342-5p, miR-646 and miR-4270 significantly decreased their viability and proliferation at 72 h posttransfection (Figure 2).MiR-342-5p decreased the cell viability and proliferation by approximately 3-fold in both culture conditions (p<0.001 relative to miR-Ctrl, and p<0.01 relative to miR-Ctrl/ CDDP), whereas miR-646 and miR-4270 had lower magnitude effects with a decrease of approximately 2-fold (p<0.05relative to respective miR-Ctrl).

Functional analysis of miRNAs reveals cytotoxic effects of miR-4270 on HOS and MG-63 osteosarcoma cell lines, but not on the SaOS-2 osteosarcoma cell line
We then analyzed the cytotoxic effects of these miRNAs (Figure 3).In addition to miR-342-5p, only miR-4270 significantly induced high cytotoxicity on HOS cells (Figure 3): 1.4-fold relative to miR-Ctrl treated cells (p<0.05), and 2.4-fold relative to miR-Ctrl/CDDP treated cells (p<0.05).In the same manner, only miR-4270 and miR-342-5p induced high cytotoxicity on MG-63 cells, by approximately 2-fold (with and without CDDP) versus approximately 3-fold with miR-342-5p (Figure 3).None of the four studied miRNAs induced cytotoxicity on SaOS-2 cells.Only the positive control miR-342-5p increased cytotoxicity by about 2-fold in the presence or absence of a sublethal dose of CDDP (Figure 3).
In the HOS cell line, the analysis of cell cycle cell progression using flow cytometry showed a significant enhancement in the number of sub-G1 events for only miR-342-5p and miR-4270 (Figure 4).Without CDDP, miR-342-5p significantly increased the number of sub-G1 events by 3.8-fold (p<0.001;11.2% of sub-G1 events with miR-Ctrl and 43.1% with miR-342-5p) and miR-4270 increased it by 3.4-fold (p<0.001;38.4% of sub-G1 events with miR-4270).CDDP increased the percentage of events in the S and G2/M phases at the expense of the G0/G1 phase.In the presence of CDDP, miR-342-5p increased cell death by 3.5-fold (p<0.001;13% of sub-G1 events with miR-Ctrl and 46.2% with miR-342-5p) and miR-4270 by 3.9-fold (p<0.01;51.4% of sub-G1 events with miR-4270).There was no significant difference in the induction of cell death between cells transfected with miRNAs alone or with miRNAs combined with CDDP treatment.The analysis of cell morphology at 72 h post-transfection showed a decrease in cell confluence and an increase in the amount of cellular debris as well as fragmented chromatinsigns typical of apoptotic cellsfor both miRNAs with or without CDDP (Figure S4).
The analysis of the cell cycle distribution on MG63 cells revealed that the treatment with a sublethal dose of CDDP induced a blockage in the S phase only (Figure 4).MiR-4270 significantly increased the number of sub-G1 events by 2.8-fold for both culture conditions (p<0.001 relative to miR-Ctrl with or without CDDP) (Figure 4).In comparison, miR-342-5p increased cell death by about 4-fold for both conditions (p<0.001) (Figure 4), as previously described (27).For both miRNAs, morphological analysis of the cells at 72 h post-transfection showed a large increase in the amount of cellular debris and in the presence of apoptotic bodies compared with the miR-Ctrl condition with or without CDDP (Figure S5).
In the SaOS-2 cell line, we observed that CDDP increased the percentage of events in the S phase at the expense of the G0/G1 phase (Figure 4).Only miR-342-5p was able to significantly increase the number of sub-G1 events by 2.9-fold in the presence or absence of a sublethal dose of CDDP (p<0.001).MiR-4270 did not significantly modulate cell death in either culture conditions: 1.6fold increase relative to miR-Ctrl and 1.9-fold relative to miR-Ctrl/ CDDP.Morphological analysis of the cells showed a reduction in confluency for both miRNAs, but fewer apoptotic bodies with miR-4270 than with miR-342-5p (Figure S6).
Overall, these results suggest that, similarly to miR-342-5p, miR-4270 may be an apoptomiR in HOS and MG63 osteosarcoma cell lines, whereas it had only a significant antiproliferative effect on the SaOS-2 osteosarcoma cell line.MiR-4270 induced cell death to the same magnitude with or without CDDP, suggesting that it is not a CDDP chemosensitizer.

MiR-4270 activates the apoptosis pathway in the HOS and MG-63 osteosarcoma cell lines
We then investigated the induction of apoptosis by miR-4270 alone in the three osteosarcoma cell lines 72 h post-transfection (Figure 5).We noted that miR-Ctrl induces cleavage of PARP (Figure 5A, S7) and increases caspase 3/7 activity in the HOS line (Figure 5).This cell line appears to be more sensitive to transfection as previously described (27).Similar to miR-342-5p, miR-4270 induced PARP cleavage in the three osteosarcoma cell lines (Figure 5).The effects were more relevant in the HOS and MG-63 cell lines than in the SaOS-2 cell line.A high amount of cleaved caspase-3 was detected with miR-4270 transfection, particularly in the HOS cell line, but this cleavage was almost absent in the SaOS-2 cell line for both miRNAs (Figure 5).In addition, miR-4270 significantly increased caspase-3/7 activity only in HOS and MG-63 cells (by 1.8-fold and by 5.7-fold relative to miR-Ctrl treated cells, respectively, p<0.001) (Figure 5).MiR-342-5p significantly induced caspase-3/7 activity, to the same magnitude as miR-4270, only in HOS and MG-63 cells as previously described ( 27)).Because the activation of caspases and the cleavage of PARP are characteristic hallmarks of apoptosis, these results suggest that miR-4270 unambiguously induces apoptosis in the HOS and MG- Cytotoxic effects of miRNAs on osteosarcoma cell lines.(A) HOS, (B) MG-63 and (C) SaOS-2 cells were transfected with miRNAs and treated with CDDP as described in Figure 2. The cytotoxicity of miRNAs was evaluated 72 h post-transfection as the mean RLU ± SD of four independent experiments for each cell line.The significance of the results between miR-Ctrl and miRNA-treated cells was evaluated using one-way ANOVA (*: p<0.05, **: p<0.01, ***: p<0.001).
63 osteosarcoma cell lines, like miR-342-5p.This is in accordance with our previous results on cell death (Figures 3, 4).

MiR-4270 modulates the expression of proteins of the Bcl-2 family in osteosarcoma cell lines
Bcl-xL, Bcl-2, Mcl-1 are all members of the anti-apoptotic Bcl-2 family and are among the predicted targets of miR-4270 and miR-342-5p.We therefore evaluated their protein expression to determine their involvement in the induction of apoptosis by this miRNA in the three osteosarcoma cell lines.
Bcl-2 protein expression did not decrease, but increased (by 1.4-fold relative to miR-Ctrl) with miR-4270 in HOS cells (Figure 6).By contrast, miR-342-5p decreased Bcl-2 protein expression by 2-fold in the HOS cell line.MiR-4270 decreased Bcl-2 protein expression by 1.7-fold and 1.2-fold, respectively, in MG-63 and SaOS-2 cells.The effects of miR-342-5p were comparable on both these two cell lines (1.4-fold decrease in Analysis of cell cycle phase distribution after miRNA treatment on osteosarcoma cell lines.(A) HOS, (B) MG-63 and (C) SaOS-2 cells were transfected with miRNAs and treated with CDDP as described in Figure 2. The cell cycle phase distribution was evaluated 72 h post-transfection as the mean ± SD of four independent experiments for HOS and MG-63 cells, and for five independent experiments for SaOS-2 cells.The significance of the results between miR-Ctrl and miRNA-treated cells was evaluated using one-way ANOVA (***: p<0.001).
MG-63 cells and 1.7-fold in SaOS-2 cells).MiR-4270 significantly decreased Bcl-xL protein expression in all three osteosarcoma cell lines (by 2.5-fold in HOS cells, 5-fold in MG-63 cells, and 1.4-fold in SaOS-2 cells) (Figure 6).We previously validated Bcl-xL as a direct target of miR-342-5p in the HOS cell line (27).Here, we confirmed its inhibitory effect on Bcl-xL protein expression in the three tested osteosarcoma cell lines with at least a 1.7-fold decrease.
MCL-1 mRNA is a putative target for both miRNAs.MiR-342-5p had no relevant effect on the level of Mcl-1 protein expression (Figure 6).In all three osteosarcoma cell lines, miR-4270 decreased Mcl-1 protein expression at different levels: by 1.4-fold in HOS cells and by 1.7-fold in MG-63 and SaOS-2 cells.
We then investigated the expression of the pro-apoptotic protein Bak, involved in the intrinsic apoptosis pathway (Figure 6).This protein participates in mitochondrial outer membrane permeabilization and subsequent induction of cell death.Both miRNAs clearly induced Bak protein expression in the three osteosarcoma cell lines.MiR-342-5p had the strongest effect (from 2.2-fold to 30-fold induction with miR-342-5p but only from 1.5-fold to 4-fold induction with miR-4270).
In summary, miR-4270 mainly decreased the expression of the anti-apoptotic proteins Bcl-xL and Mcl-1, whereas miR-342-5p mainly decreased that of Bcl-xL and also Bcl-2 in osteosarcoma cells.Both miRNAs significantly increased the expression of the pro-apoptotic protein Bak.

Direct inhibition of BCL2L1 mRNA by miR-4270 in HOS osteosarcoma cell line
We then verified that the effect of miR-342-5p and miR-4270 on Bcl-2 (BCL2 gene), Bcl-xL (BCL2L1 gene) or Mcl-1 (MCL1 gene) resulted from direct binding to their respective mRNAs in the 3'UTR region.Using HOS cells, we co-transfected luciferase reporter vectors containing the 3'UTR sequence of these mRNAs as well as miR-4270 and miR-342 mimics or their corresponding hairpin inhibitors (Figure 7).
In silico analysis identified a potential binding site for miR-4270 between nucleotides 556-562 in BCL2-3'UTR (Figure 7).MiR-4270 did not significantly decrease (16%) the luciferase activity of the WT-BCL2-3'UTR vector compared with miR-Ctrl.Co-transfection of anti-miR-4270 led to a slight increase in the luciferase activity of the WT-BCL2-3'UTR (12% relative to miR-Ctrl, ns), in favor of a specific inhibitory effect of miR-4270.Nevertheless, due to the analyses of Bcl-2 protein expression and of the poor inhibitory effects on luciferase activity, we concluded that miR-4270 does not target directly BCL2 mRNA in HOS cells.MiR-342 also had a potential binding site on the WT-BCL2-3'UTR vector and its transfection decreased Bcl-2 protein expression.Nevertheless, as previously described (27), it did not significantly decrease (22%) the luciferase activity of the WT-BCL2-3'UTR vector.Moreover, cotransfection with anti-miR-342-5p did not significantly reversed luciferase activity of the vector.As previously shown (27), despite its inhibitory effect on Bcl-2 protein expression, we cannot confirm that BCL2 mRNA is a direct target of miR-342-5p in HOS cells.
We identified two presumptive binding sites for miR-4270 in BCL2L1-3'UTR at positions 326-333 and 738-744 (Figure 7).We have already demonstrated direct binding of miR-342-5p at position 679-686 and 1407-1413 of BCL2L1-3'UTR in HOS cells (27).As expected, miR-342-5p induced a great decrease in the luciferase activity of the WT-BCL2L1-3'UTR vector (70%, p<0.01) and anti-miR342-5p reversed this decrease.The co-transfection of miR-4270 resulted in a significant decrease (35%, p<0.05) in the luciferase activity of the reporter vector.The co-transfection of anti-miR-4270 restored the luciferase activity to the miR-Ctrl level, which is consistent with an inhibitory effect of miR-4270.Therefore, in HOS cells, miR-4270 downregulated Bcl-XL protein expression through direct binding to the 3'UTR of BCL2L1 mRNA.
In silico analysis revealed that both miRNAs have a potential binding site in MCL1-3'UTR: between nucleotides 2427-2433 for miR-342-5p and between nucleotides 1585-1591 for miR-4270 (Figure 7).MiR-342-5p did not decrease the luciferase activity of the WT-MCL1-3'UTR vector, indicating no post-transcriptional regulation of MCL1 mRNA by miR-342-5p, in accordance to the analysis of Mcl-1 protein expression (Figure 6).Surprisingly, as anti-miR-Ctrl, anti-miR-342-5p increased RLU beyond the miR-Ctrl level (80% relative to miR-Ctrl).MiR-4270 reduced the luciferase activity of the MCL1-3'UTR vector by 15% compared with miR-Ctrl, but this inhibitory effect was not statistically significant.Anti-miR-4270 had also an inhibitory effect, indicating a non-specific effect of miR-4270 on luciferase activity.Therefore, despite its inhibitory effect on Mcl-1 protein expression, we cannot confirm that MCL1 mRNA is a direct target of miR-4270 in HOS cells.

Discussion
Osteosarcomas and chondrosarcomas are the most common forms of bone sarcomas.Today, osteosarcoma is a well-treated tumor given the efficacy of combination chemotherapy associated with surgery.However, for patients with metastases or recurrences, the survival rate is much lower.For chondrosarcomas, these tumors are resistant to conventional chemotherapy and radiotherapy and surgery remains the only effective therapy.Nevertheless, the location of the tumor may not allow surgery.The use of therapeutic miRNAs has the advantage of being a multi-target approach.Moreover, miRNAs may lead to the discovery of new therapeutic targets for clinical applications (14,17,30,31).
Among miRNAs decreasing SW1353 cell proliferation in realtime cell analysis, we focused on four miRNAs potentially targeting the anti-apoptotic members of the Bcl-2 family.Indeed, Bcl-xL and Bcl-2 are promising targets in the treatment of chondrosarcomas and osteosarcomas (20)(21)(22)(23)(24).In a previous study, we demonstrated a suppressive role for miR-342-5p on chondrosarcoma cells through direct inhibition of Bcl-xL and Bcl-2 (26).We also showed that miR-491-5p is an apoptomiR in chondrosarcoma cells, through the direct inhibition of Bcl-xL and the repression of epidermal growth factor receptor (EGFR) expression.In a second study performed on osteosarcoma cells, we demonstrated a direct inhibition of Bcl-xL by both miRNAs, and their concomitant induction of apoptosis (27).
RTCA and endpoint morphological analysis of highthroughput screening of miRNAs on SW1353 cells were performed in glass culture plates under normoxia.However, chondrosarcomas are predominantly hypoxic tumors.We therefore carried out an individual functional validation of miR-16-1-3p, miR-646, miR-3667-3p and miR-4270 in conventional plastic culture plates, under normoxia and hypoxia.We were also interested in the chemosensitizing effects of these miRNAs.Unfortunately, we did not confirm the potential cytotoxic effect of demonstrated for the first time, that Bcl-xL is a direct target of miR-4270.We also demonstrated that miR-4270 decreased Mcl-1 protein expression particularly in MG-63 and SaOS-2 cells.However, we did not confirm that Mcl-1 is a direct target of miR-4270 in HOS cells.Furthermore, miR-4270 was able to increase the expression of the pro-apoptotic protein Bak especially in the HOS and MG-63 cell lines.To gain insights into the mechanism of action of miR-4270, we could use selective inhibitors of Bcl-xL (WEHI-539, A-1331852) or siRNA against Bcl-xL to compare their effects to that of miR-4270.
We used three osteosarcoma cell lines from young Caucasian patients.Cell lines did not have the same aggressiveness.HOS cell line had high aggressiveness compared to SaOS-2 and MG-63 cell lines, and MG-63 cells had less invasive/migratory capacities than SaOS-2 cells (47).Genomic analyses revealed key differences in the three osteosarcoma cell lines used.Deletion and rearrangement of the p53 tumor suppressor gene were found in MG-63 and SaoS2 cell lines, with no p53 expression detected.A point mutation within the p53 coding sequence has been described in HOS cells which results in overproduction of mutant p53 (48).Furthermore, several mRNAs and miRNAs were differentially expressed in highly aggressive osteosarcoma cell lines compared with non-aggressive cell lines.Nevertheless, miR-4270 was not identified in the list of miRNAs significantly upregulated or downregulated in the aggressive cell lines (47).Further investigations could be needed to link the differential effect of miR-4270 to the genetic background of the 3 cell lines.
Numerous signaling pathways have been linked to tumor progression in osteosarcoma.Wnt/b-catenin signaling has been shown to be oncogenic in a wide range of tumor types, including osteosarcoma (49).NF-kB appears to play a causative role in the malignancy of osteosarcoma (50).Moreover, crosstalk between NF-kB and Wnt/b-catenin pathways link inflammation with tumorigenesis (51).Overexpression and abnormal activation of Raf/MEK/ERK signaling pathway may regulate tumor proliferation, migration, and metastasis in osteosarcoma as well as in other malignancies (52).PI3K/Akt is a critical driver of oncogenesis in OS and largely studied with the development of inhibitors of mTOR, the most important downstream molecule of the PI3K/AKT pathway (53).To the best of our knowledge, no study has explored the effect of miR-4270 on these signaling pathways.This could be explored in the future.However, our preliminary data suggested a possible downstream inhibition of the AKT and ERK signaling pathways on MG-63 cells (data not shown).
Overall, we demonstrated the tumor suppressive role of miR-4270 on HOS and MG-63 osteosarcoma cell lines.MiR-4270 had antiproliferative effects and some anti-apoptotic effects on SaOS-2 cell line, but these latter effects were not sufficient to induce cell death in this osteosarcoma cell line.Likewise, miR-4270 did not induce cell death in the SW1353 chondrosarcoma cell line, despite its antiproliferative effects.Moreover, we showed that miR-4270 does not affect healthy primary human articular chondrocytes (HAC) cell cycle, nor does it cause any cell death (Figure S8).As for miR-342-5p (26), these two experiments suggest its biosafety on non-cancerous cells.However further investigations would be required to verify tumor suppressive role of miR-4270 in in vivo studies or in three-dimensional (3D) OS cell culture models.Nevertheless, miR-4270 therefore appears to have a specific function according to the type of tumor, and sometimes it can act as an oncogene or as a tumor suppressor as described above.Furthermore, it may have a more or less pronounced effect depending on the cell line used, particularly according to its altered status of differentiation.

Conclusions
Our study demonstrated antiproliferative and cytotoxic effects of miR-4270 on HOS and MG-63 osteosarcoma cell lines.MiR-4270 also significantly induced apoptosis in these cell lines.Here, for the first time, Bcl-xL was identified as a direct target of miR-4270.MiR-4270 also decreased Mcl-1 protein expression, whereas it increased that of Bak.We already demonstrated the therapeutic potential of specifically targeting Bcl-xL and Bcl-2 on chondrosarcoma and osteosarcoma cells.Here, we provide additional evidence that Bcl-xL could be used as a relevant therapeutic target in combination with chemotherapy in osteosarcomas, with BH3-mimetic derivates for example to inhibit Bcl-xL family members and improve sensitivity to conventional chemotherapy.

6
FIGURE 6 Effects of miR-342 and miR-4270 on the expression of members of the Bcl-2 family in osteosarcoma cell lines.In panels (A-D), HOS, MG-63 and SaOS-2 cells were transfected with miRNAs 24 h after seeding.Bcl-2, Bcl-xL, Mcl-1 and Bak levels were analyzed 72 h post-transfection.Representative blots of three independent experiments are shown.Protein expressions are indicated at the bottom of the blots.They were normalized to b-tubulin or GAPDH and to the corresponding miR-Ctrl.

7
FIGURE 7 Analysis of inhibition of BCL2, BCL2L1 and MCL1 mRNAs by miR-342-5p and miR-4270 in HOS cells.Schematic representations of the 3'UTR sequences of BCL2, BCL2L1 and MCL1 are respectively shown in panels (A-C), with predicted sites of hsa-miR-4270 or hsa-miR-342-5p.Reporter vectors of BCL2-3'UTR, BCL2L1-3'UTR and MCL1-3'UTR were transfected concomitantly with miR or anti-miR in HOS cells.Reporter assays were carried-out 48 h post-transfection.Gaussia luciferase (GLUC) activity was normalized to Secreted alkaline phosphatase (SEAP) activity.The data show representative experiments of at least three independent experiments (mean of RLU ± SD).The significance of the values between the different treatments and miR-Ctrl was evaluated using two-tailed Student's t-test (*: p<0.05, **: p<0.01).