miR-20a regulates Fas expression in osteosarcoma cells by modulating Fas promoter activity and can be therapeutically targeted to inhibit lung metastases

Abstract

The metastatic potential of osteosarcoma cells is inversely correlated to cell surface Fas expression. Downregulation of Fas allows osteosarcoma cells to escape Fas ligand-mediated apoptosis when they enter a Fas ligand-positive microenvironment such as the lung. We have previously demonstrated that miR-20a, encoded by the miR-17-92 cluster, downregulates Fas expression in osteosarcoma. We further demonstrated an inverse correlation between Fas expression and miR-20a expression. However, the mechanism of Fas regulation by miR-20a was still unclear. The purpose of the current study was to evaluate the mechanism of Fas regulation by miR-20a in vitro and test the effect of targeting miR-20a in vivo. We investigated whether miR-20a’s downregulation of Fas was mediated by binding to the 3′ untranslated region (3′-UTR) of Fas mRNA with the consequent induction of mRNA degradation or translational suppression. We identified and mutated two miR-20a binding sites on the Fas mRNA 3′-UTR. Using luciferase reporter assays, we demonstrated that miR-20a did not bind to either the wild-type or mutated Fas 3′-UTR. By contrast, overexpression of miR-20a resulted in downregulation of Fas promoter activity. Similarly, the inhibition of miR-20a increased Fas promoter activity. The critical region identified on the Fas promoter was between −240 bp and 150 bp. Delivery of anti-miR-20a in vivo using nanoparticles in mice with established osteosarcoma lung metastases resulted in upregulation of Fas and tumor growth inhibition. Taken together, our data suggest that miR-20a regulates Fas expression through the modulation of the Fas promoter and that targeting miR-20a using anti-miR-20a has therapeutic potential.

Introduction

Osteosarcoma, the most common primary malignant bone tumor, metastasizes almost exclusively to the lung, where Fas ligand (FasL; also known as tumor necrosis factor ligand superfamily member 6) is constitutively expressed. Interaction of FasL with Fas activates the apoptosis pathway. We have previously demonstrated that Fas⁺ osteosarcoma cells are eliminated by the constitutive FasL expressed on the lung epithelium, whereas Fas⁻ osteosarcoma cells escape this surveillance mechanism and form lung metastases (1–4). Immunohistochemically staining of patient lung tumor samples shows that lung tumor nodules from more than 60 osteosarcoma patients were Fas⁻ (5). The absence of Fas or molecular changes that interfere with the expression of this cell surface receptor may therefore play an important role in the metastatic process of osteosarcoma, and understanding these molecular changes may lead the way to the development of novel therapeutic strategies.

MicroRNAs (miRNAs) are small (21–25 nucleotides) noncoding RNAs that negatively regulate gene expression (6, 7). They are frequently dysregulated in a variety of cancers, including breast cancer, colon cancer, lung cancer, hepatocellular carcinoma, and osteosarcoma (8–12). miRNAs have been shown to post-transcriptionally regulate gene expression by targeting the 3′ untranslated region (3′-UTR) of specific mRNAs or suppressing translation initiation. Death receptor signaling proteins, such as tumor necrosis factor–α (TNFα), Fas-associated death domain protein (FADD), Bcl-2 interacting mediator of cell death protein (Bim; also called Bcl-2-like protein 11), and p21, were also found to be regulated by miRNAs (13). Several miRNAs, including miR-20, miR-21, miR-200c, let-7, and miR-146, have been reported to regulate Fas or FasL expression in different tumor cells (14–19).

We previously showed that a specific miRNA cluster, miR-17-92, is upregulated in Fas⁻ osteosarcoma cells that metastasize to the lung (14). Among the five miRNAs encoded by the miR-17-92 cluster, miR-20a was found to regulate the expression of Fas and in turn, the metastatic potential of osteosarcoma cells to the lung (14). Overexpression of miR-20a in non-metastatic Fas⁺ SAOS-2 osteosarcoma cells downregulated Fas expression and decreased their sensitivity to FasL. By contrast, inhibiting miR-20a in Fas⁻ LM7 osteosarcoma cells increased Fas expression and sensitivity to FasL. An inverse correlation between Fas and miR-20a expression in 8 cell lines was demonstrated (14). However, the mechanism of Fas downregulation was not elucidated.

Traditionally, miRNAs function in post-transcriptional regulation of gene expression. The purpose of the current study was to evaluate the mechanism of Fas regulation by miR-20a in vitro and evaluated the effect of targeting miR-20a in vivo. We now show that the regulation of Fas by miR-20a was not mediated by its binding to the 3′-UTR of Fas mRNA, thereby inducing Fas mRNA degradation or protein translational suppression, but rather by an indirect effect on the Fas promoter. We demonstrate that a 90 bp region (−240 bp to −150 bp) on the Fas promoter was critical for Fas regulation by miR-20a. We further demonstrate that targeting miR-20a by administering nanoparticle-formulated anti-miR-20a oligonucleotides suppressed osteosarcoma lung metastases in mice, indicating that targeting miR-20a is a potential therapeutic strategy for osteosarcoma.

Materials and Methods

Cell lines and reagents

SAOS-2 (ATCC HTB-85) and U2OS(ATCC HTB-96) cell lines were obtained from the American Type Culture Collection (ATCC) (Manassas, VA) in 1997; 293T and HeLa cell lines were also purchased from ATCC in 2010. The metastatic LM7 cell line was derived from SAOS-2 cells in our laboratory (1, 20) in 1999. When injected LM7 cells form lung metastases within 6-8 weeks, by contrast, the parental SAOS-2 cells do not form metastases following i.v. injection; The K7M3 cell line, a subline of K7M2 murine osteosarcoma cells, was developed by injecting K7M2 cells i.v. into mice, harvesting the lung metastases, and growing these metastatic cells in culture as described previously (21) in 2006. All cells were maintained in Dulbecco modified Eagle medium supplemented with 2 mmol/L L-glutamine, 1 mmol/L sodium pyruvate, 1× nonessential amino acids, 2× minimal essential medium vitamin solution, 10% heat-inactivated fetal bovine serum, 100 U/mL penicillin, and 100 mg/mL streptomycin at 37°C in 5% CO₂ and 95% atmosphere. All cell lines were limited cultured for not more than 30 passages, mycoplasma contamination was checked every other month and all cell lines were verified to be negative for mycoplasma species using the MycoAlert Mycoplasma Detection Kit (Lonza, Inc.). Unique signature identification of SAOS-2, K7M3 and LM7 cells as well as all the LM7 single cell clones, respectively, were confirmed by Short Tandem Repeat (STR) DNA microsatellite fingerprinting analysis carried out at the CCSG CCLC core facility in the University of Texas MD Anderson Cancer Center, Houston, TX in 2012.

Oligonucleotides, plasmids, and transfection

The miR-20a precursor oligonucleotides and the scramble control oligonucleotides were purchased from Applied Biosystems. The miR-20a antisense oligonucleotides (anti-miR-20a) with matched scramble control oligonucleotides were obtained from Enzo Life Sciences or Sigma-Aldrich Corporation.

When cells were grown to 50–70% confluence, transfection was performed using Lipofectamine™ 2000 transfection reagent (Invitrogen, USA). At 8-10 hours after transfection, the medium was replaced with fresh medium containing 10% fetal bovine serum.

RNA extraction and quantitative real-time PCR

Total RNA containing miRNA and mRNA was extracted and purified using a miRNeasy Mini Kit (Qiagen Inc.). RNA was resuspended in DEPC-treated water. miRNA quantification was performed by two-step real-time PCR (RT-PCR) using TaqMan kits (Applied Biosystems). Briefly, cDNA was reverse transcribed from a 10 ng total RNA sample using specific miRNA loop RT primers from the TaqMan MicroRNA Assays kit and reagents from the TaqMan MicroRNA Reverse Transcription Kit. RT-PCR was performed with the specific miRNA PCR primer from the TaqMan MicroRNA Assays kit and TaqMan Universal PCR Master Mix. All RT-PCR analysis was conducted with the iCycler iQ RT-PCR Detection System (Bio-Rad). The levels of each miRNA were normalized to U44 controls (Applied Biosystems). The relative expression levels were evaluated using the 2^−ΔΔCt method and expressed in terms of fold change.

Flow cytometry analysis and sorting

Cultured cells were washed with 1xPBS once, trpsinized, and resuspended in 1xPBS containing 0.5% FBS. Cells were stained with phycoerythrin (PE)-conjugated Fas antibody or PE-conjugated control IgG (Biolegend, USA) for 30 min at 4°C in the dark and examined with a fluorescence-activated cell-sorting (FACS) flow cytometer (BD Biosciences, USA). Sorted single-cell clones were isolated and expanded, all single cell clones were confirmed from LM7 cells by fingerprinting analysis.

Western blotting

Cell lysates were prepared with radioimmunoprecipitation assay (RIPA) buffer in the presence of proteinase inhibitor cocktail (Roche) on ice. Equal amounts of protein lysates (30 μg each) were separated by SDS-PAGE and then electrotransferred to nitrocellulose membranes (Invitrogen, USA). The membranes were blocked with TBST (Tris-buffered saline and 0.5% Tween 20) solution containing 5% non-fat dry milk for 1 h and incubated with primary antibody against Fas (Enzo) or, as a control, β-actin (Santa Cruz Biotechnology, USA) overnight at 4°C. The membranes were washed three times with TBST and then incubated with horseradish peroxidase-conjugated anti-mouse or -rabbit secondary antibody (Santa Cruz Biotechnology) for 1 h and washed three times with TBST. The specific protein was detected by enhanced chemiluminescence (ECL) using the ECL Plus Western Blotting Detection Kit (GE Healthcare).

Luciferase reporter assay

To increase miR-20a expression, 0.2×10⁶ cells were co-transfected with 1.0 μg luciferase reporter construct with either 25 nM miR-20a precursor oligonucleotide or 25 nM matched scramble control oligonucleotide. To decrease miR-20a expression, cells were co-transfected with 1.0 μg luciferase reporter construct with either 25 nM miR-20a antisense oligonucleotide or matched antisense scramble control oligonucleotide using Lipofectamine 2000 reagent (BD Biosciences). The cells were also co-transfected with 50 ng pRL-TK vector as an internal standard for a dual-luciferase assay (firefly and Renilla luciferases). Luciferase activity was measured at 48 h using a dual luciferase reporter assay kit (Promega, USA). Results were expressed as relative luciferase activity. All experiments were repeated three times in triplicate.

Target prediction

For bioinformatics analysis, we used the PicTar (http://pictar.mdc-berlin.de/), miRanda (http://www.microrna.org), and TargetScan (http://www.targetscan.org/) programs.

Formulation of Staramine liposomes containing anti-miR-20a

Anti-miR-20a oligonucleotides (5′-CTACCTGCACTATAAGCAC-3′) and control oligonucleotides (5′-TAACACGTCTATACGCCCA-3′) were purchased from Sigma-Aldrich Corporation. PEGylated Staramine nanoparticles (Celsion Corp.) were used to formulate anti-miR-20a oligonucleotides into a nanocomplex for use in vivo. Reagents were prepared according to the manufacturer’s instructions. Briefly, Staramine nanocomplex solution was prepared by mixing the oligonucleotide solution (0.2 mg/mL of anti-miR-20a or control oligonucleotide in a volume of 200 μL at 5% dextrose) with the Staramine nanoparticle solution (102 μL in a volume of 200 μL at 5% dextrose). The mixture was incubated for 5 minutes at room temperature and concentrated to a volume of 200 μL by centrifuge using an Amicon® Ultra-4 3kDa Centrifugal Filter Device (Billerica).

In vivo anti-miR-20a study

Approval for the animal experiment was obtained from the MD Anderson Institutional Animal Care and Use Committee (00000896-RN00). Four-week-old female nu/nu nude mice and BALB/c mice were fed in a specific pathogen-free animal facility approved by the American Association American Association for Laboratory Animal Science in accordance with current regulations and standards of the United States Department of Agriculture, the Department of Health and Human Services, and the National Institutes of Health. LM7 cells were injected into nude mice through the tail vein by using a sterile 30-gauge needle. A total of 2×10⁶ cells in a 200 μL cell suspension was administered. Six weeks after injection, two mice were randomly selected for euthanization, and micro-metastases in the lung were confirmed. The remaining mice were randomly grouped (9 mice per group) and treated with anti-miR-20a or control oligonucleotide formulated with PEGylated Staramine nanoparticles by i.v. injection (200 μL per mouse), twice per week for 4 weeks. Mice were euthanized 1 week after final treatment, and lungs were removed and evaluated for tumor by counting visible nodules.

K7M3 cells (3.0 × 10⁵) were injected in the tail vein of wild-type BALB/c mice, ten days later after lung micro-metastases was confirmed, mice were treated with anti-miR-20a or control oligonucleotide formulated with PEGylated Staramine nanoparticles twice per week for 3 weeks and analyzed 10 days later the same as above LM7 mouse lung tumor model analysis.

Immunohistochemical staining and quantification

Lung tissues were fixed in formalin and embedded in paraffin, followed by the five- micrometer sections mounted on glass slides for staining analysis. Immunohistochemical staining was performed to detect Fas expression in these lung section slides obtained from mice treated with anti-miR-20a or control oligonucleotide by using rabbit anti-human Fas antibody (Santa Cruz Biotechnology, Inc). Representative images were obtained using a cooled charge-coupled device C5810 camera (Hamamatsu Corporation) and the Optimas imaging software program (Media Cybernetics, Inc). Lung sections were scanned at 20× magnification, and the positively stained lesions were counted.

TUNEL assay

The paraffin embedded mouse lung tumor section slides were warmed for 20 minutes at 56°C, deparaffinized and rehydrated. Terminal deoxynucleotidyl transferase-mediated dUTP nick end-labeling (TUNEL) assay was carried out with the DeadEnd Fluorometric TUNEL Cell Death Detection Kit (Promega), following the manufacturer’s protocol. Samples were directly photomicrographed using a cooled charge-coupled device C5810 camera (Hamamatsu Corporation) and the Optimas imaging software program (Media Cybernetics, Inc) and analyzed.

Statistical analysis

All values were expressed as the mean ± SD. We performed a two-sided Student t-test using the Excel program (Microsoft) to determine statistical significance; P ≤ 0.05 was considered significant.

Results

Effect of miR-20a on the 3′-UTR of Fas mRNA

We first confirmed that there was an inverse correlation between Fas and miR-20a expression. To create single cell clones, Fas⁻ LM7 cells were sorted and expanded. Fas protein expression in the individual clones was then evaluated and compared to the non-metastatic SAOS-2 cells. Several cloned cells had no detectable Fas protein (Figure 1A). The miR-20a expression in each of these clones was ≥ 2 times that in the parental Fas⁺ SAOS-2 cells (Figure 1B). By contrast, expression of the non-related miRNA miR-20b was unchanged (Figure 1C).

Figure 1.

Expression of miR-20a was elevated in LM7 cells and Fas⁻ LM7 single cell clones. (A) Western blot of Fas protein expression in LM7 cell clones. Fas⁺, low-miR-20a-expressing SAOS-2 cells served as the control; actin blot as a loading control. (B, C) Relative expression levels of miR-20a (B) and miR-20b (C) in cloned LM7, LM7 and SAOS-2 cells was detected by RT-PCR analysis. Each bar represents the mean of three independent experiments with standard deviation.

miRNAs are known to control gene expression by recognizing a partial complementary sequence in the 3′-UTR of the target mRNA, resulting in translational repression and/or degradation of mRNA. This function leads to the silencing of the gene. Bioinformatics analysis of the Fas mRNA 3′-UTR whole sequence revealed no typical miR-20a target sites on the Fas mRNA 3′-UTR. However, we did find two weak binding regions at positions 410 and 1134 (Figure 2A). We therefore next determined whether miR-20a was binding to the Fas mRNA 3′-UTR, resulting in Fas downregulation. The 3′-UTR sequence of Fas mRNA was cloned into a luciferase reporter construct at a position immediately downstream from the luciferase gene (pMir-Fas3′UTR; Figure 2B). We also generated three more reporter constructs in which the two predicted regions targeting miR-20a within the Fas 3′-UTR were specifically mutated individually or together, in order to abolish miR-20a binding (pMir-Fas3′UTRm1, pMir-Fas3′UTRm2, and pMir-Fas3′UTRm12; Figure 2B). Either synthesized double-stranded oligonucleotide precursors that mimic endogenous mature miR-20a function (miR-20a precursors) or scramble control oligonucleotides were transfected into SAOS-2 (Fas⁺/low level of miR-20a) cells. miR-20a expression was confirmed by RT-PCR, and there was a subsequent decrease in Fas expression in the miR-20a precursor-transfected SAOS-2 cells (supplemental Figure 1, A and B).

Figure 2.

miR-20a does not target the Fas mRNA 3′-UTR in osteosarcoma cells. (A) Two putative miR-20a binding sites on the Fas mRNA 3′-UTR were identified. (B) pMir-Luc is the original luciferase construct and serves as control. Luciferase reporter constructs for the Fas mRNA 3′-UTR (pMir-Fas3′UTR), its two single miR-20a binding site mutations (pMir-Fas3′UTRm1 and pMir-Fas3′UTRm2), and the double miR-20a binding site mutation (pMir-Fas3′UTRm12). (C, D) Luciferase reporter assay for SAOS-2 cells (C) or U2OS (low miR-20a-expressing) cells (D) co-transfected with pMir-Fas3′UTR, pMir-Fas3′UTRm1, pMir-Fas3′UTRm2, or pMir-Fas3′UTRm12, together with either a scramble control or the miR-20a precursor. Each bar represents the mean of three independent experiments with standard deviation.

Since we demonstrated that miR-20a downregulates Fas expression (14; supplemental Figure 1, A–D), a decrease in luciferase activity for the Fas 3′-UTR was anticipated following co-transfection of pMir-Fas3′UTR and miR-20a precursor. Although Fas expression was decreased following transfection of the miR-20a precursor (supplemental Figure 1B), no change in the Fas 3′-UTR was demonstrated (Figure 2C). Similarly, there was no change when cells were transfected with the mutated Fas 3′-UTR constructs (pMir-Fas 3′UTRm1, pMir-Fas 3′UTRm2, and pMir-Fas 3′UTRm12; Figure 2C). These findings were duplicated using Fas⁺, low miR-20a-expressing U2OS osteosarcoma cells (Figure 2D; supplemental Figure 1, C and D). These results demonstrated that while miR-20a downregulates Fas expression, this decreased Fas expression is not mediated by targeting the Fas mRNA 3′-UTR.

To confirm that our assay system can detect the binding of miR-20a to a 3′-UTR binding site but that miR-20a does not target the Fas mRNA 3′-UTR, we employed a sense construct, pGL3-luc-miR-20a-sense (22), which has two confirmed miR-20 binding sites on the 3′-UTR of the firefly luciferase gene, as a positive miR-20a targeting indicator. The pGL3-luc-control construct, in which the sequences of the two miR-20 binding sites were reversed, served as the negative control (22). pGL3-luc-miR-20a-sense or pGL3-luc-control constructs were co-transfected with the miR-20a precursor or scramble control into 293T cells (high level of miR-20a). When co-transfected with either the miR-20a precursor or scramble control, the luciferase activity of the pGL3-luc-miR-20a-sense-transfected cells was lower than that of pGL3-luc-control-transfected cells (Figure 3A). Meanwhile, when co-transfected with scramble control, the pGL3-luc-miR-20a-sense showed more than 70% reduction in luciferase activity compared with pGL3-luc-miR-control (Figure 3A, grey columns). As 293T cells expressing high miR-20a, these results indicate that the reduction of luciferase activity in pGL3-luc-miR-20a-sense transfected 293T cells can be caused by the endogenous miR-20a targeting miR-20a binding sites on the pGL3-luc-miR-20a-sense, but not on the pGL3-luc-miR-control (without miR-20a binding sites).

Figure 3.

miR-20a does not target the Fas mRNA 3′-UTR in 293T and HeLa cells. (A) Luciferase reporter assay for co-transfection of a pGL3-luc-control or pGL3-luc-20a-sense construct with miR-20a oligonucleotide precursors or scramble control into 293T cells. The pGL3-luc-20a-sense construct has two miR-20a binding sites at its 3′UTR. pGL3-luc-control has two nonfunctional miR-20a binding sites in reverse orientation at the 3′UTR. (B) Luciferase reporter assay for co-transfection of a pGL3-luc-control or pGL3-luc-20a-sense construct with anti-miR-20a or control oligonucleotide into 293T cells. (C) Luciferase assay for Fas 3′-UTR luciferase reporter construct pMir-Fas3′UTR co-transfected with either miR-20a precursors or scramble control into 293T cells. (D) Luciferase assay for Fas 3′-UTR luciferase reporter construct pMir-Fas3′UTR co-transfected with either anti-miR-20a or control oligonucleotide into 293T cells. (E) Luciferase reporter assay for co-transfection of pGL3-luc-control or pGL3-luc-20a-sense construct with miR-20a oligonucleotide precursors or scramble control into HeLa cells. (F) Luciferase reporter assay for co-transfection of pGL3-luc-control or pGL3-luc-20a-sense construct with anti-miR-20a oligonucleotide or control oligonucleotide into HeLa cells. (G) Luciferase assay for Fas 3′-UTR luciferase reporter construct pMir-Fas3′UTR co-transfected with either miR-20a precursors or scramble control into HeLa cells. (H) Luciferase assay for Fas 3′-UTR luciferase reporter construct pMir-Fas3′UTR co-transfected with either anti-miR-20a or control oligonucleotide into HeLa cells. * P<0.05. Each bar represents the mean of three independent experiments with standard deviation.

Furthermore, when pGL3-luc-miR-20a-sense was co-transfected with miR-20a precursors, the luciferase activity of pGL3-luc-miR-20a-sense was further reduced as compared to co-transfection with the scramble control. By contrast, transfection with additional miR-20a precursors did not change the luciferase activity of pGL3-luc-control (Figure 3A). These results demonstrate successful downregulation when miR-20a is endogenously expressed or is overexpressed using miR-20a precursors and confirm the functional activity of the pGL3-luc-miR-20a-sense construct in detecting miR-20a binding. Co-transfection with anti-miR-20a oligonucleotides, but not control oligonucleotides, reversed the pGL3-luc-miR20a-sense downregulation caused by endogenous miR-20a expression (Figure 3B), further confirming that miR-20a targets specific binding sites on the pGL3-luc-miR-20a-sense construct.

We next investigated whether miR-20a affects the Fas 3′-UTR in 293T cells. When pMir-Fas3′UTR, a luciferase reporter construct that contains the Fas 3′-UTR, was transfected into 293T cells with high levels of miR-20a, neither the co-transfected miR-20a precursors (Figure 3C) nor anti-miR-20a (Figure 3D) had any effect on luciferase activity of pMir-Fas3′UTR. The studies were repeated and confirmed using HeLa cells (low expressing miR-20a and high Fas expression, Figure 3E–H). Taken together, these data support the conclusion that miR-20a does not downregulate Fas by targeting the 3′-UTR of Fas mRNA.

Effect of miR-20a on Fas promoter activity

Another potential mechanism for Fas downregulation is through transcriptional repression. Since miRNAs only regulate post-transcriptionally, we next determined whether miR-20a regulates Fas expression through the indirect modulation of the Fas promoter. For these studies, we used a Fas promoter plasmid with luciferase, pFas-FOR-Luc (23). Serial deletion of fragments from −720 to −19 bp (pFas-PR-Δ4) down to −80 to −19 bp (pFas-PR-Δ6.5) of the Fas promoter was used to generate individual Fas promoter reporter plasmids (Figure 4A). A plasmid containing the reverse orientation of the Fas promoter insert, pFas-P-REV, served as the negative control.

Figure 4.

miR-20a indirectly modulates Fas promoter activity. (A) Diagram of serial deletion constructs of the luciferase-labeled Fas promoter. (B) Luciferase assay for the Fas promoter reporter construct pFas-P-FOR co-transfected into HeLa cells with miR-20a precursors or scramble control. pFAS-P-REV, a negative control construct for Fas promoter reporter, was also co-transfected into HeLa cells with miR-20a precursors or scramble control. (C) Luciferase assay of the Fas promoter reporter construct pFas-P-FOR co-transfected into HeLa cells with anti-miR-20a or control oligonucleotide. pFAS-P-REV control was also co-transfected into HeLa cells with anti-miR-20a or control oligonucleotide. (D, E) The Fas promoter reporter constructs created through serial deletion were contransfected into HeLa cells with miR-20a precursors or scramble control. (F, G) The Fas promoter constructs created through serial deletion were contransfected into HeLa cells with anti-miR-20a or control oligonucleotide. * P<0.05. Each bar represents the mean of three independent experiments with standard deviation.

HeLa cells (low level of miR-20a) were transiently co-transfected with pFas-FOR-luc or pFas-P-REV (control) and either miR-20a precursors or anti-miR-20a oligonucleotides. As expected, co-transfection of HeLa cells with miR-20a precursors reduced pFas-FOR-luc activity by 39% compared with the scramble control (P < 0.01) (Figure 4B). By contrast, co-transfection of pFas-FOR-luc with anti-miR-20a oligonucleotides increased pFas-FOR-luc activity by 63% relative to co-transfection with control oligonucleotides (P < 0.01) (Figure 4C).

HeLa cells were then co-transfected with the individual plasmid for each Fas promoter deletion construct and either miR-20a precursors or the scramble control (Figure 4D, 4E). Deletions between −1739 and −240 bp on the Fas promoter (pFas-PR-Δ4, -Δ5, -Δ5.5, -Δ5.7, -Δ5.8, and -Δ6) had no effect on the ability of miR-20a to downregulate Fas promoter activity (Figure 4D). Further deletion after −240 bp (pFas-PR-Δ6.2 and pFas-PR-Δ6.3), however, resulted in the loss of response of Fas promoter to miR-20a but not scramble control, indicating the importance of this region for miR-20a having an indirect effect on the Fas promoter. No activity was seen after deletion of −150 to −120 bp (pFas-PR-Δ6.4 and pFas-PR-Δ6.5; Figure 4E). This lack of activity for construct pFas-PR-Δ6.4 may be explained by the deletion of the first transcription start site at the −129 bp position.

Consistently, anti-miR-20a oligonucleotides had an increased effect on Fas promotor activity when sequence −1739 to −240 bp was deleted (Figure 4F), but this effect was abolished when sequences after −240 bp were deleted (Figure 4G). Taking these findings together, we conclude that the 90 bp region on the Fas promoter from −240 to −150 bp just before the transcription start site is critical for the regulation of the Fas promoter by miR-20a.

In vivo delivery of nanoparticle-formulated anti-miR-20a oligonucleotides suppressed osteosarcoma lung metastasis

Our data show that miR-20a regulates Fas expression by indirectly affecting the transcription of the promoter. The majority of miRNAs that have been used in therapeutic approaches regulate gene expression via mediating mRNA 3′UTRs. However, since suppression of miR-20a in LM-7 cells upregulated Fas expression, and reduced their metastatic potential and the formation of lung metastases in vivo (14), we wished to determine whether targeting miR-20a has therapeutic potential against Fas⁻ LM7 osteosarcoma lung metastases. To determine whether inhibiting miR-20a in established osteosarcoma lung metastases would inhibit tumor growth, mice with confirmed micrometastases were treated with nanoparticle-formulated anti-miR-20a or control oligonucleotides by i.v. injection, twice a week for 4 weeks. Compared with the control nanoparticles, treatment with nanoparticles containing anti-miR-20a oligonucleotides resulted in a significant decrease in the number of LM7 lung nodules (Figure 5A). Meanwhile, the treatment of anti-miR-20a oligonucleotides resulted in an increase of Fas expression on remained lung metastatic cells of OS (Figure 5 B and C), demonstrating the effective function of anti-miR-20a oligonucleotides in vivo to upregulate Fas expression. TUNEL assay further demonstrates that increased Fas expression would induce more apoptosis on the lung metastatic nodules, as shown in Figure 5D and 5E, level of apoptosis in anti-miR-20a oligo treated mouse lung tumor nodules is significantly greater than that control oligo treated mice indicating that the increase expression of Fas by anti-miR-20a treatment may result to elevated apoptosis in metastasis lung tumor nodules.

Figure 5.

Anti-miR-20a oligonucleotides suppress osteosarcoma lung metastasis. LM7 cells were injected into nude mice through the tail vein. Six weeks after injection, micro-metastases in the lung were confirmed as detailed in Material and Methods. Mice were randomly grouped (9 mice per group) and treated with anti-miR-20a or control anti-miR formulated with PEGylated Staramine nanoparticles by i.v. injection (200 μL), twice per week for 4 weeks. Mice were euthanized 1 week after the final treatment. Lungs were extracted and fixed. (A) Lungs were examined under the microscope to confirm the response of OS lung metastases and the visible OS tumor nodules were quantified. (B) Immunohistochemistry was performed to detect the expression of Fas on OS lung metastases. Fas expression on representative OS lung metastases was shown (20× magnification); (C) Fas expression on the lung metastatic nodules was analyzed and quantified. (D) Detection of apoptosis by TUNEL assay on the mouse lung metastatic nodules from control oligo treated mice and anti-miR-20a oligo treated mice. Apoptosis cells stained by fluorescein (green) on representative OS lung metastases was shown (20× magnification); (E) apoptosis cells on the lung metastatic nodules was analyzed and quantified. (F) mouse K7M3 cells were injected in the tail vein of wild-type BALB/c mice, ten days later after lung micro-metastases was confirmed, mice were treated with anti-miR-20a or control oligonucleotide formulated with PEGylated Staramine nanoparticles twice per week for 3 weeks and lungs were extracted and fixed 10 days later. Lungs were examined under the microscope to confirm the response of OS lung metastases and the visible OS tumor nodules were quantified. P value was calculated using Student t-test.

We also generate another cell line xenograft lung metastasis model using a mouse osteosarcoma metastasis cell line K7M3. By using the same nano particle delivered anti-miR-20a oligo, we do see increased miR-20a expression can inhibit the growth of K7M3 lung metastases as we have seen in LM7 lung met model (Figure 5F). These results indicate that the suppression of OS lung metastases by the treatment with nanoparticles containing anti-miR-20a oligonucleotides is mediated by the upreguatlion of Fas.

Discussion

We previously demonstrated that the downregulation of the Fas gene in osteosarcoma cells plays a major role in their ability to form lung metastases where the lung epithelium constitutively expresses FasL (1, 2, 21, 24–26). While primary osteosarcoma tumors in the bone contained both Fas⁺ and Fas⁻ cells, those in the lung were almost exclusively Fas⁻ (4). We also demonstrated that the downregulation of Fas on the surface of osteosarcoma cells was via an epigenetic mechanism (27, 28). Several agents were able to induce the re-expression of Fas on the cell surface both in vitro and in vivo in established osteosarcoma lung metastases, which was followed by tumor regression (29–33). We went on to demonstrate that CpG methylation was not responsible for Fas gene silencing (34). Rather, this downregulation was mediated by the miRNA cluster miR-17-92, which encodes miR-20a (14).

In this study, we investigated the target of miR-17-92 and miR-20a in an effort to understand the mechanism of Fas silencing. The most common way miRNA negatively regulates gene expression is through the degradation or translational repression of the targeted mRNA, accomplished by interacting with the target’s 3′-UTR. Bioinformatics analysis of the Fas mRNA 3′-UTR revealed no miR-20a binding sequences. However, there were two weak regions where the match threshold was decreased. In keeping with the bioinformatics analysis, while miR-20a transfection led to decreased Fas mRNA, direct targeting of the Fas 3′-UTR was not seen using the luciferase reporter assay. Endogenous miR-20a did not decrease luciferase activity of the transfected Fas mRNA 3′-UTR. Similarly, no change was seen in response to miR-20a precursor oligonucleotides and anti-miR-20a oligonucleotides.

By contrast, reporter assays using Fas promoter-driven luciferase expression showed that the activity of the Fas promoter was affected by miR-20a. Serial deletions from the 5′ end revealed a 90 bp region (−240 to −150 bp) that was the critical region. While the transcriptional proteins responsible for this downregulation of the Fas promoter have not yet been identified, this is the first report showing that miR-20a gene regulation of Fas is mediated by an effect on the promoter and not by binding to the Fas 3′-UTR mRNA. Recent evidence supports our findings that members of the miR-17-92 cluster such as miR-20a may regulate non-protein coding transcripts and have targets in addition to or instead of the 3′ UTR (35). miR-20a is one of the oncogenic miRNAs found to cause cancer in multiple tissues and organs (35–37). miR-20a has been shown to target more than a dozen mRNAs, including the transcription factor E2F1 (38). Targeting sequence bioinformatics analysis predicts that miR-20a can target more than 1200 genes by affecting the mRNA 3′-UTRs (39). More than 100 of these genes encode for transcriptional factors. Identifying the transcriptional factor regulated by miR-20a that binds to the Fas promoter merits further investigation.

More than 25 years have passed since a new therapeutic approach has been identified for the treatment of children and adolescents with osteosarcoma. Different combinations of chemotherapy have failed to improve the 60-65% long-term survival rate of these patients. Once a patient relapses or fails to respond to chemotherapy, effective therapeutic options are limited as planned intensification treatment with combined chemotherapy has made no impact on long-term survival (40). A meta-analysis published in 2011 showed that while treating newly diagnosed osteosarcoma patients with a 3-drug was superior to using 2 drugs, there was no benefit in using 4 drugs (41). Randomized clinical trials confirmed this, showing that the addition of ifosfamide (standard or high dose with or without etoposide) to the 3-drug regimen of cisplatin, doxorubicin, and high-dose methotrexate added considerable toxicity but no patient benefit in terms of improving either event-free or overall survival (40,42–44). Finally, a recent review of pediatric and adult relapsed osteosarcoma patients treated in 14 unique Phase I clinical trials showed no response to any of the cytotoxic, targeted, or biologic agents used (45). Unconventional and more creative approaches are thus a priority.

We have demonstrated that cell surface Fas expression and an intact Fas signaling pathway play an important role in the potential of osteosarcoma cells to metastasize to the lung (21, 25–29). Downregulation of Fas expression or upregulation of c-FLIP (which blocks the Fas signaling pathway) promotes the metastatic phenotype. Conversely, inducing Fas expression in established Fas⁻ osteosarcoma lung metastases resulted in tumor regression (21, 23, 26, 31, 32). Therefore, identifying the mechanism by which the expression of Fas and other death-receptor proteins on the cell surface of osteosarcoma cells is regulated not only will help our understanding of the metastatic process but also may facilitate the identification of new therapeutic approaches for tumors that metastasize to the lung. The concept of altering the expression of Fas on the cell surface of osteosarcoma cells and harnessing the lung microenvironment to aid in tumor killing is a novel therapeutic concept worth pursuing. For maximum effectiveness and to ensure the selection of the correct patient population, a better understanding of the tumor biology is required.

Our previous work demonstrated that the Fas receptor is downregulated in osteosarcoma cells, which in turn allows them to survive in the FasL⁺ lung microenvironment (4). This epigenetic downregulation is mediated by miR-20a encoded by the miR-17-92 cluster (14). The expression of miR-20a has been shown to be upregulated in osteosarcoma patient tumors compared with matched adjacent normal tissues (46). Furthermore, overexpression of miR-20a promoted osteosarcoma cell proliferation by suppressing EGR2, accelerating the G1/S phase transition, upregulating cyclin D1, and downregulating p21 (46). Our data on miR-20a–mediated downregulation of Fas, together with the data cited above, provide compelling evidence that miR-20a plays a critical role in osteosarcoma tumorigenesis, growth, and metastasis and that targeting this miRNA may block several different pathways and thus have multiple beneficial and therapeutic effects.

The oligonucleotides therapeutic application research are rapidly advanced in recent years and many pre-clinical data has been generated, but oligonucleotide-based therapeutics has not yet delivered a clinical drug to the market in the cancer field due to all kinds of different challenges presented for cancer targeting in a clinical setting, these include oligonucleotides poor stability in vivo, inadequate targeting, potential off-target effects, potential toxicity and immunostimulation as well as inefficient intracellular delivery to target cells or tissues (47–49). Different chemical modifications of the oligonucleotides have been widely applied to achieve increased resistance to degradation, increased affinity to target sequence. A well-designed oligonucleotide with different types of nucleotide modifications directed by detailed bioinformatics analysis can also minimize the off-target effects, potential toxicity and unwanted immune activation. The efficient and targeted delivery of the oligonucleotides could be achieved by oligonucleotides associated with nanoparticle to enhance permeability and retention effect (47, 50, 51).

Here, we demonstrated that miR-20a regulates Fas expression by modulating Fas promoter activity. Furthermore, we demonstrated that the i.v. injection of nanoparticles formulated with anti-miR-20a oligonucleotides (but not control oligonucleotides) suppressed established osteosarcoma lung metastases. Nucleic acid-based therapeutics are dependent upon an efficient method to specifically deliver nucleic acids (such as an anti-miR or siRNA) to the tumor cells. It has been shown that the cationic liposomes have strong potential as a therapeutic delivery system (51). PEGylated Staramine, which is lipid based delivery systems that has been functionalized by the addition of methoxypolyethylene glycol, has been shown to efficiently deliver inhibitory oligonucleotides to the lung in mice and rats resulting durable target specific knockdown (52, 53). Our results suggest that an anti-miR-20a oligonucleotide formulated with PEGylated Staramine nanoparticles can significantly inhibit established osteosarcoma lung metastases in mice mediated by upreregulation of Fas expression of OS and may therefore be a novel therapeutic approach to treat relapsed patients that were unresponsive to chemotherapy.