The present study was approved by the Ethics Committee of the First Affiliated Hospital of Anhui Medical University (Hefei, China). Tumor tissues and adjacent noncancerous tissues (normal tissues) were collected from 27 patients with melanoma (aged 29-71 years; 15 male and 12 female), who were diagnosed at the First Affiliated Hospital of Anhui Medical University between March 2014 and January 2018. No patients received chemotherapy or any other types of therapy before surgery. Inclusion criteria included histologically confirmed melanoma. Exclusion criteria included: i) Current or previous history of any other severe or uncontrolled diseases, including previous history of cancer, bleeding dyscrasia and immune system disease; ii) currently on medications that could interfere with assessment of biological outcomes, including estrogens, progestogens, androgens, prednisone or psychoactive drugs; and iii) aged <16 years. All patients signed a written informed consent. Fresh tissues were immediately frozen and stored in liquid nitrogen.

Human melanoma cells (M14 and A375) and human primary normal epidermal melanocytes HEMa-LP were purchased from the American Type Culture Collection and cultured in DMEM (Gibco; Thermo Fisher Scientific, Inc.) with 10% FBS (Invitrogen; Thermo Fisher Scientific, Inc.) in a humidified atmosphere with 5% CO₂ at 37˚C, unless otherwise stated.

Specific small interfering (si)-RNAs against TUG1 (si-TUG1#1, si-TUG1#2 and si-TUG1#3; final concentration, 20 nM), the scrambled negative control (NC; si-NC; final concentration, 20 nM), miR-145-5p mimic (final concentration, 15 nM), miR-NC (final concentration, 15 nM), pcDNA-TUG1 (final concentration, 1.5 µg/ml), si-SOX2 (final concentration, 20 nM), miR-145-5p inhibitor (in-miR-145-5p; final concentration, 25 nM), its negative control (in-miR-NC; final concentration, 25 nM), pcDNA-SOX2 (final concentration, 1 µg/ml), specific short hairpin (sh)RNA against TUG1 (sh-TUG1) and its negative control sh-NC were all synthesized by Shanghai GenePharma Co., Ltd. The pcDNA plasmid was purchased from Thermo Fisher Scientific, Inc. The sequences for the siRNAs, miRNA mimics and inhibitors used in the present study were shown in Table I.

The aforementioned oligonucleotides or plasmids were transfected into M14 and A375 cells using Lipofectamine^™ 2000 reagent (Invitrogen; Thermo Fisher Scientific, Inc.) following the manufacturer's instructions. Briefly, M14 and A375 cells were cultured for 24 h, before the constructs were transfected into the cells using 500 µl Opti-MEM containing 20 µl Lipofectamine^™ 2000 (1 mg/ml) in a humidified atmosphere with 5% CO₂ at 37˚C. After incubating for 6 h, culture medium was replaced by fresh medium. Subsequently, 24, 48 or 72 h after transfection, the cells were collected for further analysis.

For the detection of mRNA expression levels, RNA was isolated from melanoma tissues or cells using TRIzol^® Reagent (Invitrogen; Thermo Fisher Scientific, Inc.). RNA was reverse-transcribed into cDNA using M-MLV Reverse Transcriptase (cat. no. 28025021; Invitrogen; Thermo Fisher Scientific, Inc.) with RNaseOUT^™ Recombinant Ribonuclease Inhibitor (cat. no. 10777; Thermo Fisher Scientific, Inc.), dNTPs (cat. no. 10297117; Thermo Fisher Scientific, Inc.) and the oligo (dT) 12-18 primer (cat. no. 18418012; Thermo Fisher Scientific, Inc.) following the standard protocols: 65˚C for 5 min; 50 min at 37˚C; and 70˚C for 15 min. Subsequent qPCR was conducted using iQ^™ SYBR^® Green Supermix (cat. no. 1708882; Bio-Rad Laboratories, Inc.). For the detection of miR-145-5p expression levels, miRNA was extracted using a miRNeasy Mini Kit (Qiagen, GmbH). Reverse transcription and RT-qPCR were conducted using a TaqMan^™ Reverse Transcription kit and TaqMan^™ MicroRNA Assays (Applied Biosystems; Thermo Fisher Scientific, Inc.). β-actin was used as the internal control for TUG1 and SOX2, whereas U6 was used as the internal control for miR-145-5p. The primers for TUG1, SOX2, β-actin, miR-145-5p and U6 are as follows: TUG1 forward, 5'-GGACCTGGAACCCGAAAGAG-3' and reverse, 5'-TGGTGGTAGTGCTTGCTCAG-3'; SOX2 forward, 5'-CAGCGCATGGACAGTTACG-3' and reverse, 5'-TTCATGTAGGTCTGCGAGCTG-3'; β-actin forward, 5'-TGGACTTCGAGCAAGAGATGG-3' and reverse, 5'-ACGTCACACTTCATGATGGAG-3'; miR-145-5p forward, 5'-CAGTCTTGTCCAGTTTTCCCAG-3' and reverse, 5'-TATGCTTGTTCTCGTCTCTGTGTC-3'; U6 forward, 5'-CTCGCTTCGGCAGCACA-3' and reverse, 5'-AACGCTTCACGAATTTGCGT-3'. RT-qPCR was performed on a StepOnePlus^™ Real-time PCR Systems (Applied Biosystems; Thermo Fisher Scientific, Inc.) with the following thermocycling conditions: Initial denaturation for 10 min at 95˚C; 40 cycles of 95˚C for 15 sec and 60˚C for 1 min; followed by final elongation at 72˚C for 10 min. The expression levels of TUG1, SOX2 and miR-145-5p were analyzed using the 2^-ΔΔCq method (20).

For the detection of cell proliferative ability, cells were seeded in 96-well plates containing DMEM with 10% FBS (1,000 cells per well) for 0, 24, 48 and 72 h. The CCK-8 reagent (MedChemExpress) was used for this assay. After collection, 10 µl CCK-8 reagent was added into each well. In total, after incubation at 37˚C for 2 h, the proliferative ability was determined through measuring the absorbance at 450 nm of each well using a spectrophotometer (NanoDrop Technologies; Thermo Fisher Scientific, Inc.).

Transwell migration and invasion assay was performed using a transwell chamber (8 µm pore size; EMD Millipore). For migration assays, 3x10⁴ transiently transfected M14 and A375 cells in serum-free DMEM were seeded into the upper chamber, whilst DMEM supplemented with 10% FBS was added to the lower chamber. Subsequently, the chamber was incubated at 37˚C. In total, 24 h later, the migrated cells were fixed with 4% paraformaldehyde for 15 min at room temperature, before being stained with 0.1% crystal violet for 30 mins at room temperature and counted under an inverted light microscope (Nikon Eclipse TS100; Nikon Corporation) at x20 magnification. For invasion assays, the protocol used was identical, but the upper chamber was precoated with 50 µl Matrigel (8 mg/ml; 1:8 dilution; Corning, Inc.) at 4˚C overnight.

The downstream target miRNAs of TUG1 were predicted using the online website LncBase Predicted v.2 (http://carolina.imis.athenainnova-tion.gr/diana_tools/web/index.php?r=lncbasev2%2Findex-predicted). Direct targets of miR-145-5p were predicted by online databse miRTarBase (http://mirtarbase.mbc.nctu.edu.tw/php/detail.php?mirtid=MIRT000307). The wild-type (WT) or mutant (MUT) sequences of TUG1 and SOX2 were amplified and cloned into pGL3 luciferase promoter vector (Promega Corporation) to synthesize TUG1 WT, TUG1 MUT, SOX2 WT and SOX2 MUT. Renilla luciferase reporter plasmids were used as the control for normalization. In total, 1x10⁶ M14 and A375 cells were co-transfected with 8 µg recombinant luciferase reporter plasmids, 0.16 µg pRL-TK plasmids and miR-145-5p mimics (final concentration, 15 nM) or miR-NC (final concentration, 15 nM) using Lipofectamine^TM 2000 (Invitrogen; Thermo Fisher Scientific, Inc.) according to the manufacturer's instruction. The luciferase activity was evaluated using the Dual-Luciferase^® reporter system (Beyotime Institute of Biotechnology) according to the protocols supplied by the manufacturer.

The RIP assay was performed using the extracts obtained from the melanoma M14 and A375 cells using the EZ-Magna RIP kit (cat. no. 17-701; EMD Millipore) according to the manufacturer's instructions, as previously described (21). In brief, the extracted supernatant (150 µg protein) was incubated with 1 mg magnetic beads that had been pre-coated with human anti-argonaute2 (Ago2) antibodies (cat. no. ab186733, 1:50, Abcam) or IgG (cat. no. PP64B, 1:20, EMD Millipore) at 4˚C overnight, which was used as a negative control. After overnight incubation at 4˚C, the RNA complex were incubate with proteinase K (RIP wash buffer, 10% SDS and 10 mg/ml proteinase K; EMD Millipore) for 30 min at 55˚C to digest the protein remaining on the beads. The RNA was then purified from the samples using phenol: chloroform: isoamyl alcohol separation. RNA detection was conducted using RT-qPCR as aforementioned.

Cell lysates of melanoma cells (M14 and A375) were prepared using the RIPA Reagent (Cell Signaling Technology, Inc.). The protein concentration was quantified with bicinchoninic acid reagent (Sigma-Aldrich; Merck KGaA). A total of 30 µg protein samples were fractionated on 15% SDS-polyacrylamide gels and then transferred onto PVDF membranes. Thereafter, the membranes were blocked with 5% skimmed milk at room temperature for 2 h and probed with primary antibodies targeting SOX2 (1:1,000, cat. no. cat. no. ab171380; Abcam) or β-actin (1:1,000; cat. no. ab8227; Abcam) at 4˚C overnight. Subsequently, the membranes were incubated with horseradish peroxidase-conjugated goat anti-rabbit secondary antibodies (1:1,000; cat. no. ab205718, Abcam) at room temperature for 2 h. The bands were visualized using an ECL detection kit (Thermo Fisher Scientific, Inc.) and quantified using the ImageJ 1.8.0 software (National Institute of Health). The aforementioned antibodies used for this investigation were purchased from Abcam (Cambridge, MA, USA).

The animal experiments, which lasted 28 days, were approved by the Animal Care and Scientific Committee of the First Affiliated Hospital of Anhui Medical University (Hefei, China). A total of 12 male BALB/c nude mice (5-week-old; 16-20 g weight range) were obtained from Shanghai SLAC Laboratory Animal Co., Ltd. (grouped into a sh-TUG1 group and sh-NC group, n=6 per group) and maintained under temperature (20-22˚C) and humidity-controlled (55±10%) pathogen-free specific conditions with a 12-h light/dark cycle and given water and food ad libitum. The stable A375 cells expressing sh-NC or sh-TUG1 were produced and maintained in medium containing 3 µg/ml puromycin (Sigma-Aldrich; Merck KGaA). Prior to injection, the stable A375 cells were resuscitated and grown for two passages. A total of 4x10⁶ stable A375 cells expressing sh-NC or sh-TUG1 were collected and suspended in 100 µl sterile PBS solution and then subcutaneously inoculated into the right flank of the nude mice. Tumor volume was measured every 7 days for 28 days and calculated according to the following equation: Volume=(length x width²)/2. The health and behavior of nude mice involved in the animal experiment were also monitored every 7 days after inoculation. All mice were sacrificed at day 28 following inoculation, for a pre-set 28-day test cycle. The 12 nude mice were anesthetized using 2% methoxyflurane (an inhalant anesthetic) (22). Subsequently, the animals were sacrificed by cervical dislocation. The xenograft tumor tissues were then collected for measurement of weight and used for RT-qPCR, western blot analyses and immunohistochemical assays.

The IHC assays were conducted to detect the expression profile of SOX2 in resected tumors from the sh-NC group and sh-TUG1 group, following the protocols from a previous report (23). In brief, resected tumors from the xenograft models were fixed in 3% formaldehyde overnight at 4˚C and embedded in paraffin and then cut into 5-µm sections. The slices were then incubated with 0.3% hydrogen peroxide (H₂O₂) solution in methanol for 20 min at room temperature to block the activity of endogenous peroxidases. Sections were blocked in 10% horse serum at room temperature for 1 h and then incubated with goat anti-SOX2 antibodies (1:100; cat. no. GT15098; Neuromics) at 4˚C overnight, followed by incubation with peroxidase-conjugated horse anti-goat IgG (cat. no. PI-9500-1; 1:2,000; Vector Laboratories, Inc.) at room temperature for 30 min. Immunoreactivity was measured using NovaRed peroxidase substrate (Vector Laboratories, Inc.) under an Eclipse TS100 fluorescence microscope (Nikon Corporation) at x20 magnification.

All data analyses were conducted using SPSS 22.0 software (IBM Corp.). Data from three independent experiments are presented as the mean ± SD. Difference between means was analyzed using paired Student's t-tests (for 2 groups) and one-way ANOVAs (for ≥3 groups) followed by Tukey's post hoc tests. Pearson correlation coefficient was used to analyze the correlation among TUG1, miR-145-5p and SOX2. Differences were considered significant at P<0.05.

To investigate the role of TUG1 in melanoma, the expression levels of TUG1 were analyzed using RT-qPCR. As shown in Fig. 1A, TUG1 expression was significantly upregulated in melanoma tissues compared with that in normal tissues. In addition, the expression levels of TUG1 were significantly raised in the melanoma M14 and A375 cell lines compared with that in normal melanocyte cell line, HEMa-LP cells (Fig. 1B).

siRNAs against TUG1 (si-TUG1#1, si-TUG1#2 and si-TUG1#3) were transfected into M14 and A375 cells to construct TUG1-silenced melanoma cells to explore the effect of TUG1 on the proliferative, migratory and invasive abilities of melanoma cells. RT-qPCR assays were conducted to detect the silencing efficiency of the 3 siRNAs (si-TUG1#1, si-TUG1#2 and si-TUG1#3) on the TUG1 expression in the transfected M14 and A375 cells. The mRNA expression levels of TUG1 in M14 and A375 cells were decreased after the indicated transfections, with the silencing efficiency of si-TUG1#3 being the most significant (Fig. 2A and B). As such, si-TUG1#3 was used for subsequent functional experiments. CCK-8 assays were employed to detect the proliferative ability of transfected M14 and A375 cells, which indicated that a reduced TUG1 expression inhibited the proliferative ability of the two cell lines (Fig. 2C and D). Transwell assay indicated that downregulation of TUG1 also repressed the migration and invasion of M14 and A375 cells (Fig. 2E and F).

LncBase Predicted v.2 was applied to seek the downstream target of TUG1 and it was hypothesized that miR-145-5p could bind to TUG1 (Fig. 3A). Dual-luciferase assays were performed to validate the relationship between TUG1 and miR-145-5p. RT-qPCR assay indicated that the expression of miR-145-5p was significantly upregulated by transfection of miR-145-5p mimic (Fig. S1A), but was significantly suppressed more by transfection with the miR-145-5p inhibitor (Fig. S1B). In addition, both transfection of TUG1 wt or TUG1 mut significantly elevated TUG1 expression (Fig. S1C). As depicted in Fig. 3B and C, the luciferase activity of M14 and A375 cells co-transfected with TUG1 WT and miR-145-5p mimics was obviously lower than that of cells co-transfected with TUG1 WT and miR-NC mimics. By contrast, the luciferase activity of M14 and A375 cells co-transfected with TUG1 MUT and miR-145-5p mimics or miR-NC mimics showed no significant difference. A RIP assay was performed to confirm the interaction between TUG1 and miR-145-5p, which showed that TUG1 mRNA could be specifically recruited to the miRNA complexes isolated using anti-Ago2 antibody following miR-145-5p transfection, suggesting that TUG1 could interact with miR-145-5p (Fig. 3D). RT-qPCR was performed to analyze the expression level of miR-145-5p in melanoma tissues and cells. miR-145-5p was found to be downregulated in melanoma tissues compared with normal tissues (Fig. 3E). In addition, downregulation of miR-145-5p was also detected in melanoma M14 and A375 cells compared with HEMa-LP cells (Fig. 3F). Introduction of TUG1 repressed the expression levels of miR-145-5p in both M14 and A375 cells relative to cells transfected with pcDNA; while silencing of TUG1 promoted miR-145-5p expression in the two cell lines compared with cells transfected with si-NC, indicating that TUG1 inversely regulated miR-145-5p expression levels (Fig. 3G).

To clarify whether TUG1 regulates the proliferative, migratory and invasive abilities of melanoma cells by targeting miR-145-5p, recovery experiments were performed. RT-qPCR assays showed that miR-145-5p mimics significantly elevated miR-145-5p expression levels in both M14 and A375 cells, while introduction of TUG1 counteracted the promoting impact of miR-145-5p mimics on miR-145-5p expression (Fig. 4A and B). CCK-8 assays also indicated that upregulation of miR-145-5p inhibited the proliferative ability of M14 and A375 cells, whereas elevated TUG1 almost reversed the repressive effect of miR-145-5p mimics on the proliferative ability of the two cell lines (Fig. 4C and D). Subsequently, the impact of miR-145-5p alone or miR-145-5p combined with TUG1, on the migratory and invasive abilities of melanoma cells was assessed using Transwell assays. As displayed in Fig. 4E and F, overexpression of miR-145-5p significantly inhibited the migratory and invasive abilities of M14 and A375 cells, while raised expression levels of TUG1 significantly weakened the repressive impact of miR-145-5p on the migratory and invasive abilities of M14 and A375 cells.

The online software mirTarBase identified SOX2 as a direct target of miR-145-5p. The putative binding site and its MUT for miR-145-5p on the 3'-UTR of SOX2 mRNA are displayed in Fig. 5A. RT-qPCR assay indicated that SOX2 expression was significantly elevated by SOX2 wt or SOX2 mut transfection (Fig. S1E). Subsequently, the targeted relationship between miR-145-5p and SOX2 was validated using dual-luciferase reporter assays. The assay indicated that upregulation of miR-145-5p significantly constrained the luciferase activity of SOX2 WT reporter in co-transfected M14 and A375 cells, compared to the miR-NC transfected cells. However, neither the miR-145-5p nor the miR-NC had little impact on the luciferase activity of the SOX2 MUT reporter in co-transfected M14 and A375 cells (Fig. 5B and C). Subsequently, RIP assays were also implemented to confirm the direct interaction between miR-145-5p and SOX2, and it was suggested that transfection with miR-145-5p triggered SOX2 enrichment in the RIP-Ago2 group in melanoma M14 and A375 cells compared with that in the RIP-IgG group, supporting the notion that SOX2 was a direct target of miR-145-5p (Fig. 5D). RT-qPCR assays indicated that SOX2 was significantly upregulated in melanoma tissues in comparison to normal tissues (Fig. 5E). Additionally, western blot assays suggested that the protein expression levels of SOX2 were higher in melanoma M14 and A375 cells than that in HEMa-LP cells (Fig. 5F). Western blot assays that the revealed introduction of miR-145-5p reduced the protein expression levels of SOX2 and that silencing of TUG1 also reduced the SOX2 protein expression levels (Fig. 5G and H). Pearson analysis illuminated a negative correlation between miR-145-5p and SOX2 expression levels (Fig. 5I), as well as between TUG1 and miR-145-5p expression levels (Fig. 5J), while a positive correlation was found between TUG1 and SOX2 expression levels (Fig. 5K) in melanoma tissues.

To further clarify that TUG1 mediates the proliferative, migratory and invasive abilities of melanoma cells by targeting the miR-145-5p/SOX2 axis, restoration experiments were conducted. As shown in Fig. S1D, the expression of SOX2 was significantly by transfection with si-SOX2. The effects of SOX2 on the proliferative ability of melanoma cells were subsequently evaluated. CCK-8 assays found that knockdown of SOX2 repressed the proliferative ability of M14 and A375 cells; however, coinstantaneous knockdown of miR-145-5p weakened the inhibitory impact (Fig. 6A and B). Similarly, knockdown of SOX2 blocked the migratory and invasive capacities of M14 and A375 cells, while simultaneous knockdown of miR-145-5p relieved the inhibitory effect (Fig. 6C and D). The CCK-8 assay also suggested that introduction of SOX2 reversed the inhibitory impact of si-TUG1 on the proliferative ability of M14 and A375 cells (Fig. 6E and F). Upregulation of SOX2 also rescued the suppressive effect of si-TUG1 on the migratory and invasive capacities of M14 and A375 cells (Fig. 6G and H).

Finally, A375 cells stably transfected with sh-NC or sh-TUG1 were constructed, with transfections being confirmed using RT-qPCR (Fig. 7A). Downregulation of TUG1 repressed the tumor volume (Fig. 7B) and weight (Fig. 7C) in nude mice compared with the sh-NC group. In addition, the expression levels of TUG1 were detected in tumor tissues of nude mice and the results showed that TUG1 was significantly reduced in the sh-TUG1 group compared with the sh-NC group (Fig. 7D). In addition, TUG1 silencing significantly elevated miR-145-5p expression in xenograft tumor tissues (Fig. 7E). IHC staining assay for SOX2 indicated that SOX2-positive cells in sh-TUG1 group were decreased compared with that in the sh-NC group (Fig. 7F). Additionally, downregulation of TUG1 also inhibited SOX2 protein expression levels in the tumors from nude mice (Fig. 7G). In summary, TUG1 regulated the proliferation, migration and invasion of melanoma cells by modulating the miR-145-5p/SOX2 axis (Fig. 8).