STAT3-induced ZBED3-AS1 promotes the malignant phenotypes of melanoma cells by activating PI3K/AKT signaling pathway

ABSTRACT

Melanoma is considered as the most frequent primary malignancy occurring in skin. Accumulating studies have suggested that long non-coding RNAs (lncRNAs) play critical parts in multiple cancers. In this study, we explored the molecular mechanism of ZBED3 antisense RNA 1 (ZBED3-AS1) in melanoma. We observed that ZBED3-AS1 expression was remarkably up-regulated in melanoma tissues, and high ZBED3-AS1 level was linked to unsatisfactory survival of melanoma patients. Then, we discovered that ZBED3-AS1 was overexpressed in melanoma cells compared with human epidermal melanocytes. In addition, loss-of-function assays verified that ZBED3-AS1 knockdown restrained cell proliferation, migration, epithelial–mesenchymal transition (EMT), and stemness in melanoma. In addition, signal transducer and activator of transcription 3 (STAT3), which also showed tumour-facilitating functions in melanoma, was confirmed as a transcriptional activator of ZBED3-AS1. Moreover, ZBED3-AS1 enhanced the expression of AT-rich interaction domain 4B (ARID4B) through sequestering miR-381-3p. Importantly, we further confirmed that ZBED3-AS1 promoted the malignant progression of melanoma by regulating miR-381-3p/ARID4B axis to activate the phosphatidylinositol 3-kinase/AKT serine/threonine kinase (PI3K/AKT) signalling pathway. In a word, our research might provide a novel therapeutic target for melanoma.

Introduction

Melanoma is identified as the most common skin cancer and its morbidity accounts for about 4% of skin cancer cases around the world [1,2]. Because of the high metastatic potential of melanoma, the survival rate of patients with advanced melanoma is no more than 15% [3,4]. Accordingly, it is necessary to delve into the cellular mechanisms affecting the tumorigenesis and progression of melanoma, which may supply novel targets for melanoma treatment.

Long non-coding RNAs (lncRNAs) are a group of RNAs with over 200 nucleotides in length and without protein-coding ability [5]. Increasing evidence demonstrates that lncRNAs participates in the progression of diverse cancers. For instance, lncRNA CCAT1 regulates miR-148a/PIK3IP1 signalling to facilitate osteosarcoma cell proliferation and migration [6]. LncRNA HOXA11-AS prompts glioma cell growth and metastasis via the regulation of miR-130a-5p/HMGB2 axis [7]. LncRNA AK027294 contributes to cell growth in non-small cell lung cancer via modulating STAT3 expression [8]. In melanoma, numerous lncRNAs have also been verified to modulate cancer initiation and progression, such as MHENCR [9], LINC00963 [10], HEIH [11] and ILF3-AS1 [12]. In our study, ZBED3 antisense RNA 1 (ZBED3-AS1) is a novel lncRNA whose role has not yet been recognized in melanoma. According to GEPIA data, we found the aberrant expression of ZBED3-AS1 in melanoma tissues in contrast to normal skin tissues. TCGA data also suggested that abnormally high ZBED3-AS1 expression was correlated with poor overall survival of melanoma patients. On the basis of these data, we aimed to further investigate the functional role of ZBED3-AS1 in melanoma.

Furthermore, it has been reported that transcription factors are pivotal regulators of lncRNAs in many cancers [13]. For instance, RUNX2 up-regulated PVT1 promotes the progression of colorectal cancer [14]. CREB1-induced FOXD2-AS1 exerts a promoting effect on cell proliferation and metastasis in glioma through miR-185/AKT1 axis [15]. SOX2-activated PVT1 enhances breast cancer cell growth and invasion [16]. However, transcription factor which could induce ZBED3-AS1 up-regulation in melanoma has never been discovered so far. Signal transducer and activator of transcription 3 (STAT3) is a member of the signal transducer and activator of transcription (STAT) family. Functionally, STAT3 transduces signals from multiple extracellular stimuli, making it a vital mediator of inflammation, cell survival, differentiation, and proliferation [17]. Accumulating evidence reveals that STAT3 acts as a transcription factor to play vital parts in various cancers, such as hepatocellular carcinoma [18], glioma [19] and oesophageal adenocarcinoma [20]. However, STAT3 has never been linked with ZBED3-AS1 before. Thus, we proposed to further explore whether STAT3 served as a transcription factor to affect ZBED3-AS1 expression in melanoma.

Moreover, lncRNAs are identified to act as competing endogenous RNAs (ceRNAs) to sponge microRNAs (miRNAs) and modulate protein-coding genes at the post-transcriptional level. As previously reported, lncRNA GAS5 inhibits the metastasis of hepatocellular carcinoma by miR-182/ANGPTL1 axis [21]. HOXD-AS1 promotes cell proliferative and migratory capabilities in ovarian cancer by targeting miR-608 to regulate FZD4 expression [22]. Aforementioned studies indicated that miRNAs make a difference to cancer by cooperating with lncRNAs. As for our research, we attempted to determine if ZBED3-AS1 could positively affect the proliferative, migratory abilities, EMT as well as stemness of melanoma cells by targeting miR-381-3p to augment AT-rich interaction domain 4B (ARID4B) expression. More importantly, it has been reported that ARID4B regulates the tumorigenesis of prostate cancer through the PTEN-PI3K pathway [23]. Additionally, the phosphatidylinositol 3-kinase/AKT serine/threonine kinase (PI3K/AKT) pathway has been reported to exert pivotal impact on cell survival, cell cycle progression and cell growth in multiple cancers, including ovarian cancer [24], lung cancer [25] and breast cancer [26]. However, ZBED3-AS1 has never been shown to regulate PI3K/AKT pathway in melanoma.

In current research, we aimed to expound the role and molecular mechanism of ZBED3-AS1 in melanoma, which might offer a potential therapeutic target for melanoma.

Materials and methods

Cell lines

Human melanoma cell lines (A2058, A375, M21, and SK-MEL-1) were available from the ATCC (Manassas, VA, USA). The human epidermal melanocytes HEMa-LP were procured from Invitrogen (Carlsbad CA, USA). All cells were cultivated in DMEM (Gibco, Rockville, MD, USA) in 5% CO₂ at 37°C, in the presence of 10% FBS and 1% antibiotics (both, Gibco).

Total RNA extraction and RT-qPCR

Total RNA was isolated by applying TRIzol Reagent (Invitrogen), followed by cDNA synthesis using Reverse Transcription Kit (Takara, Kyoto, Japan) according to the standard method. SYBR Green PCR Kit (Takara) and ABI 7500 real-time polymerase chain reaction system (Applied Biosystems, Foster City, CA, USA) were utilized for quantitative analysis. Gene expression relative to GAPDH or U6 was calculated by means of the 2^−ΔΔCt method. The primer sequences used here were listed in Supplementary Table 1.

Subcellular fractionation

Subcellular fractionation assay in M21 and SK-MEL-1 cells was undertaken with PARIS™ Kit (Invitrogen) in light of the guidebook. Cells were treated with cell fractionation buffer to separate cell cytoplasm and cell nuclei. ZBED3-AS1 content in both fractions was finally determined by RT-qPCR, while GAPDH and U6 were utilized respectively as the cytoplasmic and nuclear controls.

Fluorescence in situ hybridization (FISH)

The ZBED3-AS1-specific RNA FISH probe was designed by Ribobio (Guangzhou, China) for conducting FISH assay. Cell nuclei were detected using DAPI dye (Beyotime, Shanghai, China). After staining, all samples were observed under Olympus fluorescence microscope (Olympus Corp., Tokyo, Japan).

Cell transfection

M21 and SK-MEL-1 cells were cultured in 48-well plates for 48 hours of transfection using Lipofectamine 3000 (Invitrogen). The ZBED3-AS1-specific shRNAs and control-shRNAs were procured from GenePharma (Shanghai, China). To overexpress STAT3 or ARID4B, their full-length cDNA sequences were individually inserted into pcDNA3.1 vector (Invitrogen), and the empty vector served as the negative control. The miRNA mimics, miR-381-3p inhibitor, and their relevant controls were acquired from Ribobio. The sequences of shRNAs targeting ZBED3-AS1 were as below: sh/Ctrl: 5ʹ-CCGGAACAAAATCCAGCTTTTCACACTCGAGTTTTTTTADDTGGAAAATGCCTTTTTG-3ʹ; sh/ZBED3-AS1#1: 5ʹ-CCGGCTGCAAGAAACTGAATGTTCTCTCGAGAGAACATTCAGTTTCTTGCAGTTTTTG-3ʹ; sh/ZBED3-AS1#2: 5ʹ-CCGGTTCATTATGGAGCTAAACACACTCGAGTGTGTTTAGCTCCATAATGAATTTTTG-3ʹ.

Colony formation assay

M21 and SK-MEL-1 cells were plated into 6-well plates (800 cells per well) for 14 days of culture, followed by fixation in formaldehyde for 30 minutes. After being stained in 0.5% crystal violet for 5 minutes, the visible colonies in three separate assays were imaged and counted manually.

5-ethynyl-2ʹ-deoxyuridine (EdU) assay

EdU assay was undertaken with the help of BeyoClick™ EdU Cell Proliferation Kit with Alexa Fluor 594 (Beyotime). After being rinsed by PBS, cells were collected and then incubated with 50 nM of EdU medium for 2 hours. Afterwards, cells were fixed for DAPI staining and assayed using Olympus fluorescence microscope. EdU assay was performed in triplicate.

Flow cytometry analysis

Cell apoptosis was studied by flow cytometry as required by the supplier (BD Biosciences, Franklin Lakes, NJ, USA) using double Annexin V/PI staining kit (Invitrogen). After 15-minute double-staining in darkness, cells in 6-well plates were rinsed with PBS and then analysed by flow cytometry. This assay was undertaken individually for three times.

TdT-mediated dUTP nick-end labelling (TUNEL) assay

M21 and SK-MEL-1 cells were first rinsed with the application of PBS and then fixed for TUNEL assay. The TUNEL assay reagent (Merck KGaA, Darmstadt, Germany) was utilized for the staining of apoptotic cells. Following the staining of nuclei via DAPI, cells were then imaged and analysed by using Olympus optical microscopy. The experiment was repeated for three times.

JC-1 assay

Cell apoptosis was also measured by detecting the changes in mitochondrial membrane potential (Δψm) by means of JC-1 assay. M21 and SK-MEL-1 cells were prepared in 12-well plates for loading with 2.5 µg/ml of JC-1 dye (Beyotime) at 37°C for 30 min. After washing in PBS, cells were observed via Olympus fluorescence microscope. This assay was performed in three independent repeats.

Transwell migration assay

The transfected M21 and SK-MEL-1 cells suspended in serum-free medium were transferred into the top compartment of transwell insert (Corning, Corning, NY, USA), with the lower compartment filled with complete culture medium. After 24-hour incubation, the migrated cells in lower chamber were treated with formaldehyde and crystal violet, and then the cells in 5 random fields were counted under the Olympus microscope. Three independent transwell assays were conducted to assess cell migration.

Immunofluorescence (IF) staining

M21 and SK-MEL-1 cells were cultured for 24 hours on culture slides for adhering to the slides, and then washed with PBS and fixed for IF assay. In short, slides were blocked in 5% BSA and treated with primary antibody against E-cadherin (ab40772; Abcam, Cambridge, UK) N-cadherin (ab18203; Abcam), or IgG followed by processing with corresponding secondary antibodies (ab7086; Abcam). All antibodies used were acquired from Abcam (Cambridge, MA, USA). After washing and DAPI staining, cells in the slides were imaged via Olympus fluorescence microscope. IgG antibody was utilized as a control and the secondary antibody against E-cadherin or N-cadherin was incubated to determine the validity of staining. Three IF assays were conducted individually.

Western blot

Cells were treated with RIPA lysis buffer to obtain total protein, and then the protein samples were separated on 10% SDS-PAGE and shifted on PVDF membranes. After being blocked, the membranes were incubated with prepared primary antibodies against NANOG (ab21624, Abcam), OCT4 (ab19857, Abcam), SOX2 (ab97959, Abcam), STAT3 (ab119352, Abcam), ARID4B (A302-233A, Bethyl Laboratories, Texas, USA), PIK3CA (ab40776, Abcam), PIK3R2 (ab28356, Abcam), p-PI3K (ab138364, Abcam), PI3K (#4292, Cell Signalling Technology, Massachusetts, USA), p-AKT (#9271, Cell Signalling Technology), AKT (#9272, Cell Signalling Technology) and the loading control GAPDH (ab181602, Abcam) all night at 4°C, followed by the incubation with HRP-tagged secondary antibodies (ab205718, Abcam). After washing in TBST, the protein signals were finally detected by ECL luminous liquid (Pierce, Rockford, IL, USA). The western blot analysis was undertaken independently for three times.

Sphere formation assay

1 × 10⁴ cells were planted into indicated well of 6-well plates containing sphere medium, and then cells were incubated for 12-day until the spheres with over 20 cells. Thereafter, the spheres were observed and measured under a light microscope. This assay was implemented in triplicate.

Chromatin immunoprecipitation (ChIP)

ChIP assay was undertaken with the assistance of the EZ ChIP™ Chromatin Immunoprecipitation Kit (Millipore, Bedford, MA). After fixation, the cross-linked samples were treated with ultrasonic to acquire about 500-bp DNA fragments. The immunoprecipitation was conducted for 6 hours with anti-STAT3 antibody (ab119352, Abcam) or control anti-IgG antibody (30000-0-AP; Proteintech, Chicago, USA). In total, 30 μl of magnetic beads were then added for collecting the precipitated chromatin, which was then subjected to RT-qPCR analysis.

Luciferase reporter assay

The ZBED3-AS1 promoter sequences with wild-type (WT) or mutant (MUT) STAT3 binding sites were inserted to pGL3-basic vectors (Promega, Madison, WI, USA). After that the constructs were co-transfected with STAT3 overexpression vectors for 48 hours into M21 and SK-MEL-1 cells by use of Lipofectamine 3000. In addition, the full-length ZBED3-AS1 or ARID4B 3ʹUTR sequences with wild-type or mutant miR-381-3p binding sites were inserted into pmirGLO vectors (Promega) to obtain pmirGLO-ZBED3-AS1-WT/MUT or pmirGLO-ARID4B-WT/MUT. Thereafter, abovementioned plasmids were subjected to co-transfection with miR-381-3p mimics or NC mimics into M21 and SK-MEL-1 cells. Relative luciferase activity of aforementioned luciferase reporter plasmids was estimated by means of luciferase reporter assay system (Promega).

RNA pull down assay

RNA pull-down experiments were implemented in line with the previous protocol [27]. In short, after lysing M21 and SK-MEL-1 cells in RIPA lysis buffer, the cell lysates were collected and then subjected to incubation with biotinylated RNA probes for ZBED3-AS1 or miR-381-3p, or biotin-labelled wild type or mutant type of miR-381-3p (Bio-miR-381-3p-WT or Bio-miR-381-3p-MUT). After adding the magnetic beads, RNA-RNA mixtures were pulled down and these RNAs were then analysed by RT-qPCR.

RNA immunoprecipitation (RIP)

RIP assay was conducted in M21 and SK-MEL-1 cells following the former protocol [28], by the use of Magna RIP™ RNA-Binding Protein Immunoprecipitation Kit (Millipore). Briefly, cell lysates were obtained from two cells above via the RIPA buffer with RNase inhibitor and protease inhibitor cocktail. After that, cell lysates were processed with magnetic beads which were conjoined with control IgG antibody (30000-0-AP; Proteintech) or human Ago2 antibody (10686-1-AP; Proteintech). Finally, RNAs captured by beads were purified via proteinase K and then analysed by RT-qPCR.

Statistical analyses

Each experiment was implemented for no less than three times. Prism 6.0 (GraphPad Software, Inc., La Jolla, CA) was applied for assessing group differences. The statistical analysis was performed in form of Student’s t-test, one-way or two-way ANOVA, with a p-value below 0.05 as the significant threshold. All data were exhibited as mean ± SD.

Results

ZBED3-AS1 is overexpressed in melanoma tissues and cells

To figure out whether ZBED3-AS1 was linked with melanoma, GEPIA 2 (http://gepia2.cancer-pku.cn/#analysis) was applied to evaluate its expression in melanoma tissues and the correlation with the survival of melanoma patients. As displayed in Figure 1(a), ZBED3-AS1 expression was obviously augmented in 461 specimens from melanoma patients compared with 558 specimens from normal controls. Simultaneously, the overall survival of melanoma patients was stratified by ZBED3-AS1 expression. The overall survival of melanoma patients with high ZBED3-AS1 expression (n = 227) was less satisfactory relative to that in the low-expression group (n = 228) (Figure 1(b)). In addition, UCSC (http://genome.ucsc.edu/) database was utilized for analysing the expression of ZBED3-AS1 in different kinds of normal tissues. The results showed that ZBED3-AS1 was expressed at a relatively low level in most normal human tissues including normal skin tissues (Figure 1(c)). Based on above results, we speculated that ZBED3-AS1 expression might be related to melanoma development. To verify our assumption, we further adopted RT-qPCR to detect the expression of ZBED3-AS1 in melanoma and corresponding normal cells. As anticipated, ZBED3-AS1 expression was apparently elevated in melanoma cell lines including A2058, A375, M21, and SK-MEL-1, in contrast to human normal epidermal melanocytes HEMa-LP (Figure 1(d)). We then selected M21 and SK-MEL-1 cells to conduct the following experiments in view of the highest expression of ZBED3-AS1 in these cells. As we know, the biological function of lncRNAs depends on their subcellular distribution to some extent [29,30]. To ascertain the cellular location of ZBED3-AS1, we conducted subcellular fractionation and FISH assays in melanoma cells. The results displayed that ZBED3-AS1 principally existed in the cytoplasm of melanoma cells (Figure S1A-S1B), suggesting that ZBED3-AS1 might participate in post-transcriptional regulation in melanoma. Totally, ZBED3-AS1 is overexpressed in melanoma tissues and cells.

Figure 1.

ZBED3-AS1 was overexpressed in melanoma tissues and cells. (a) Data from GEPIA 2 revealed the expression of ZBED3-AS1 in 461 melanoma tumour tissues (T) compared to 558 normal ones (N). (b) The overall survival of melanoma patients with high or low expression of ZBED3-AS1 and ZBED3-AS1 expression in these two groups of melanoma tissues were obtained from TCGA. (c) UCSC data showed the expression of ZBED3-AS1 in different kinds of normal tissues. (d) RT-qPCR was used to detect the fold change of ZBED3-AS1 expression in melanoma cells (A2058, A375, M21 and SK-MEL-1) relative to normal human epidermal melancocytes (HEMa-LP). *P < 0.05, **P < 0.01

ZBED3-AS1 knockdown inhibits cell proliferation and promotes apoptosis in melanoma

To further determine the role of ZBED3-AS1 in melanoma cells, we first transfected two shRNAs targeting ZBED3-AS1 into M21 and SK-MEL-1 cells. RT-qPCR analysis verified the effective knockdown of ZBED3-AS1 in both cells by sh-ZBED3-AS1#1 or sh-ZBED3-AS1#2 (Figure S1C). Furthermore, FISH assay validated the weakened signals of ZBED3-AS1 in ZBED3-AS1-silenced groups (Figure S1D). Subsequently, we tested the impact of silenced ZBED3-AS1 on melanoma cell proliferation. Results indicated that the quantity of colonies diminished due to ZBED3-AS1 silencing (Figure 2(a)). Similarly, we found that the absence of ZBED3-AS1 in melanoma cells caused a decrease of EdU positive cells (Figure S2A). Based on previous literature, dysregulated cell apoptosis is also an important symbol of cancers [31]. Accordingly, we conducted a succession of assays to examine whether ZBED3-AS1 influenced melanoma cell apoptosis. Shown by the data from flow cytometry analysis, the apoptosis rate significantly increased in melanoma cells responding to ZBED3-AS1 knockdown (Figure 2(b)). Likewise, the results from TUNEL assay manifested that TUNEL positive cell percent rose after ZBED3-AS1 silencing in melanoma cells (Figure S2B). Furthermore, the outcomes of JC-1 assay further verified that inhibiting ZBED3-AS1 elevated the apoptosis capability of melanoma cells (Figure 2(c)). Totally, ZBED3-AS1 knockdown hinders cell proliferation, whereas facilitates cell apoptosis in melanoma.

Figure 2.

ZBED3-AS1 knockdown inhibited cell proliferation and promoted cell apoptosis in melanoma. (a) Pictures from colony formation assay of M21 and SK-MEL-1 cells with ZBED3-AS1 silencing were taken and colony numbers of each group were counted manually. (b). Flow cytometry analysis assessed the influence of ZBED3-AS1 down-regulation on the apoptosis rate of M21 and SK-MEL-1 cells. (c) JC-1 experiment detected the apoptosis of melanoma cells with ZBED3-AS1 silencing versus control. **P < 0.01

Silencing of ZBED3-AS1 inhibits the migration, EMT and stemness of melanoma cells

Further, we planned to elucidate whether ZBED3-AS1 affected melanoma cell migration. As displayed in Figure 3(a), ZBED3-AS1 depletion dramatically suppressed the migration ability of M21 and SK-MEL-1 cells. Considering EMT process always occurs during cancer development [32], we adopted IF assay to examine the positivity of E-cadherin and N-cadherin in melanoma cells with ZBED3-AS1 knockdown. It was illustrated in M21 and SK-MEL-1 cells that ZBED3-AS1 inhibition increased E-cadherin expression while reducing N-cadherin expression. In IgG group, no fluorescent signal was captured after incubation with secondary antibody against E-cadherin/N-cadherin (Figure 3(b)). Furthermore, we applied RT-qPCR and western blot to assess the mRNA and protein levels of stemness-related markers in melanoma cells. As revealed in Figure S2C-S2D, ZBED3-AS1 reduction remarkably lessened the levels of NANOG, OCT4, and SOX2, suggesting that ZBED3-AS1 might be related to the stem cell characteristics of melanoma. Sphere formation assay results further testified that ZBED3-AS1 depletion weakened the sphere formation ability of melanoma cells (Figure S2E). Together, silencing of ZBED3-AS1 restrains the migration, EMT, and stemness of melanoma cells.

Figure 3.

Silencing of ZBED3-AS1 inhibited cell migration and EMT in melanoma. (a) Transwell assays were conducted to display the migration capacity of M21 and SK-MEL-1 cells with the transfection of sh/Ctrl or shRNAs targeting ZBED3-AS1. (b) IF assay examined the expression level of E-cadherin and N-cadherin in M21 and SK-MEL-1 cells transfected with different plasmids (sh/Ctrl, sh/ZBED3-AS1#1 or sh/ZBED3-AS1#2). Anti-IgG was utilized as control and the secondary antibodies against E-cadherin and N-cadherin were incubated to determine the validity of staining. **P < 0.01

STAT3 elevates ZBED3-AS1 expression through transcriptional activation

To find out the upstream molecule responsible for ZBED3-AS1 overexpression in melanoma cells, we focused on certain transcription factor that might affect ZBED3-AS1 transcription. First, we analysed the alterations on ZBED3-AS1 expression caused by shRNAs targeting the potential transcription factors. Seen from Figure S3A, we observed that the most evident downregulation of ZBED3-AS1 expression was caused by the silencing of STAT3 but not others. Moreover, we found that melanoma cells possessed higher STAT3 level than the normal HEMa-LP cells (Figure 4(a)). Melanoma cell lines also exhibited more STAT3 protein and activated p-STAT3 levels (Figure 4(b)). Thereafter, it was proved that along with the elevated level of STAT3 induced by pcDNA3.1/STAT3 (Figure S3B), the expression of ZBED3-AS1 was also up-regulated in M21 and SK-MEL-1 cells (Figure 4(c)). On the contrary, STAT3 knockdown led to reduced ZBED3-AS1 expression in these two cells (Figure S3C and 4D). Furthermore, we observed that depleting STAT3 could hamper melanoma cell proliferation (Figure S3D-S3E), stimulate cell apoptosis (Figure S3F), and also hinder cell migration (Figure S3G) and stemness (Figure S3H). Moreover, we predicted that there were four binding sites for STAT3 in ZBED3-AS1 promoter using JASPAR website (http://jaspar.genereg.net/) (Figure 4(e–f)). To confirm which site was responsible for the binding of STAT3 to ZBED3-AS1 promoter, ChIP assays were performed. As a result, only P1 fragment of ZBED3-AS1 promoter was positively enriched in anti-STAT3 groups (Figure 4(g)). Also, the outcomes of luciferase reporter assays indicated that the luciferase activity of reporters with P1 fragment was apparently augmented, while that of P2 fragment unaffected under STAT3 up-regulation (Figure 4(h)). Furthermore, it manifested that the luciferase activity of reporters with wild-type ZBED3-AS1 promoter was obviously enhanced due to up-regulated STAT3 in M21 and SK-MEL-1 cells, and that of reporters containing site 1 or site 2 mutated ZBED3-AS1 promoter was also slightly augmented after STAT3 enhancement, whereas that of reporters covering promoter with both mutated site 1 and site 2 was nearly unaffected in response to STAT3 elevation (Figure 4(i)). All these results suggested that STAT3 serves as a transcription activator of ZBED3-AS1 in melanoma cells.

Figure 4.

STAT3 served as transcription factor to activate ZBED3-AS1 expression. (a) STAT3 level in melanoma cells versus normal HEMa-LP cells was detected by RT-qPCR. (b) STAT3 and p-STAT3 proteins in melanoma cells versus normal HEMa-LP cells were detected by western blot. (c) ZBED3-AS1 level in M21 and SK-MEL-1 cells with STAT3 overexpression was detected by RT-qPCR. (d) ZBED3-AS1 level in M21 and SK-MEL-1 cells with STAT3 knockdown was detected by RT-qPCR. (e,f). JASPAR website predicted the binding motif of STAT3 and four binding sites for STAT3 in ZBED3-AS1 promoter. (g) ChIP assays tested the enrichment of P1 or P2 fragment of ZBED3-AS1 promoter in anti-STAT3 groups and control anti-IgG groups. (h) Luciferase reporter assays analysed the influence of STAT3 augmentation on the luciferase activity of reporters with P1 or P2 fragment of ZBED3-AS1 promoter. (i) Luciferase reporter assays measured the luciferase activity of different types of ZBED3-AS1 promoter (including WT, Site 1-MUT, Site2-MUT and Site 1 + 2-MUT) in M21 and SK-MEL-1 cells with or without STAT3 up-regulation. *P < 0.05, **P < 0.01, ***P < 0.001. n.s.: no significance

ZBED3-AS1 promotes cell proliferation, migration, and stemness in melanoma through a miR-381-3p-dependent way

In view of the main distribution of ZBED3-AS1 in cytoplasmic fraction, it was postulated that ZBED3-AS1 might be a ceRNA in melanoma cells. Bioinformatics analysis revealed 15 miRNAs that had potential to combine with ZBED3-AS1 (Figure S4A). However, RNA pull down assay data manifested that only five of them (including miR-381-3p, miR-3118, miR-382-5p, miR-506-5p and miR-362-5p) were significantly pulled down by biotin-labelled ZBED3-AS1 in both M21 and SK-MEL-1 cells (Figure 5(a) and S4B). Furthermore, the luciferase activity of ZBED3-AS1 was obviously reduced by miR-381-3p mimics while no change of that was observed when upregulating other miRNAs (Figure 5(b) and S4C). Then, starBase (http://starbase.sysu.edu.cn) predicted the highly matched sequences between miR-381-3p and ZBED3-AS1 (Figure 5(c)). Moreover, the outcomes of luciferase reporter assays certified that the luciferase activity was apparently lessened in M21 and SK-MEL-1 cells co-transfected with miR-381-3p mimics and ZBED3-AS1-WT, whereas no change of that was observed in M21 and SK-MEL-1 cells co-transfected with miR-381-3p mimics and ZBED3-AS1-MUT (Figure 5(d) and S4D). Meanwhile, RNA pull-down assay data revealed that ZBED3-AS1 was largely abundant in Bio-miR-381-3p-WT group (Figures 5(e) and S4E), indicating the interaction between these two molecules in melanoma cells. Further, rescue experiments were implemented to corroborate the function of ZBED3-AS1/miR-381-3p axis in melanoma cells. As demonstrated in Figure 5(f) and S4F, the number of colonies declined after silencing ZBED3-AS1 while such decrease was thoroughly rescued by miR-381-3p inhibitor. Similar results were also observed via EdU assays (Figure 5(g) and S4G). In contrast, flow cytometry depicted that miR-381-3p inhibition completely reversed the apoptosis facilitated by ZBED3-AS1 knockdown in M21 and SK-MEL-1 cells (Figure 5(h) and S4H). In addition, transwell assay results displayed that ZBED3-AS1 knockdown suppressed the migration ability of M21 and SK-MEL-1 cells, but this effect was totally counteracted by miR-381-3p inhibitor (Figure 5(i) and S4I). We also observed that the sphere formation efficiency in melanoma cells was reduced under ZBED3-AS1 deficiency, whereas such phenomena were absolutely reversed by suppressing miR-381-3p (Figure 5(j) and S4J). Together, ZBED3-AS1 prompts cell proliferative and migratory capabilities as well as stemness in melanoma through targeting miR-381-3p.

Figure 5.

ZBED3-AS1 promoted the proliferation, migration and stemness of melanoma cells through targeting miR-381-3p. (a) RNA pull down assay were implemented to quantify the enrichment of 15 miRNAs in ZBED3-AS1 probe groups. (b) Luciferase reporter assays examined the luciferase activity of ZBED3-AS1 in M21 cells transfected with mimics targeting five indicated miRNAs (miR-381-3p, miR-3118, miR-382-5p, miR-506-5p and miR-362-5p). (c) StarBase predicted the binding sites between ZBED3-AS1 and miR-381-3p. (d) Luciferase reporter assays assessed the luciferase activity in M21 cells co-transfected with miR-381-3p/NC mimics and ZBED3-AS1-WT/MUT. (e) RNA pull down assay depicted the enrichment of ZBED3-AS1 in Bio-miR-381-3p-WT groups. (f) Colony formation assay detected the proliferation of M21 cells transfected with sh/Ctrl, sh/ZBED3-AS1#1, sh/ZBED3-AS1#1+ NC inhibitor or sh/ZBED3-AS1#1+ miR-381-3p inhibitor. (g) EdU assays assessed the proliferation of M21 cells in four groups above. (h) Flow cytometry analysis detected the apoptosis rate of M21 cells in indicated groups. (i) Transwell assay tested the migration ability of M21 cells with different transfections. J. Sphere formation assay examined the stemness of indicated M21 cells. **P < 0.01, ***P < 0.001. n.s.: no significance

ARID4B is the target of miR-381-3p in melanoma cells

Subsequently, we aimed at unveiling the downstream mediator of ZBED3-AS1/miR-381-3p signalling in melanoma cells. In this regard, the targets of miR-381-3p were searched. Several bioinformatics programmes in starBase including miRmap, TargetScan, PicTar, miRanda and microT revealed 13 mRNAs that might be targeted by miR-381-3p (Figure S5A). Then, RNA pull-down assay was performed to figure out which mRNA combined with miR-381-3p in melanoma cells. The results attested that ARID4B, ZBED6 and CCNY were significantly harvested in miR-381-3p probe groups in both M21 and SK-MEL-1 cells (Figure 6(a)). We selected ARID4B for following experiments due to the highest enrichment of ARID4B in miR-381-3p probe groups. Thereafter, RT-qPCR and western blot depicted that the level of ARID4B was evidently reduced by ZBED3-AS1 silencing or miR-381-3p up-regulation (Figure 6(b–c)). Then, we obtained the binding sites between miR-381-3p and ARID4B using starBase website (Figure 6(d)). As expected, the luciferase activity noticeably decreased in M21 and SK-MEL-1 cells with the transfection of miR-381-3p mimics and ARID4B-WT while no apparent alteration of that was not observed in miR-381-3p mimics and ARID4B-MUT co-transfected group (Figure 6(e)). Furthermore, the strong abundance of ARID4B was undoubtedly observed in Bio-miR-381-3p-WT groups (Figure 6(f)). Additionally, Ago2-RIP assays affirmed that ZBED3-AS1, miR-381-3p, and ARID4B were highly enriched in Ago2 groups (Figure 6(g)), indicating that these three RNAs co-existed in Ago2-related RNA-induced silencing complexes (RISCs). All these results suggested that ARID4B is the downstream target of ZBED3-AS1/miR-381-3p axis in melanoma cells.

Figure 6.

ARID4B was the target of miR-381-3p in melanoma cells. (a) RNA pull down assay detected the enrichment of 13 mRNA candidates in miR-381-3p probe and control groups. (b) RT-qPCR and western blot detected the mRNA and protein levels of ARID4B in M21 and SK-MEL-1 cells with or without ZBED3-AS1 silence. (c) RT-qPCR and western blot results of ARID4B expression in two melanoma cells with or without miR-381-3p up-regulation. (d) StarBase website revealed the binding sites between miR-381-3p and ARID4B. (e) Luciferase reporter assay detected the luciferase activity in M21 and SK-MEL-1 cells co-transfected with miR-381-3p mimics (or NC mimics) and ARID4B-WT (or ARID4B-MUT). (f) RNA pull down assay examined the abundance of ARID4B in indicated groups. (g) Ago2-RIP assays detected the enrichment of ZBED3-AS1, miR-381-3p and ARID4B in Ago2 groups relative to IgG groups. *P < 0.05, **P < 0.01, ***P < 0.001. n.s.: no significance

ZBED3-AS1 promotes the proliferation, migration, and stemness of melanoma cells through miR-381-3p/ARID4B to activate PI3K/AKT pathway

To further determine whether ARID4B mediated the regulation of ZBED3-AS1 on melanoma cell growth, a set of rescue experiments were performed. As presented in Figure S5B, the expression of ARID4B evidently rose in M21 and SK-MEL-1 cells transfected with plasmids covering ARID4B. Then, the results of colony formation assays revealed that the impeded proliferation ability led by ZBED3-AS1 depletion was completely rescued by ARID4B overexpression (Figure 7(a)). Also, the results of EdU assay showed that the proportion of EdU positive cells lowered by ZBED3-AS1 deficiency was totally recovered by ARID4B up-regulation (Figure 7(b)). Flow cytometry analysis displayed that the apoptosis rate was enhanced as a result of ZBED3-AS1 inhibition, while such enhancement was completely reversed by ARID4B elevation (Figure 7(c)). Similarly, TUNEL assays showed that the percent of TUNEL positive cells increased by ZBED3-AS1 knockdown was drastically recovered by ARID4B overexpression (Figure S5C). Transwell assay testified that ARID4B up-regulation totally enhanced the migration capacity hindered by silenced ZBED3-AS1 in SK-MEL-1 cells (Figure 7(d)). RT-qPCR and western blot analyses indicated that the suppression of ZBED3-AS1 knockdown on the expressions of NANOG, OCT4 and SOX2 was drastically offset by ARID4B up-regulation (Figure S5D-S5E). Sphere formation assay results indicated that the abrogated sphere formation ability of ZBED3-AS1-depleted cells was fully recovered after overexpressing ARID4B (Figure S5F). In addition, since it has been reported that ARID4B can activate PI3K pathway, we also analysed the effect of ZBED3-AS1/ARID4B axis on proteins related to PI3K/AKT pathway in M21 and SK-MEL-1 cells. As indicated in Figure S6A, the levels of PIK3CA, PIK3R2, p-PI3K, and p-AKT were obviously reduced by ZBED3-AS1 deficiency whereas were normalized under ARID4B overexpression. All these results indicated that ZBED3-AS1 contributes to the malignant behaviours of melanoma cells through miR-381-3p/ARID4B axis to activate PI3K/AKT signalling.

Figure 7.

ZBED3-AS1 promoted the proliferation and migration of melanoma cells through miR-381-3p-targeted ARID4B. Rescue assays were performed in SK-MEL-1 cells with the transfection of sh/Ctrl, sh/ZBED3-AS1#1, sh/ZBED3-AS1#1+ Vector or sh/ZBED3-AS1#1+ ARID4B. (a,b) Colony formation and EdU assays examined the proliferation of cells in these four groups. (c) Flow cytometry detected cell apoptosis rate in these four groups. (d) Transwell assay tested the migration capacity of cells in these four groups. **P < 0.01

Discussion

Melanoma is a common death-related cancer in the world [33]. There is still no completely satisfactory treatment for melanoma patients developing into advanced stages [34]. Thus, it is necessary to explore effective biomarkers and therapeutic targets for melanoma. Recently, the significant role of lncRNAs in melanoma has gradually come into people’s view. For instance, lncRNA CASC15 accelerates melanoma cell proliferation and invasion through regulating PDCD4 [35]. LncRNA-UCA1 boosts the proliferative and migratory capabilities of melanoma cells via binding with miR-28-5p to elevate HOXB3 expression [36]. LINC00888 contributes to the tumorigenesis of melanoma via targeting miR-126 to upregulate CRK [37]. The main subject of this study was lncRNA ZBED3-AS1. This lncRNA has been identified to be closely linked to various biological processes, including the chemoresistance of ovarian cancer cells [38], and human bone marrow mesenchymal stem cell differentiation [39]. Different from previous studies, our research first found that ZBED3-AS1 was overexpressed in melanoma tissues and cells. Significantly, high expression of ZBED3-AS1 suggested poor overall survival of melanoma patients. In addition, ZBED3-AS1 silencing apparently suppressed cell proliferation, migration, EMT and stemness in melanoma. Thus, ZEBD3-AS1 was first demonstrated as a promoter for melanoma progression.

Since transcription factors can affect the expression of lncRNAs via transcriptional regulation [40], we further explored the potential transcription factor upstream of ZBED3-AS1 in melanoma. STAT3 has been discovered to serve as a transcription factor of diverse lncRNAs in multiple cancers. As an example, STAT3 activated HOXD-AS1 boosts SOX4 expression to facilitate liver cancer metastasis [27]. Similarly, our study first found that STAT3 bound to ZBED3-AS1 promoter to activate ZBED3-AS1 expression in melanoma cells, so as to facilitate cell proliferation, migration, EMT and stemness.

Moreover, since ZBED3-AS1 was mainly located in the cytoplasm of melanoma cells, we further explored the post-transcriptional control of ZBED3-AS1 in melanoma. Previous reports have demonstrated that lncRNAs can act as a ceRNA to sequester miRNAs in cancer. As reported by former works, lncRNA DDX11-AS1 promotes colorectal cancer progression through targeting miR-873 to enhance CLDN7 level [41]. LncRNA SNHG16 acts as the sponge of miR-98-5p to affect cellular processes in osteosarcoma [42]. PTENP1 acts as a ceRNA to upregulate PTEN by sequestering miR-19b in breast cancer [43]. In line with these reports, our study found that ZBED3-AS1 sponged miR-381-3p in melanoma cells. As reported previously, miR-381-3p plays a crucial part in the progression of several cancers, such as papillary thyroid carcinoma [44], breast cancer [45], and cervical cancer [46]. Importantly, miRNAs usually function through negatively modulating the expression of target mRNAs in post-transcriptional regulation [47]. Present study unveiled that there was a binding site between miR-381-3p and ARID4B. For all we know, ARID4B is a good biomarker correlated with WHO grades in glioma and meningioma [48]. Herein, we found that ZBED3-AS1 promoted ARID4B expression to accelerate cell proliferation, migration and stemness in melanoma through sequestering miR-381-3p. Meanwhile, ARID4B has been identified to be a positive regulator of PI3K/AKT signalling pathway, which was in line with the previous findings in prostate cancer reported by Wu, R.-C., et al. [23]. Furthermore, we confirmed that ZBED3-AS1 activated PI3K/AKT pathway by upregulating ARID4B.

In conclusion, this research discovered that STAT3-induced ZBED3-AS1 aggravates cell proliferation, migration, EMT and stemness in melanoma through activating PI3K/AKT pathway via miR-381-3p/ARID4B signalling (Figure 8). Significantly, the findings of this research may shed light on a novel therapeutic target for melanoma.

Figure 8.

Concept map illustrated STAT3-activated ZBED3-AS1 sponges miR-381-3p to upregulate ARID4B to drive proliferation, migration, stemness, activate PI3K pathway, and impede apoptosis in melanoma cells

Data availability statement

Research data and material have been shown within this manuscript and the supplementary files.The data will be available from the corresponding author on reasonable request.

Disclosure statement

No conflicts of interest exist.

Supplementary material

Supplemental data for this article can be accessed here.