RETRACTED ARTICLE: Long non-coding RNA SNHG3, induced by IL-6/STAT3 transactivation, promotes stem cell-like properties of gastric cancer cells by regulating the miR-3619-5p/ARL2 axis

Abstract

Background

Chemotherapy is, next to surgery and radiotherapy, the mainstay regimen for the clinical management of gastric cancer. This therapy is, however, heavily compromised by the acquisition of resistance. Here, we aimed to clarify the potential involvement of long non-coding RNA SNGH3 in the acquisition of cisplatin resistance and stemness in gastric cancer.

Methods

Cell viability and proliferation were measured using Cell Counting Kit-8 and colony formation assays, respectively. Stem cell-like cell growth was evaluated using a mammosphere formation assay. RNA levels of SNHG2, OCT-4, SOX-2, CD44, miR-3619-5p and ARL2 were determined using qRT-PCR, whereas protein levels of OCT-4, SOX-2, CD44, ARL2, STAT3 and pSTAT3 were determined using Western blotting. Dual luciferase reporter assays were employed to interrogate regulatory interactions between STAT3, SNHG3, miR-3619-5p and ARL2, respectively. Direct binding of STAT3 to the SNHG3 promoter was investigated using a chromatin immunoprecipitation assay.

Results

We found that IL-6 triggered stem cell-like properties in cisplatin-treated gastric cancer cells and activated STAT3, which in turn transcriptionally regulated SNHG3 expression. SNHG3 expression up-regulation positively correlated with cisplatin resistance and stemness of gastric cancer cells, while SNHG3 down-regulation inhibited stem cell-like properties. In addition, we found that SNHG3 up-regulated ARL2 expression through sponging miR-3619-5p, which predominantly mediated the oncogenic properties of SNHG3 in this disease.

Conclusions

Our data indicate an involvement of aberrant SNHG3 over-expression in the acquisition of both cisplatin resistance and stemness of gastric cancer cells, and of the IL-6/STAT3/SNHG3/miR-3619-5p/ARL2 signaling cascade in the oncogenic properties of SNHG3.

Introduction

Gastric cancer is the fifth leading type of cancer and the third leading cause of cancer-related death worldwide [1]. According to the Cancer Statistics, more than one million new cases are diagnosed and 783,000 deaths claimed by gastric cancer annually [2]. Epidemiological investigations show that the majority of the cases occur in East Asia and Eastern Europe, with a gender tendency to man. Risk factors associated with gastric cancer include smoking, diet, genetic variation and Helicobacter pylori infection, which account for more than 60% of gastric cancer cases [3]. The clinical management of this disease is mainly based on a combination of surgery, chemotherapy and radiotherapy [4]. In addition, potential benefits of targeted therapy and immunotherapy are currently under intensive investigation [5]. Despite advances that have been made in the diagnosis and therapy of gastric cancer, the prognosis is still unsatisfactory, with a 5-year overall survival rate of < 10% [6]. Chemotherapy is the mainstay regimen for the clinical management of gastric cancer, which is however heavily compromised by the acquisition of resistance. Therefore, insight into the occurrence of drug resistance is of critical importance.

Cancer stem cells represent a subpopulation intra-tumoral cells with stem cell-like properties, and are believed to be the origin of multiple types of human cancer [7, 8]. Notwithstanding their relatively low abundance, cancer stem cells have been shown to play critical roles in cancer cell self-renewal, drug resistance and metastasis, and they have been linked to relapse and therapy failure [9].

Long non-coding RNAs (lncRNAs) comprise a class of RNAs longer than 200 nucleotides without protein-coding potential [10]. Accumulating evidence indicates an involvement of lncRNAs in different aspects of tumor biology, including differentiation, proliferation, apoptosis, metastasis and stemness. LncRNA SNHG3 has been shown to be involved in lung cancer [11], liver cancer [12], ovarian cancer [13, 14] and colon cancer [15]. Previously, we found that high SNHG3 expression levels are associated with a poor prognosis in gastric cancer [16]. We also found that SNHG3 deficiency inhibited gastric cancer cell proliferation and viability in vitro, and xenograft tumor progression in vivo. Whether SNHG3 dysregulation is involved in other aspects of gastric cancer cell biology still remains to be addressed.

Here, we focused on cisplatin resistance of gastric cancer cells in the context of IL-6 exposure and related stem cell-like features, and attempted to clarify the potential role of SNHG3 in this phenomenon, as well as the underlying molecular mechanisms and signaling pathways involved. In addition, as the master transcription factor STAT3 was previously found to be implicated in the stemness and/or chemoresistance of various other cancers [17–22], we also aimed to clarify its involvement in gastric cancer. Our data indicate that high SNHG3 expression may be involved in the acquisition of stem cell-like characteristics of cisplatin-resistant gastric cancer cells. Our data also indicate that the IL-6/STAT3/SNHG3/miR-3619-5p/ARL2 signaling cascade may sustain the oncogenic properties of SNHG3 in this disease. These results offer new mechanistic insight into the acquisition of cisplatin resistance in gastric cancer, and suggest a therapeutic value of miR-3619-5p.

Materials and methods

Clinical samples

In total, 60 gastric tumor samples (28 cisplatin sensitive and 32 resistant) were collected from the Fudan University Shanghai Cancer Center, and independently confirmed by three experienced pathologists. Written informed consent was obtained from all enrolled patients. The study was approved by the Institutional Ethics Committee of the Fudan University Shanghai Cancer Center and performed in strict accordance to NIH guidelines.

Cell culture conditions

Human gastric cancer cell lines SGC7901 and BGC823, two of the most commonly gastric cancer cell lines used as in vitro models for stemness [23–25], were purchased from the American Type Culture Collection (ATCC, VA, USA) and authenticated by STR profiling. The cells were maintained in DMEM medium supplemented with 10% FBS and 1% penicillin/streptomycin in a humidified CO₂ incubator (5%) at 37 °C. For mammosphere formation assays, petri dishes were pre-coated with polyHEMA (18 mg/ml), after which cells were cultured in serum-deprived DMEM/F12 medium containing 20 ng/ml EGF, 10 ng/ml bFGF, 1% N2 and 2% B27 (Gibco, MA, USA). Cisplatin-resistant cell lines were established through the application of gradually increased doses of cisplatin for 3 months. shRNAs and siRNAs were purchased from RiboBio (Guangzhou, China) and cell transfections were conducted using Lipofectamine 2000 (Invitrogen, MA, USA). The shRNA/siRNA sequences used were: SNHG3 shRNA1: 5′-GGGCACTTCGTAAGGTTTAAA-3′; SNHG3 shRNA2: 5′-GGTTGAGTGCAAGATGAGTTA-3′; STAT3 siRNA: 5′-CTCAGAGGATCCCGGAAATTT-3′;

Cytotoxicity assay

Cells were seeded in 96-well plates (5000 cells per well) for 24 hours and subsequently treated with cisplatin (0 µM; 0.5 µM; 1 µM; 1.5 µM; 2 µM; 2.5 µM; 3 µM; 3.5 µM; 4 µM; 4.5 µM) for 2 days. Cell viability was determined using Cell Counting Kit-8 (CCK-8, Dojindo, Kumamoto, Japan) according to the manufacturer’s recommendations.

Colony formation assay

Cells were seeded in 6-well plates (500 cells per well) and cultured in the presence of 1.5 µM cisplatin and 40 ng/ml IL-6 for naïve SGC7901 and BGC823 cells, or 10 µM cisplatin for resistant SGC7901 and BGC823 cells, respectively, for 3 hours, Subsequently, the respective media were replaced with fresh complete medium and the cells were cultured for another two weeks. Next, the colonies that were formed were fixed with 4% PFA, stained with crystal violet for 10 min and counted.

Quantitative RT-PCR

Total RNA was extracted from the indicated cells using TRIzol reagent (Invitrogen, MA, USA), after which cDNA was prepared by reverse transcription using a High Capacity cDNA Reverse Transcription Kit (Thermo). Subsequent real-time PCR was performed using a SYBR Green PCR Master Mix (Applied Biosystems, MA, USA) on a CFX96 Touch Real-Time PCR Detection System (Bio-Rad, CA, USA). The relative expression levels of target genes were calculated using the 2^−ΔΔCt method with GAPDH as internal reference. The primer sequences used were: miR-3619-5p Forward: 5′-CTGTGGTGGTTTACAAAGTAATT-3′; miR-3619-5p Reverse: 5′-CTGTGGTGGTTTACAAAGTAATT-3′; U6 Forward: 5′-CTGTGGTGGTTTACAAAGTAATT-3′; U6 Reverse: 5′-CTGTGGTGGTTTACAAAGTAATT-3′; SNHG3 Forward: 5′-TTCAAGCGATTCTCGTGCC-3′; SNHG3 Reverse: 5′-AAGATTGTCAAACCCTCCCTGT-3′; SOX-2 Forward: 5′-ACACCAATCCCATCCACACT-3′; SOX-2 Reverse: 5′-GCAAACTTCCTGCAAAGCTC-3′; OCT-4 Forward: 5′-TTGAGGCTCTGCAGCTTAG-3′; OCT-4 Reverse: 5′-GCCGGTTACAGAACCACAC-3′; CD44 Forward: 5′-TGCCGCTTTGCAGGTGTATT-3′; CD44 Reverse: 5′-CCGATGCTCAGAGCTTTCTCC-3′; ARL2 Forward: 5′-GAAGCAGAAAGAGCGGGA-3′; ARL2 Reverse: 5′-CTGTGAAAATGCGGCTGGA-3′; GAPDH Forward: 5′-ACAACTTTGGTATCGTGGAAGG-3′; GAPDH Reverse: 5′-GCCATCACGCCACAGTTTC-3′.

Western blotting

Cells were lysed using RIPA buffer containing protease and phosphatase inhibitor cocktails (Roche, Basel, Switzerland). Next, proteins were resolved using 10% SDS-PAGE and transferred to PVDF membranes (Millipore, MA, USA) for hybridization with primary antibodies (anti-OCT-4, #2750, 1:1000; anti-SOX-2, #2748, 1:1000; anti-CD44, #3578, 1:1000; anti-STAT3, #4904, 1:1000; anti-pSTAT3, #9131, 1:1000, from Cell Signaling Technology, MA, USA, and anti-ARL2, sab2500106, 1:2000 from Sigma-Aldrich, MO, USA). After incubation with appropriate HRP-conjugated secondary antibodies (anti-rabbit, #7074, 1:5000; anti-moue, #7076, 1:5000; Cell Signaling Technology), protein bands were visualized using an enhance chemiluminescence kit (ECL, Millipore, MA, USA).

Dual luciferase reporter assay

The SNHG3 promoter region, its full-length cDNA and the ARL2 3’UTR region were amplified by PCR and cloned into a psiCHECK2 vector, respectively. Cells were seeded in 96-well plates (5000 cells per well) and subsequently co-transfected as indicated in the figure legends. After 24 hours, luciferase activity was measured using a Dual Luciferase Reporter Assay System (Promega, WI, USA) according to the manufacturer’s instructions.

Chromatin immunoprecipitation (ChIP) assay

Direct binding of STAT3 to the SNHG3 promoter was analyzed by ChIP using an EZ-ChIP Kit (Millipore) in accordance with the provider’s protocol. Sonicated chromatin was immunoprecipitated using an anti-STAT3 antibody (#9139 from Cell Signaling Technology), after which enriched fragments were detected by real-time PCR as described above. The primers used for ChIP were: SNHG3 Forward: 5′-GTAGGAGGAACACCATAA-3′; SNHG3 Reverse: 5′-TGAAGTAGGCAAGTAAAGA-3′.

Statistical analysis

Data processing and analysis were performed using GraphPad 7.0. Student’s t-test (two-tailed) and one- or two-way ANOVA analyses with appropriate post hoc tests were employed for statistical comparisons, and p < 0.05 was considered statistically significant.

Results

IL-6 induces stem cell-like properties and STAT3 activation in gastric cancer cells

First, we found that treatment with IL-6 significantly induced cisplatin resistance as indicated by CCK-8 cell viability assays in both SGC7901 (IC50: 2.42 ± 0.31 vs. 1.45 ± 0.16, p < 0.01, Fig. 1A) and BGC823 (IC50: 2.58 ± 0.35 vs. 1.56 ± 0.15, p < 0.01, Fig. 1B) cells. Likewise, we found that IL-6 greatly increased the colony formation (Fig. 1C) and mammosphere formation (Fig. 1D) capacities of these cells in the presence of cisplatin. To next understand the molecular mechanism underlying the acquired cisplatin resistance upon IL-6 exposure, we assessed the expression profiles of key stemness factors in response to IL-6 treatment. We found that both the transcript (Fig. 1E, F) and protein (Fig. 1G) levels of OCT-4, SOX-2 and CD44 were markedly elevated in SGC-7901 and BGC823 cells, which suggests a potential contribution of these factors to cisplatin resistance. Since the master transcription factor STAT3 has previously been implicated in tumor stemness, we next set out to test this option in our system. In doing so, we found that the phosphorylation of STAT3 dramatically increased in an IL-6 dose-dependent manner, while the total STAT3 level remained unaltered (Fig. 1H). Therefore, we preliminarily conclude that IL-6 exposure may stimulate cisplatin resistance in gastric cancer cells, and may be correlated with alterations in OCT-4, SOX-2 and CD44 expression. STAT3 signaling was significantly activated in this system, which may consequently account for the acquired stem cell-like properties.

Fig. 1.

IL-6 triggers stem cell-like properties and STAT3 activation in gastric cancer cells. (A and B) SGC7901 and BGC823 cells were treated with 40 ng/ml IL-6 (IL-6 (+)) or PBS (IL-6 (-)) for 24 h after which cell viability was determined usinga CCK-8 assay with treatment of the indicated concentrations of CDDP for 48 h. Average IC₅₀ values of three independent experiments were calculated. (C) Colony forming capacities of SGC7901 and BGC823 cells treated with 40 ng/ml IL-6 (IL-6 (+)) or PBS (IL-6 (-)) determined after treatment with 1.5 µM CDDP. (D) IL-6 treatment-improved mammosphere formation in SGC7901 and BGC823 cells. (E-G) Upregulation of mRNA and protein expression levels of stem cell markers (OCT-4, SOX-2 and CD44) in SGC7901 and BGC823 cells after IL-6 treatment, as assessed by qRT-PCR and Western blotting. (H) IL-6-triggered STAT3 phosphorylation in SGC7901 and BGC823 cells, as determined by Western blotting. ACTIN was used as a loading control. The data represent the mean ± SD. * p < 0.05; ** p < 0.01; *** p < 0.001

SNHG3 is transcriptionally regulated by STAT3

Previously, we observed STAT3 activation upstream of up-regulated OCT-4, SOX-2 and CD44 in response to IL-6 exposure in gastric cancer cells, with the intermediate effectors yet to be defined. Here, we found that SNHG3 transcripts were markedly induced by IL-6, which could subsequently be abrogated by co-transfection with a STAT3-specific siRNA (Fig. 2A, B). Concordantly, we found that SNHG3 promoter-fused luciferase activity was inhibited upon STAT3 knockdown in both SGC7901 and BGC823 cells (Fig. 2C, D). In addition, we assessed direct binding of STAT3 to the SNHG3 promoter by ChIP, and noted significant enrichment of SNHG3 promoter fragments after STAT3 immunoprecipitation in both cell lines (Fig. 2E). Taken together, these results suggest that SNHG3 is transcriptionally regulated by IL-6-activated STAT3.

Fig. 2.

SNHG3 is transcriptionally regulated by STAT3. (A) qRT-PCR analysis of SNHG3 expression in SGC7901 and BGC823 cells transfected with STAT3 siRNAs or negative control siRNAs, followed by treatment with 40 ng/ml IL-6 (IL-6 (+)) or PBS (IL-6 (-)) for 24 h. (B) Knockdown efficiencies of STAT3 in SGC7901 and BGC823 cells confirmed by Western blotting. (C and D) STAT3 knockdown-reduced SNHG3 promoter-derived luciferase activity in SGC7901 and BGC823 cells. (E) ChIP assays in SGC7901 and BGC823 cells using a STAT3 antibody, followed by qRT-PCR analysis of SNHG3 promoter enrichment. The data represent the mean ± SD. * p < 0.05; ** p < 0.01

SNHG3 expression positively correlates with cisplatin resistance and stemness of gastric cancer cells

Next, we sought to clarify whether SNHG3 up-regulation by the IL-6/STAT3 signal cascade may be correlated with cisplatin resistance and stemness in gastric cancer cells. We started by analyzing SNHG3 expression in cisplatin-resistant cell lines derived from both SGC7901 and BGC823. We found that SNHG3 expression was markedly higher in the resistant cell lines than the parental ones (Fig. 3A). Similar changes in SNHG3 expression were observed in mammospheres compared to monolayer SGC7901 and BGC823 cells (Fig. 3B). More importantly, we found that the SNHG3 transcript levels were higher in cisplatin resistant gastric cancer patient samples (n = 32) compared to the sensitive ones (n = 28, Fig. 3C). We next divided the gastric cancer patients into SNHG3-high and SNHG3-low groups, and found by Kaplan-Meier analysis a significantly better overall survival in the SNHG3-low group (Fig. 3D).

Fig. 3.

SNHG3 expression is positively associated with cisplatin resistance and stemness of gastric cancer cells. (A) qRT-PCR analysis of SNHG3 expression in CDDP-resistant SGC7901/BGC823 cell lines (CDDP) and parental CDDP-sensitive BGC823/SGC7901 cell lines (Parental). (B) Expression of SNHG3 in SGC7901/BGC823-derived mammospheres (Mammosphere) and parental BGC823/SGC7901 cell lines (Parental) determined by qRT-PCR. (C) SNHG3 expression in 28 cisplatin-sensitive (CS) and 32 cisplatin-resistant (CR) gastric cancer (GC) tissues determined by qRT-PCR. (D) Kaplan–Meier plots of relapse-free survival rates based on SNHG3 expression in GC patients. * p < 0.05; ** p < 0.01

SNHG3 knockdown decreases stem cell-like properties of gastric cancer cells

Since the above analysis of primary gastric cancer samples suggested a role of SNHG3 in cisplatin resistance and stemness, we next sought to provide experimental evidence in support of this notion. To this end, we generated SNHG3-knockdown SGC7901 and BGC823 cells using two independent shRNAs (Fig. 4A). We subsequently found that the viability was markedly compromised by SNHG3 silencing in the cisplatin-resistant gastric cancer cells when exposed to different doses of cisplatin (Fig. 4B, C, D). Similarly, cell proliferation in cisplatin-containing medium (10 µM) was evidently suppressed by SNGH3 silencing (Fig. 4E). Concordantly, we found that the mammosphere forming capacity was greatly compromised in SNHG3 silenced cells compared to its SNHG3 wild type counterparts (Fig. 4F). Stemness factor profiling showed that OCT-4, SOX-2 and CD44 were all down-regulated upon SNHG3 silencing at both the transcript (Fig. 4G, H) and protein (Fig. 4I) levels, which implicates an important contribution of SNHG3 to cisplatin resistance and stemness via modulating OCT-4, SOX-2 and CD44 expression.

Fig. 4.

SNHG3 knockdown inhibits stem cell-like properties of gastric cancer cells. (A) SNHG3 expression levels determined by qRT-PCR in SGC7901/DDP and BGC823/DDP cells stably transfected with SNHG3 shRNAs (sh-SNHG3-1 and sh-SNHG3-2) or empty vector (sh-CTR). (B-D) Viabilities of SGC7901/DDP and BGC823/DDP cells stably transfected with SNHG3 shRNAs (sh-SNHG3-1 and sh-SNHG3-2) or empty vector (sh-CTR) determined by CCK-8 assay after CDDP treatment at the indicated concentrations for 48 h. Average IC₅₀ values of three independent experiments were calculated. (E) Colony forming capacities of SGC7901/DDP and BGC823/DDP cells stably transfected with SNHG3 shRNAs (sh-SNHG3-1 and sh-SNHG3-2) or empty vector (sh-CTR) determined after with treatment of 10 µM CDDP. (F) SNHG3 knockdown-inhibited mammosphere formation in SGC7901 and BGC823 cells. (G-I) mRNA and protein expression of stem cell markers (OCT-4, SOX-2 and CD44) in SGC7901 and BGC823 cells stably transfected with SNHG3 shRNAs (sh-SNHG3-1 and sh-SNHG3-2) assessed by qRT-PCR and Western blotting. The data represent the mean ± SD. * p < 0.05; ** p < 0.01; *** p < 0.001

SNHG3 knockdown decreases cisplatin resistance and stem cell-like properties of gastric cancer cells in vivo

We next asked whether SNHG3 exerts similar functions in vivo. To this end, SNHG3-knockdown SGC7901 cells and control cells were used for tumorigenesis assays in mice (Fig. 5). Following 5 mg/kg cisplatin treatment, the tumor volumes formed by SNHG3 silenced SGC7901 cells (sh-SNGH3-1) were considerably smaller than those formed by control shRNA cells (sh-CTR) throughout the experimental course of 25 days (Fig. 5A). At the endpoint, tumors were recovered from the two groups of mice for weight comparison, after which a significant reduction in tumor weight was observed in the sh-SNGH3-1 group compared to the sh-CTR group (Fig. 5B). In addition, an in vivo limiting dilution assay revealed that mice harboring SNHG3-knockdown SGC7901 cells showed impaired tumor-initiating abilities (Table 1). These data underscore that SNHG3 may suppress cisplatin resistance and stem cell-like properties of gastric cancer cells in vivo.

Fig. 5.

SNHG3 knockdown suppresses cisplatin resistance of gastric cancer cells in vivo. SNHG3-knockdown SGC7901 (sh-SNHG3-1) cells and control (sh-CTR) cells were used for tumorigenesis assays. Cisplatin (5 mg/kg) was intraperitoneally injected into the nude mice every 5 days (5 mice in each group). The tumor volumes were measured at the indicated days (A). After 25 days, the mice were euthanized, after which the tumors were harvested and weighed (B). * p < 0.05; ** p < 0.01; *** p < 0.001

Table 1.

Stem cell frequency of SNHG3-silenced SGC-7901 cells

Cell Dose	sh-CTR	sh-SNHG3-1
1 × 106	6/6	6/6
5 × 105	6/6	4/6
1 × 105	4/6	1/6
5 × 104	2/6	0/6

Limiting dilution assay of gastric cancer stem cells after inoculation with sh-CTR- or sh-SNHG3-1-SGC-7901 cells in nude mice (n = 6 for each group)

SNHG3 up-regulates ARL2 expression through sponging miR-3619-5p

We next set out to identify SNHG3 downstream targets in gastric cancer cells. Using the miRcode algorithm, miR-3619-5p was predicted as a candidate with a potential to bind SNHG3, after which a putative binding site was indeed identified using RNAhybrid (Fig. 6A). Using a dual luciferase assay, we found that co-transfection with miR-3619-5p dramatically inhibited SNHG3-promoter reporter activity, which could readily be abolished through the introduction of a mutation into the putative binding site (Fig. 6B, C). Endogenous miR-3619-59 levels were also increased in SNHG3 silenced cells compared to parental control cells (Fig. 6D), and decreased in response to exogenous SNHG3 overexpression (Fig. 6E). We subsequently predicted miR-3619-5p target genes using TargetScan, and an alignment between miR-3619-5p and ARL2 is illustrated in Fig. 6F. As expected, co-transfection with miR-3619-5p greatly inhibited the ARL2-promoter reporter activity, which was completely abolished by a mutation disrupting the binding between miR-3619-5p and ARL2 (Fig. 6G, H). The ARL2 transcript level was decreased after miR-3619-5p mimic transfection and up-regulated after miR-3619-5p-specific inhibition (Fig. 6I). Western blot analysis revealed changes in endogenous ARL2 protein levels, in accordance with its mRNA levels, in response to either miR-3619-5p transfection or its inhibition (Fig. 6J). We further examined ARL2 expression in SNHG3 silenced cells and found, in line with the competitive regulatory effect of SNHG3 on miR-3619-5p, that ARL2 was markedly down-regulated in these cells (Fig. 6K). Conversely, we found that overexpression of SNHG3 sponged miR-3619-5p and alleviated its inhibitory effect on ARL2 expression, which consequently led to ARL2 up-regulation in both SGC7901 and BGC823 cells (Fig. 6L). Western blot analysis confirmed these observations at the protein level (Fig. 6M). In summary, we conclude that SNGH3 up-regulates ARL2 expression in gastric cancer cells through sponging miR-3619-5p.

Fig. 6.

SNHG3 up-regulates ARL2 through sponging miR-3619-5p. (A) Prediction of miR-3619-5p binding sites in SNHG3 and schematic luciferase reporter SNHG3 wild-type (SNHG3 wt) and the miR-3619-5p-binding-site mutated (SNHG9 mut) vector constructs. (B and C) Luciferase reporter constructs (SNHG3 wt and SNHG9 mut) were transfected into SGC7901 and BGC823 cells together with miR-3619-5p or miR-NC, after which relative luciferase activities were determined 48 h post transfection. (D and E) Relative expression level of miR-3619-5p in SGC7901 and BGC823 cells stably transfected with SNHG3 shRNAs (sh-SNHG3-1 and sh-SNHG3-2) or empty vector (sh-CTR) and SNHG3 plasmid (pSin-SNHG3) or empty vector (pSin-VEC) determined by qRT-PCR. (F) Schematic diagram of miR-3619-5p binding sites in the ARL2 3’UTR. Sequences were compared between putative miR-3619-5p and wild-type (ARL2 3’UTR wt) or mutant (ARL2 3’UTR mut) target sites in the 3’UTR of ARL2. (G and H) Dual luciferase reporter assay performed in SGC7901 and BGC823 cells co-transfected with ARL2 3’UTR wt or ARL2 3’UTR mut and miR-3619-5p or miR-NC plasmids. (I and J) Relative expression levels of ARL2 in SGC7901 and BGC823 cells transfected with miR-3619-5p mimics and miR-3619-5p inhibitors or their respective negative controls determined by qRT-PCR and Western blotting. (K-M) Relative expression level of ARL2 in SGC7901 and BGC823 cells stably transfected with SNHG3 shRNAs (sh-SNHG3-1 and sh-SNHG3-2) or empty vector (sh-CTR) and SNHG3 plasmid (pSin-SNHG3) or empty vector (pSin-VEC) determined by qRT-PCR and Western blotting. The data represent the mean ± SD. * p < 0.05; ** p < 0.01; ns = not significant

SNHG3 regulates stem cell-like properties of gastric cancer cells via the miR-3619-5p/ARL2 axis

Since the relevance of the SNHG3/miR-3619-5p/ARL2 axis in cisplatin resistance and stemness of gastric cancer cells was still elusive, we next performed co-transfections of pSin-SNHG3 and miR-3619-5p into both SGC7901 and BGC823 cells. We found that ARL2 up-regulation by SNHG3 was completely abrogated by simultaneous transfection with miR-3619-5p (Fig. 7A, B). Cell viability was enhanced by SNHG3 overexpression under cisplatin treatment, which was subsequently abolished by miR-3619-5p transfection (Fig. 7C, D and E). This observation highlighted a critical role of miR-3619-5p in modulating cisplatin resistance downstream of SNHG3 in gastric cancer cells. A similar conclusion was drawn from colony formation assays, showing significantly increased colony numbers in SNHG3-proficient cells, which was restored by the application of miR-3619-5p mimics (Fig. 7F). Similarly, mammosphere formation was stimulated by exogenous SNHG3 expression and compromised by concurrent miR-3619-5p transfection (Fig. 7G). In line with these observations, we found that the expression levels of OCT-4, SOX-2 and CD44 were markedly increased after SNHG3 overexpression and subsequently decreased by miR-3619-5p co-transfection (Fig. 7H, I and J). These results support a role of SNHG3 in regulating stem cell-like properties in gastric cancer cells via the miR-3619-5p/ARL2 axis.

Fig. 7.

SNHG3 regulates stem cell-like properties of gastric cancer cells via miR-3619-5p/ARL2 axis. SGC7901 and BGC823 cells were co-transfected with empty vector and negative control miRNA (pSin-VEC + miR-NC), SNHG3 overexpression plasmid and negative control miRNA (pSin-SNHG3 + miR-NC) or SNHG3 overexpression plasmid and miR-3619-5p mimics (pSin-SNHG3 + miR-3619-5p). (A and B) Expression levels of ARL2 in the co-transfected cells measured by qRT-PCR and Western blotting. (C-E) Cell viabilities of the co-transfected cells determined by CCK-8 assay after treatment of the indicated concentrations of CDDP for 48 h. Average IC₅₀ values of three independent experiments were calculated. (F) Colony forming capacities of the co-transfected cells determined after treatment with 1.5 µM CDDP. (G) Mammosphere formation of the co-transfected cells. (H-J) mRNA and protein expression of stem cell markers (OCT-4, SOX-2 and CD44) in the co-transfected cells determined by qRT-PCR and Western blotting. The data represent the mean ± SD. * p < 0.05; ** p < 0.01; *** p < 0.001

Discussion

Cancer stem cells are considered to constitute subpopulations of intra-tumoral cells with stem cell-like characteristics, and they are believed to be the origin of multiple types of human cancer [7, 8]. Notwithstanding their small proportion within the tumor mass, stem cells have been reported to play critical roles in cancer cell self-renewal, drug resistance and metastasis, and to be linked to relapse and failure of cancer therapy [26, 27]. Based on our previous observation that SNHG3 promotes the proliferation and metastasis of gastric cancer by regulating its neighboring gene MED18 [16], we hypothesized that SNHG3 may be linked to the stemness of gastric cancer and, thus, continued to investigate the oncogenic role of SNHG3 in gastric cancer development. First, we found that IL-6 exposure induced cisplatin resistance. Concordantly, we found that the stemness markers OCT-4, SOX-2 and CD44 were significantly up-regulated, which was accompanied by increased STAT3 phosphorylation. More importantly, we found that IL-6-activated STAT3 induced SNHG3 expression via direct binding to its promoter region. SNHG3 was also found to be significantly up-regulated in cisplatin-resistant cells and mammospheres derived from gastric cancer cells. Clinically, aberrantly high SNHG3 expression was noticed in cisplatin-resistant gastric cancer patients compared to sensitive patients, which was associated with a poorer overall survival. shRNA-mediated knockdown of SNHG3 re-sensitized resistant cells to cisplatin as indicated by a decreased viability, proliferation and mammosphere formation capacity. In addition, we found that the stemness markers were down-regulated in SNHG3 silenced cells. We also bioinformatically predicted and experimentally verified miR-3619-5p as a downstream target of SNHG3, and that SNHG3 competitively sponges miR-3619-5p to up-regulate ARL2. Co-transfection with miR-3619-5p completely abolished the stimulatory effect of SNHG3 on ARL2 expression, and consequently restored SNHG3-compromised cell viability, proliferation and mammosphere formation capacities. We also found that expression of the stemness markers OCT-4, SOX-2 and CD44 was markedly decreased. Together, our data support a central role of SNHG3 in stemness transition and cisplatin resistance. We therefore propose a novel IL-6/STAT3/SNHG3/miR-3619-5p/ARL2 pathway that underlies the acquisition of stem cell-like properties in cisplatin-resistant gastric cancer cells (Fig. 8).

Fig. 8.

Schematic working model

Competition with endogenous RNAs has been increasingly recognized as a mode-of-action of lncRNAs [28], and multiple microRNAs have been found to be involved in the complex regulatory network of SNHG3. Zheng et al. showed, for example, that a SNHG3/miR-151a-3p/RAB22A signaling pathway may modulate osteosarcoma cell migration and invasion [29], whereas Chen et al. showed that SNHG3 may induce cell proliferation via sponging miR-196a-5p, which in turn may be linked to an unfavorable clinical outcome in this disease [30]. Wang et al. suggested that SNHG3 may stimulate cell migration and proliferation in laryngeal carcinoma through regulating the miR-384/WEE axis [31]. In colorectal carcinoma, Huang et al. proposed a competing endogenous RNA (ceRNA) feature of SNGH3, which may promote malignant progression of this disease [15]. Zhao et al. further provided evidence that SNHG3 may contribute to hepatocellular tumorigenesis by interacting with miR-326 [32]. Here, we identified miR-3619-5p as a novel target of SNHG3 in gastric cancer, which competitively and convergently regulates ARL2 expression and, consequently, functions as a negative modulator of cisplatin resistance and stemness. In agreement with the tumor suppressor role of miR-3619-5p proposed here, multiple studies have previously addressed its anti-tumor properties in a number of human cancers. An initial study by Niu et al. uncovered a previously unnoticed role of miR-3619-5p in suppressing β-catenin-mediated invasion and proliferation in non-small-cell lung cancer [33], which was supported by studies performed by Zhang et al. [34] and Tan et al. [35]. Subsequently, Fite et al. showed that down-regulation of miR-203, miR-887, miR-182 and miR-3619 effectively blocked vimentin-triggered and phospholipase D-dependent cancer cell invasion [36]. Li et al. reported that miR-3619-5p may suppress prostate cancer cell proliferation through upregulating the expression of the tumor suppressor CDKN1A [37]. In cutaneous squamous-cell carcinoma, the relevance of miR-3619-5p to drug response was addressed by Zhang et al., suggesting that miR-3619-5p may hamper cell growth and cisplatin resistance via KPNA4 [38].

ARL2 is a small GTP-binding protein of the RAS superfamily and functions as an ADP-ribosylation factor. Evidence is emerging that ARL2 may act as an oncogenic factor, as also suggested by our results. Peng et al. showed that miR-214 may negatively regulate ARL2 expression and, by doing so, inhibit the invasion and proliferation of cervical cancer cells [39]. Taniuchi et al. reported that BART may suppress pancreatic cancer cell migration through blocking ARL2-mediated RhoA inactivation [40]. However, opposite roles of ARL2 in human cancer have also been proposed. Wang et al. found, for example, that overexpression of ARL2 attenuated glioma cell proliferation and tumorigenicity through suppressing AXL [41], whereas Beghin et al. found that ARL2 may modulate breast tumor aggressiveness in immuno-deficient mice [42]. As a consequence, the specific roles played by ARL2 appear to be context dependent.

In summary, our data indicate an involvement of aberrant SNHG3 overexpression in both cisplatin resistance and stemness of gastric cancer cells, and uncovered the IL-6/STAT3/SNHG3/miR-3619-5p/ARL2 signaling cascade as being involved in the oncogenic properties of SNHG3.

Abbreviations

lncRNA

Long non-coding RNA

ChIP

Chromatinimmunoprecipitation

ATCC

American Type CultureCollection

CCK-8

Cell Counting Kit-8

Authors’ contributions

Bo Sun, Yang Han, Hong Cai, Hua Huang and Yi Xuan collected and assembled the data. Bo Sun and Yang Han performed the statistical analyses. Hong Cai and Yi Xuan conceived and designed the study and wrote the manuscript. All authors read and approved the final manuscript.

Funding

This work was supported by Shanghai Anticancer Association “aoxiang” Project (SACA-AX201901).

Data Availability

Not applicable.

Compliance with ethical standards

Conflict of interest

No conflicts of interest, financial or otherwise, are declared by the authors.

Consent for publication

All of the authors have approved publication of this work.

Ethics approval and consent to participate

The study was approved by the Institutional Ethics Committee of Fudan University Shanghai Cancer Center and performed in strict accordance to NIH guidelines.