The lncRNA MEG3 mediates renal cell cancer progression by regulating ST3Gal1 transcription and EGFR sialylation

Renal cell carcinoma (RCC), arising from the epithelium of renal tubules, is a common cancer worldwide (Lopez-Beltran et al., 2006). Once metastasis occurs in RCC, the median survival of patients is ∼8 months (Chen and Kuo, 2016); half of patients who present with metastasis die within the first year and only 10% of such patients survive longer than five years (Motzer et al., 2004; Rini, 2006). Radiation therapy and surgical excision remain the main method of therapy in RCC (Banumathy and Cairns, 2010; Lopez-Beltran et al., 2006). Thus, it is essential to identify novel diagnostic markers and effective therapeutic targets for RCC patients.

Long noncoding RNAs (lncRNAs), a type of non-protein-coding RNA, are reported to participate in diverse biological processes, including transcriptional regulation, cell differentiation, cell apoptosis, cancer cell metastasis, chemotherapy drug resistance, and RNA alternative splicing (Dinger et al., 2008; Ponting et al., 2009; Seles et al., 2016). Several lncRNAs have been well characterized in RCC, such as HOTAIR (Chiyomaru et al., 2014). MALAT (Hirata et al., 2015), H19 (Raveh et al., 2015), and maternally expressed gene (MEG3) (Wang et al., 2015b). HOTAIR knockdown reduces the proliferation, migration and tumorigenicity of RCC (Wu et al., 2014). MALAT1 expression is higher in RCC tissue compared to normal tissue, which enhances RCC cell progression and can be a prognostic clinical parameter (Ji et al., 2003). Knockdown of H19 results in decreased growth, migration, and invasion of RCC. Higher expression of H19 is associated with significantly shorter overall survival of RCC patients (Wang et al., 2015a). For MEG3, however, it has been only reported that the apoptotic rate is increased in RCC cell lines after transfection with this lncRNA (Wang et al., 2015b), and the biological function and regulatory mechanism of MEG3’s involvement in RCC progression are unknown.

Sialic acids are widely distributed in nature as terminal sugars of oligosaccharide chains of glycoconjugates (glycoproteins and glycolipids), which are important posttranslational modifications during cancer progression (Harduin-Lepers et al., 1995; Warren et al., 1972). Sialyltransferases (STs) are key enzymes in the biosynthesis of sialic acid, which contains oligosaccharides and glycoconjugates (Pilatte et al., 1993). It has been reported that β-galactoside α-2,3-sialyltransferase 1 (ST3Gal1) is an independent adverse prognostic factor for recurrence in RCC patients who underwent nephrectomy (Bai et al., 2015). However, relevant studies of untreated RCC patients is rare. The transcription factor c-Jun (also known as JUN) has been reported to play an important role in regulating cancer progression. However, there is no idea about the relationship between c-Jun and ST3Gal1.

Epidermal growth factor receptor (EGFR), which possesses tyrosine kinase activity, is a 170 kDa transmembrane glycoprotein with an extracellular ligand-binding domain and an intracellular region (Schlessinger, 2014; Wee and Wang, 2017). EGFR can activate several signaling pathways and highly correlates with cell proliferation, differentiation, drug sensitivity, and angiogenesis (Veale et al., 1993). Furthermore, the overexpression of EGFR is associated with shorter survival in non-small cell lung cancer (Mendelsohn and Baselga, 2006). EGFR is also phosphorylated in RCC, and stimulates the activation of phosphoinositide 3-kinase (PI3K), which leads to AKT translocation and localization. The EGFR–PI3K–AKT pathway is frequently overexpressed in RCC (Cancer Genome Atlas Research Network et al., 2016; Stumm et al., 1996). EGFR sialylation suppresses its phosphorylation in lung cancer cells (Liu et al., 2011). However, the molecular mechanism of EGFR sialylation in RCC progression requires further investigation.

In this study, the expression of MEG3 was assessed in RCC tissues and cell lines. Alterations in MEG3 expression were associated with the proliferative and invasive abilities of RCC cells. Furthermore, the underlying mechanism involved in MEG3–ST3Gal1–EGFR-mediated EGFR sialylation and PI3K–AKT activation in RCC was explored.

To determine the expression pattern of MEG3 in RCC tissues and cell lines, reverse-transcription qPCR (RT-qPCR) was performed. As shown in Fig. 1A, MEG3 expression was downregulated in RCC tissues (2.99±1.45, mean±s.d.) compared with related adjacent normal tissues (5.06±1.67, mean±s.d.). Fluorescence in situ hybridization (FISH) was also performed in RCC tissues and adjacent tissues to identify the location of MEG3 expression. According to the fluorescence signal, RCC tissues exhibited lower fluorescence intensity of MEG3 than adjacent tissues, and MEG3 showed high expression in cell nuclei (Fig. 1B). As shown in Fig. 1C, 293T cells exhibited high levels of MEG3, while RCC cell lines 786-O, ACHN and Caki-1 presented lower MEG3 relative expression levels (0.70±0.05, 0.49±0.138, and 0.39±0.11, respectively; mean±s.d.). These results indicated that a low level of MEG3 might be associated with RCC.

Given the downregulated expression of MEG3 in RCC tissues and cell lines, we performed loss-of-function and gain-of-function studies in 786-O cells. A CCK-8 cell-counting assay was performed after cells were transfected with MEG3 or siRNA targeting MEG3 (siMEG3). The results indicated that MEG3 overexpression impaired cell proliferation, whereas knockdown of MEG3 promoted cell proliferation in vitro (Fig. 2A,B). As shown in Fig. 2C and D, the migratory ability was decreased after cells were transfected with MEG3, while the ability was recovered in MEG3 knockdown cells, as determined by wound healing assays. The effect of MEG3 on RCC cell invasion was reversed when cells were transfected with siMEG3, as determined using Transwell assays (Fig. 2E,F). In addition, ACHN cells showed the same trend (Fig. S1). All data indicated that MEG3 might play an important role in RCC progression.

We performed tumorigenicity assays and tail-vein injection assays in nude mice. To confirm whether MEG3 affected RCC tumorigenesis, 786-O cells were transfected with vector, MEG3, control scrambled shRNA (shSCR), or an shRNA targeting MEG3 (shMEG3), and injected into the right flank of nude mice. After 28 d, the mice were killed and xenograft tumors were taken for further research. Tumor volume was smaller in the mice treated with MEG3 than in those treated with the vector control, whereas the shMEG3 group showed opposite result (Fig. 3A). Ki67 levels were measured by immunohistochemistry (IHC) staining (Fig. 3B,C). To further verify that MEG3 affected cell metastasis, 786-O cells transfected with MEG3 or shMEG3 were injected into tail veins of nude mice. As expected, the MEG3 group showed lower near-infrared fluorescence intensity than the vector group, whereas MEG3 knockdown resulted in increased fluorescence intensity (Fig. 3D). The nude mice bearing 786-O cells transfected with MEG3 showed attenuated lung metastasis, whereas knockdown of MEG3 facilitated lung metastasis of 786-O cells (Fig. 3E). These data indicated promising therapeutic targets for RCC malignancy.

Aberrant sialylation is closely associated with the malignant phenotype of cancer cells, including metastatic potential and invasiveness (Miyagi, 2008). To further research the relationship between STs and RCC, the expression of ST3Gal1 was detected in RCC. We found that ST3Gal1 mRNA and protein expression was lower in RCC tissues than in adjacent tissues by IHC staining (Fig. 4A) and RT-qPCR (Fig. 4B). A significant positive correlation was observed between MEG3 and ST3Gal1 (r=0.4736, P<0.0001; Fig. S2). In addition, the mRNA and protein levels of ST3Gal1 were detected after cells were transfected with MEG3 or siMEG3. The aberrant expression of MEG3 regulated ST3Gal1 mRNA and protein levels (Fig. 4C,D). According to FISH, MEG3 was mostly located in nuclei (Fig. 4E). Moreover, overexpression of MEG3 obviously elevated transcriptional activity of the ST3Gal1 promoter, as determined using a luciferase reporter assay (Fig. 4F).

To investigate the potential regulators involved in ST3Gal1 expression, bioinformatics software was used to scan for potential transcription-factor-binding sites in the ST3Gal1 promoter. According to the bioinformatics predictions, c-Jun was identified as a candidate transcription factor that regulates the ST3Gal1 promoter through direct binding (Fig. 4G). Knockdown of c-Jun promoted ST3Gal1 mRNA and protein levels (Fig. 4H). A chromatin immunoprecipitation (ChIP) assay was performed to prove that c-Jun directly bound to the ST3Gal1 promoter (Fig. 4I).

It has been reported that c-Jun transactivation activity is inhibited by its interaction with MEG3 (Zhou et al., 2017). To verify this conclusion, an RNA immunoprecipitation assay (RIP) and western blot analysis were performed to figure out the relationship between MEG3 and c-Jun in RCC cells. As shown in Fig. 4J, altered MEG3 levels regulated the expression of c-Jun. The RIP assay further evaluated the specific interaction between MEG3 and c-Jun. The results suggested that MEG3 positively regulated ST3Gal1 mRNA transcription by interacting with c-Jun (Fig. 4K).

It has been reported that EGFR sialylation suppresses EGFR phosphorylation by inhibiting EGF binding and EGFR dimerization (Yen et al., 2015). To verify this conclusion, phosphorylation of EGFR was detected following treatment with EGF in 786-O cells transfected with ST3Gal1 or siRNA targeting ST3Gal1 (siST3Gal1). According to Fig. 5A, EGFR phosphorylation was reduced in 786-O cells transfected with ST3Gal1. In contrast, the level of EGFR phosphorylation was increased in siST3Gal1 cells (Fig. 5B). In the presence of EGF, the EGFR phosphorylation level was inhibited after ST3Gal1 overexpression, whereas the effect was reversed when 786-O cells transfected with siST3Gal1 were treated with EGF (Fig. 5A,B).

Considering that c-Jun regulated ST3Gal1, which was mediated by MEG3, we supposed that a MEG3–c-Jun–ST3Gal1 axis might regulate phosphorylated EGFR. To confirm this supposition, EGFR was measured after transfecting cells with MEG3 or siMEG3 upon EGF stimulation. As shown in Fig. 5C, overexpression of MEG3 increased EGFR affinity for the lectin MAL by elevating ST3Gal1 levels in 786-O cells compared to the control groups, as measured by lectin affinity assay. Conversely, treatment of 786-O cells with siMEG3 reduced EGFR affinity for the lectin MAL, by inhibiting ST3Gal1 expression.

Seeing that disruption of EGFR activity can regulate downstream signaling, including the AKT–PI3K–mTOR pathway (Bai et al., 2015), we examined the effect of MEG3 on this EGFR-mediated intracellular signaling pathway in 786-O cells. The protein levels of phosphorylated EGFR, PI3K p110α (encoded by PIK3CA) and phosphorylated AKT were reduced upon EGF stimulation in 786-O cells transfected with MEG3 compared with levels in control cells (Fig. 5D). The total level of EGFR and AKT proteins remained unchanged. Upon EGF stimulation, 786-O cells treated with siMEG3 showed stronger phosphorylation of PI3K–Akt pathway components compared to levels in control cells (Fig. 5D). Furthermore, the same trend was observed in ACHN cells (Fig. S3). The data suggested that MEG3 influences EGFR phosphorylation and leads to the activation of the PI3K–AKT pathway in RCC cells by modulating ST3Gal1 levels.

The lncRNA MEG3, as a tumor suppressor, has been confirmed to take part in cancer progression. MEG3 was found to inhibit the proliferation of RCC cells by inducing G2/M cell-cycle arrest and apoptosis (Qin et al., 2013). MEG3 is downregulated in most RCC tissues and lowly expressed in RCC cell lines 786-O and SN12. The viability of 786-O cells is reduced, and the apoptosis rate is enhanced after overexpression of MEG3 (Qin et al., 2013). In this study, differential expression of MEG3 was shown in RCC samples and RCC cell lines. Moreover, MEG3 mediated the proliferation, migration, invasion and metastasis of RCC cells both in vitro and in vivo. Therefore, MEG3 is implicated as a tumor suppressor in RCC progression.

Sialylation, a common glycosylation posttranslational modification, plays roles in regulating cellular interactions and signal transduction. ST3Gal1 mediates the transfer of sialic acid with an α2,3-linkage to terminal galactose residues. High expression of ST3Gal1 is associated with advanced stage epithelial ovarian cancer (Wen et al., 2017). Altered levels of ST3GAL1 correspond to the drug-resistant phenotype of chronic myeloid leukemia (CML) cell lines (Li et al., 2016). However, similar studies of the role of ST3Gal1 in RCC are rare. According to The Cancer Genome Atlas (https://portal.gdc.cancer.gov/) and our research, we found that ST3Gal1 showed higher expression in adjacent tissues than in RCC tissues. The ST3Gal1 level was regulated by MEG3, and a positive correlation between MEG3 and ST3Gal1 expression was shown.

The c-Jun transcription factor has been reported as a positive regulator of cell proliferation. The growth-promoting activity of c-Jun is regulated by repression of tumor suppressors, such as tumor-suppressor PTEN (Shaulian and Karin, 2002). Other studies also demonstrated that c-Jun is a negative transcription factor, which negatively regulates PHLPP1 transcription by binding directly to the PHLPP1 promoter (Zhou et al., 2017). Here, MEG3 mediated c-Jun activity, which negatively regulated ST3Gal1 transcription by binding directly to the ST3Gal1 promoter. We reported that the lncRNA MEG3 participated in the protein sialylation process by regulating ST3Gal1 at the transcriptional level.

EGFR is commonly upregulated in cancers such as metastatic colorectal cancer, glioblastoma, pancreatic cancer, and breast cancer (Li et al., 2016; Shaulian and Karin, 2002; Wen et al., 2017). EGF binds to the exposed dimerization interface of EGFR for receptor activation. After autophosphorylation, dimeric EGFR recruits and activates various downstream signaling proteins. It has been reported that EGFR phosphorylation is suppressed after EGFR sialylation (Shaulian and Karin, 2002). However, EGFR phosphorylation has garnered considerably less research attention in RCC. In this study, the level of phosphorylated EGFR was assayed in ST3Gal1-regulated cells. The results consistently showed that downregulation of ST3Gal1 promoted EGFR phosphorylation. Additionally, phosphorylated EGFR expression was regulated through sialylation by ST3Gal1, mediated by MEG3. According to the outcome, sialylation of EGFR might be potential marker for RCC diagnosis.

EGFR signaling is a complicated network regulated by its phosphorylation. Upon activation of EGFR, various adaptor and effector molecules are activated, triggering several signaling cascades, such as the PI3K–AKT pathway. Furthermore, phosphorylated (Y920) EGFR can dock with PI3K, which inactivates the apoptotic cascade via BAD (Bcl2-associated death factor) and caspase-9 (Datta et al., 1997). Here, we focused on the effect of the MEG3–c-Jun–ST3Gal1 axis on phosphorylated EGFR and the signaling molecules AKT and PI3K in RCC cell lines. Alteration of MEG3 mediated the activation of EGFR, as well as inducing PI3K–AKT pathway activation. It is reasonable to conclude that the promotional effects of MEG3 on RCC progression could at least be partially mediated via elevated EGFR-mediated PI3K–AKT signaling.

In summary, our study defines a mechanism in which MEG3 acts as a tumor suppressor that regulates ST3Gal1 by interacting with c-Jun. Sialylation of EGFR inhibits EGFR phosphorylation and PI3K–AKT pathway activation. However, the mechanisms of RCC progression are multiple and complicated, and further investigations should be made to explore the MEG3–c-Jun–ST3Gal1–EGFR crosstalk in RCC progression.

Human RCC cell lines 786-O, Caki-1, ACHN were purchased from KeyGEN Company (Nanjing, China). ACHN and 293T cells were cultured in Modified Eagle's medium (MEM) supplemented with 10% fetal bovine serum (FBS, Gibco, Grand Island, NY). 786-O and Caki-1 were maintained in RPMI-1640 supplemented with 10% FBS in a humidified atmosphere of 5% CO₂, maintained at 37°C. MEG3 was cloned into the pcDNA3.1 vector (Invitrogen) for expression. siMEG3, siSCR, shMEG3 and shSCR were synthesized by GenePharma. Lipofectamine 3000 reagents (Invitrogen Co., Carlsbad, CA, USA) were used for cell transfection.

Thirty pairs of RCC tissues and adjacent non-tumor tissues were collected from patients who underwent surgical resections at the First Affiliated Hospital of Dalian Medical University from January 2016 to September 2018, after obtaining informed consent. The investigation project and the informed consent have been approved by the Ethics Committee of the First Affiliated Hospital of Dalian Medical University (Ethics Reference No: YJ-KY-FB-2016-37) and all investigations were conducted according to the principles expressed in the Declaration of Helsinki. The specimens were frozen in liquid nitrogen and stored at −80°C.

TRIZOL reagent (Invitrogen, Carlsbad, CA, USA) was used to isolate total RNA. RT-qPCR analysis was performed using SYBR-Green quantitative RT-PCR Master Mix kit (TIANGEN BIOTECH, Beijing, China). The sequences of the upstream and downstream primers were as follows: 5′-GCCAGCTGTCCCTCTTACCTAA-3′ (F) and 5′-TCCAGTTTGCTAGCAGGTGAAC-3′ (R) for MEG3, 5′-ATGGTGACCCTGCGGAAGA-3′ (F) and 5′-TCATCTCCCCTTGAAGATCC-3′ (R) for ST3Gal1, and 5′-ACCACAGTCCATGCCATCAC-3′ (F) and 5′-TCCACCACCCTGTTGCTGTA-3′ (R) for GAPDH. The 2^−ΔΔCt method was used for quantification, and fold change for target genes was normalized using an internal control. Each experiment was repeated three times.

The activation of EGFR was tested in serum-deprived cells by adding 20 ng/ml EGF for 5 or 10 min. Ice-cold 10% trichloroacetic acid was used to stop the reaction. The cellular precipitates were scraped and collected by centrifugation at 14,000 g for 5 min. Cells were washed with PBS, scraped and solubilized on ice for 15 min in RIPA buffer (50 mM Tris-HCl pH 7.4, 150 mM NaCl, 1% Triton X-100, 1% sodium deoxycholate and 0.1% SDS) and protease inhibitor cocktail (200 mM AEBSF, 30 μM Aprotinin, 13 mM Bestatin, 1.4 mM E64 and 1 mM Leupeptin). The protein concentration of the cell extract was tested using a BCA kit.

Cell extracts were electrophoresed by SDS–PAGE. Proteins were transferred to a PVDF membrane. Non-specific binding was blocked with 5% skimmed milk for 2 h at 37°C. The membrane was incubated with anti-EGFR (1:1000 dilution; Abcam), anti-phospho-EGFR (1:1000 dilution; Abcam), anti-ST3Gal1 (1:1000 dilution; Abcam), anti-PI3K, anti-phospho-AKT, anti-AKT (all 1:1000 dilution; Abcam) or Maackia amurensis lectin (MAL, 1:2000, Vector, California, USA, Catalog. B-1315) at 4°C overnight. GAPDH antibody (1:1000 dilution; Abcam) was used as a control. ɑ2,3-Sialylated EGFR was detected using a lectin affinity assay by immunoprecipitating lysates with the anti-EGFR antibody. The membrane was incubated with peroxidase-conjugated anti-rabbit IgG (1:2000 dilution; Abcam). Images were obtained using a multifunctional gel imaging system (Image Quant LAS 500, General Electric, Fairfield, CT, USA).

The ST3Gal1 promoter region was synthesized and inserted into a pGL3-basic vector (Promega, Madison, WI, USA). 786-O cells were co-transfected with the ST3Gal1 promoter-driven luciferase reporter together with pGL3-MEG3 or pGL3- basic vector. A Dual-Luciferase Assay Kit (Promega, Madison, WI, USA) was used to assess luciferase activities.

The invasion assays were performed using 24-well transwell chambers (8 μm; Corning, Shanghai, China), and the upper chambers were covered with 1 mg/ml Matrigel at first. Tumor cells were resuspended in serum-free RPMI-1640 medium and 1.0×10⁴ cells were seeded into the upper chambers. 0.5 ml RPMI-1640 containing 10% FBS was added to the bottom chambers. After 24-h incubation, cells on the upper surface of the membrane were scrubbed off, and the migrated cells were fixed in 75% methanol, stained with 0.1% crystal violet for 25 min, and counted under a light microscope (40× objective, 10× eyepiece magnification).

Four-week-old nude mice were purchased from the Model Animal Research Institute of Nanjing University. 5×10⁶ cells stably transfected with MEG3, shMEG3 or control vector were injected into the right flank of every nude mouse by subcutaneous injection. Tumor size was determined every 7 d. After 28 d, the nude mice were killed, and tumors were isolated and weighed. Tumor volume was calculated as the following formula: (length×width²)/2.

The GFP-labeled 786-O cells with different treatments were cultured in RPMI-1640 medium, resuspended at a concentration of 1×10⁷ cells/200 µl, and injected subcutaneously into each mouse tail vein. After 3 weeks, mice were killed and lungs were harvested for fluorescence analysis. All lungs were imaged within 20–30 min after euthanasia. The representative X-ray and fluorescence images were captured to reveal the degree of lung metastasis in vivo. The images were taken using the in vivo Imaging System of Dalian Medical University. All animal experiments were performed according to approved guidelines.

The cell viability was performed using a Cell Counting Kit-8 (CCK-8; KeyGEN, Nanjing, China). 5×10³ cells were seeded into 96-well plates with 100 μl of RPMI-1640 medium containing 10% FBS. 10 μl CCK-8 solution was added to the plate after incubation. The absorbance was measured in a microplate reader at 450 nm (168-1000 Model 680, Bio-Rad). Each experiment was performed three times.

4×10⁵ cells were seeded in a 12-well plate. The following day, when adherent cells were observed, the cell layer was scratched using a pipette tip. Next, images of cell morphology were captured at the initiation time and for 24 h under an Olympus microscope (40× objective, 10× eyepiece magnification). The migratory ability was quantified and normalized to relative gap distance. The results of the experiments were analyzed using IPP6.0 software (MEDIA CYBERNETICS, USA).

Chromatin immunoprecipitation (ChIP) assays were performed using a ChIP chromatin immunoprecipitation kit (Beyotime, Shanghai, China). Immunoselections of cross-linked protein–DNA were performed using anti-c-Jun antibody (1:1000 dilution; Abcam) and proteinA/G agarose beads, at 4°C overnight. Anti-rabbit IgG was used as a negative control. Purified DNAs were analyzed by PCR.

The 786-O cells were lysed in RIP lysis buffer (N653-100 ml; Shanghai Hao Ran Biotechnology Co. Ltd., Shanghai, China). Then, the extract was incubated with magnetic beads conjugated with c-Jun antibody or control IgG for 6–8 h at 4°C. The beads were washed with washing buffer (EHJ-BVIS08102; Xiamen Huijia Biological Technology Co. Ltd., Xiamen, China) and incubated with proteinase K at 55°C for 30 min to remove the proteins. RNA was extracted using Trizol, and qPCR was performed to detect MEG3 RNA present in the immune complex.

The RCC tissues, adjacent normal tissues and RCC cells were fixed in 10% formaldehyde, dehydrated in a dehydration series of alcohol and then embedded in paraffin. 0.5 mm thick slices were cut, dewaxed and rehydrated. The samples were digested with proteinase K for 15 min in 37°C, then denatured with formamide. The cells were incubated with Cy3-labeled MEG3 probes at 73°C for 15 min and incubated at 37°C overnight. The slices were washed with 2×SSC for 3 min. DAPI (4′,6-diamidino-2-phenylindole) was used to counterstain the nuclei for 20 min in darkness. The samples were then scanned and photographed.

Immunohistochemistry (IHC) staining was performed on paraffin-embedded sections of tissue samples. Serial sections (5 μm each) were deparaffinized, rehydrated and immersed in 3% hydrogen peroxide for 10 min to block the endogenous peroxidase activity. The slices were incubated with the corresponding primary antibodies overnight at 4°C. After extensive washing with PBS, the secondary streptavidin–horseradish peroxidase-conjugated antibody staining was performed at room temperature, visualized using 3,3′-diaminobenzidine (DAB; ZLI9018, ZSGBBIO, China), then counterstained with Hematoxylin, dehydrated and cover-slipped.

Data are presented as the mean±s.d. Differences between two groups were assessed using an unpaired two-tailed t-test. P<0.05 was considered to be statistically significant. All results were reproduced across triplicate experiments. Statistical analyses were carried out using GraphPad Prism (GraphPad Software, Inc., USA).

The peer review history is available online at https://jcs.biologists.org/lookup/doi/10.1242/jcs.244020.reviewer-comments.pdf