The lncRNA UCA1 promotes proliferation, migration, immune escape and inhibits apoptosis in gastric cancer by sponging anti-tumor miRNAs

1As the fourth most common cancer worldwide, gastric cancer (GC) is still incurable and induces about 700,000 mortalities each year. What is worse, non-resectable or metastatic GC is associated with poor prognosis, and systemic chemotherapeutic approaches provide minimal benefit. Thus, novel biomarkers for improving GC, early diagnosis, prognostic evaluation and tumor grading is urgently needed.

2Owing to the development of continued advances in high throughput sequencing, scientists realized that the most part of human genome is transcribed to non-protein coding RNAs, which indicates that a large group of RNA regulators is dedicated to regulating a relatively small amount of effectors. Among the newly discovered RNA elements, long non-coding RNAs (lncRNAs) have been identified to function as key regulators of diverse cellular processes, such as development, differentiation, and cell fate as well as disease pathogenesis. lncRNAs can serve as signal mediators, molecular decoys and scaffold or enhancers of transcription and what is intriguing is that a large group of lncRNAs function as competing endogenous RNA (ceRNA) that regulate gene expression through absorbing miRNAs.

3Until now, different groups have screened the expression profile of lncRNAs in GC tumors and found several disordered lncRNAs relates to GC carcinogenesis such as PVT1, H19, LNC00152 and so on. However, the underling mechanisms are still not well understood.

4UCA1 (urothelial carcinoma associated 1) is a lncRNA that was first identified in human bladder carcinoma, whose expression is found to be up-regulated in many other cancers till now. In gastric cancer, overexpressed UCA1 has been found in tumor tissues compared with paired non-cancerous samples, and the function of UCA1 have been confirmed to contribute to the cancer cell migration, invasion and drug resistance.

5In this study, we examined the level of UCA1 in different GC subtype tissue samples and unveiled the mechanism of how UCA1 modulates GC cells proliferation, migration and immune escape.

6Human gastric cancer cell line AGS, SGC-7901, BGC-823, MGC-803 and SNU-1 were obtained from National Infrastructure of Cell Line Resource (Beijing, China). Primary stomach epithelial cells from two healthy individuals were purchased from Abace-biology (Beijing, China) (www.abace-biology.com/primary-cell.htm) All the cells were maintained in humidified incubator at 37 °C in a CO2 incubator in RPMI 1640 medium supplemented with 10% fetal bovine serum (FBS) and 1% penicillin-streptomycin.

7A total of 40 primary gastric adenocarcinoma tumor tissues and corresponding adjacent non-tumorous gastric tissue samples were obtained between 2013 and 2015 at Ren Ji Hospital, Shanghai Jiao Tong University. A part of each tissue samples was subject to formalin fixation and paraffin-embedment. And another part of each tissue samples was stored in a refrigerator at − 80 °C. All the 40 paired tumor and non-tumor tissues were subjected to following study including immunoblotting, RNA extraction and qRT-PCR.

8Full length of 2314 UCA1 sequence was cloned into pCR3.1 vector to construct the UCA1 overexpression vector. For dual-luciferase assay, full length of UCA1 was cloned into pmirGLO vector, following firefly luciferase coding region. A 621 bp segment of PDL1 3’UTR was cloned into pmirGLO vector to form PDL1 3’UTR luciferase reporter vector.

9For UCA1 knockdown construction, two guide RNAs targeting the promoter region of UCA1 were constructed into lenti-guide-puro vector (Addgene: 52963), named pUCA1-KD. For UCA1 knockout cell establishment, AGS or SGC-7901 cells were co-transfected with pUCA1-KD and plentiCas9-Blast (Addgene: 52962) followed by selection for more than 5 days with blasticidin (4μg/ml) and puromycin(2μg/ml) for 1 week.

10Quantitive RT-PCR analysis was used to determine the relative expression level of mRNAs and miRNAs. Total RNA was extracted from clinical samples and cells, using Trizol Reagent (Invitrogen, Carlsbad, CA, USA) according to the manufacturer’s instructions. The RNA samples were treated using TURBO DNA-free kit to remove the DNA contamination and then the DNA contamination was assessed by agarose gel. RevertAid First Strand cDNA Synthesis Kit (Thermo Scientific) was employed for the reverse-transcription, and the level of specific RNAs were determined using SYBR® Green real-time PCR kit (Invitrogen, Carlsbad, CA, USA) with GADPH as loading control. The level of β-actin was also quantified to confirm the relative expression of UCA1.

11The expression level of miRNAs was detected by TaqMan miRNA RT-Real Time PCR. Single-stranded cDNA was synthesized by using TaqMan MicroRNA Reverse Transcription Kit (Applied Biosystems, Foster City, CA, USA) and then amplified by using TaqMan Universal PCR Master Mix (Applied Biosystems, Foster City, CA, USA) together with miRNA-specific TaqMan MGB probes (Applied Biosystems, Foster City, CA, USA). The U6 snRNA was used for loading control. The level of small nucleolar RNA, C/D box 58B (RNU58B) was also quantified to confirm the relative expression of miRNAs. Each sample in each group was measured in triplicate and the experiment was repeated at least three times for the detection of miRNAs.

12Slides were deparaffinized and rehydrated through immersion in xylene and an ethanol gradient and then digested with proteinase K (20 μg/mL) for 20 min at 37 °C. Slides were firstly fixed in 4% paraformaldehyde for 5 min at room temperature, and then dehydrated by immersion in an ethanol gradient and air dried; slides were pre-hybridized using DIG Easy Hyb (Roche, Mannheim, Germany) at 50 °C for 1 h. The 10 pmol digoxin-labeled UCA1 DNA probe was denatured in hybridization buffer at 95 °C for 2 min and then chilled on ice. The UCA1 probe was diluted in 250 μL pre-warmed in hybridization buffer. Each sample was covered with 50–100 μL diluted probe and incubated in a humidified hybridization chamber at 50 °C overnight. Slides were washed twice in 50% formamide in 4 × SSC at 37 °C for 30 min, and then washed three times in 2 × SSC at 37 °C for 15 min. After twice washing with maleic acid buffer containing Tween-20 (MABT), slides were blocked using blocking buffer (Roche, Mannheim, Germany) at room temperature for 30 min, the blocking buffer was drained off, and the samples were incubated with 1:250 diluted Anti-Digoxigenin-AP Fab fragments (Roche, Mannheim, Germany) at 37 °C for 1 h. After washing twice in MABT and once in detection solution (0.1 M Tris-HCl, 0.1 M NaCl, pH 9.5), the slides were stained with freshly diluted NBT/BCIP detection solution (Roche, Mannheim, Germany) and incubated at 37 °C for 30 min. Slides were washed in PBS twice, air dried for 30 min, and then mounted with Eukitt quick-hardening mounting medium (Sigma Aldrich, St. Louis, MO, USA). Images were obtained using a microscopy (Leica DM2000 LED) and a digital camera (Leica DMC 2900). Three different random images were captured for each sample at 400× magnification and the relative density of UCA1 signal was quantified by Image J. The results were analyzed using paired t-test and p < 0.05 was statistically significant.

13AGS cells were rinsed by PBS and then fixed in 4% formaldehyde for 10 min at RT. Further, the cells were permeabilized in PBS containing 0.5% Triton X-100 and 5 mM VRC(New England Biolabs Inc., USA) on ice for 10 min, then washed with PBS. After rinsed once in 2 × SSC, hybridization was carried out using FAM labelled UCA1 DNA probe at 37 °C for 12 h. After three times washing by 2 × SSC, one of the candidate miRNAs was detected by incubation with Cy5 labelled miRNA specific probe at 37 °C for 12 h. After nucleus staining by DAPI, images were captured by confocal microscopy (Leica TCS SP8) at 630× magnification.

14Cell extracts were incubated with biotin labelled UCA1 probe at 4 °C for 2 h, with biotin labelled sequence scrambled DNA oligo as control. Hybridized material was captured with magnetic streptavidin beads (Thermo Fisher Scientific). Beads were subjected to RNA extraction followed by RT-qPCR to quantify miRNAs levels.

15The number of apoptotic cells were evaluated by assaying for annexin V. Cells were stained with FITC labeled Annexin V antibody and PI (BioLegend), and then cells were analyzed using flow cytometry. Results were analyzed using Flowjo software (Ashland).

16AGS and SGC-7901 cells were seeded in 96-well plates at density of 5 × 103 cells per well in RPMI 1640 culture, and allowed to attach overnight. pCR3.1-UCA1 was transfected into cells for 48 h, with empty vector as control. Twenty microliters MTT (5 mg/ml) (Sigma, St. Louis, MO, USA) were added into each well and the cells were incubated for further 4 h. The absorbance was recorded at A570nm with a 96-well plate reader after the DMSO addition.

17Indicated cells were deprived of serum overnight, treated with mitomycin-C, seed into the upper chamber at density of 1 × 105/well (8 μm pore size; BD Bioscience). The chemoattractant in the lower chamber was filled with medium supplemented with 10% FBS. After 12 h, the lower chambers cells were fixed with 4% paraformaldehyde and then stained with crystal violet. The migrated cells were counted in 5 randomly different fields with an inverted microscope. The experiments were performed in triplicate wells and each experiment was performed in triplicate.

18The experiments including animals were approved by Ren Ji Hospital Animal Care and Use Committee. Female severe combined immunodeficient (SCID) beige mice were purchased from Charles River Laboratories at age 8 weeks with a weight between 20 g and 25 g. A total of 5 × 106SGC-7901 cells in 100 μL PBS together with an equal volume of Matrigel basement membrane matrix were injected subcutaneously into the shoulder to establish a human GC cell xenograft model.

19Protein samples were boiled in SDS/β-mercaptoethanol sample buffer, and 20 μg of each sample was loaded into each lane of 4–12% polyacrylamide gels. After separation by electrophoresis, the proteins in the gels were transferred onto PVDF membranes (Amersham Pharmacia Biotech, St. Albans, Herts, UK). The membrane was incubated with rabbit anti-CDK6 polyclonal antibody (ab151247, Abcam, Cambridge, MA, USA) or rabbit anti-EZH2 monoclonal antibody (ab191080, Abcam, Cambridge, MA, USA) or rabbit anti-PDL1 polyclonal antibody (ab205921, Abcam, Cambridge, MA, USA) or rabbit anti-TAK1 polyclonal antibody (ab109526, Abcam, Cambridge, MA, USA) or rabbit anti-KRAS polyclonal antibody (ab180772, Abcam, Cambridge, MA, USA), or rabbit anti-FGF9 polyclonal antibody (ab71395, Abcam, Cambridge, MA, USA) or mouse anti-β-actin monoclonal antibody (Santa Cruz Biotechnology Inc., Santa Cruz, CA, USA) over night at 4 °C. The specific protein-antibody complex was detected by using horseradish peroxidase conjugated goat anti-rabbit or rabbit anti-mouse IgG. Detection by the chemiluminescence reaction was carried using the ECL kit (Pierce, Appleton, WI, USA). The β-actin signal was used as a loading control.

20Paraffin-embedded sections were firstly deparaffinized and then incubated with rabbit polyclonal anti-Ki-67 (ab15580, Abcam, Cambridge, MA, USA) or rabbit polyclonal anti-PDL1 (ab205921, Abcam, Cambridge, MA, USA) primary antibody at 4 °C overnight. After three times wash by TBST, the sections were then incubation with HRP conjugated goat anti-mouse or goat anti-rabbit secondary antibody. The sections were washed by TBST for three times and the signal was detected using DAB Substrate kit following the manufacture’s instruction. Images were obtained using a microscopy (Leica).

21The peripheral blood was obtained from patients using heparin as the anticoagulant. The peripheral blood mononuclear cells (PBMCs) were isolated by Ficoll-Conray density gradient centrifugation, as described previously.The PBMCs (2.0 × 106/ml) were plated onto 6-well dishes and cultured with RPMI 1640 in the presence of human IFNγ (1.0 × 106 U/L); recombinant human interleukin 2 (5.0 × 105 U/L); 10% inactivated human serum; 25 mM HEPES; and 2 mM l-glutamine. After 24 h, monoclonal antibody against CD3 (100 μg/L, Antibody Diagnostic Inc., New York, NY, USA) and interleukin-1 alpha (1.0 × 105 U/L) were added. 2 days after incubation, the cells were cultured in medium without IFNγ and the medium was replaced every 3 days.

22For the cytotoxicity analysis, UCA1-KO and control GC cells were used as the target cells, while CIK cells were used as the effector cells. The effector and target cells were co-cultured in triplicate at ratios of 10:1, 20:1 and 40:1 with the wells containing just effector cells or target cells as controls. The cells were maintained in a humidified incubator at 37 °C in 5% CO2 for 24 h. 100 μl medium from the co-cultured cells was mixed with 20 μl CCK-8 solutions, and then incubated for another 4 h. The optical density (OD) was recorded at 450 nm for each well by a microplate reader. The survival (%) = (effector target cell mixture-effector group)/target cell*100%.

23Two-tailed Student’s t test was used to calculate statistical significance between two comparator groups. The differences of lncRNA expressions between paired tissue sample was evaluated with Wilcox matched pairs signed ranks test. The correlation analysis was analyzed by χ2-analysis. ALL the data were analyzed by using SPSS Statistical Package version 16. The survival times of different groups of patients were analyzed using the Kaplan-Meier method. A p value < 0.05 was considered as statistically significant.

24UCA1 is a lncRNA whose expression is found to related to the initiation and progression of many kinds of cancers. To explore its function in GC, we first analyzed the expression of UCA1 in GC tissues and noncancerous control tissues by using two published microarray datasets, GSE54129 (111 GC patients and 21 control) and GSE65801 (32 paired GC and noncancerous tissues). We found a significantly overexpressed UCA1 in patients with GC (Fig. 1a). We also collected 40 paired GC and noncancerous tissue samples (21 intestinal, 13 diffused and 7 mixed) and examined the UCA1 level by qRT-PCR. We found UCA1 level only up-regulated in the intestinal GC subtype tissues significantly (Fig. 1b). To further confirm up-regulation of UCA1 in GC tissues and identify its subcellular location, we detected the UCA1 in GC and control tissue sections using in situ hybridization (ISH). As shown in Fig. 1c, dense UCA1 signal only was found in GC tissues and mainly existed in the cytoplasm. After quantification and statistical analysis, the ISH results indicated that UCA1 was overexpressed in GC tissues. Finally, we analyzed the data reported by Szász AM et al. (320 intestinal GC, 241 diffused GC and 32 mixed), and found intestinal GC patients with higher UCA1 level have a poor overall survival, suggesting UCA1 has a potential oncogenic role during the carcinogenesis of GC(Fig. 1d).

25To further explore the biological function of UCA1 in the GC cells, we first constructed UCA1 overexpression vector and established UCA1 overexpression GC cell lines (Fig. 2a). As shown in Fig. 2b, GC cells with overexpressed UCA1 have significantly increased cell viability. Meanwhile, we constructed UCA1 knockdown vector which expresses two guide RNAs targeting the promoter region of UCA1 and then established UCA1 knockdown cell lines (Fig. 2c, d). The UCA1 promoter region knockout was confirmed by genotyping and sequencing(Additional file 1: Figure S1). As shown in Additional file 1: Figure S1A, The wildtype PCR product is 1705 bp which between bands of 1500 bp and 2000 bp of the marker lane. Meanwhile, the expected UCA1 knockdown PCR products are around 1103 bp which located between bands of 1000 bp and 1500 bp of the marker lane.

26We found significantly reduced viability and increased apoptosis in the UCA1 knockdown cells (Fig. 2e, f). Subsequently, examined by transwell assay, UCA1 overexpression GC cells exhibited increased motility and UCA1 knockout GC cells have reduced migration cell numbers (Fig. 2g, h).

27Accumulating evidence indicates that lncRNAs can exert “sponge-like” effects on various miRNAs, which subsequently inhibits miRNA-mediated functions. So, we firstly analyzed the potential of interaction between UCA1 and miRNAs. We found miR-26a and -b may bind with UCA1 directly, and this binding was subsequently confirmed by dual-luciferase assay (Fig. 3a, b), RAP assay (Additional file 3: Figure S3A) and FISH (Additional file 3: Figure S3B). Meanwhile, when UCA1 was overexpressed in the GC cells, the expression of endogenous miR-26a and -b was significantly downregulated Fig. 3c and d. We also detected the protein level of two miR-26 targets EZH2 and CDK6 that were confirmed by other researchers. As shown in the right panel of Fig. 3c and d, the level of EZH2 and CDK6 was significantly increased in the UCA1 overexpression cells, suggesting the function of miR-26 was repressed. To further determine miR-26a and -b level in vivo, we detected miR-26a and -b expression using qRT-PCR. We found decreased miR-26a in all the GC tissues and decreased miR-26b in intestinal GC tissues (Fig. 3e). We also found a significant strong negative correlation between UCA1 and miR-26b (Fig. 3e).

28To verify that UCA1 regulate EZH2 and CDK6 expression through sponging miR-26a and miR-26b, we first quantified the EZH2 and CDK6 protein level in tumour and control tissues by immunoblotting. Results indicated that both EZH2 and CDK6 protein were increased in GC tissues (Additional file 2: Figure S2A) and negatively correlated with UCA1 level(Additional file 2: Figure S2B and C). Meanwhile, in the UCA1 siRNA treated cells, EZH2 and CDK6 expressions were repressed, which can be rescued by miR-26a/b antagonists(Additional file 2: Figure S2D). These results indicated that UCA1 regulated EZH2 and CDK6 expression through sponging miR-26a/b.

29In addition, we also found miR-214 and -193 have the potential to be ‘sponged’ by UCA1 (Fig. 4a), and both were reported function as tumor suppressors. Dual luciferase assay was processed to confirm the interaction between miRNAs and UCA1. As shown in Fig. 4b, luciferase activities were repressed by miR-193a and − 214 mimics, and up-regulated by miR-193a and − 214 inhibitors, meanwhile, the level of miR-193a and − 214 was reduced in UCA1 overexpression cells (Fig. 4c). In the GC tissues, both of miR-193 and miR-214 were reduced and negatively correlated with UCA1 level (Additional file 4: Figure S4A), indicating the direct interaction between miRNAs and UCA1. The direct interaction between UCA1 and miR-193 or miR-214 was confirmed by RAP assay (Additional file 3: Figure S3A) and FISH (Additional file 3: Figure S3B) in GC cells.

30PDL1 is an important member of the B7 family molecules which can interact with PD-1 to inhibit the T-cell activation, induce apoptosis of effector T cells, and finally impair the anti-tumor immunity. Increasing evidences indicated that PDL1 overexpression in many human cancers is significantly associated with tumor progression and patient’s prognosis. Predicted by online bioinformatics tools, we surprisingly found miR-214 and -193a target sites in the 3’UTR of PDL1 Mrna (Fig. 4d). Subsequent dual-luciferase assay indicated that miR-214 and -193a can repress luciferase activity by targeting the 3’UTR of PDL1(Fig. 4e). Meanwhile, in the gastric cancer cells, the endogenous PD-L1 expression was also modulated by miR-214 and -193a (Fig. 4f). In intestinal GC tissues, PD-L1 expression was found increased and positively correlated with UCA1 level(Additional file 4: Figure S4B).

31To further determine whether UCA1 knock out can induce PDL1 downregulation in GC cells by sponging miRNAs, and subsequently alleviate T cell suppression, we first analyzed the PDL1 in the UCA1 knock out GC cells by western blot and flow cytometry. We found a significant reduced PDL1 expression and left shift peak of the UCA1-KO GC cells (Fig. 5a,b). Meanwhile, the peak left shift can partially rescued by antagonists for miR-193a and miR-214 suggesting the cell surface PDL1 level was regulated by UCA1-miR-193a/214 axis.

32Subsequently, we evaluated if modulation of PDL1 expression on GC cell lines could alter PHA-stimulated activation of healthy donor PBMC as measured by the released supernatant IFNγ level. As shown in Fig. 5c, IFNγ release by PHA stimulated PBMC was partially suppressed by the AGS-NC or SGC7901-NC cells when compared with PHA-stimulated PBMC alone suggesting a significant PDL1 mediated immune suppressive effect. When PBMCs co-cultured with UCA-KO GC cells, the IFNγ levels were increased, indicated that PDL1 downregulation induced by UCA1 knock out alleviated the immune suppressive effect. Subsequently, we analyzed the cytotoxicity of CIK cells on both AGS and SGC-7901 cells when the effector-target ratio was in the range 10:1 to 40:1. As shown in Fig. 5d, the cancer cell survival rate was obviously related to the effector-target ratio, and decreased with increase of the effector-target ratio. The cytotoxic activity of the CIK cells against AGS-UCA1-KO cells was markedly higher than that of AGS-NC cells under the same conditions (Fig. 5d, p < 0.05), and the same in SGC-7901 cells.

33In order to explore the effect of UCA1 knockout in the regulation of tumor growth in vivo, we further examined the function of UCA1 using AGS cell subcutaneous xenograft mouse model. The AGS-NC and AGS-UCA1-KO cells were injected subcutaneously into SCID mice. The tumor volumes were measured every 7 days for 28 days. As shown in Fig. 6a, knockout of UCA1 significantly inhibit tumor growth in mice. After 30 days, the tumor volume and weight of AGS-UCA1-KO group were significant smaller and lighter than those of AGS-NC group (Fig. 6b, c). Meanwhile, the level of anti-tumor miRNAs (miR-26a, miR-26b, miR-193a and miR-214) that directly interact with UCA1, was increased in the UCA1-KO tumors (Fig. 6d). Decreased Ki-67, PDL1 and increased cleaved PARP1 and caspase-3 were also found in the UCA1-KO tumors (Fig. 6e and f).

34LncRNAs have been identified to function as key regulators of diverse cellular processes including the initiation and progression of cancer. As an lncRNA, although UCA1 was found overexpressed in gastric cancer, the role of UCA1 in GC is still not fully understood. In this study, we first analyzed two newly released microarray datasets, and found the consistent results showing overexpressed UCA1 in the GC tissues. According to Lauren’s criteria, gastric cancer is classified into two major histological subtypes: intestinal type and diffuse type adenocarcinoma. After analyzing the UCA1 level in 40 paired GC and adjacent non-tumor tissue samples, we found that UCA1 especially overexpressed in intestinal GC tissues. Since the two histological subtypes of gastric cancer have lots of distinct clinical and molecular characteristics, further analysis combine the UCA1 expression and other factors including molecular characteristics, drug sensitivity and biological behaviors, should be helpful to guide the GC clinical treatment.

35MiR-26a and miR-26b are the two multiple functional miRNAs which were found dysregulated in diverse cancers and are involved in various biological processes, including proliferation, migration, invasion, angiogenesis, and metabolism by modulating different signaling pathways. MiR-26a level has been found to be reduced in the plasma and cancer tissue samples of GC patients, but the mechanism are still not known. We identified that UCA1 functions as a miRNA “sponge” modulating miR-26a and miR-26b level by direct interaction. Furthermore, we also identified the negative correlation between UCA1 level and miR-26a or miR-26b in GC tissue samples, which may provide an answer of how UCA1 modulate the level of miR-26a or miR-26b in GC.

36PDL1 is a transmembrane protein that usually overexpressed in cancer cells, and plays a major role in suppressing the immune system by binding with PD1, which was found on activated T cells, B cells, and myeloid cells. In gastric cancer, increasing evidences indicates that upregulated PDL1 level associates cancer progression and worse prognosis. In this study, we identified that miR-193a and − 214 repressed PDL1 expression by targeting 3’UTR. Meanwhile, we confirmed that UCA1 functioned as a “miRNA sponge” absorbs miR-193a and miR-214 and protect the PDL1 expression at RNA level. These results indicated that UCA1 targeting treatment may not only a treatment targeting the tumor cells, but also a therapy activates the host immune system. Furthermore, our research indicated that UCA1-KO GC cells could induced a higher IFNγ expression when co-cultured with PBMCs, and have a lower survival rate when co-cultured with CIK cells in vitro, suggesting UCA1-KO GC cells may be more easily to be eliminated by the immune system.

37In conclusion, we present here that UCA1 functions as an onco-lncRNA promotes GC cells proliferation, migration and inhibits GC cells apoptosis by repressing anti-tumor miRNAs as miR-26a and miR-26a. Meanwhile, UCA1 also represses the host immune systems by up-regulating the PDL1 level of GC cells. UCA1-KO GC cells formed smaller tumors, and had higher miR-26a, −26b, −193a and − 214 level, reduced cell proliferation and increased apoptosis in vivo, indicating that UCA1 could serve as a potential novel therapeutic target for GC treatment.