|Home | About | Journals | Submit | Contact Us | Français|
O-GlcNAc transferase (OGT) is the only enzyme in mammals that catalyzes the attachment of β-D-N-acetylglucosamine (GlcNAc) to serine or threonine residues of target proteins. Hyper-O-GlcNAcylation is becoming increasingly realized as a general feature of cancer and contributes to rapid proliferation of cancer cells. In this study, we demonstrated that O-GlcNAc and OGT levels were increased in all six gastric cancer (GC) cell lines as compared with immortal gastric epithelial cells. Downregulation of the O-GlcNAcylation level by silencing OGT inhibited cell viability and growth rate via the cdk-2, cyclin D1 and ERK 1/2 pathways. In vivo xenograft assays also demonstrated that the hyper-O-GlcNAc level markedly promoted the proliferation of tumors. Moreover, compared with noncancerous tissues, the O-GlcNAcylation level was increased in cancerous tissues. GC patients with higher levels of O-GlcNAcylation exhibited large tumor sizes (≥5 cm), deep tumor invasion (T3 and T4), high AJCC stages (stage III and IV), more lymph node metastases and lower overall survival. Notably, the phosphorylation level of ERK 1/2 was increased progressively with the increase of O-GlcNAcylation in both SGC 7901 and AGS cells. Consistently, human GC tissue arrays also revealed that ERK 1/2 signaling was positively correlated to O-GlcNAcylation (r = 0.348; P = 0.015). Taken together, here we reported that hyper-O-GlcNAcylation significantly promotes GC cells proliferation by modulating cell cycle related proteins and ERK 1/2 signaling, suggesting that inhibition of OGT may be a potential novel therapeutic target of GC.
Gastric cancer (GC) is the third most common cancer in China . Established cell lines and animal models have been vividly intimated the important aspects of GC and provided significant clues to the critical molecular events during the process of the GC development. However, the mechanism behind GC development is still largely unknown.
O-GlcNAcylation is known as a reversible post-translational modification (PTM) regulated by O-GlcNAc transferase (OGT) and O-GlcNAcase (OGA) enzymes [2–6]. One of the remarkable features of cancer cells is the increased in aerobic glycolysis and increased in the rate of glucose uptake and utilization, which is termed the Warburg effect . UDP-GlcNAc, a product of the hexosamine biosynthetic pathway (HBP), is the activated form of GlcNAc and substrate of OGT. A high rate of aerobic glycolysis and glucose uptake would increase the HBP flux, ultimately leading to the elevation of O-GlcNAcylation in cancerous tissues [6–8]. Extensively investigations had confirmed that elevated protein O-GlcNAcylation and changes of OGT and/or OGA expression has a clear functional role in many types of malignancies, including bladder cancer , lung cancer , prostate cancer [11, 12], hepatocellular carcinoma  and colon cancers [14, 15]. O-GlcNAcylation regulates the activities of a wide panel of proteins involved in almost all aspects of cell biology, such as cell proliferation, survival, energy metabolism, migration and invasion [5, 6, 16, 17].
Previous studies revealed that hyper O-GlcNAcylation is associated with the development and progression of GC [18–20]. Recent studies have shown that O-GlcNAc expression levels were progressively increased during the carcinogenesis and progression of GC. Meanwhile, the expression of O-GlcNAcylation in H. pylori-infected chronic gastritis were higher than that in chronic gastritis without H. pylori infection . Yet, the potential effect and mechanism of O-GlcNAcylation in GC are still unclear. Moreover, no evidence is available on whether O-GlcNAcylation can be a hallmark of GC. In this study, we investigated the correlation between the O-GlcNAcylation level and clinicopathological parameters of GC patients. Importantly, we also elucidated the potential role of O-GlcNAcylation in GC proliferation by regulating cell cycle related proteins and activating the extracellular signal regulated kinase (ERK) signaling, which may provide a new sight for the diagnostic and therapeutic strategies of GC in human.
To evaluate the relationship between O-GlcNAcylation and gastric cancer, the OGT and O-GlcNAcylation levels were analyzed in various gastric cancer cell lines and immortal gastric epithelial cells (GES). The results showed that the OGT and O-GlcNAcylation levels were increased in all six gastric cancer cell lines compared with those in GES cells (Figure (Figure1A).1A). Consistently, the OGT mRNA level in all six GC cell lines was increased (Figure (Figure1B).1B). In addition, we found that OGA levels slightly increased in several GC cell lines as compared with GES cells (Supplementary Figure S1). The increase of OGA in GC may be a kind of compensatory change which attempt to “normalize” changing O-GlcNAc levels.
To investigate the function of O-GlcNAcylation in gastric cancer, we used siRNA to knock down endogenous OGT expression in SGC 7901 and AGS cell lines. The efficiency of siRNA silencing was confirmed by Quantitative Real-Time PCR (RT-) and Western blotting (Figure (Figure2A).2A). As expected, silencing OGT led to a significant reduction in overall O-GlcNAcylation compared with control cells (Figure (Figure2A).2A). In addition, to prove that an elevated overall O-GlcNAcylation level in SGC 7901 and AGS cell lines can stimulate cell proliferation, the OGA inhibitors PUGNAc and Thiamet-G (TMG) were used. The drug efficiency was confirmed by Western blotting (Figure (Figure2B2B).
Next, cell viability was measured by the CCK-8 assay for 6 days. Compared with scrambled negative control siRNA, OGT siRNA significantly inhibited cell proliferation (Figure (Figure2C,2C, P < 0.01). By contrast, the upregulation of O-GlcNAcylation caused by the OGA inhibitors PUGNAc and Thiamet-G (TMG) enhanced cell proliferation (Figure (Figure2D,2D, P < 0.05). Meanwhile, a clone formation assay of SGC 7901 cells showed similar results (Figure (Figure2E).2E). These results suggest that O-GlcNAcylation can promote gastric cell proliferation.
Reduced growth can be attributed to cell cycle defects. Thus, we wondered whether O-GlcNAcylation could modulate cell cycle progression. SGC 7901 cells were prepared for flow cytometry analysis after Thiamet-G (TMG, 10 μmol/L) treatment or after OGT siRNA transfection. We synchronized all of the cells in G0 phase by 24-hour serum starvation. The cells were then collected, stained with propidium iodide (PI) and analyzed by flow cytometry after 8 hours of serum stimulation (Figure (Figure3A).3A). As a result, the reduction of OGT in SGC 7901 cells caused a significant accumulation of cells in the G1 phase compared with control siRNA-infected cells: 60.2 ± 1.9% G1 content in OGT siRNA-infected cells relative to 48.7 ±1.2% in control siRNA cells (Figure (Figure3B).3B). Accordingly, the upregulation of the O-GlcNAcylation level by TMG led to increased cell proportion in the S- and G2/M-phases compared with control (Figure (Figure3B3B).
Next, we explored the mechanism by which O-GlcNAcylation could regulate GC cell proliferation. Cyclin D1 is the key molecule regulating the G0/G1 transition. The levels of O-GlcNAcylation and cyclin D1 in 6 GC cell lines after 12 hours of treatment with Thiamet-G (10 μmol/L) or isometric DMSO were detected by immunoblotting assay. The results showed that O-GlcNAcylation upregulation by Thiamet-G could increase cyclin D1 levels in 5 GC cell lines (Figure 4A and 4B). We then selected SGC 7901 and AGS for further investigation. The immunoblotting assay showed that the expression levels of cdk-2 and cyclin D1, proteins involved in the regulation of the mitotic cell cycle, were decreased when the OGT gene was downregulated by siRNA (Figure 4C and 4D). Abnormal activation of extracellular signal-regulated kinase (ERK) signaling pathway leads to inappropriate cell proliferation and survival. In our studies, we found that the phosphorylation level of ERK1/2 was significantly decreased by OGT siRNA, indicating that the O-GlcNAcylation level may regulate GC cell proliferation through the ER signaling pathway (Figure (Figure4D).4D). Immunofluorescence analysis of O-GlcNAc, Ki-67 and cyclin D1 further validated that O-GlcNAcylation could promote GC cell proliferation (Figure (Figure4E4E)
SGC 7901 cells were infected with recombinant lentivirus expressing OGT (LV-OGT), small hairpin RNA (Sh-OGT) or a vector control lentivirus (Vector), and stable clones were established (named SGC 7901 LV-OGT, SGC 7901 Sh-OGT and SGC 7901 Vector, respectively). The upregulation or downregulation of OGT expression was confirmed by qPCR and Western blot analyses (Figure 5A and 5B). To further explore the role of O-GlcNAcylation, we inoculated equal amounts of SGC 7901 LV-OGT, SGC 7901 Sh-OGT and SGC 7901 Vector cells into the front flank of female athymic nude mice. Compared with the SGC 7901 Vector groups, the size and weight of the tumors were significantly lower in the SGC 7901 Sh-OGT group (P < 0.01, Figure Figure5).5). Additionally, the size and weight of the SGC 7901 LV-OGT group were the highest (P < 0.05, Figure Figure55).
To characterize the clinicopathological relevance of O-GlcNAcylation in GC patients, we analyzed the relationships between the O-GlcNAcylation and clinicopathological parameters. Compared with noncancerous tissues in 90 cases of gastric cancer patients, the O-GlcNAcylation level was increased in cancerous tissues (Figure 6A and 6B). Then cancerous tissues and matched normal counterparts were divided into the high O-GlcNAcylation group (score: 5–12) and low O-GlcNAcylation group (score: 0–4), respectively (Figure (Figure6B).6B). High expression of O-GlcNAcylation was detected in 47 (52.22%) cases of GC tissues and in 21 (23.33%) of the matched normal counterparts (P < 0.01, Figure Figure6A).6A). Consistent with other findings [11, 21], the overall survival in patients with a high O-GlcNAcylation level was remarkably reduced (Figure (Figure6C,6C, P < 0.01). Meanwhile, the tumor size, invasion depth, lymph node metastases and AJCC stage also affected the overall survival of GC patients as assessed by univariate analysis (Table (Table1,1, P < 0.05). In addition, GC patients with a large tumor size (≥ 5 cm), deep tumor invasion (T3 and T4), high AJCC stage (stage III and IV), lymph node metastases, or lymphatic and/or vascular invasion had significantly higher levels of O-GlcNAcylation (Table (Table2,2, P < 0.05). Multivariate analyses also revealed that a high O-GlcNAcylation level was an independent risk factor for the poor prognosis of GC (Table (Table3,3, P < 0.05).
To further investigate the correlation between p-ERK 1/2 and O-GlcNAcylation, we detected the level of p-ERK 1/2 in SGC 7901 and AGS cells after treatment with TMG (10 μmol/L) at different times (0 h, 1 h, 3 h, 6 h, 9 h or 12 h). The results showed that the level of p-ERK 1/2 was also increased progressively with the increase in O-GlcNAcylation in both SGC 7901 and AGS cells (Figure (Figure7A).7A). To further verify the correlation between the levels of p-ERK 1/2 and O-GlcNAcylation in vivo, we analyzed the levels of p-ERK 1/2 and O-GlcNAcylation in two identical gastric cancer TMAs (provided by Xijing Hospital of Digestive Diseases, Fourth Military Medical University), which were constructed with formalin-fixed paraffin-embedded gastric neoplastic tissues of 48 gastric cancer patients. We divided the levels of p-ERK 1/2 and O-GlcNAcylation into three groups (Figure (Figure7B).7B). The results showed that there was a positive correlation between the levels of p-ERK 1/2 and O-GlcNAcylation in vivo by Pearson correlation analysis. (r = 0.348; P = 0.015) (Figure (Figure7C7C).
Studies in recent years have shown that a wide panel of oncogenic factors can be modified by O-GlcNAc and up-regulation of O-GlcNAcylation is believed to participate in tumor malignancy . Such as the modification of p53 and c-Myc with O-GlcNAc would increase their stabilities by blocking ubiquitin degradation [12, 22]. In addition to stabilizing oncogenic factors, O-GlcNAcylation could regulate their transcriptional activities. For example, the O-GlcNAcylation of p65 at Thr322 and Thr352 is critical for p65 promoter binding . Furthermore, the OGT and O-GlcNAcylation levels were high in late S phase and peaked at the M phase, indicating it may participate in the regulation of cell cycle progression [22, 23]. In our study, we found that the upregulation of O-GlcNAcylation promoted cell proliferation and depletion of OGT suppressed cell proliferation, which were accompanied by consistent alterations of the cell cycle and cell cycle-related proteins such as cdk-2, cyclin D1 and ERK 1/2. All of these results point to the ability of O-GlcNAcylation to enhance proliferation by modulating cell cycle progression and the ERK 1/2 pathway.
Clinical studies have also found that the high level of O-GlcNAcylation participate in the malignant clinicopathological features of various cancer patients [11, 13]. In breast ductal carcinomas, poorly differentiated tumors (grade II and III) displayed significantly higher OGT expression than grade I tumors . Additionally, the OGT mRNA level was even higher in the urine of bladder cancer patients with invasive or advanced stages (grade II and III) compared with those with early stages (grade I) . In human laryngeal cancer, the increased level of OGT mRNA was also related to a larger tumor size, nodal metastases, and a higher grade as well as the incidence of disease recurrence . An independent cohort of colorectal cancer patients from the Moffitt Cancer Center (GSE17536)  also showed that patients with high levels of OGT exhibited worse prognoses. In this study, we explored the clinical significance of the O-GlcNAcylation level in numerous GC tissues and found that its level was positively correlated with a large tumor size (≥ 5 cm), deep tumor invasion (T3 and T4), high AJCC stage (stage III and IV), lymph node metastasis and the overall survival of patients with a high O-GlcNAcylation level was significantly shortened. Meanwhile, we detected the expression of OGT and ERK 1/2 in five cases of gastric cancerous tissues and matched normal counterparts. The result also showed that the level of OGT was increased in cancerous tissues (Supplementary Figure S3).
The extracellular signal-regulated kinases ERK1 and ERK2 (ERK 1/2) cascade was first discovered as a classical signal-transduction pathway of the MAPK family, which regulates various cellular processes in normal and cancer tissues [26, 27]. In our study, we have elucidated that phosphorylated ERK 1/2 was decreased when OGT was silenced in GC cell lines. Similarly, high level of O-GlcNAcylation promoted high glucose-induced cardiac hypertrophy via ERK 1/2 , indicating that the regulation of the ERK 1/2 pathway by O-GlcNAcylation plays an important role in many diseases. Importantly, a recent study has revealed that activated ERK 1/2 signaling leads to increased expression of OGT and O-GlcNAcylation levels in various cancers , indicating that there is possibly positive feedback between the ERK 1/2 signaling pathway and the O-GlcNAcylation level in cancer cells. Although many colleagues have reported the correlation between O-GlcNAcylation and the ERK 1/2 pathway [28–31], the mechanism underlying the O-GlcNAc regulation of ERK 1/2 activation in gastric cancer is unclear. We predicted that there are O-GlcNAc sites in ERK 2 using the YinOYang 1.2 Server (www.cbs.dtu.dk/services/YinOYang) (Supplementary Figure S2A), which indicates that ERK 1/2 may be directly modified by O-GlcNAc. It is the accepted conclusion that O-GlcNAcylation has extensive crosstalk with phosphorylation to regulate signaling, transcription and the cytoskeleton in response to intracellular or extracellular stress [8, 32]. Thus, we hypothesized that the O-GlcNAcylation of ERK 1/2 can promote its phosphorylation and then activate this pathway, a finding that needs further research to confirm. In addition to direct O-GlcNAc modification, activation of this pathway may occur through the regulation of upstream ERK signaling components such as MEKs. Skorobogatko et al. proved that MEK2, a signaling component of the ERK 1/2 pathway, can be modified by O-GlcNAc and promotes the activation of the ERK 1/2 pathway at neuronal synapses. 
There is also a close relationship between O-GlcNAcylation and cell cycle regulation. A high level of glucosamine, which would increase the level of O-GlcNAcylation, could increase cell cycle regulatory proteins, including cyclin D1, CDK4, cyclin E, and CDK2, in mouse embryonic stem cells . In addition, the reduction of O-GlcNAcylation in breast cancer cells leads to the inhibition of tumor growth and is associated with decreased cell cycle progression . Meanwhile, OGT interference blocking cyclin D1 expression and the catalytic activity of OGT are necessary for the G0/G1 transition . Similar phenomena were also confirmed in our experiments. Our group has found that the upregulation of O-GlcNAcylation by TMG could increase cyclin D1 in various GC cell lines. Cyclin D1 and cdk-2 were decreased when OGT was silenced in SGC 7901 and AGS cell lines. It is known that O-GlcNAcylation can stabilize β-catenin , a transcription factor that can promote cyclin D1 transcription, which may have increased the cyclin D1 level as observed in our study. Interestingly, using the YinOYang 1.2 Server (www.cbs.dtu.dk/services/YinOYang), we predicted that there are two O-GlcNAc sites in cyclin D1 (Supplementary Figure S2B). However, further studies are needed to clarify whether the levels of these proteins are regulated by O-GlcNAcylation through the above-mentioned mechanisms.
In summary, our results have demonstrated that hyper-O-GlcNAc level markedly promoted the proliferation of tumors and indicated poor prognosis of GC patients. In this study, we have proven that O-GlcNAcylation can enhance GC cell proliferation mechanistically by orchestrating the expression of cell cycle-related proteins and ERK 1/2 pathway. And thus we propose that inhibition of hyper-O-GlcNAcylation may as a potential novel therapeutic target for cancer treatment.
GES, GC9811, MKN-45, MKN-28, BGC 823, SGC 7901 and AGS cells were maintained in DMEM medium (Gibco, Grand Island, NY, USA) supplemented with 10% fetal bovine serum (Bioind, Kibbutz Beit Haemek, Israel) and 1% penicillin/streptomycin (Gibco, Grand Island, NY, USA) at 37°C in a humidified incubator with 5% CO2.
PUGNAc (Sigma-Aldrich, St. Louis, USA) was dissolved in dimethyl sulfoxide (Sigma-Aldrich, St. Louis, USA) to a concentration of 200 mmol/L and was diluted to a final concentration of 100 μmol/L in RPMI-1640 medium. Thiamet-G (Calbiochem, San Diego, USA) was dissolved in DMSO to a concentration of 40 mmol/L and was diluted to a final concentration of 10 μmol/L in DMEM medium. Prior to PUGNAc or Thiamet-G treatment, the cells were incubated with DMEM medium containing 10% FBS and 1% penicillin/streptomycin to reach 80% cell confluency. Next, cells were treated with PUGNAc (100 μmol/L) or Thiamet-G (10 μmol/L) in the absence of FBS for 12 h.
Cells were seeded in 96-well plates (1 × 104 cells/well), cultivated at 37°C in a 5% CO2 humidified incubator for 24 h. Ten microliters of Cell Counting Kit-8 solution (Dojindo, Kumanoto, Japan) were added to each well and were incubated at 37°C in a 5% CO2-humidified incubator for 2 h. Spectrometer Varioskan® Flash (ThermoFisher, Waltham, USA) was used to measure the 450-nm absorbance. A proliferation curve was drawn with the time as the abscissa and average absorbance value in each group as the ordinate. Triplicate reactions were performed for the experiment.
Cells were seeded in 6-well plates (1000 cells/well) and were incubated at 37°C in a 5% CO2-humidified incubator. After 2 weeks, the cells were stained with Gentian violet (Beyotime Biotechnology, Shanghai, China). The experiments were repeated three times.
Cell immunofluorescence was performed as previously reported .
Stable transfectant SGC 7901 cell lines overexpressing OGT (SGC 7901 LV-OGT) or with OGT knockdown by small hairpin RNA (SGC 7901 Sh-OGT) were generated by lentiviral transduction using a GV341 vector (GeneChem Co., Ltd, Shanghai, China). An empty vector was used as the negative control (SGC 7901 Vector). Sequences are available from the authors on request.
Eighteen female BALB/c nude mice (5 to 6 weeks old; The Fourth Military Medical University, Xi'an, China) were randomly divided into three groups (LV-OGT, Sh-OGT and Vector groups). All of the mice were maintained under specific pathogen-free conditions. This study was carried out in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health. All of the procedures in the studies involving animals were performed in accordance with the ethical standards of The Fourth Military Medical University. The cells were cultured, and single-cell suspensions were prepared with cells at the log phase. Cells (1 × 107 cells) were subcutaneously injected into the right side of the back of the nude mice to establish the subcutaneous xenograft tumor model. The mice were checked daily, and the body weight and tumor size were measured every week. After 5 weeks, all of the mice were sacrificed under deep anesthesia. The final volume and weight of each tumor were recorded. The tumor volumes were calculated using the following formula: tumor volume (mm3) = [length (mm)×width (mm)2]π/6.
We used Prism 5 software (San Diego, CA, USA) and SPSS 18.0 (Chicago, IL, USA) for statistical analysis. All of the values are expressed as the means ± SD. The usefulness of the O-GlcNAc level to predict the risk of GC development was assessed using the Kaplan–Meier and Cox proportional hazards model analyses. The results were compared using Student's t-test. A P value < 0.05 was considered to indicate a significant difference.
We acknowledge Dr. Jiaqiang Dong from the Fourth Military Medical University for selfless help in the experiment. We appreciate Dr. Gang Wang from the Fourth Military Medical University for the pathological analyses and excellent statistical assistance.
CONFLICTS OF INTEREST
The authors declare that they have no conflicts of interest.
This work was supported by the National Natural Science Foundation of China (Nos. 81421003, 81430072, 81120108005, 81322037 and 81472300, 2015BAI13B07, CBSKL2015Z12, 81572676).