The Association of EBV and Gastric Cancer

The present study aimed to investigate the clinicopathological features of EBV positive and EBV negative gastric cancer patients and the expression levels of proliferative, apoptotic and cell signaling proteins in tissues samples from these patients. The biological role of EBV infection was assessed in gastric cancer. Results: EBV was localized in the nuclei of gastric cancer cells (positive rate 6.86%). The infection rate of EBV in normal gastric mucosal cells, which were adjacent to cancer tissues, was 0. The difference noted was signicant (P = 0. 023). The expression levels of caspase-3 (P = 0.0423), FASL (P = 0.00297) and cyclin D1 (P = 0.0345) proteins were signicantly different in EBV positive and negative gastric cancer tissues. When the parameters gender, age, Lauren classication, histological grade, early and advanced tumor stage, vascular and nerve invasion, TNM grade and survival status were compared, the maximum tumor diameter, number of lymph node metastasis, caspase-8, Ki67 and P53 protein expression did not reveal signicant differences. Bcl-2 protein expression was positive in only one gastric cancer cell sample and negative in the other gastric cancer cell samples as well as in the corresponding normal gastric mucosal epithelial cells. However, signicant differences were noted with regard to the positive expression of Bcl-2 in the immune cells of gastric cancer and adjacent tissues (P = 1.17749E-39). The expression levels of Bcl-2 in the immune cells of EBV positive and EBV negative gastric cancer tissues were not signicantly different. The expression levels of caspase-8, caspase-3, FASL, Ki67, cyclin D1 and P53 proteins in gastric cancer cells were signicantly different compared to those of normal gastric mucosal cells derived from adjacent tissues (P < 0.05). These ndings were noted in both EBV positive and/or EBV negative gastric cancer cases (P < 0.05). The survival time of the patients Following analysis of 7 EBV positive gastric cancer tissues and the corresponding adjacent tissues, no differences were noted in the expression levels of these signaling molecules. The results indicated that Bcl-2 was negative in all EBV positive gastric cancer cells and in the corresponding normal gastric mucosal epithelial cells, whereas the positive expression of Bcl-2 in immune cells of gastric cancer tissues was lower than that noted in normal adjacent tissues. The difference was statistically signicant (P = 0.004061027). The expression levels of caspase-3 (P = 0.000212546), caspase-8 (P = 0.00096) and FASL (P = 9.06334E-05) in cancer were lower than noted

which initially leads to chronic atrophic gastritis, intestinal metaplasia, atypical hyperplasia and nally early and advanced gastric cancers 2 . The continuous development of speci c treatment methods has improved the survival rate of gastric cancer patients. However, the pathogenesis of gastric cancer remains unclear. A high number of gastric cancer patients have been identi ed with poor prognosis resulting in a high incidence rate of this disease. Further investigation of the speci c pathogenesis and factors that may affect disease prognosis is required to provide theoretical and experimental basis for successful prevention and identi cation of more effective treatment methods for gastric cancer 2 .
In 1990, Burke et al. detected the expression of Epstein-Barr virus (EBV) in gastric adenocarcinoma tissues by PCR. Since then, several reports have been published indicating that EBV expression can be detected in gastric cancer tissues [3][4][5] . The positive rate of EBV expression was basically stable and estimated at approximately 10% [3][4][5] . However, its role in gastric cancer and the identi cation of the speci c signaling molecules involved in the development of this disease remain unclear. The positive rate of EBV infection in gastric cancer cases was estimated to approximately 10%. Moreover, the analysis of large sample data indicates that the prognosis of EBV positive gastric cancer patients is better than that of EBV negative gastric cancer patients, whereas the speci c mechanism associated with these effects remains unclear [3][4][5] . Previous studies that have utilized a small sample size indicated no signi cant differences in the prognosis between EBV positive and EBV negative gastric cancer patients 5 . The low positive rate of EBV in gastric cancer reduces the sample size and limits further the validity of the statistical analysis. However, the factors that may improve the prognosis of EBV positive gastric cancer patients are yet to be identi ed and the results suggesting improvement in prognosis are contradictory 5 .
The possible mechanism may involve the activation of lymphocytes following EBV infection and the elimination of gastric cancer cells. In EBV positive gastric cancer, tumor suppressor gene expression is upregulated and these cells are more sensitive to chemotherapy. However, controversial studies have been published examining this mechanism of action 5 .
The present study investigated whether EBV initially infected normal gastric mucosal epithelial cells and subsequently cancer cells, or whether it acted by directly infecting gastric cancer cells. Based on the assumption that EBV positive gastric cancer patients exhibit an optimal prognosis, we addressed whether this is caused by EBV infection of gastric cancer cells or whether it is associated with other factors, such as the proteins involved in proliferation, apoptosis and signaling of gastric cancer and peripheral immune cells as well as the clinicopathological features of the patients.
To address this hypothesis, the present experimental study was designed. We investigated whether EBV infected initially normal gastric mucosal epithelial cells or directly gastric cancer cells. Secondly, the present study investigated whether the expression of associated signaling proteins, such as those involved in proliferation, apoptosis and immunosuppression of EBV positive gastric cancer cells was different from that of adjacent cells and EBV negative gastric cancer cells. Finally, the correlation between clinicopathological diagnosis and prognosis of EBV positive and EBV negative gastric cancer patients was explored to provide theoretical and experimental basis for Meta big data analysis.

Materials And Methods
Statement All methods were performed in accordance with the relevant guidelines.
Subjects: A total of 102 cases of gastric cancer with follow-up data were collected from January 1, 2016 to December 31, 2017 in the Department of Pathology, A liated Hospital of Medical College of Guilin.
The inclusion criteria were the following: (1) Cases of gastric adenocarcinoma diagnosed by two experienced pathologists. The diagnostic criteria were based on the 2019 NCCN guidelines; (2) cases with incomplete follow-up data and clinicopathological data were excluded; (3) complete information of TNM staging and tumor samples was obtained from all cases. The patients were followed up until December 31, 2020. The present study was approved by the Medical Ethics Committee of the A liated Hospital of Medical College of Guilin and informed consent was obtained for all patients who were willing to participate in this study.
Preparation of para n-embedded tissue and tissue microarray samples and hematoxylin eosin staining (H&E staining): The tissue samples were xed with 10% neutral formalin at room temperature for 24-48 h. Following sampling, conventional xation, dehydration and para n embedding, wax blocks were made, sectioned continuously to a thickness of 4 μm, dewaxed to water and stained with H&E. Two experienced pathologists veri ed the selection of the typical lesion sites (cancer and normal gastric mucosal tissue more than 5 cm away from the cancer lesion). The core points were selected according to a diameter of 1.5 mm and the tissue chips were made (Fig. 1). Each chip was provided with two marking points. The tissue microarray was continuously sectioned with a thickness of 4 μm, dewaxed to water and stained with H&E. The sections were incubated with alcohol and xylene to produce the H&E sections.
Immunohistochemical experiment and data analysis: The tissue microarray was continuously sectioned for 4 μm and conventionally dewaxed to water. Following antigen thermal repair (100°C boiling water EDTA high temperature repair for 10 min), 1% hydrogen peroxide was added dropwise to block the samples at room temperature for 10 min. PBS was used for washing (5 min/3 times). Blocking was performed using 2% fetal bovine serum at room temperature for 10 min and the aforementioned primary antibodies (i.e. Bcl-2 antibody 1:1, P53 antibody 1:1, Ki67 antibody 1:1, cylinD1 antibody 1:1, anti-CASP8 antibody 1:200, anti-cleaved-caspase-3 rabbit antibody 1:200 and anti-Fas Ligand rabbit pAb 1:200) were added. The samples were incubated at 37°C for 1.5 h and washed three times with PBS containing 1/1,000 Tween 20 for 5 min each. Moreover, biotin-labeled sheep anti-mouse/rabbit IgG polymer was added to the samples, which were incubated at room temperature for 30 min and washed 3 times with PBS containing 1/1,000 Tween 20 for 5 min each. The DAB color developing solution was added dropwise and the color development was controlled under the microscope. Finally, the nucleus was stained with hematoxylin, dehydrated and sealed. The sections were prepared by a digital pathological section scanner (KF-PRO-005) and the immunohistochemical results were interpreted. The evaluation of the nuclear positive markers (Ki67, cyclin D1, P53) was performed using the hot spot counting method to select the site with the highest positive expression and the positive percentage was counted. The expression of the cytoplasmic and membrane positive markers (caspase-8, caspase-3, FASL, Bcl-2) was assessed based on a speci c score according to the following methods: Detection and evaluation methods of EBV infection: The EBER (Epstein-Barr virus encoded RNAs) in situ hybridization staining was used to detect whether the tissues in the aforementioned microarray were infected with EBV. The tissue chip was continuously sectioned with a thickness of 4 μm, the para n section was placed in fresh xylene solution, soaked 3 times for 10 min each and then placed in absolute ethanol. Subsequently, the tissues were soaked 3 times for 3 min each and dried in air for 10 min. A total of 300 μl gastric enzyme working solution was added dropwise to the sections of the tissue microarray, which were incubated at 37°C for 10 min. The gastric enzyme solution was discarded and the tissues were dehydrated with gradient ethanol (75%, 95% and 100% for 2 min each) and nally dried with air. A total of 100 μl digoxigenin labeled probe or blank control reagent was added dropwise to the tissue, which was sealed with silici ed cover glass and rubber cement. The samples were incubated overnight at 37°C (in wet box). The slides were immersed in PBS buffer for 10 min, rinsed 3 times with PBS buffer for 2 min each and rinsed further 2 times with deionized water for 2 min each. A total of 100 μl of freshly prepared DAB color solution was added to the slides, which were incubated at room temperature for 10 min, washed with water, re-stained with hematoxylin for 10 sec, differentiated and colored blue. Finally, the slices were dehydrated and sealed. The brown-yellow marker appeared in the nucleus of the target cell as positive. The margins of > 5% positive area and above medium positive were considered as reliable positive criteria. Each tissue chip was assessed with a positive control tissue.
Clinicopathological analysis: The aforementioned tissue samples were re-examined by two experienced pathologists according to the 2019 NCCN guidelines to con rm the basic data, such as histological classi cation, tumor classi cation and staging, nerve and vascular invasion and TNM staging. All cases were followed up by telephone and information, such as survival and time of death were retrieved.
Statistical processing: All data were statistically processed by the SPSS 23 software. The differences in the qualitative data were analyzed by the chi-square test (comparison of two independent sample rate composition ratio), whereas the differences in the unpaired quantitative data were compared using the T test (double sample equal variance test). The differences in the paired quantitative data were assessed by the T test (paired two-sample analysis of mean value). Survival analysis was performed by Kaplan-Meier analysis and Life table analysis. P < 0.05 was considered to indicate statistically signi cant differences.

Correlation of EBV status (positive or negative) in gastric cancer with clinicopathological parameters
In the present study, 102 wax specimens of gastric cancer and normal tissues (localized over 5 cm adjacent to the cancer tissue) were collected from the Department of Pathology, A liated Hospital of Medical College of Guilin from January 1, 2016 to December 31, 2017. All cases had follow-up data and the follow-up ended on December 31, 2020. Among them, 77 males and 25 females were included. The male:female ratio was 3.08; the age ranged from 37 to 78 years, with an average age of 59.27 ± 9.08 years (mean ± standard deviation); a total of 16 cases of early gastric cancer and 86 cases of advanced gastric cancer were noted, whereas 7 cases of EBV positive gastric cancer and 95 cases of EBV negative gastric cancer were included. The positive rate of EBV infection was 6.86%. The Chi-squared test was used for analysis and the data indicated no signi cant differences between EBV positive and EBV negative gastric cancer with regard to gender (P = 0. with regard to stage and histological type. However, the EBV virus was only present in gastric cancer cells and did not exist in normal gastric mucosal cells, suggesting that EBV may play a certain role in the development and progression of this disease. Normal gastric mucosal cells will not be infected by EBV and gastric cells may be more susceptible to EBV infection following carcinogenesis of the gastric mucosa.   (Fig. 3).
The difference was not statistically signi cant (Log Rank. P = 0.974; Breslow. P = 0.927; Tarone-Ware. P = 0.924), which may be associated with the small number of cases (Table 3). 2. The differences in the expression levels of signaling molecules associated with proliferation and apoptosis of EBV positive and EBV negative gastric cancer and normal gastric mucosal (adjacent to cancer tissues) tissues.
Immunohistochemical analysis was used to detect the differences in expression levels of cell proliferation-and apoptosis-associated signaling molecules, such as Bcl-2, caspase-8, caspase-3, FASL, Ki67, cyclin D1 and P53 in all 102 cases of gastric cancer and normal tissues (adjacent to the cancer epithelia). The results indicated that the expression levels of these molecules in gastric cancer specimens were signi cantly different than those noted in adjacent tissues. Among them, Bcl-2 (P = 1.17749E-39) (positive expression site in cytoplasm and cell membrane) was expressed at low levels in immune cells found in gastric cancer tissues and at high levels in immune cells found in adjacent tissues (Table 4 mucosal tissues (adjacent to cancer tissue) (Fig. 5). The mutant P53 gene (P = 2.00994E-17) (positive expression site in nucleus) was highly expressed in gastric cancer tissues, whereas it was expressed at low levels in normal gastric mucosal tissues (adjacent to cancer tissues) (Fig. 6). However, Ki67 (P = 2.68778E-18) (positive expression site in the nucleus) and cyclin D1 (P = 6.21448E-23) (positive expression site in the nucleus) were highly expressed in gastric cancer tissues and expressed at low levels in normal gastric mucosal tissues (Table 4) (Fig. 6). Following analysis of 7 EBV positive gastric cancer tissues and the corresponding adjacent tissues, no differences were noted in the expression levels of these signaling molecules. The results indicated that Bcl-2 was negative in all EBV positive gastric cancer cells and in the corresponding normal gastric mucosal epithelial cells, whereas the positive expression of Bcl-2 in immune cells of gastric cancer tissues was lower than that noted in normal adjacent tissues. The difference was statistically signi cant (P = 0.004061027). The expression levels of caspase-3 (P = 0.000212546), caspase-8 (P = 0.00096) and FASL (P = 9.06334E-05) in cancer tissues were lower than those noted in adjacent gastric mucosal tissues. The expression levels of Ki67 (P = 0.001600915), cyclin D1 (P = 0.00174) and P53 (P = 0.0127) were higher in cancer tissues than those noted in adjacent gastric mucosal tissues and the differences were statistically signi cant. The results suggested that in EBV positive gastric cancer, Bcl-2, caspase-8, caspase-3, FASL, Ki67, cyclin D1 and P53 and other cell proliferation and apoptosis-associated signaling molecules were all involved in the occurrence and development of this disease (Table 5). Further analysis demonstrated differences in the expression levels of these signaling molecules in 95 EBV negative gastric cancer tissues and in the corresponding adjacent tissues. The results indicated that Bcl-2 expression was positive in cancer cells, with the exception of 1 case of gastric cancer. Bcl-2 expression was negative in the remaining 94 gastric cancer cell specimens and in 95 normal gastric mucosal epithelial samples (adjacent to cancer tissues). The expression of this protein in immune cells was signi cantly lower in EBV negative gastric cancer tissues than that noted in normal tissues (P = 1.92083E-37). Caspase-8 (P = 2.4811E-12), caspase-3 (P = 1.65E-11) and FASL (P = 1.33E-24) were expressed at lower levels in gastric cancer tissues than in adjacent tissues, while Ki67 (P = 4.28609E-17), cyclin D1 (P = 1.04761E-20) and P53 (P = 5.65185E-16) were expressed at higher levels in gastric cancer tissues than in adjacent tissues. The results were signi cantly different (Table 6).  were more closely associated with EBV positive gastric cancer. However, the speci c mechanism requires further experimental con rmation (Table 7).  5 . The detection rate of EBV virus in gastric cancer ranges from 7 to 15%, but the majority of these patients are stable cases (approximately 10%). The results of the present study indicated that this index was 6.86%, which is close to the previously reported ndings 6-10 .
EBV is a double-stranded linear DNA virus, which mainly infects lymphocytes and speci c epithelial cells.
Its infection rate in humans can reach 90% or higher 5,8,[10][11][12][13][14] . At present, the mechanism of the EBVmediated gastric mucosal epithelium infection has been described as follows: Initially, EBV infects B lymphocytes and oral epithelial cells; subsequently, EBV is reactivated in B lymphocytes in the stomach and released to infect epithelial cells 5,8,[10][11][12][13][14] . In the current study, EBV was not detected in normal gastric epithelial cells. However, the expression levels of EBV were detected only in less than 7% of gastric cancer cells, suggesting that it was unlikely for the virus to directly infect gastric mucosal cells. It is likely that gastric mucosal cells will exhibit an altered phenotype after their initial carcinogenic transformation, which facilitates EBV infection. However, this hypothesis requires further experimental con rmation. In the present study, we proposed a hypothesis for EBV-positive gastric cancer, suggesting that antibodies against EBV may play a therapeutic role with lower side effects. Previous meta-analysis data demonstrated that the prognosis of EBV positive gastric cancer patients is improved compared to that of EBV negative gastric cancer patients. Previous research studies with small sample sizes have revealed similar ndings. However, no signi cant differences were observed 6, 10,11 . The present study demonstrated that the survival time of EBV positive patients was longer than that of EBV negative patients. However, the difference was not statistically signi cant and additional samples have to be collected for further analysis. [15][16][17] . The present study indicated that EBV positive and negative gastric cancer cases did not reveal signi cant differences with regard to gender (P = 0. This phenomenon has been noted in all gastric cancer samples and EBV positive and negative samples 15,21 . It is suggested that EBV positive gastric cancer also evades apoptosis and promotes proliferation through this signaling pathway. However, the expression levels of caspase-3 and FASL signaling molecules in EBV positive gastric cancer cases were slightly lower than those in EBV negative gastric cancer samples, suggesting that EBV positive gastric cancer cells may be able to escape apoptosis more e ciently. However, additional experiments are required to con rm these ndings. 22,23 .
Ki67 is a protein that promotes proliferation. It is mainly located in the nucleus and is overexpressed in gastric cancer and various malignant tumors. The higher the positive rate of expression, the stronger the cell proliferative activity 16 . Cyclin D1 is a regulator of the CDK4/6 enzymes that promote cell cycle progression. Its expression is often increased in malignant tumors and its function is to promote the G1 to S transition of the cell cycle 22,[24][25][26] . The present study indicated that the expression levels of Ki67 and cyclin D1 in cancer tissues were higher than those in normal tissues, suggesting that the proliferative activity of gastric cancer cells was higher than that of normal gastric mucosal cells. The follow-up time period was short. The shortest follow up time period was 38 months and the longest was 60 months. The mean ± standard deviation was 49.63 ± 6.18 months. The comparison of the aforementioned parameters with the majority of the indices used in the study led to non-signi cant comparisons. Additional EBV positive and negative gastric cancer cases have to be collected for further analysis in subsequent studies. However, the data in the present study may provide detailed information for the design of subsequent large sample meta-analyses.
In conclusion, the positive detection rate of EBV in gastric cancer tissues was 6.86%. All these cases were expressed in cancer cells; the expression levels of the apoptosis inhibitor Bcl-2 was almost absent in epithelial cells of cancer tissues and in normal samples derived from specimens adjacent to cancer tissues. However, Bcl-2 expression was positive in immune cells of cancer and normal tissues. Its expression levels in immune cells from cancer tissues was lower than that noted in adjacent tissues; the expression levels of apoptosis-associated proteins, such as caspase-3, caspase-8 and FASL in cancer tissues were lower than those noted in normal gastric mucosal cells adjacent to cancer tissues. However, the expression of the proliferation-associated proteins in cancer tissues was higher than that noted in the corresponding adjacent tissues, suggesting that gastric cancer cells exhibited higher proliferative activity and lower apoptotic rate than normal cells adjacent to cancer tissues. Moreover, the expression levels of caspase-3, FASL and cyclin D1 noted in EBV-positive gastric cancer tissues were signi cantly different compared to those of EBV-negative gastric cancer tissues. The follow-up data indicated that the survival time period of EBV-positive gastric cancer patients was longer, whereas the difference was not statistically signi cant and consequently additional data need to be collected for further analysis.   Life table analysis indicating that the survival time period of EBV positive gastric cancer is longer than that of EBV negative gastric cancer cases and that the disease prognosis is better (Fig. 3). However, the differences were not statistically signi cant.   Immunohistochemical staining of P53, Ki67 and cyclin D1. (A and B) P53 was highly expressed in gastric cancer cells and expressed at low levels in the corresponding normal tissues, which were adjacent to cancer. (C and D) Ki67 was highly expressed in gastric cancer cells and expressed at low levels in the corresponding normal tissues, which were adjacent to cancer tissues. (E and F) Cyclin D1 was highly