Implication of Leptin-Signaling Proteins and Epstein-Barr Virus in Gastric Carcinomas

We investigated the clinicopathological implications of leptin-signaling proteins and Epstein-Barr virus (EBV)-infection status in gastric carcinomas. Immunohistochemistry for leptin signalling-related proteins (leptin, leptin-receptor, pSTAT3, ERK, pAkt, mTOR and HIF-1 alpha), and in situ hybridization for EBV-encoded small RNAs was performed in 343 cases of gastric carcinomas. The siRNA against leptin-receptor was transfected into three stomach cancer cell lines, and western blot for caspase 3 was performed. The TNM stage was a prognostic factor in all 343 patients, and was negatively correlated with expression of leptin, pSTAT3, ERK, pAkt, mTOR and HIF-1 alpha (P < 0.05). Leptin-receptor expression was correlated with poor survival in 207 patients of the advanced gastric cancer (AGC) subgroup, 139 of the Lauren diffuse group, and in 160 patients with lymph node metastasis (P < 0.05, respectively). Additionally, in stomach cancer cells, cleaved caspase 3 level increased by leptin-receptor inhibition, that is, apoptosis increased. Interestingly, EBV-positive AGC (n = 29) tended to show better survival of patients than EBV-negative AGC (n = 178) (P = 0.06). pAkt expression was related with a good survival of 32 patients (9%) in the EBV-positive subgroup, but was not an independent prognostic factor. Among, leptin signaling-related proteins, expressions of leptin-receptor and mTOR were different between EBV-positive subgroup and EBV-negative subgroup (P < 0.05, respectively). In conclusion, leptin-signaling proteins and EBV status show different significance on patient survival, according to subsets of gastric carcinomas. The leptin-receptor may predict poor patient prognosis in the AGC, Lauren diffuse and lymph node metastasis subgroups, while EBV-positive status can show a good prognosis in the AGC. Each leptin signaling-related protein may be differently involved in carcinogenesis of EBV-negative and EBV-positive subsets.


Introduction
Gastric carcinoma is the most common malignancy except for thyroid carcinoma, and ranks as the third cause of cancer death in Korea [1]. The worldwide occurrence of gastric carcinomas is estimated as 8%, and the mortality accounts for 10% of all cancer cases [2]. Despite current advances in the understanding of risk factors associated with gastric carcinoma, its etiology is not fully understood. It is commonly accepted that multiple stepwise molecular alterations contribute to gastric carcinogenesis [3]. Hence, it is important to identify molecular biomarkers that can be used as prognostic factors for effective management of gastric carcinoma patients.
Leptin is a hormone with a molecular mass of 16kDa. It is mainly produced by adipose tissue and secreted into plasma [4], and detected in variable tissues, including placenta, skeletal muscle, ovaries, mammary epithelial cells and the gastric mucosa [5]. Leptin was initially known to play a critical role in controlling food intake and energy balance [4,6]. Leptin functions via the leptin-receptor, a single-transmembrane-domain receptor of the class I cytokine-receptor family, localized to the cell membrane [7]. Recently, it was suggested that leptin and its receptor act as a potential tumorigenic factor by decreasing cell apoptosis, increasing cell proliferation, regulating cell differentiation and promoting angiogenesis in various carcinomas [8][9][10][11]. Its tumorigenic action is mediated via multiple signaling pathway activation, including JAK/STAT, Raf/MER/extracellular-regulated kinase (ERK), PI3K/Phosphatase and the tensin homolog/Akt/mTOR (PI3K/PTEN/Akt/mTOR) pathways [7,12,13]. Leptin expression can be up-regulated through hypoxia-inducible factor-1(HIF-1) alpha, which controls tumor angiogenesis in many solid tumors [8]. Furthermore, leptin-receptor is activated directly by HIF-1 alpha [14]. There have been controversial results with respect to its prognostic effects on gastric carcinomas, although HIF-1 alpha is generally thought to be involved in gastric carcinogenesis [15][16][17].
We evaluated expression of leptin-signaling-related proteins (leptin, leptin-receptor, pSTAT3, ERK, pAkt, mTOR and HIF-1 alpha) and the EBV infection status within cancer cells, and determined their clinicopathological significance in gastric carcinomas.

Materials and Methods Patients Ethics Statement
All human tissue specimens were obtained during diagnostic and therapeutic surgery. The participants did not provide the written consent to participate in the present study. The retrospective study was performed using the stored paraffin blocks containing tissue samples after the pathologic diagnosis, and all of the samples were anonymized before the study. This retrospective study was approved by the Institutional Review Board of Seoul National University Boramae Hospital under the condition of the anonymization (IRB No. 26-2014-13/022, 06-2011-40/106).

Tissue microarray
After histological review of all tumor sections, one of the representative portions was chosen from the individual donor block and core tissue with 2.0 mm diameter was punched-out by a trephine apparatus. All tissue cores were then inserted in a new recipient block that included 59 tissue cores and one ink core that was used as a direct marker. A total of six tissue array blocks were thus prepared.

Immunohistochemistry
Immunohistochemistry was performed using an automated immunostainer (Ventana Bench-Mark XT, Ventana Medical Systems Inc., Tucson, AZ, USA) according to the manufacturer's protocol. We performed immunohistochemical staining for leptin, leptin-receptor, phosphorylated STAT3 (pSTAT3), ERK, phosphorylated Akt (pAkt), mTOR, and HIF-1 alpha (Table 1). Four pathologists (EC, S-jB, SHK &, MSC,) reviewed the staining results. The expressions of leptin-signaling proteins were sufficiently homogenous in a tumor that one core tissue can be considered to be representative (S1 Fig Immunohistochemical staining for leptin and leptin-receptor was scored according to the intensity of positively stained tumor cells as follows: negative (score = 0), weak (score = 1), moderate (score = 2), and strong (score = 3). Scores of 1, 2 and 3 were then reclassified as positive.
In situ hybridization (ISH) for EBV-encoded small RNAs EBV-encoded small RNAs (EBER) were detected by in situ hybridization with the EBER detection kit according to the manufacturer's instructions. Next, the sections were deparaffinized, enzymatically digested with proteinase K, and hybridization solution was applied to the sections, warmed up and incubated. Hybridization detection was performed using a fluoresceinconjugated EBV oligonucleotide probe targeting EBER (Novocastra, Newcastle-upon-Tyne, UK). The slides were then counterstained with Nuclear Fast Red and coverslipped. Positive staining was identified as blue nuclear dots. Nuclear staining in tumor cells was considered positive, and the absence of nuclear staining was considered negative.

Statistical analysis
The relationship between the clinicopathological features and immunohistochemical staining results was examined by Pearson's χ 2 test and Spearman rank correlation coefficient. Patient survival data were obtained by the Kaplan-Meier method and log-rank test. Univariate and multivariate analysis (forward stepwise) were performed by the Cox proportional hazards model. P-values of < 0.05 were considered as statistically significant. All statistical calculations were carried out with SPSS/PC version 21.0 for Windows (IBM SPSS Inc., Chicago, IL, USA).

Cell culture and small interfering RNA (siRNA) transfection
Three gastric carcinoma cell lines, SNU719, SNU484 and SNU601, which were purchased from the Korean Cell Line Bank (Seoul, Korea), and cells were maintained at 37°C in RPMI 1640 (Gibco BRL, Rockville, MD, USA) supplemented with 10% fetal bovine serum, 2 mmol/L L-glutamine, and antibiotics (100 units/mL penicillin and streptomycin) in a humidified 5% CO 2 /95% air atmosphere. The leptin-receptor-specific siRNA (Ob-R siRNA (h) sc-36116, Santa Cruz, CA, USA) was transfected into gastric carcinoma cells. A scrambled siRNA (Dharmacon, Lafayette, CO, USA) containing a random sequence of nucleotides with no known specificity was used as a negative control. Cells were plated at 50% confluence in 6-well tissue culture plates. After 24 hours, siRNAs were transfected using Lipofectamine 2000 (Invitrogen, Carlsbad, CA, USA) at a final RNA concentration of 50 nM per well. Cells were harvested 72 hours after siRNA transfection.

Western blot analysis for apoptosis
Cellular extracts from cell lines were prepared by dissociation in lysis buffer, and protein was measured using a BCA protein assay kit. Proteins were resolved by SDS-PAGE using the laemmli buffering system with 15% polyacrylamide running gels and 5% stacking gels, and then transferred to reinforced PVDF membranes (Millipore, Bedford, MA, USA). After blocking with 3% skim milk for 1 hour, the membranes were incubated with primary antibodies against leptin-receptor (sc-8391, 1:500, Santa Cruz,) and caspase-3 (H-227) (sc-7148, 1:1,000, Santa Cruz) for 2 hours at room temperature. After washing, the blots were incubated for 45 minutes at room temperature with a horseradish peroxidase (HRP)-conjugated anti-mouse secondary antibody (Abcam, Cambridge, UK). After extensive washing, the antigen-antibody complexes were visualized by exposing the membranes to X-ray film (Fuji, Japan). An antibody against beta-actin antibody (AC -15, 1:10000, Abcam, Cambridge, UK) was used to verify equal loading and transfer of total proteins.

Associations between clinicopathological features and immunohistochemical findings
The overall features of all 343 patients were summarized in Table 2. Based on tumor invasion depth, there were 136 patients (40%) with early gastric carcinoma (EGC) (mucosa or submucosa invasion) and 207 patients (60%) with advanced gastric carcinoma (AGC) (proper muscle or deeper invasion). When grouped by cancer stage (tumor-lymph node-distant metastasis, TNM stage), 144 patients (42%) belonged to the early cancer stage (TNM stage I), and the remaining 199 patients (58%) to the advanced cancer stage (TNM stage II, III or IV).

Clinicopathological characteristics of EBV-positive gastric carcinomas
There were 32 cases of the total 343 gastric carcinomas (9%) in the EBV-positive group. The EBV-positive group revealed a male predominance, a predilection of tumor location in the upper 1/3 of the stomach, more frequent diffuse Lauren subtype and more advanced cancer stage (TNM stage) (P < 0.05, respectively) ( Tables 2 & 3). In terms of leptin signaling-related protein, the EBV-positive tumors have a more abundant expression of leptin-receptor and a lower expression of mTOR compared to non EBV-related tumors, but a putative correlation with EBV cannot be shown in this way (Table 3).

Different prognostic implication of leptin signaling-related proteins according to clinicopathological subsets of gastric carcinomas
Patient outcome (median follow-up period: 77 months) was mortality in 148 of the total 343 (43%) patients, which included 10 of the 32 EBV-positive group (31%) patients. pTNM stage was the absolute independent prognostic factor, on univariate and multivariate analyses in all 343 patients (P < 0.05) (Fig 2A). None of the examined protein markers (leptin, leptin-receptor, pSTAT3, ERK, pAkt, mTOR and HIF-1 alpha) alone, had a statistical significance on overall survival in all 343 patients on multivariate analysis.
We classified all 343 gastric carcinomas based on clinicopathological criteria, to identify the correlation between expression of the examined markers and patient survival according to subsets of gastric carcinomas. All patients were initially categorized into two groups based upon tumor invasion depth; early gastric cancer (EGC: tumor limited to the mucosal and submucosal invasion) and advanced gastric cancer (AGC). There were 136 cases of EGC and 207 cases of AGC. In 207 advanced gastric carcinomas, leptin-receptor expression was correlated with a poor survival of patients in univariate and multivariate analyses (P < 0.05) (Fig 2B). Furthermore, the EBV-positive AGC subgroup (n = 29) tended to have better patient survival than the EBV-negative AGC subgroup (n = 178), but there was no statistical significance (P = 0.06). All patients were divided into intestinal, diffuse or mixed subtype according to the Lauren classification based on microscopic histopathology. There were 139 (41%) cases of diffuse type, 160 (47%) cases of intestinal type, and 44 (13%) cases of mixed type. Leptin-receptor positivity in the Lauren diffuse type was correlated with a poor survival of patients in univariate and multivariate analysis (P < 0.05) (Fig 2C). Additionally, leptin-receptor expression was compatible with poor survival outcome in 160 (47%) patients with lymph node metastasis, on univariate and multivariate analysis (P < 0.05) (Fig 2D).
Expression of pAkt in 32 (9% = 32/343) cases of EBV-positive gastric carcinomas appeared to be correlated with better survival of patients on univariate analysis, but lost significance on multivariate analysis.

Effect by leptin-receptor inhibition in stomach cancer cells
In three gastric carcinoma cell lines, transfection of siRNA against leptin-receptor resulted in a remarkable increase of cleaved caspase 3 (p11, p17 and p20 subunits), comparing to each original cells and each scrambled siRNA-transfected cells (Fig 3). In other words, leptin-receptor is suggested to prevent apoptosis in gastric carcinoma cells.

Discussion
The present study highlighted that leptin-receptor expression can be an independent poor prognostic factor in specific subsets of gastric carcinomas. Expression of leptin-receptor was correlated with adverse clinicopathological parameters, including lymph node metastasis and Lauren diffuse subtype. Leptin-receptor positivity was related with poor survival of patients, especially in the AGC group, Lauren diffuse type group, and lymph node metastasis group, on univariate and multivariate analyses (P < 0.05). Our results were consistent with the report of Ishikawa et al that lymph node metastasis and distant metastasis were frequently observed in the leptin-receptor positive group among gastric carcinoma patients [27]. As one of explanations for that leptin-receptor expression group showed poor survival rate of patients, we suggest that leptin-receptor counteracts apoptosis in cancer cells, since leptin-receptor inhibition induced an increased apoptosis in the present study. Antiapoptotic effect of leptin-receptor has been described in other human cancers [28]. Leptin-receptor may thus play a pivotal role in the progression of gastric carcinoma. However, to the best of our knowledge, apart from the study of Ishikawa et al, there have been few studies on the prognostic significance of the leptinreceptor in gastric carcinoma. Therefore, further studies are required to clearly define the prognostic value of leptin-receptor expression in gastric carcinomas. The present study indicated that leptin signaling proteins may contribute to HIF-1 alphamediated angiogenesis in gastric carcinoma. HIF-1 alpha expression was positively correlated with the expression of pSTAT3, ERK, pAkt and mTOR (P < 0.05). This result was consistent with previous reports of a positive interrelationship between molecules involved in leptin-signaling and HIF-1 alpha in gastric carcinoma [29] and renal cell carcinoma [30].
The present study determined the clinicopathological characteristics of EBV-positive gastric carcinomas. We found that the EBV-infection rate in gastric carcinomas was 9%; EBV-  positive gastric carcinomas were more prevalent in male patients; tumor location in the upper 1/3 stomach; and the Lauren diffuse type. These results reiterated the findings of previous reports [18,22,[31][32][33][34]. EBV-infection may have only limited prognostic value, at most, only in the advanced gastric cancer group. There was a tendency that EBV-positive AGC subgroup had a better patient survival compared with EBV-negative AGC subgroup (P = 0.06). There are different opinions on EBV-infection as a prognostic factor in gastric carcinomas. Most authors suggest that EBV is not an independent prognostic factor [32][33][34]. On the other hand, some authors claim that EBV is a good prognostic factor for survival of patients with gastric carcinomas through meta-analysis [35]. Global methylation of genes is known as a basic and initial mechanism in EBV-associated carcinogenesis [23], and multistep alteration of genes occur subsequently. Therefore, EBV-infection itself may not act as an independent prognostic factor. While leptin-receptor expression was more frequently observed in EBV-positive subgroup than in EBV-negative subgroup, mTOR was less expressed in EBVpositive subgroup than in EBV-negative subgroup (P < 0.05, respectively). Moreover, leptin expression was quite less observed in EBV-positive group. Possibly, leptin signaling may not be associated with EBV-induced carcinogenesis. Surprisingly, expression of pAkt was related with a good survival of patients with 32 EBV-positive gastric carcinomas (P < 0.05). Previous reports have shown that expression of pAkt correlates with poor prognosis in conventional gastric carcinoma [36][37][38][39]. We are unaware of any study that analyzes the prognostic impact of pAkt limited to EBV-positive gastric carcinoma cases. The analysis of more EBV-positive gastric carcinoma cases is needed to clarify the association between pAkt expression and prognosis of EBV-positive gastric carcinoma. Besides, our research corroborated a previous study by Sukawa et al that showed a lack of statistical difference in pAkt expression between EBV-positive gastric carcinoma and EBV-negative gastric carcinoma.
In conclusion, leptin signaling-related proteins and EBV status show different significance on patient survival according to subsets of gastric carcinoma. Expression of leptin-receptor can be an independent poor prognostic indicator in the advanced gastric cancer group, Lauren diffuse group, and the lymph node metastasis group. EBV-positive may have a limited prognostic value only in the advanced gastric cancer group, and leptin signaling may not be involved in EBV-positive gastric carcinogenesis. Further studies are required to understand how leptin signaling-related proteins and the EBV status play a role in regulating carcinogenesis, and ultimately to establish them as therapeutic targets in gastric carcinomas.