Identification and validation of PGLS as a metabolic target for early screening and prognostic monitoring of gastric cancer

Abstract Background Gastric cancer is the third leading cause of cancer‐related death in the world. The purpose of the present study is to investigate the expression and prognostic significance of 6‐phosphogluconolactonase (PGLS) in gastric cancer. Methods The protein extracted from a panel of four pairs of gastric cancer tissues and adjacent tissues, labeled with iTRAQ (8‐plex) reagents, and followed by LC‐ESI‐MS/MS. The expressions of proteins were further validated by immunohistochemistry analysis. The expression levels of mRNA were analyzed and validated in the Oncomine database. The correlations of PGLS with prognostic outcomes were evaluated with Kaplan‐Meier plotter database. Results The present study found that PGLS was significantly up‐regulated in gastric cancer by using iTRAQ‐based proteomics and immunohistochemistry analysis. The sensitivity of PGLS in gastric cancer was 72.9%. The high expression of PGLS was significantly correlated with TNM staging in gastric cancer (p = 0.02). The overexpression of PGLS predicts worse overall survival (OS) and post‐progression survival (PPS) for gastric cancer (OS, HR = 1.48, p = 2.1e‐05; PPS, HR = 1.35, p = 0.015). Specifically, the high PGLS expression predicts poor OS, PPS in male gastric cancer patients, in patients with lymph node metastasis and in patients with Her‐2 (‐). Conclusions These findings suggested that PGLS was aberrantly expressed in gastric cancer and predicts poor overall survival, post‐progression survival for gastric cancer patients. The present study collectively supported that PGLS is an important target for early determining and follow‐up monitoring for gastric cancer.

Although the advancement of diagnostic and therapeutic methods has been made in recent years, a number of patients remain have poor prognosis partially because it was almost advanced stage when patients were diagnosed with gastric cancer. 2,3 Previous studies have demonstrated that gastric cancer patients diagnosed in early stage have a survival rate of up to 61%. However, if they were diagnosed in advanced stage, they only have a 5-year survival rate of 24%. 4 In order to screen early-stage gastric cancer patients, several methods have been widely used in clinical practice, such as H. pylori infection testing, the serum pepsinogen test, and the gastrin 17 test. In addition, serum CA-199 and carcinoembryonic antigen (CEA) were widely used for screening early gastric cancer. However, these methods and biomarkers yield low sensitivity and specificity. 5 Therefore, the identification of novel diagnostic biomarkers which can sensitively diagnose primary tumor and metastatic cancer is urgently needed.
Cancer metabolism is a complex process where the cancer cells can acquire specific traits that enable them to survive from extremely microenvironments. A number of metabolic enzymes which were found to be aberrantly expressed in cancer cells have profound impact on tumor progression and metastasis. 6,7 The reaction catalyzed by these metabolic enzymes is closely involved in tumor oxidation reduction and microenvironments which promote the cancer cell proliferation, survival, angiogenesis, and evasion of host immune response, etc. 8,9 Hence, studying the key metabolic enzymes expressed by the tumor may yield a new range of potential biomarkers and therapeutic targets.
Proteomics approaches are powerful tools for identifying biomarkers in tissue specimens of malignant tumors. Traditional 2-DE-based proteomic yields low sensitivity and specificity outcome. Isobaric tags for relative and absolute quantification (iTRAQ) technology and LC-ESI-MS/MS were currently the most widely used methods for high-throughput protein quantification. [10][11][12] In this study, we performed iTRAQ-based LC-ESI-MS/MS to analyze the gastric cancer tissues proteome compared with adjacent cancer tissues. We aim to identify a range of novel metabolic signatures in gastric cancer and to evaluate the specificity and sensitivity of these candidates for gastric cancer diagnosis.

| Samples
In the present study, 70 gastric cancer tissues, surrounding adjacent gastric tissues and 25 benign lesions were selected from the department of gastroenterology surgery of affiliated Hospital of North Sichuan Medical College. All of the patients did not undergo radiotherapy or chemotherapy prior to the surgery. The adjacent cancer tissues were obtained at least 5 cm away from the tumor center and pathologically confirmed as normal gastric mucosa. The partial tissues were fixed with 4% formaldehyde and the rest was stored in liquid nitrogen for the following protein extraction and proteomics analysis. All of the experiments were approved by the Ethics Committee of North Sichuan Medical College.

| Protein extraction
Thawed gastric cancer tissues (150 mg) were cut into pieces with scissors. RIPA lysis buffer and 10μl PMSF (Thermo Fisher Scientific) were added to the tissues. The sample was placed on ice and the suspension was treated by ultrasound (80 W, 15 s, 20 times), followed by centrifugation at 15,180 g for 15 min. The suspension was filtered by 0.22μm filter membrane and the filtrate was collected. The total protein concentration was determined using a Bradford protein assay kit (Bio-Rad). The total protein samples were stored at −80°C.
the eluted peptides were pooled into 20 fractions. The peptides were processed to desalt and then evaporated to dryness using a SpeedVac.

| LC-ESI-MS/MS analysis
The samples were resuspended in buffer A (2% CAN, 0.1% FA) and make the final concentration of peptide 0.25 μg/μl. With the LTQ Orbitraq Velos (Thermo) system, the sample volume was 10 μl per injection. Use a blank to clear the system after each sample. A data-dependent procedure was applied to the MS scanning. The threshold ion count is 5,000.
The mass spectrometer m/z scan range was 350-2,000 Da.

| Database searches and bioinformatics
The identification of protein and relative iTRAQ quantification were performed with Mascot software (Version 2.2). Each MS/MS spectrum was searched against the International Protein Index (IPI) human database. The search parameters considered cysteine modification and biological modification. All of the proteins were grouped to minimize redundant. To estimate the false discovery rate (FDR), a decoy database search strategy was adopted for peptide identification. Correspondingly, a randomized database was generated. The data obtained above were then exported into Excel for manual data interpretation. A 1.2-fold change threshold for all iTRAQ ratios was adopted to identify differentially expressed proteins between gastric cancer and adjacent tissue. The Blast 2 GO software and Kyoto Encyclopedia of Genes and Genomes (KEGG) database were used to perform ontology analysis and identify the tumor-associated pathways with the differently expressed proteins.

| Immunohistochemistry (IHC)
The paraffin-embedded gastric cancer tissues were cut into 4μm
The mRNA expression level of PGLS gene in gastric cancer and normal tissues was analyzed according to the following parameters, data type of mRNA, gene rank of all, fold change of 1.5, and p value of ≦ 0.05. The study of D'Errico was selected to analyze the expression of PGLS genes in gastric cancer.

| Statistical analysis
The fold changes, ranks, and P values were analyzed and displayed on Oncomine database analysis. The survival curves were analyzed by Kaplan-Meier plotter database, with the HR, 95% CI and p values were also calculated. The correlations of PGLS expression with clinical parameters were calculated with Spearman's correlation and statistics significance. Chi-square test was used to analyze enumeration data.
Statistical analysis was performed with SPSS 22.0 (SPSS) for windows.

| PGLS was found to be up-regulated in gastric cancer
The present study identified 431 proteins which were aberrantly expressed in gastric cancer including 224 proteins and 207 proteins that were increased and decreased expressed in gastric cancer tissues, respectively. 13 Compared with the surrounding normal gastric tissues, PGLS was up-regulated with 1.379-fold changes in gastric F I G U R E 1 PGLS was increased expressed in gastric cancer compared with that in gastric mucosa with fold change of 1.523 from Oncomine database. p = 0.013. (1, Gastric mucosa; 2, Gastric mixed adenocarcinoma;) cancer tissues. Moreover, from Oncomine database analysis, PGLS was overexpressed in gastric cancer compared with that in normal gastric mucosa with 1.523-fold change (p = 0.013) (Figure 1).

| Immunohistochemistry analysis of PGLS expression in gastric cancer
Immunohistochemistry was performed to determine PGLS expression in gastric cancer tissues, adjacent tissues, and in benign lesions tissues. The results of immunohistochemistry showed that PGLS was strongly expressed in gastric cancer tissues compared with those in adjacent tissues ( Figure 2 Table 2 shows the different pathological background parameters

| Clinicopathologic parameters correlated with PGLS in gastric tissues by IHC study
To further analyze whether the PGLS expression correlated with clinicopathological variables, we divided the subjects into several groups according to their clinicopathological variables. Chi-square test was used to check the difference between subjects with PGLSpositive expression and those without PGLS expression, with the significant level of p ≤ 0.05. Table 3 shows the correlation of PGLS expression with various clinicopathological characteristics in gastric tissues. The results showed that the PGLS expression was significantly differed between gastric cancer patients with TNM Ⅰ staging and those with TNM Ⅱ, Ⅲ , and Ⅳ staging (p = 0.02).

| The PGLS expression and clinic prognosis of gastric cancer
The potential prognostic values of PGLS in gastric cancer were analyzed through Kaplan-Meier plotter database. Figure 3A

| The PGLS high expression predicts poor OS in patients with Her-2 (-)
In patients with Her-2(-), the high PGLS expression showed shorter

| DISCUSS ION
In the present iTRAQ work scheme, a total of 431 proteins were identified in gastric cancer tissues. Bioinformatics analysis showed that those aberrantly expressed proteins in gastric cancer tissues were related to cell proliferation, differentiation, cellular movement, and cell death. Several proteins have been used earlier to predict cancer development and metastasis. 14    to form [1-13 C] 6PG, they could significantly differentiate the glioblastoma tumor tissues from normal brain tissues in vivo. 20 To explore the site-based differentiate expressed proteins in PPP pathway, Cha et al found that 6PGL (PGLS) was positively expressed in bone metastasis, with a shorter overall survival rates for breast cancer patients. 21 Further, to explore potential cellular targets for mycoepoxydiene in cervical cancer, Jin found that PGLS was significantly down-regulated after treatment of mycoepoxydiene, suggesting that PGLS is a potential molecular target of mycoepoxydiene for treatment of cervical cancer. 22 In the present study, we found that PGLS expression was significantly higher in gastric cancer tissues than that of PGLS in adjusted cancer tissues. Further analysis found that the PGLS positivity in cancer tissues with T3 and T4 staging was significantly higher than that of patients with T1 and T2 staging, which indicated that the PGLS expression has strong impact on tumor stating. Moreover, we found that PGLS expression was strongly correlated with TNM staging Ⅱ-Ⅳ, suggesting the poor prognosis of gastric cancer. Consistent with Ou's study, by using the methods of integrated proteomics, they found that 6PGL was strongly expressed in breast cancer tissues, while it was lowly expressed in all the normal breast tissues. 23 Both in normal cells and tumor cells, PPP is the basic metabolic pathway. It produces ribulose-5-phosphates for nucleotide biosynthesis and reducing power in the form of NADPH that is needed for macromolecular biosynthesis and redox maintenance. [26][27][28][29][30][31] Many studies have targeted on glucose-6-phosphate dehydrogenase and 6-phosphogluconate dehydrogenase for therapeutic treatments.
However, the function of PGLS was not well studied partially due to the unstable property of 6-phosphoδ-gluconolactone and rapid hy- provides new important evidence of PGLS in cancer development. 24 Choi et al. 25 detected the proteins of pentose phosphate pathway in breast cancer. They found that G6PDH, PGLS, and 6PGDH were significantly increased in breast cancer. Notably, the expression of PGLS was higher in Her-2 (-) breast cancer. They also found that PGLS was significantly associated with tumor staging, suggesting that PGLS may predict poor prognosis of gastric cancer.
Consistent with Choi's study, we found that PGLS expression was significantly correlated with OS and PPS in gastric cancer patients. The mOS of patients with PGLS high expression was significantly

months shorter than those with low expression in patients with
Her-2 (-). These results suggested the converse correlations of PGLS expression with Her-2 expression in gastric cancer.
In the present study, with the methods of quantitative proteomics analysis, we found that PGLS was significantly overexpressed in patients with gastric cancer. The PGLS positivity was strongly associated with TNM staging and predicts poor prognosis of gastric cancer. The study provided new evidence that the PGLS might be a potential diagnostic and therapeutic target for gastric cancer.

CO N FLI C T O F I NTE R E S T
The authors declare that they have no conflict of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
Data are available in the article.