Glycogen Synthase Kinase-3β Is Associated with the Prognosis of Hepatocellular Carcinoma and May Mediate the Influence of Type 2 Diabetes Mellitus on Hepatocellular Carcinoma

Background Although many studies have shown glycogen synthase kinase-3β(GSK-3β) was associated with type 2 diabetes mellitus(T2DM) and implicated with a wide range of cancers, the role of GSK-3β in hepatocellular carcinoma(HCC) and the correlation among GSK-3β, T2DM and HCC remains unclear. Our objectives were to identify the effect of p-Ser9-GSK-3β on the prognosis of patients with HCC and to learn more about the interaction among T2DM, GSK-3β and the prognosis of HCC. Methods Firstly we used reverse transcriptase-PCR(RT-PCR) and western blotting to determine the expression levels of GSK-3β and p-Ser9-GSK-3β in human HCC samples. We then used immunohistochemical staining to evaluate the expression pattern of p-Ser9-GSK-3β in 178 patients with HCC after curative partial hepatectomy. Finally we statistically analyzed the association of p-Ser9-GSK-3β and T2DM with the prognosis of patients with HCC. Results P-Ser9-GSK-3β was over-expressed in tumor tissues compared with their normal counterparts. Correlation and regression analysis indicated that the over-expression of p-Ser9-GSK-3β was significantly associated with T2DM, and the correlation coefficient was 0.259 (P = 0.001). Multivariate analysis showed that the over-expression of p-Ser9-GSK-3β(P<0.001) and T2DM(P = 0.008) were independently associated with poor prognosis of HCC, respectively. Further analysis demonstrated that these two variables are closely related with each other. Conclusion The over-expression of p-Ser9-GSK-3β and T2DM are strongly correlated with worse surgical outcome of HCC. P-Ser9-GSK-3β may play a significant role in mediating the influence of T2DM on the prognosis of HCC.


Introduction
Hepatocellular carcinoma(HCC) is the fifth most common malignancy worldwide and the second most frequent cause of cancer-related mortality [1]. Partial hepatectomy remains the standard curative treatment, with a 5-year survival rate of 50%-70% for early HCC [2,3]. Accurately identifying the prognostic factors is important to facilitate screening of high risk patients and make decisions on the adjuvant therapy.
Glycogen synthase kinase 3 (GSK3), a constitutive multifunctional serine threonine kinase, was involved in diversely physiological pathways ranging from insulin signaling pathways which associated with the disease of type 2 diabetes mellitus (T2DM) and Wnt signaling pathways which related to oncogenesis [4][5][6][7]. GSK-3b (47 kDa), an isoform of GSK-3, could be inactivated by insulin through phosphorylation of an N-terminal domain serine residue(Ser9 of GSK-3b) [8]. Phosphorylation of Ser9-GSK-3b (p-Ser9-GSK-3b) was associated with the activation of Wnt signaling which could prevent phosphorylation and lead to the stabilization of b-catenin. This mechanism had been reported in a wide range of cancers [9]. However, few researchers reported the significance of GSK-3b or p-Ser9-GSK-3b in the prognostic evaluation of patients with HCC [10].
Currently, epidemiology studies revealed that diabetes mellitus(DM) increased the incidence of HCC [11,12]. Although it remained controversial about the influence of DM on the prognosis of patients with HCC after partial hepatectomy [13,14], accumulating evidence showed that DM was an independent prognostic indicator for patients with HCC after curative resection [15,16]. The purpose of the present study is to identify if the expression level of p-Ser9-GSK-3b is an effective prognostic factor for HCC patients and learn more about the interaction among the T2DM, expression of p-Ser9-GSK-3b and the prognosis of HCC.  [17]. Of these surgical HCC specimens, 60 samples which were collected between July 1 st and December 15 th , 2003, were subjected to RNA and protein extraction that could be done for RT-PCR and western blot analysis. 178 patients collected between December 16 th , 2003 and May 15 th , 2005 met the following inclusion criteria and then underwent immunohistochemical staining (IHC) and prognostic analyses: had WHO performance status 0-1; did not receive any preoperative anti-cancer therapy; did not have any distant metastasis, image visible ascites, or severe esophageal varices; had the preoperative Child-Pugh class A; received curative hepatectomy; had pathologically proven hepatocellular carcinoma. Patients were excluded from this study if they had a history of other solid tumors, or died from severe postoperative complications. The clinical characteristics of the 178 patients were listed in Table 1.
The definition of curative hepatectomy was described as our previous reports [18]. The histological grade of tumor differentiation was assigned by the Edmondson Steiner grading system [19]. This study was approved by the Institutional Review Board of Eastern Hepatobiliary Surgery Hospital. All patients gave written informed consent to participate. The parents wrote the informed consent on behalf of the patients whose age,18. The ethics committees approved this consent procedure. The data did not contain any information that could identify the patients.

Immunohistochemistry and evaluation of immunostaining
Immunohistochemical staining was performed with the Dako Envision Plus System (Dako, Carpinteria, CA) according to the manufacturer's instructions. The primary antibodies were anti-GSK-3b and anti-p-Ser9-GSK-3b (Cell Signaling Technology Inc.,Beverly, MA, 1:50). The tissues were evaluated as positive for GSK-3b and p-Ser9-GSK-3b staining when there were more than 10% of tumor cells demonstrating cytoplasmic and/or nucleus immunoreaction deposits. The sections were scored with a fourtier scale: 0 = negative (0-10%), 1 = weak signal (10-20%), 2 = intermediate signal (20-50%) and 3 = strong signal (. 50%). 0 and 1 were defined as low exprression, while 2 and 3 were defined as over-expression. All sections were scored independently by two observers who did not have any prior knowledge of the clinic-pathologic data. The concordance between scores from different sections of the same tumor was greater than 90%. All discrepancies in scoring were reviewed and a consensus was reached.

Western blotting analysis
Fresh surgical specimens were snap frozen in liquid nitrogen and stored in deep freezer. The normal and the tumor tissues were lysed in T-PER Tissue Protein Extraction Reagent(Pierce, Rockford, IL) containing proteinase inhibitors(CalBiochem, San Diego, CA). The extractions were collected and centrifuged at 12,0006g for 5 min. The protein concentrations were determined using the BCA Protein Assay (Pierce) according to the manufacturer's instructions. The following antibodies were used: anti-GSK-3b and anti-p-Ser9-GSK-3b (Cell Signaling Technology Inc., Beverly, MA), we also used b -actin as a loading control.

RNA extraction and real-time reverse transcriptase-PCR
Total RNA of tissue samples were isolated using TRIzol (Life Technologies, Inc., Rockville, MD) according to the manufacturer's instructions. cDNA was generated from 1 ug of each RNA sample and a reverse transcribed using a transcription kit (Takara, Kyoto, Japan). Real-time quantitative reverse transcriptase-PCR (RT-PCR) was done in the 7300 Real Time PCR System (Applied Biosystems) to detect GSK-3b gene expression in paired liver samples from 30 HCC patients. The primers were as follows:

Diagnosis of Type 2 diabetes
Diagnosis of diabetes mellitus was based on the 1999 World Health Organization criteria. Patients who were found to have a fasting blood sugar level between 5.6 and 6.1 mmol/L were defined as having impaired fasting glucose and were referred for confirmation of their diabetes status. Those with a fasting blood sugar level of at least 7.0 mmol/L or those diagnosed with Type 2 diabetes before entering the hospital were defined as having Type 2 diabetes and were referred for further diabetic care. Patients with prior diagnosis of Type 1 diabetes were excluded from the study.

Follow-up
The follow-up ended on May 2012 and the median follow-up duration was 48.2 months (3.8-114.5 months). Contrast-enhanced CT scan or MRI of the abdomen was carried out once every 3 months in the first two years after surgery, and then once every 6 months thereafter. The diagnostic criteria for HCC recurrence was the same as used for the initial diagnosis.

Statistical analysis
Overall Survival (OS) and Time To Recurrence (TTR) were used as primary endpoints. OS was defined as the interval between hepatectomy and death or the date of last follow-up. TTR was calculated from surgery to the date when recurrence was diagnosed. All statistical calculations were carried out using the SPSS. 16.0 software. The x2 test or Fisher's exact test were used to compare qualitative variables, while continuous variables were compared using Student's t-test or Mann-Whitney test for variables with an abnormal distribution. Receiver operating characteristic curve analysis was used to determine the optimal cut-offs of continuous variables. Survival curves were calculated by the Kaplan-Meier method and compared using the log-rank test. The Cox proportional hazards model was used to determine the independent factors for tumor recurrence and patient's survival based on the variables selected by univariate analysis. Differences were considered statically significant at p, 0.05.

Association of p-Ser9-GSK-3b expression with T2DM
We examined the association between the expression level of p-Ser9-GSK-3b in the tumor tissues with T2DM and other clinicpathological characteristics of patients (Table 1). Correlation and regression analysis indicated that the over-expression of p-Ser9-GSK-3b was significantly associated with T2DM, and the correlation coefficient was 0.259 (P = 0.001).
These significant variables were further subjected to multivariate Cox proportional hazards model, which shown that tumor size(P = 0.008), tumor number(P = 0.032), microvascular invasion(P = 0.019), p-Ser9-GSK-3b expression level(P,0.001) were independent risk factors that could affect the survival of patients with HCC(table 2). We speculated that some interaction or colinearity existed between the variables of p-Ser9-GSK-3b and T2DM. Thus, we further conducted multivariate Cox proportional hazards analysis excluding the variables of p-Ser9-GSK-3b or T2DM respectively (table 2). The results also demonstrated that these two variables were closely related with each other.

Discussion
In the present study, we showed that the total GSK-3b and the p-Ser9-GSK-3b were increased at both mRNA and protein levels in HCC specimens. The over-expression of p-Ser9-GSK-3b was associated with the poor prognosis of patients with HCC after curative resection. We also found that T2DM was related with the poor prognosis of patients with HCC, and the expression level of p-Ser9-GSK-3b was significantly related with T2DM. We speculated that the regulation of GSK-3b might play a significant role in mediating the impact of T2DM on HCC prognosis. To the best of our knowledge, no previous study reported the impact of p-Ser9-GSK-3b expression on the prognosis of HCC and the relationship between the expression of p-Ser9-GSK-3b and T2DM.
GSK-3b was a crucial signaling molecule involved in glycogen synthesis, gene expression, cell migration, cell cycle, cellular architectural pathways and oncogenesis [20,21]. It was a downstream target of insulin and insulin-like growth factor(IGF) receptor-dependent pathways which could result in GSK-3b phosphorylation on the serine 9 residue(p-Ser9-GSK-3b), and subsequent phosphorylation of insulin receptor substrate-1(IRS1) [22,23]. Insulin receptor substrate (IRS) proteins were major docking proteins that mediated insulin/insulin-like growth factor 1 (IGF1) signaling in various insulin-sensitive tissues [24]. Therefore, dysregulation of GSK-3b could result in attenuating insulin signaling, which was thought to induce insulin resistance. Insulin resistance was a major pathophysiological problem underlying the initiation and progression of type II diabetes mellitus (T2DM) [25]. Moreover, Activation of Wnt signaling could lead to GSK-3b inactivation and prevented phosphorylation of b-catenin protein in the cytosol. The increased levels of b-catenin transported to the nuclear, and interacted with transcription factors LEF/TCF (lymphoid enhancer factor/T cell factor) to activate the expression of target genes like c-myc and cyclin D1, and lead to cell proliferation [21,26]. This process was implicated with a wide range of cancers. With respect to HCC, apart from the Wnt-bcatenin pathway, insulin and insulin-like growth factor(IGF) pathways were also implicated with the development and/or progression of HCC [27][28][29]. GSK-3b was involved in these pathways and recent findings indicated that hyperphosphorylation of Ser9-GSK-3b was a hallmark of hepatocarcinogenesis. And also, GSK-3b was showed as a potential therapeutic target in various cancer involved in pancreatic, colon, renal, prostate, thyroid, ovarian cancer and leukemia [30][31][32][33][34][35][36][37].
Increasing evidence showed that DM was an independent risk factor for the prognosis of patients with HCC [15,38]. We focused on T2DM because type 1 and type 2 DM might have different impacts on HCC prognosis and their pathophysiology mechanisms were different. In this study, our findings showed that T2DM was associated with poor survival of patients with HCC. In the subgroup analysis, T2DM was a significant independent risk factor in early HCC (tumor size, or = 5cm, P = 0.026 for recurrence, P = 0.028 for survival). While in the group of HCC with tumor size.5cm, T2DM showed no significant difference (P = 0.094 for recurrence, P = 0.125 for survival). The mechanism underlying the impact of DM on HCC prognosis was still unclear. Insulin resistance and hyperinsulinemia could result in increasing of IGF-1. Both insulin and IGF-1 promoted proliferation and inhibited apoptosis by receptor-mediated pathways, which may lead to carcinogenesis [39].
This study focused on the impact of p-Ser9-GSK-3b on the prognosis of patients with HCC and the interaction among the expression level of p-Ser9-GSK-3b, T2DM and the prognosis of HCC. Further studies are needed to confirm the interconnection or explain the complex mechanisms among them. We have began a new study aimed at finding this potential mechanism by detecting the expression levels of other related molecules in multiple pathways involved in IGF, NF-kB, PI 3-kinase/Akt and Wnt/b-catenin pathways [40,41].
In conclusion, we firstly reported p-Ser9-GSK-3b was an independent prognostic factor of patients with HCC and it might mediate the influence of type 2 diabetes mellitus on the prognosis of patients with HCC.