CDK5RAP3 suppresses Wnt/β-catenin signaling by inhibiting AKT phosphorylation in gastric cancer

Human gastric tumor tissues with detailed clinic pathologic parameters were obtained from 295 patients at Fujian Medical University Union Hospital (Fujian, China). All patients underwent radical gastrectomy from 2010 to 2015. None of the patients underwent preoperative chemotherapy or radiotherapy. Postoperative adjuvant chemotherapy was performed with 5-fluorouracil-based drugs plus oxaliplatin in advanced cases. The pathologic stage of the tumor was reassessed according to the 2010 International Union Against Cancer (UICC) TNM classification of gastric cancer (seventh edition) [11]. Adjacent non-tumor tissues were obtained from an area at least 5 cm away from the gastric tumor. All fresh specimens were stored in liquid nitrogen after resection until protein or RNA extraction. The 209 paraffin-embedded gastric tumor and adjacent non-tumor tissues were collected for immunohistochemistry (IHC) from 2010 to 2014. The 86 fresh gastric tumor tissues and adjacent non-tumor tissues were subjected to western blotting between 2014 and 2015. This study was approved by the ethics committee of Fujian Medical University Union Hospital, and written consent was obtained from all involved patients.

All patients were systematically followed up by trained doctors following the institutional follow-up protocol via several approaches, including outpatient services, letters, telephone, mail or visits. Follow-up was conducted every 3 months during the first year and every 6 months after the second year, and all surviving patients were followed for more than three years. Survival time was defined as the time from the date of surgery until the last follow-up or the date of death. All 209 patients involved in the IHC analysis completed the follow-up.

A series of TMAs containing gastric cancer samples were constructed. Briefly, all the gastric cancer tissues were reviewed by a pathologist, and representative areas free from necrotic and hemorrhagic materials were premarked in the paraffin blocks. For each sample, a 1.5-mm core was punched from the donor blocks and transferred to the recipient paraffin block at defined array positions using a tissue microarray instrument. Several serial sections (4 μm in thickness) were cut from all TMAs, and one section from each TMA was stained with hematoxylin-eosin and served as a reference.

Paraffin blocks that contained sufficient formalin-fixed tumor specimens were serially sectioned at 4 μm and mounted on silane-coated slides for IHC analysis. The sections were deparaffinized with dimethylbenzene and rehydrated through an ethanol gradient, including 100, 95, 85 and 75% ethanol. Antigen retrieval was performed with 0.01 mol/L sodium citrate buffer (pH 6.0) in an autoclave at 121 °C for 2 min, and endogenous peroxidase was blocked by incubation with 3% hydrogen peroxide for 10 min at room temperature. The slides were then washed in phosphate-buffered saline (PBS), blocked with 10% goat serum (ZhongShan Biotechnology, China) for 30 min and incubated with antibodies in a humidified chamber at 4 °C overnight. Following three washes in PBS, the sections were incubated with HRP-conjugated secondary antibody for 30 min at room temperature. Next, the signal was developed with diaminobenzidine (DAB) solution, and all the slides were counterstained with 20% hematoxylin. Last, the slides were dehydrated and mounted with cover slips. For negative controls, the primary antibody diluent was used instead of the primary antibody. The staining intensity was scored as 0 to 3. The heterogeneity of staining was scored as 0 to 3, depending on the percentage of tumor cells that were positively stained (3). To obtain an IHC score that considers the IHC signal intensity and the frequency of positive cells, we generated a composite expression score (CES) ranging from 0 to 9. A CES of 0, 1, 2 and 3 was defined as low expression, but a CES of 4, 6 and 9 was defined as high expression.

The methods were described as previously described [3]. Anti-GSK-3β (#9832), AKT (pan) (#4685), phosphorylated AKT (Ser473) (#4060), GSK-3β (Ser9) (#12456) and phosphorylated GSK-3β (Ser9) (#5558) were purchased from Cell Signaling Technology. Anti-CDK5RAP3 (ab157203), E-cadherin (ab1416), N-cadherin (ab76057), Vimentin (ab8978), and GAPDH (ab181602) antibodies were purchased from Abcam. AKT siRNA (#6211) and control siRNA (#6568) were purchased from Cell Signaling Technology. We corrected the loading error based on loading controls and compared the expression levels of target proteins in gastric tumor and adjacent non-tumor tissues. The protein expression in tumors was defined as high level when it was higher than that in normal tissue but was defined as low level when it was lower than that in normal tissue.

Human gastric cancer cell lines AGS and HGC-27 were obtained from the Cell Line Bank at the Chinese Academy of Sciences. All the cell lines were confirmed to be free of mycoplasma contamination by PCR and culture. The species origin was confirmed with PCR. The identity of the cell line was authenticated with short tandem repeat (STR) profiling. These cell lines were cultured in RPMI 1640 (Gibco, Grand Island, NY) supplemented with 10% fetal bovine serum (FBS) (Gibco, Grand Island, NY) and incubated at 37 °C in a humidified atmosphere containing 5% CO₂.

The cell lines were established as previously described [3].

Cell proliferation was assessed using the CCK8 assay and colony formation assay. Details of cell proliferation are described in the Additional file 1.

Cell migration and invasion assays were assessed in Transwell chambers (polycarbonate filters of 8-μm porosity, BD Bioscience) and by the wound healing assay. The details are described in the Additional file 1.

All statistical analyses were performed using SPSS 18.0 for Windows (SPSS, Chicago, IL) and Prism 5.0 software (GraphPad). The chi-squared test was used to evaluate the difference in proportions, and Student’s t-tests were used to evaluate continuous variables. Overall survival curves were calculated with the Kaplan-Meier method; the log-rank test was used to detect differences between groups. Independent prognostic factors were identified with a Cox proportional hazards regression model. P < 0.05 was considered statistically significant, and all P values were two-sided.

Our previous study had confirmed that CKD5RAP3 suppressed tumor cells through the inhibition of β-catenin signaling by regulating GSK-3β phosphorylation. To study the interactions among AKT, CDK5RAP3 and GSK-3β, CDK5RAP3 was stably overexpressed or knocked down in AGS cells. CDK5RAP3 expression was confirmed using Western Blotting, and the effect of AKT on CDK5RAP3 function in gastric cancer cells was assessed using AKT siRNA. In CDK5RAP3-overexpressing AGS cells, AKT siRNA reduced AKT levels and also reduced phosphorylated GSK-3β (Ser9) and β-catenin levels, even in the presence of Lenti-shC53 (Fig. 1a). However, PKA and PKC silencing did not affect GSK-3β phosphorylation (Ser9) (Fig. 1b-c), which meant that CDK5RAP3 controlled GSK-3β phosphorylation without phosphorylating PKA and PKC. In addition, knockdown of CDK5RAP3 in gastric cancer cells led to upregulated AKT phosphorylation at Ser473, GSK-3β phosphorylation at Ser9 and β-catenin expression, whereas overexpression of CDK5RAP3 led to the opposite effects (Fig. 1d), indicating that CDK5RAP3 may regulate β-catenin activity by suppressing AKT phosphorylation at Ser473.

Inhibition of AKT in CDK5RAP3 stably knockdown AGS and HGC-27 cells abrogated accelerated proliferation (Fig. 2a-b, d-e), colony formation (Fig. 2c, f), and cell migration and invasion (Fig. 3a-b). We also investigated the role of AKT in the CDK5RAP3-induced epithelial–mesenchymal transition (EMT) process. When AKT was knocked down, the protein levels of EMT markers (E-cadherin, N-cadherin and Vimentin) were not significantly changed whether CDK5RAP3 was knocked down or overexpressed (Fig. 3c). These results indicate that the regulation of CDK5RAP3 on EMT depends on the activity of AKT in gastric cancer.

CDK5RAP3, p-AKT (Ser473) and p-GSK-3β (Ser9) protein levels were detected in gastric tumor tissues and adjacent non-tumor tissues from an additional 86 patients using Western Blotting. The 24 representative samples were shown in Fig. 4a. In cancer samples with low CDK5RAP3 protein levels, the proportion of samples exhibiting p-AKT (Ser473) expression was significantly higher than that exhibiting high CDK5RAP3 expression (77.0% vs. 24.0%) (Fig. 4b). In addition, in cancer samples with low CDK5RAP3 protein levels, the proportion of samples exhibiting p-GSK-3β (Ser9) expression was also significantly higher than that exhibiting high CDK5RAP3 expression (64.0% vs. 32.0%) (Fig. 4c).

To further verify the associations among CDK5RAP3, p-GSK-3β (Ser9) and p-AKT (Ser473), we detected the expression of CDK5RAP3, p-GSK-3β (Ser9) and p-AKT (Ser473) in the 209 gastric cancer samples using IHC. In cancer samples with low CDK5RAP3 expression, the staining of phosphorylated AKT (S473) was significantly darker than in those with high CDK5RAP3 expression; this was also observed regarding phosphorylated GSK-3β (Ser9) (Fig. 4d). In low CDK5RAP3 IHC scoring samples, phosphorylated AKT (S473) protein expression was significantly higher than that in high CDK5RAP3 IHC scoring samples (65.2% vs. 39.2%) (Fig. 4e). The trend of phosphorylated GSK-3β (Ser9) was the same (57.6% vs. 39.2%) (Fig. 4f). These results suggest that CDK5RAP3 regulates tumor cells by regulating AKT phosphorylation and GSK-3β phosphorylation.

To evaluate the prognostic value of combined CDK5RAP3 and p-AKT (Ser473) expression, the Cox proportional hazards regression model was used (Table 1). Univariate analysis revealed that tumor size, histology, depth of invasion, lymph node metastasis, distant metastasis and CDK5RAP3 and p-AKT (Ser473) expression were associated with overall survival. Multivariate Cox regression analyses showed that T, N and M stages as well as CDK5RAP3 and p-AKT (Ser473) expression remained independent prognostic factors.

Next, we compared the overall survival of patients categorized by CDK5RAP3 and p-AKT (Ser473) expression. The overall survival rate of patients with low CDK5RAP3 expression was significantly lower than that of patients with high CDK5RAP3 expression (Fig. 5a). When p-AKT (S473) expression was low, a significant difference in overall survival was not observed between patients with low or high CDK5RAP3 expression (Fig. 5b). However, when p-AKT (Ser473) expression was high, gastric cancer patients with low CDK5RAP3 expression had a poorer prognosis than those with high CDK5RAP3 expression (P < 0.05, Fig. 5c). On the other hand, CDK5RAP3 expression did not affect the prognostic role of p-AKT (S473) expression (Fig. 5e-g). These results confirm that the clinical value of CDK5RAP3 is dependent on AKT activity.