Background
Gastric cancer (GC) is the fourth most fatal malignant cancer worldwide and the second leading cause of cancer-related occurrence and death in China, which has had a severe impact on national health [
1,
2]. Gastric carcinogenesis/development is a complicated multi-stage process, and during this process, the accumulation of genetic alterations plays an important role, such as the activation of an oncogene or the deletion of a tumor suppressor gene [
3].
Phosphatase and tensin homolog (
PTEN) is an important tumor suppressor gene that was discovered in 1997 [
4]. Its encoded protein has activities of both a protein phosphatase and a lipid phosphatase and can regulate a complex network system dependent on phosphatase or non-phosphatase activity to affect biological functions [
5‐
7]. The C124S mutation of PTEN can completely ablate its protein phosphatase and lipid phosphatase activity but does not affect its non-phosphatase function. The G129E mutation was discovered from a family with Cowden disease, and this mutation only abrogates its lipid phosphatase activity while maintaining its protein phosphatase activity [
8,
9]. However, the substitution effect of protein phosphatase activity after the loss of lipid phosphatase activity remains unclear. Therefore, we investigated the effects of the loss of PTEN phosphatase function and the loss of its lipid phosphatase activity on GC.
The Hippo pathway is a highly conserved growth control signaling pathway identified in Drosophila in 1995 [
10]. In recent years, the inactivation or abnormal regulation of Hippo has been found in lung adenocarcinoma, esophageal squamous cell carcinoma, liver cancer, ovarian serous cystadenoma, and bladder cancer [
11‐
15]. The mammalian Hippo pathway is composed of three parts: multiple upstream signal input factors; the core kinase cascade reaction chain, including mammalian STE20-like protein kinase 1/2 (MST1/2), large tumor suppressor 1/2 (LATS1/2), human salvador homolog 1 (SAV1) and MOB kinase activator 1 (MOB1); and downstream transcription coactivators, including Yes-associated protein (YAP) and transcriptional coactivator with PDZ binding motif (TAZ) [
16]. YAP and TAZ are the key effectors downstream of the pathway. YAP is localized at human chromosome 11q22 and can promote gene expression by enhancing transcription factor activity [
17]. TAZ, located at chromosome 3q25, is homologous to YAP, and these proteins have similar functions [
18]. Once Hippo has been activated by upstream signal molecules, MST1/2 binds to SAV1 and thus activates LATS1/2, which leads to YAP/TAZ phosphorylation and its stagnation in cytoplasm. If the pathway is blocked or inactivated, the nonphosphorylated YAP/TAZ will translocate into the nucleus. Because YAP/TAZ do not have DNA-binding sequences, they mainly combine with transcription factors in the nucleus, such as TEA domain family members (TEADs), to regulate the expression of downstream genes, inducing those related to cell transformation, proliferation, invasion and metastasis [
19]. However, the exact regulatory mechanism of Hippo has not been fully explored and is still unclear.
At present, the most in-depth study of PTEN regards its negative regulation of the PI3K/Akt pathway in a phosphatase-dependent manner [
20]. However, recent studies have found that PTEN can play a role that is independent of the PI3K/Akt pathway; therefore, the identification of other possible molecular mechanisms involved in PTEN function has important clinical significance [
8,
21,
22]. Furthermore, the Hippo and PI3K/Akt signaling pathways have a joint effect in regulating cell proliferation and apoptosis [
23]. How does PTEN phosphatase inactivation lead to the occurrence and development of GC? Are the Hippo and PI3K/Akt signaling pathways involved in PTEN phosphatase inactivation mediated gastric tumorigenesis? And what is the relationship between the Hippo and PI3K/Akt signaling pathways in this system? In view of these questions, this study investigated PTEN phosphatase inactivation and the PI3K/Akt and Hippo pathways and conducted in-depth studies at the human, animal and cell levels to provide new insights into the diagnosis and treatment of GC.
Methods
Tissue specimens
A GC tissue microarray containing 90 cases of GC and paired adjacent non-tumor tissues (ANTTs) was purchased from Shanghai Outdo Biotech (HStm-Ade180Sur-06). The study was approved by the Medical Research Ethics Committee and the Institutional Review Board of the First Affiliated Hospital of Nanchang University.
Cell lines
BGC-823 and SGC-7901 cells were respectively cultured in Dulbecco’s Modified Eagle Medium and RPMI-1640 (Thermo Scientific HyClone, Beijing, China) supplemented with 10% fetal bovine serum (FBS), 100 U penicillin, and 100 μg/ml streptomycin (Gibco of Thermo Fisher Scientific Inc., Waltham, MA, USA) at 37 °C in an atmosphere of 5% CO2.
Reagents and lentivirus
YAP-TEAD-driven transcription was disrupted with verteporfin (VP, SML0534; Sigma-Aldrich, St. Louis, MO, USA) [
24]. The cells were incubated with different concentrations of VP for 72 h for cell quantification and motility detection. Wild-type PTEN, dominant-negative mutant type PTEN (C124S and G129E) and empty lentiviral supernatants were purchased from Invitrogen (of Thermo Fisher Scientific Inc.). For lentivirus infection, BGC-823 and SGC-7901 cells were grown to approximately 80% confluence and incubated with viral supernatants and hexadimethrine bromide (Sigma-Aldrich, St. Louis, MO, USA) for 6 h. Forty-eight hours later, the cells were split and cultured in selection media containing blasticidin (Sigma-Aldrich) for an additional 2 weeks. To isolate single cell lines, the cells were diluted and plated into wells at a ratio of 0.3 cells per well in 96-well plates in media without puromycin. Stable cell lines expressing wild-type or dominant-negative mutant PTEN were then established.
Immunoblotting
Western blotting was performed according to standard method described previously [
25] using the following antibodies: anti-PTEN (#9559; 1:1000), anti-phosphorylated (p)-PTEN (Ser380/Thr382/383) (#9554; 1:1000), anti-Akt (#4691; 1:1000), anti-p-Akt (Ser473) (#4060; 1:1000), anti-p-Akt (Thr308) (#13088; 1:1000), anti-Bad (#9239; 1:1000), anti-p-Bad (Ser136) (#4366; 1:1000), anti-p-FoxO1 (Ser256) (#9461; 1:1000), anti-FoxO1(#2880; 1:1000), anti-p27 Kip1 (#3686; 1:1000), anti-p-GSK-3β (#9323; 1:1000), anti-GSK-3β (#9315; 1:1000), anti-p-YAP (Ser 127) (#13008; 1:1000), anti-YAP/TAZ(#8418; 1:1000), anti-SAV1 (#13301; 1:1000), anti-LATS1 (#3477; 1:1000), anti-LATS2 (#5888; 1:1000), anti-MOB1 (#8699; 1:1000), anti-histone H3 (#4499; 1:1000), and anti-TEAD (#13295; 1:1000) from Cell Signaling Technology (Danvers, MA, USA) and anti-GAPDH (BA2913, 1:1000) from Boster Biological Technology (Preston, CA, USA).
Immunohistochemistry
Immunohistochemistry was performed on 90 cases of GC and their paired ANTTs according to standard method described previously [
25] using the following antibodies: anti-p-PTEN (Ser380/Thr382/383) (ab47332; 1:800) antibody (Abcam, Cambridge, UK), anti-YAP (#8418; 1:200) antibody and anti-p-Akt (Ser473) antibody (#4060; 1:100) (Cell Signaling Technology, Danvers, MA, USA). A total of 100 cells were counted in five random fields at 200×, and scores were evaluated based on the ratio and intensity of stained cells as described previously. Then, we categorized the expression into two groups based on score: score 0–3, low expression; score 4–12, high expression.
Cell survival was detected using the MTT assay. The stable cell lines expressing wild-type PTEN, dominant-negative mutant PTEN (C124S and G129E) or empty vector were seeded at a density of 2 × 10
3 cells/well in a 96-well plate, and 20 μl of MTT solution (Sigma-Aldrich, St. Louis, MO, USA) was added 4 h before detection as described previously [
20]. Cell proliferation was also detected using a BrdU assay kit (EMD Millipore, Billerica, MA, USA). For colony-formation assay, stable cell lines were trypsinized into a single-cell suspension and seeded in 6-cm dishes at 10
3 per plate. After 2–3 weeks of incubation, the colonies were stained with crystal violet dye. For the migration assay
, stable cell lines were seeded at a density of 3 × 10
4 cells/well in transwell chambers (8.0-μm pore size; Corning Inc., Corning, NY, USA) with medium containing 20% FBS in the lower chamber as the chemoattractant and incubated for 48 h at 37 °C with 5% CO
2. Cells that did not migrate through the pores were mechanically removed with a cotton swab. Cells attached to the bottom of the membrane insert were fixed in methanol at room temperature for 5 min and stained with hematoxylin. The number of invaded cells on the lower surface of the membrane was counted under a microscope at 400× magnification. For the high content analysis, stable cell lines were seeded at a density of 2 × 10
3 cells/well in a 96-well plate, and after adherence, the cells were changed to medium without FBS and incubated overnight. The cells were then were placed into the high content analysis system to detect the cell numbers and relative motility distance.
Mouse studies
Female BALB/c nude mice were provided by the Experimental Animal Center of the Fourth Military Medical University. All animals were housed and maintained in pathogen-free conditions. All animal studies complied with the Fourth Military Medical University animal use guidelines, and the protocol was approved by the Medical Research Ethics Committee and the Institutional Review Board of the First Affiliated Hospital of Nanchang University. Thirty-two nude mice were randomly divided into four groups according to the expression of the target genes: Control (with empty vector), WT (expressing WT PTEN), Mut 1 (expressing PTEN C124S), and Mut 2 (expressing PTEN G129E). The stable cell lines were trypsinized into a single-cell suspension and diluted to 1 × 107/ml. A total of 200 μl of cells was injected subcutaneously into each flank of the nude mice. The tumor volumes were monitored using a living imaging system, and the growth curves of the tumors were plotted accordingly. After approximately 4 weeks, the nude mice were sacrificed, and the tumors were weighed.
Coimmunoprecipitation
The appropriate cell numbers lysed for coimmunoprecipitation were determined by the same expression levels of LATS1 and LATS2 in the total lysate among groups. Coimmunoprecipitation was accomplished by incubating lysates with anti-MOB1 antibody (sc-161,867, Santa Cruz, Dallas, TX, USA) for 2 h and then with Protein A/G beads overnight. The immunoprecipitates were washed three times with lysis buffer. The immunoprecipitated proteins and input lysates were resolved by SDS-PAGE and then immunoblotted with the indicated antibodies.
Immunofluorescence
The stable cell lines were seeded into an immunofluorescence culture chamber at the same density and cultured using standard protocols. The medium was removed, and the cells were washed with PBS and fixed in 4% formaldehyde solution for 15 min. The cells were then permeabilized in 0.5% Triton for 15 min, blocked with 10% serum for 30 min, and stained using indicated antibodies overnight at 4 °C. The cells were stained with the secondary antibody for 1 h and DAPI for 15 min and then imaged under a confocal microscope.
Nuclear-cytoplasm separation
The nucleus-cytoplasm separation assay was performed using the Nucleus-cytoplasm Protein Extraction Kit (Beyotime Biotechnology, Shanghai, China). Extracted nuclear and cytoplasmic proteins were resolved by SDS-PAGE and then immunoblotted with the indicated antibodies.
Statistical analyses
The data are summarized as the means±standard deviations (SDs) or percentages of the control. The chi-square test was performed to evaluate differences in categorical variables. One-way analysis of variance (ANOVA) was used to determine the differences in numerical variables. Kruskal-Wallis or Mann-Whitney tests were used to determine the differences in numerical variables between differently defined groups. Growth curves were plotted using the Kaplan-Meier method. An independent factor analysis for the prognosis of GC was performed using univariate and multivariate COX regression models; p < 0.05 was considered significant.
Discussion
PTEN deletion has been recognized as a driver of many sporadic cancers in the clinic [
26‐
31]. PTEN has dual activity of a protein phosphatase and a lipid phosphatase and can thus regulate a complex set of signaling pathways. However, the substitution effect of protein phosphatase activity remains unclear. Therefore, we investigated the role of the loss of PTEN phosphatase and lipid phosphatase functions in GC development. We established stable cell lines that expressed wild-type PTEN (with robust phosphatase activity), dominant-negative mutant PTEN C124S (loss of complete phosphatase activity), or G129E (lacks only lipid phosphatase activity) to conduct functional and mechanistic studies in vitro and in vivo. Our results showed that PTEN lipid phosphatase inactivation increased the malignant biological behavior of GC cells, especially cell proliferation and migration. The analysis of in vivo tumor growth in nude mice also confirmed that PTEN lipid phosphatase inactivation was positively related to GC progression. Another study on transgenic mice also showed that compared with heterozygous mice carrying one wild-type allele and one null allele, mice carrying both a wild-type PTEN allele and a mutant PTEN (C124S/G129E) allele had a higher risk of breast cancer [
22,
32]. These findings suggest the significance of PTEN lipid phosphatase inactivation on tumorigenesis.
PTEN phosphorylation at Ser380/Thr382/383 can deprive PTEN of phosphatase activity and maintain PTEN stability, similarly to dominant-negative mutant PTEN (C124S and G129E). Therefore, in our study, we detected the expression of p-PTEN (Ser380/Thr382/383) in human tissues to reflect the role of PTEN phosphatase inactivation in GC and found that the expression of p-PTEN in GC tissues was significantly higher than that in ANTT. One study showed that PTEN inactivation may occur as a relatively late event during cancer development [
33]. In this study, we found that p-PTEN expression was positively correlated with the depth of tumor invasion, lymph node metastasis, clinical stage and a poor prognosis in GC patients and is an independent prognostic factor for GC. Based on these findings, p-PTEN (PTEN phosphatase inactivation) is expected to become an effective biomarker for predicting the risk of GC occurrence and prognosis.
To date, the known functions of the Hippo signaling pathway are mainly the following: (1) to regulate the size and growth of tissues and organs, (2) to regulate cell proliferation and apoptosis, (3) to regulate the epithelial-mesenchymal transition and intercellular contact, and (4) to maintain stem cell self-renewal and versatility [
34]. Dysregulation of the Hippo pathway may cause uncontrolled phenotypic changes, which result in disease [
35]. YAP is the key effector downstream of the Hippo pathway. However, there are some controversies regarding the suppression or promoting role of YAP in cancer. Barry et al. [
36] found that YAP expression was decreased in colon cancer patients, thereby inhibiting the Wnt signaling pathway in the colon. However, Wang et al. [
37] reported that 53.5% of colon cancer patients show YAP overexpression, and this overexpression was found to be associated with lymph node metastasis, TNM staging, and a short overall survival rate. In addition, another study on breast cancer found that YAP showed higher nuclear expression in normal breast tissues and was absent in breast cancer tissues [
38]. Furthermore, although the role of YAP in human glioma cells is difficult to judge, there is an indication for YAP overexpression [
39]. These reports suggest that YAP may play a dual role in suppressing and promoting cancer in these types of tumor.
What is the exact role of YAP in gastric cancer? To address this question, we detected YAP expression in 90 cases of GC and their paired ANTTs and found that YAP expression was significantly higher in GC tissues than in ANTTs. The total and nuclear YAP levels were positively correlated with the tumor size, depth of tumor invasion, lymph node metastasis, clinical stage and poor prognosis in GC patients, whereas the cytoplasmic YAP levels did not correlate with these clinicopathologic features. In addition, both the total and nuclear YAP levels are independent prognostic factors for GC. Based on the above-described results, it can be noted that YAP may play an important role in promoting GC development, and the localization of YAP in GC cells may lead to different risks of disease. However, YAP can exert different functions in the cytoplasm and nucleus, which depend on the tissue type. For example, a previous study found that YAP expression was elevated in cervical squamous cell carcinomas and that the cytoplasmic YAP level was positively correlated with lymph node metastasis, high-grade tissue type, and early recurrence. In contrast, nuclear YAP was overexpressed in cervical adenocarcinoma, and increased nuclear YAP led to a reduction in the overall survival rate [
40]. These results suggest that YAP may serve as a potential anti-cancer target, which has important clinical significance.
However, the regulatory mechanism of Hippo/YAP remains unclear. In our study, we found that the expression levels of nuclear YAP were positively related to p-PTEN in GC tissues. Interestingly, our study also demonstrated that PTEN lipid phosphatase inactivation decreased the expression of upstream kinases in the Hippo pathway, increased YAP phosphorylation and promoted YAP translocation into the nucleus in stable GC cell lines. Furthermore, previous studies emphasized that increased binding of MOB1 with LATS1/2 promotes further activation of the upstream-kinase-activation-loop, thus increasing YAP phosphorylation and inhibiting the nuclear translocation of YAP [
41,
42], also in our study, we found that PTEN lipid phosphatase inactivation abolished the MOB1-LATS1/2 complexes, which may contribute to the observed decrease in YAP phosphorylation and the enhancement of nuclear localization.
We then added VP, which can effectively decrease YAP/TEAD expression and disrupt YAP-TEAD-driven transcription, to the stable cell lines to determine how Hippo/YAP contributes to the gastric tumorigenesis mediated by PTEN lipid phosphatase inactivation. We found that cell proliferation and motility were significantly inhibited in a dose-dependent manner in the stable cell lines, especially in the cells expressing dominant-negative mutant PTEN. The analysis of in vivo tumor growth in nude mice also confirmed the important role of Hippo/YAP in PTEN lipid phosphatase inactivation induced GC progression. In summary, PTEN lipid phosphatase inactivation regulates the Hippo signaling pathway, reduces YAP phosphorylation, promotes YAP nuclear translocation, and leads to an enhanced synergistic effect of YAP-TEAD, thus inducing GC proliferation and migration.
The PI3K/Akt signaling pathway is important in controlling cell survival, promoting proliferation and inhibiting apoptosis [
20]. Our studies demonstrated that PTEN lipid phosphatase inactivation activated PI3K/Akt and regulated downstream molecules, such as FoxO, GSK-3β, β-catenin, p27 and Bad. And also p-Akt (Ser473) expression was positively correlated with expression of p-PTEN in human GC tissues. Because Hippo/YAP and PI3K/Akt are both involved in the regulation of PTEN lipid phosphatase activity, we further explored their link. Interestingly, we found expression of p-Akt (Ser473) was positively correlated with that of nuclear YAP in human GC tissues. Strikingly, we further found that VP could decrease Akt activation in a dose-dependent manner. This finding was consistent with previous studies, as some reports have confirmed the synergistic effects of Hippo/YAP and P13K/Akt on cell proliferation and apoptosis. Some scholars found that Yki, a homolog of YAP in Drosophila, positively regulates the expression and activity of Akt, and activation of the Hippo signaling pathway can reduce Akt expression [
43]. Tumaneng et al. [
23] found that overexpression of YAP increased phosphorylation of Akt and conversely, p-Akt was reduced in YAP knockdown cells. In addition, a study on medulloblastoma found that YAP could increase the radioresistance of tumor cells by inducing the expression of insulin-like growth factor and Akt activation, thereby promoting the growth of tumor cells after radiotherapy [
44]. Another in vivo study has shown that the binding of YAP-TEAD directly promotes the transcription of P13K, thereby activating the PI3K/Akt signaling pathway and promoting myocardial cell proliferation in rats [
45,
46].