Activation of sonic hedgehog (Shh) signaling has been implicated in progression of a variety of tumors. In this study, we elucidated a role for Shh in the invasion of gastric tumors and determined the mechanism by which Shh is regulated. Immunohistochemical analysis of 178 primary human gastric tumor biopsies indicated that Shh expression was positively correlated with lymph node metastasis, high lymphatic vessel density, and poor prognosis. In mouse xenograft models of human gastric cancer, enforced expression of Shh significantly enhanced the incidence of lung metastasis compared with nonexpressing controls. Mechanistic investigations revealed that phosphoinositide 3-kinase (PI3K)/Akt inhibition blocked Shh-induced epithelial–mesenchyme transition, the activity of matrix metalloproteinase 9 (MMP-9), and lymphangiogenesis, reducing tumor invasiveness and metastasis. Taken together, our findings establish that Shh signaling promotes the metastasis of gastric cancer through activation of the PI3K/Akt pathway, which leads to mesenchymal transition and MMP-9 activation. These findings offer preclinical validation of Shh as a candidate therapeutic target for treatment of metastatic gastric cancers. Cancer Res; 71(22); 7061–70. ©2011 AACR.

The sonic hedgehog (Shh) signaling pathway plays a critical role in stem cell maintenance and specifying patterns of cell growth and differentiation during embryonic development (1). Shh activates its signaling by binding to the receptor, Patched (Ptch), which relieves Ptch-mediated repression of a downstream membrane protein related to G protein-coupled receptors, Smoothened (Smo). Upon activation, Smo promotes nuclear translocation of a family of transcription factors [Ci in Drosophila and Glis (Gli1, Gli2, and Gli3) in vertebrates], and subsequently activates target genes through Glis (2).

Recently, the Shh signaling pathway was shown to not only be involved in induction of angiogenesis in an adult mammalian system but also in the control of the motility and migration of multiple cell types (3). Studies conducted by Hochman and colleagues have suggested that the components of the Shh signaling pathway participate directly in cell migration and angiogenesis, whereas inhibition of this pathway blocks Shh-induced cell migration and angiogenesis (4). In addition, Shh and Gli1 expression has also been shown to affect epithelial-to-mesenchymal transition (EMT) in pancreatic cancer cell lines and lymphatic metastasis in esophageal squamous cell carcinoma (5, 6). These observations imply that disregulated Shh signaling is correlated with the severity of the associated tumor and contributes to maintain metastatic behavior. Therefore, understanding these pathways in tumor metastasis will provide great promise for the discovery of novel therapeutics and the treatment of metastatic disease. Nevertheless, the exact role and regulatory mechanisms of Shh signaling in tumorigenesis and metastatic progression of gastric cancer are controversial and need further classification. To clarify the functional role of Shh signaling in gastric cancer progression, we used multiple approaches in testing the gene, both in vitro and in vivo, and validated the outcome results in a clinical setting. In this study, we found that the Shh signaling pathway, acting through the phosphoinositide 3-kinase (PI3K)/Akt pathway, may contribute to tumor metastasis by inducing EMT and mediating matrix metalloproteinase 9 (MMP-9) activity.

Cell lines and cultures

AGS and KATOIII were obtained from the American Type Culture Collection and MRC-5, SNU-5, SNU-16, SNU-216, SNU-638, MKN-28, MKN-45, and MKN-74 were purchased from the Korea Cell Line Bank. Adult human dermal lymphatic microvascular endothelial cells (HMVEC-dLyAd) were obtained from Lonza and the cells were maintained according to manufacturer's instructions.

Authentication and characterization of cell lines

Each cell was characterized by cytogenetic analysis. SNU-5, SNU-16, SNU-216, and SNU-638 were initially characterized by Park and colleagues (7). MKN-28, MKN-45, and MKN-74 were characterized by Kitano and Kitamura (8). Origin of MRC-5 was confirmed by nucleoside phosphorylase, glucose-6-phosphate dehydrogenase, and lactate dehydrogenase isoenzyme electrophoresis. HMVECs stain positive for acetylated LDL and von Willebrand's (factor VIII) antigen and stain negative for smooth muscle α-actin.

Clinical samples

The archival paraffin-embedded tissues were used from 178 patients who had undergone surgical excision for gastric cancer at Korea University Guro Hospital between March 2002 and April 2004. Tissue samples were obtained from Korea Lung Tissue Bank, assigned and supported by the Korea Science & Engineering Foundation in the Ministry of Science & Technology. Patient clinical characteristics according to the International Union Against Cancer, as revised in 2002, are shown in Table 1. This protocol was reviewed and permitted by Institutional Review Board of Guro Hospital.

Table 1.

Relationship between Shh expression and other clinicopathologic features in gastric cancer

Clinicopathologic feature (number of cases examined)Number of casesShh expressionPa
PositiveNegative
Age (174) 
 >60 79 37 (46.8) 42 (53.2) 0.071 
 ≤60 95 68 (71.6) 25 (28.4)  
Gender (173) 
 Male 124 82 (66.1) 42 (33.9) 0.095 
 Female 49 23 (46.9) 26 (53.1)  
Lauren's classification (162) 
 Intestinal type 57 41 (72) 16 (28) 0.021 
 Diffuse type 105 61 (58.1) 44 (41.9)  
Histologic grade (73) 
 WDb 4 (80) 1 (20) 0.018 
 MDc 47 36 (76.6) 11 (23.4)  
 PDd 21 8 (38) 13 (62)  
Lymph node status (174) 
 No metastasis 46 19 (41.3) 27 (58.7) 0.042 
 Metastasis 128 88 (68.8) 40 (31.2)  
TNM staging (178) 
 I 31 13 (41.9) 18 (58.1) 0.039 
 II–IV 147 92 (62.6) 55 (37.4)  
Clinicopathologic feature (number of cases examined)Number of casesShh expressionPa
PositiveNegative
Age (174) 
 >60 79 37 (46.8) 42 (53.2) 0.071 
 ≤60 95 68 (71.6) 25 (28.4)  
Gender (173) 
 Male 124 82 (66.1) 42 (33.9) 0.095 
 Female 49 23 (46.9) 26 (53.1)  
Lauren's classification (162) 
 Intestinal type 57 41 (72) 16 (28) 0.021 
 Diffuse type 105 61 (58.1) 44 (41.9)  
Histologic grade (73) 
 WDb 4 (80) 1 (20) 0.018 
 MDc 47 36 (76.6) 11 (23.4)  
 PDd 21 8 (38) 13 (62)  
Lymph node status (174) 
 No metastasis 46 19 (41.3) 27 (58.7) 0.042 
 Metastasis 128 88 (68.8) 40 (31.2)  
TNM staging (178) 
 I 31 13 (41.9) 18 (58.1) 0.039 
 II–IV 147 92 (62.6) 55 (37.4)  

NOTE: The values in parenthesis are expressed in percentage.

aStatistical analysis was conducted using Student t test. P < 0.05 was considered statically significant.

bWell differentiated.

cModerately differentiated.

dPoorly differentiated.

Animal studies using a subcutaneous and tail vein injection model

Exponentially growing cells (1 × 106) in 0.1 mL of PBS were prepared and then injected into the subcutis or tail veins of 6-week-old BALB/c nu/nu mice (Charles River Laboratories). Mice were sacrificed 7 weeks after inoculation of the cells, and metastatic lesions on the lungs were counted macroscopically. Primary tumor tissues were harvested, weighed, and processed for further analyses.

Reverse transcriptase PCR analysis, luciferase assay, immunohistochemical staining, cell proliferation assay, wound healing, Matrigel invasion assays, Western blotting, immunofluorescence staining, and gelatin zymography

The above analyses were conducted as described earlier (9, 10). Antibodies, sources, clones, and dilutions used in immunofluorescence staining are listed in Supplementary Table S1. The sequences of primers used for reverse transcriptase PCR (RT-PCR) analysis are shown in Supplementary Table S2.

In vitro kinase assay and in vitro tube formation assay

Nonradioactive Akt kinase assays were carried out according to the manufacturer's protocol (Cell Signaling). Quantitative analysis of lymphatic vessel density (LVD) was conducted in sections, which were single stained for D2-40.

Statistical analysis

The similar methods were used for statistical analysis of in vitro and in vivo data described previously (9). The relationships between variables were assessed with a Spearman rank correlation coefficient. Difference in the incidence of lung metastasis was analyzed by Fischer exact test. Kaplan–Meier survival curves were constructed for patients positive and negative for Shh. Differences between survival curves were tested for statistical significance by a log-rank test. P values less than 0.05 were defined as statistically significant.

Elevated expression of Shh is correlated with advanced clinical stage, lymph node metastasis, and poor prognosis in patients with gastric cancer

To investigate the role of Shh pathway in gastric cancer metastasis, we first conducted RT-PCR for components of the Shh pathway in a panel of gastric cells. As expected, the comparison of various gastric cancer cell lines or tissues revealed that all of the components of this pathway were expressed in diverse gastric tumor cells to a different extent (Supplementary Fig. S1A and B). Interestingly, we detected significantly higher expression of Shh and Gli1 in gastric tumor tissues obtained from patients with lymph node metastasis compared with patients without lymph node metastasis (Fig. 1A). Furthermore, we showed that there was a significant positive correlation between the levels of Shh and Gli1 transcripts (P < 0.01, r = 0.766, Fig. 1A).

Figure 1.

Shh pathway plays a crucial role in invasiveness and metastatic potential of gastric cancer cells in vivo and in vitro. A, the gastric cancer samples were divided into 2 groups according to the lymph node status (LN−, no metastasis; LN+, metastasis), and then association between expression of Shh mRNA or Gli1 mRNA and gastric cancer metastasis was observed by RT-PCR in the left. The correlation between the levels of Shh mRNA and Gli1 mRNA was assessed with the Spearman rank correlation coefficient (right). B, the 5-year survival rate was calculated in 147 patients with gastric cancer at stages II to IV relative to the expression of the Shh (Shh positive vs. negative, left) or stages (stage I vs. II–IV, right) by the Kaplan–Meier method. P = 0.002 and P = 0.02 were determined by the log-rank test, respectively. C, gastric cancer cells were treated N-Shh (0.5 μg/mL) alone or together with KAAD-cyclopamine (cyclopamine, 10 μmol/L) and then assayed for their motility (top) and invasiveness (bottom) by wound healing and Matrigel invasion assay as described in the Materials and Methods section. D, the cells transfected with Gli1 siRNA or nonspecific (Con) siRNA were then evaluated in a performance migration (top) and invasion assay (bottom). Relative mRNA level was normalized to the corresponding β-actin mRNA expression. Bars represent the SD of 3 independent experiments conducted in triplicate. *, P < 0.05; **, P < 0.01. E, the metastatic potential of either AGS cells expressing Shh/or control vector (Con) or SNU-16 cells Shh shRNA/or nonspecific control shRNA (con shRNA) was assessed by measuring colonization of lung surfaces following subcutaneous (s.c.) or tail vein injection (i.v.) of these cells into athymic mice. Seven weeks later, the mice were sacrificed and the number of mice with metastatic lesions to lung was counted.

Figure 1.

Shh pathway plays a crucial role in invasiveness and metastatic potential of gastric cancer cells in vivo and in vitro. A, the gastric cancer samples were divided into 2 groups according to the lymph node status (LN−, no metastasis; LN+, metastasis), and then association between expression of Shh mRNA or Gli1 mRNA and gastric cancer metastasis was observed by RT-PCR in the left. The correlation between the levels of Shh mRNA and Gli1 mRNA was assessed with the Spearman rank correlation coefficient (right). B, the 5-year survival rate was calculated in 147 patients with gastric cancer at stages II to IV relative to the expression of the Shh (Shh positive vs. negative, left) or stages (stage I vs. II–IV, right) by the Kaplan–Meier method. P = 0.002 and P = 0.02 were determined by the log-rank test, respectively. C, gastric cancer cells were treated N-Shh (0.5 μg/mL) alone or together with KAAD-cyclopamine (cyclopamine, 10 μmol/L) and then assayed for their motility (top) and invasiveness (bottom) by wound healing and Matrigel invasion assay as described in the Materials and Methods section. D, the cells transfected with Gli1 siRNA or nonspecific (Con) siRNA were then evaluated in a performance migration (top) and invasion assay (bottom). Relative mRNA level was normalized to the corresponding β-actin mRNA expression. Bars represent the SD of 3 independent experiments conducted in triplicate. *, P < 0.05; **, P < 0.01. E, the metastatic potential of either AGS cells expressing Shh/or control vector (Con) or SNU-16 cells Shh shRNA/or nonspecific control shRNA (con shRNA) was assessed by measuring colonization of lung surfaces following subcutaneous (s.c.) or tail vein injection (i.v.) of these cells into athymic mice. Seven weeks later, the mice were sacrificed and the number of mice with metastatic lesions to lung was counted.

Close modal

Next, we conducted immunohistochemical staining for Shh in 178 gastric cancer specimens. Clinicopathologic analysis revealed that the expression of Shh was strongly elevated in the cases with advanced stages (stages II–IV) and with lymph node metastasis compared with early stage (stage I, P = 0.039) and nonmetastatic cases (P = 0.042; Table 1 and Supplementary Fig. S1C). The results also indicate that in advanced stages (stages II–IV) of gastric cancer (n = 147), patients with positive expression of Shh had a significantly worse rate of survival than the patients with low expression of the gene (P = 0.002 by log-tank test; Fig. 1B). On the basis of univariate Cox regression analysis, the death risk of patients with increased Shh expression was 3.2 times higher than the risk of patients with a reduced expression of the gene. Indeed, the overall survival rate was significantly lower in patients with gastric cancer at stages II to IV than those at stage I (P = 0.02 by log-tank test; Fig. 1B). Taken together, these results clearly show that elevated expression of Shh is correlated with advanced clinical stage, lymph node metastasis, and poor prognosis in patients with gastric cancer.

Crucial function of Shh signaling on invasiveness and metastatic potential of gastric cancer cells

Next, we tested whether or not the Shh signaling pathway is involved in progression to metastatic disease in gastric cancer, especially in late tumorigenesis, including migration and invasion. Our results showed that N-Shh treatment significantly stimulated the migration and invasiveness of these cells (Fig. 1C, and D). In contrast, the stimulatory effects of N-Shh on cell migration and invasion were completely abolished by cyclopamine (a Shh signaling inhibitor) or Gli1-specific siRNA treatment (Supplementary Fig. S2 A).

The AGS cell line expressing Shh and containing only vector was subcutaneously injected into athymic mice, then the formation and growth rate of the tumor was monitored every 4 days for 24 days. All of the vector only and Shh expressing cells had similar growth rates during the indicated period, suggesting that overexpression of Shh did not significantly affect tumor growth rate (Supplementary Fig. S2B). This result is consistent with data obtained from in vitro BrdU labeling and fluorescence-activated cell sorting analysis that assesses the proliferation and the cell-cycle distribution of the vector only and Shh expressing AGS cells (Supplementary Fig. S2C). On the contrary, tumors stably expressing Shh showed nearly 5 times more colony formation on lung surfaces compared with the tumors containing only vector (Supplementary Fig. S2D). In addition, Shh significantly increased the number of mice with metastasis to the lung, and similar results were also obtained in a tail vein injection model (Fig. 1E).

To further determine whether there is a causal correlation between Shh expression and metastatic function of gastric cancer, Shh shRNA expressing highly metastatic SNU-16 cells were compared with vector only expressing cells for the progression of lung metastases. Lung metastasis developed in 9 of the 10 (90%) mice inoculated with nonspecific shRNA only expressing cells, whereas only 1 of the 10 (10%) mice that inoculated with Shh shRNA expressing cells developed lung metastases (Fig. 1E). These results strongly suggest that Shh signaling enhances the metastatic potential of gastric cancer cells.

Shh signaling pathway induces lymphangiogenesis and morphologic changes of gastric cancer cells through the EMT process and MMP-9 activity

In the context of tumor progression, EMT is an early event in tumor invasion and increased motility and invasion are positively correlated with EMT (11). Here, we found that treatment of SNU-216 and MKN-74 cells with N-Shh for 72 hours exhibited dramatic changes in cell morphology, from a cuboid, epithelial-like shape to a spindle, fibroblastic-like appearance, consistent with EMT (Supplementary Fig. S3A). Concomitant with the change in phenotype, the cells treated with N-Shh showed strong suppression of E-cadherin 72 hours after addition, whereas strong expression of Snail, which is known to repress expression of the E-cadherin gene, were detected in cells treated with N-Shh. These alterations were completely recovered by pretreatment with cyclopamine before N-Shh stimulation or by knocked down Gli1 (Fig. 2A).

Figure 2.

Effect of Shh signaling pathway on EMT, MMP-9 activity, and lymphangiogenesis. A, the cells were treated with N-Shh (0.5 μg/mL) alone, cyclopamine (10 μmol/L), nonspecific Con siRNA, or Gli1 siRNA for 72 hours. Western blotting with anti–E-cadherin or anti-Snail was carried out. B, left, the cells treated with N-Shh (0.5 μg/mL) alone or together with cyclopamine (10 μmol/L) for 24 hours were collected and RNA extract was subjected to quantitative real-time PCR using specific primers for MMP-9 (top). The conditioned culture media of the same set of samples were subjected to zymography assay for MMP-2 and MMP-9 (bottom). Bars represent the SD of 3 independent experiments conducted in triplicate. †, <0.05, ††, <0.01, compared with cells treated with control vehicle. *, P < 0.05, **, P < 0.01, compared with cells treated with N-Shh. Right, AGS cells were transfected with MMP-9 siRNA or nonspecific (Con) siRNA for 24 hours, followed by stimulation with N-Shh (0.5 μg/mL). Right, the migration and invasive ability were evaluated by wound healing (top) and the Matrigel invasion assay (bottom) after 24 hours. The data are expressed as the means of 3 independent experiments ± SD. *, P < 0.05; **, P < 0.01. C, the metastatic potential of cells expressing Shh, control vector (Con), MMP-9 shRNA, or nonspecific control shRNA (con shRNA) was assessed by measuring colonization of lung surfaces following tail vein injection (i.v.) of these cells into athymic mice. Seven weeks later, the mice were sacrificed and the number of mice with metastatic lesions to lung was counted. D, AGS cells were treated with N-Shh (0.5 μg/mL) alone or a combination of either cyclopamine (10 μmol/L) or Gli1 siRNA and RNA extract was subjected to RT-PCR using the indicated primers. Relative mRNA level was normalized to the corresponding β-actin mRNA expression. E, HMVECs were treated with N-Shh (0.5 μg/mL) alone or together with cyclopamine (10 μmol/L), and then the migration ability was evaluated by wound healing assay after 24 hours (top panel). Cyclopamine-pretreated HMVECs were seeded on the top of growth factor-reduced Matrigel and treated with or without N-Shh (0.5 μg/mL) for 9 hours (bottom panel).

Figure 2.

Effect of Shh signaling pathway on EMT, MMP-9 activity, and lymphangiogenesis. A, the cells were treated with N-Shh (0.5 μg/mL) alone, cyclopamine (10 μmol/L), nonspecific Con siRNA, or Gli1 siRNA for 72 hours. Western blotting with anti–E-cadherin or anti-Snail was carried out. B, left, the cells treated with N-Shh (0.5 μg/mL) alone or together with cyclopamine (10 μmol/L) for 24 hours were collected and RNA extract was subjected to quantitative real-time PCR using specific primers for MMP-9 (top). The conditioned culture media of the same set of samples were subjected to zymography assay for MMP-2 and MMP-9 (bottom). Bars represent the SD of 3 independent experiments conducted in triplicate. †, <0.05, ††, <0.01, compared with cells treated with control vehicle. *, P < 0.05, **, P < 0.01, compared with cells treated with N-Shh. Right, AGS cells were transfected with MMP-9 siRNA or nonspecific (Con) siRNA for 24 hours, followed by stimulation with N-Shh (0.5 μg/mL). Right, the migration and invasive ability were evaluated by wound healing (top) and the Matrigel invasion assay (bottom) after 24 hours. The data are expressed as the means of 3 independent experiments ± SD. *, P < 0.05; **, P < 0.01. C, the metastatic potential of cells expressing Shh, control vector (Con), MMP-9 shRNA, or nonspecific control shRNA (con shRNA) was assessed by measuring colonization of lung surfaces following tail vein injection (i.v.) of these cells into athymic mice. Seven weeks later, the mice were sacrificed and the number of mice with metastatic lesions to lung was counted. D, AGS cells were treated with N-Shh (0.5 μg/mL) alone or a combination of either cyclopamine (10 μmol/L) or Gli1 siRNA and RNA extract was subjected to RT-PCR using the indicated primers. Relative mRNA level was normalized to the corresponding β-actin mRNA expression. E, HMVECs were treated with N-Shh (0.5 μg/mL) alone or together with cyclopamine (10 μmol/L), and then the migration ability was evaluated by wound healing assay after 24 hours (top panel). Cyclopamine-pretreated HMVECs were seeded on the top of growth factor-reduced Matrigel and treated with or without N-Shh (0.5 μg/mL) for 9 hours (bottom panel).

Close modal

Next, we evaluated Shh-mediated cell migration and invasion to determine whether it resulted from elevated levels of MMPs, which are well-documented ECM-degrading enzymes, the activity of which is associated with tumor invasiveness. RT-PCR analysis showed that N-Shh significantly increased the mRNA levels of MMP-9 in AGS and MKN-28 cells (Fig. 2B). In contrast, abrogating the Shh pathway by cyclopamine or Gli1 siRNA reduced induction of MMP-9 by N-Shh. Similar results were obtained in Western blot (data not shown) and gelatin zymographic analysis of conditioned medium obtained from AGS and MKN-28 cells (Fig. 2B). As expected, overexpression of MMP-9 siRNA reduced the ability of Shh to stimulate migration and invasion of those cells compared with the cells expressing control siRNA (Supplementary Fig. S3B and Fig. 2B). Importantly, in the tail vein experiments, MMP-9 knockdown in the cells expressing Shh also resulted in significant reduction in the lung metastasis compared with cells expressing only Shh (Fig. 2C). Collectively, these findings suggest that Shh-mediated induction of EMT or MMP-9 contributes to Shh pathway-related migration and invasive ability of gastric cancer cells.

The metastatic spread of tumors through lymphatic vessels to local or regional lymph nodes is an early event in metastatic disease for a variety of human tumors, and nodal metastasis is considered to be dependent on the tumor-induced lymphangiogenesis (12–14). These findings suggest that Shh signaling may influence the metastatic capacity of gastric cancer to lymph nodes by inducing lymphangiogenesis. Our results showed that cells treated with N-Shh displayed significantly higher levels of expression of Lyve-1, Podoplanin, and VEGF-D than cells treated without N-Shh (Fig. 2D). The levels of expression of these molecules in cells treated with cyclopamine or Gli1 siRNA were significantly lower than those in the cells treated with control vehicle. Next, we evaluated the responses in terms of migration and tubule formation in HMVECs, both of which are essential steps for lymphangiogenesis in vivo. As shown in Fig. 2E, N-Shh significantly increased the motility and tubule-like structure formation of HMVECs compared with control vehicle. In contrast, pretreatment of HMVECs with cyclopamine completely inhibited Shh-induced cell migration and the generation of tubular networks.

PI3K/Akt signaling pathway is required for Shh-mediated EMT, MMP-9 activity, and lymphangiogenesis in gastric cancer cells

Recently, numerous studies have shown that PI3K/Akt pathway is an essential component of Shh signaling and inhibition of PI3K/Akt pathway with the PI3K/AKT inhibitor, LY294002, results in impairment of Shh morphogenic activity (15–17). Western blot analysis showed that the activation of PI3K-dependent Akt phosphorylation in AGS, MKN-28, and SNU-216 cells was induced as early as 30 minutes after N-Shh stimulation (Supplementary Fig. S4A and Fig. 3A). However, blockade of Shh signaling by cyclopamine completely reversed the enhancement of Akt phosphorylation (Fig. 3A). In vitro Akt kinase assay showed that N-Shh treatment enhanced GSK-3α/β phosphorylation, whereas the cyclopamine blocked N-Shh–activated an increase in the phosphorylated GSK-3α/β (Supplementary Fig. S4B). Notably, nonphosphorylatable Akt mutant (DN-Akt) significantly reduced the ability of N-Shh to stimulate activity of the Gli-luciferase reporter (Fig. 3A). Similar results were also obtained in real-time RT-PCR analysis for Gli1 target gene (Ptch and Gli1; data not shown). In addition, in spite of the presence of added N-Shh, treatment of these cells with LY294002 or DN-Akt suppressed Shh-mediated Akt activation, as well as cell migration and invasion (Supplementary Fig. S4C and Fig. 3B). Similar inhibition was observed with Akt1 siRNA (Supplementary Fig. S4D).

Figure 3.

Shh promotes tumor metastasis and lymphangiogenesis via PI3K/Akt pathway followed by upregulation of EMT and MMP-9 in the in vitro model. A, the cells were treated with N-Shh (0.5 μg/mL) alone or together with cyclopamine (10 μmol/L), then Western blotting with anti–p-Akt, anti-Akt, or anti–β-actin was carried out (left). AGS cells were cotransfected with 0.5 μg of Con vector or 0.5 μg of DN-Akt and 0.5 μg of Gli-Luc reporter plasmid or 0.1 μg of pRL-TK vector. Twenty-four hours after transfection, cells were treated with N-Shh (0.5 μg/mL) or control vehicle. After incubating for 24 hours, pGli1-Luc activity was measured by luciferase assay (right). The amount of luciferase activity was normalized to Renilla luciferase activity. B, the cells were pretreated with LY294002 (20 μmol/L) for 30 minutes (left) or transfected with either control vector (Con), CA-Akt, or DN-Akt (right), followed by stimulation with N-Shh (0.5 μg/mL). The migration and invasive ability of those cells were evaluated by wound healing (top) and the Matrigel invasion assay (bottom) after 24 hours. C, protein was extracted from SNU-216 cells and resolved by SDS-PAGE and then Western blotting with anti–E-cadherin or anti-Snail antibodies was carried out. β-Actin was used as a control. D, the levels of expression (top) and enzymatic activity of MMP-9 (bottom) in AGS cells were determined by RT-PCR and gelatin zymography, respectively. E, AGS cells were pretreated with LY294002 (20 μmol/L) for 30 minutes, followed by stimulation with N-Shh (0.5 μg/mL), and then RNA extract was subjected to RT-PCR using specific primers for the indicated primers (left). Relative mRNA level was normalized to the corresponding β-actin mRNA expression. LY294002-pretreated HMVECs were cultured on Matrigel-coated 24-well plates with or without N-Shh stimulation and then assayed for their motility and tube formation by wound hearing (middle) and in vitro tube formation assay (right), respectively. In all cases, the data are expressed as the means of 3 independent experiments ±SD. *, P < 0.05; **, P < 0.01.

Figure 3.

Shh promotes tumor metastasis and lymphangiogenesis via PI3K/Akt pathway followed by upregulation of EMT and MMP-9 in the in vitro model. A, the cells were treated with N-Shh (0.5 μg/mL) alone or together with cyclopamine (10 μmol/L), then Western blotting with anti–p-Akt, anti-Akt, or anti–β-actin was carried out (left). AGS cells were cotransfected with 0.5 μg of Con vector or 0.5 μg of DN-Akt and 0.5 μg of Gli-Luc reporter plasmid or 0.1 μg of pRL-TK vector. Twenty-four hours after transfection, cells were treated with N-Shh (0.5 μg/mL) or control vehicle. After incubating for 24 hours, pGli1-Luc activity was measured by luciferase assay (right). The amount of luciferase activity was normalized to Renilla luciferase activity. B, the cells were pretreated with LY294002 (20 μmol/L) for 30 minutes (left) or transfected with either control vector (Con), CA-Akt, or DN-Akt (right), followed by stimulation with N-Shh (0.5 μg/mL). The migration and invasive ability of those cells were evaluated by wound healing (top) and the Matrigel invasion assay (bottom) after 24 hours. C, protein was extracted from SNU-216 cells and resolved by SDS-PAGE and then Western blotting with anti–E-cadherin or anti-Snail antibodies was carried out. β-Actin was used as a control. D, the levels of expression (top) and enzymatic activity of MMP-9 (bottom) in AGS cells were determined by RT-PCR and gelatin zymography, respectively. E, AGS cells were pretreated with LY294002 (20 μmol/L) for 30 minutes, followed by stimulation with N-Shh (0.5 μg/mL), and then RNA extract was subjected to RT-PCR using specific primers for the indicated primers (left). Relative mRNA level was normalized to the corresponding β-actin mRNA expression. LY294002-pretreated HMVECs were cultured on Matrigel-coated 24-well plates with or without N-Shh stimulation and then assayed for their motility and tube formation by wound hearing (middle) and in vitro tube formation assay (right), respectively. In all cases, the data are expressed as the means of 3 independent experiments ±SD. *, P < 0.05; **, P < 0.01.

Close modal

We subsequently examined whether or not Shh-induced EMT, MMP-9 activity, or lymphangiogenesis is mediated through the PI3K/Akt pathway in gastric cancer cells. Immunofluorescence staining and Western blotting showed that the loss of E-cadherin and gain of Snail by N-Shh was completely reversed by LY294002 or DN-Akt (Fig. 3C and Supplementary Fig. S5A). RT-PCR analysis and gelatin zymography also showed that MMP-9 activity and the expression of lymphangiogenic factor (Lyve-1, Podoplanin, and VEGF-D) were completely abolished by exposure of the cells to LY294002 or DN-Akt, even in the presence of N-Shh (Fig. 3D and E). After treating HMVECs with N-Shh and LY294002, we also found that LY294002 abrogated the Shh-induced cellular motility and tube formation of HMVECs (Fig. 3E). Together, these results showed that PI3K/Akt signaling pathway is necessary for Shh-mediated EMT, MMP-9 activity, and lymphangiogenesis in gastric cancer cells.

Shh signaling pathway promotes metastatic potential of gastric cancer cells by activation of the PI3K/Akt pathway in vivo

The effects of Akt activation on Shh-mediated metastatic capacity of human gastric cancers were further investigated using a mouse xenograft model. The metastatic potential of AGS cells expressing vector only, Shh, DN-Akt, or Shh/DN-Akt was evaluated by measuring colonization of lung surfaces by tumor cells after subcutaneous or tail vein injection. The tumor growth did not reveal significant differences among all 4 tumor xenografts (Supplementary Fig. S5B). In both the tail vein and subcutaneous experiments, the Shh cells expressing DN-Akt also showed a significant reduction in the number of lung metastatic colonies or mice with lung metastasis compared with cells expressing only Shh (Fig. 4A). In addition, expression of those EMT- and mesenchymal-related proteins was evidently decreased in Shh/DN-Akt tumor cells compared with Shh tumor cells (Fig. 4B). As expected, the levels of expression of MMP-9, Lyve-1, Podoplanin, and VEGF-D mRNA were also significantly decreased in the Shh/DN-Akt tumor cells than in the Shh tumor cells (Fig. 4C and D). Collectively, these results revealed that the Shh signaling pathway promotes metastatic potential of gastric cancer cells by sequential activation of the PI3K/Akt pathway, followed by upregulation of the EMT pathway and MMP-9.

Figure 4.

Shh promotes tumor metastasis and lymphangiogenesis via PI3K/Akt pathway followed by upregulation of EMT and MMP-9 in the in vivo model. A, AGS cells expressing vector only (Con), Shh, DN-Akt, or Shh/DN-Akt were injected s.c. or i.v. into athymic mice. Mice were sacrificed 7 weeks after the inoculation of the cells, and tumor colonies on the lung surfaces were counted macroscopically (left). The chart of right panel represents the incidence of lung metastasis in mice that received intravenous tail injections of each cell. The expressions of EMT marker (N-cadherin, vimentin, and Snail protein; B), MMP-9 mRNA (C), and lymphangiogenic marker (Lyve-1, Podoplanin, and VEGF-D; D) in subcutaneous tumors were evaluated via Western blot and RT-PCR analysis, respectively. In all cases, the data are expressed as the means of 3 independent experiments ±SD. *, P < 0.05, **, P < 0.01.

Figure 4.

Shh promotes tumor metastasis and lymphangiogenesis via PI3K/Akt pathway followed by upregulation of EMT and MMP-9 in the in vivo model. A, AGS cells expressing vector only (Con), Shh, DN-Akt, or Shh/DN-Akt were injected s.c. or i.v. into athymic mice. Mice were sacrificed 7 weeks after the inoculation of the cells, and tumor colonies on the lung surfaces were counted macroscopically (left). The chart of right panel represents the incidence of lung metastasis in mice that received intravenous tail injections of each cell. The expressions of EMT marker (N-cadherin, vimentin, and Snail protein; B), MMP-9 mRNA (C), and lymphangiogenic marker (Lyve-1, Podoplanin, and VEGF-D; D) in subcutaneous tumors were evaluated via Western blot and RT-PCR analysis, respectively. In all cases, the data are expressed as the means of 3 independent experiments ±SD. *, P < 0.05, **, P < 0.01.

Close modal

Correlation between Shh expression and Akt activity, EMT, or lymphangiogenesis in gastric tumor specimens

Clinicopathologic analysis revealed that phospho-Akt, E-cadherin, or MMP-9 expression levels in tumors with lymph node metastases was higher than that in tumors without lymph node metastases (E-cadherin: P = 0.024; MMP-9: P = 0.019; Phospho-Akt: P = 0.056; Supplementary Table S3). Therefore, we examined whether or not expression of phospho-Akt, E-cadherin, or MMP-9 was associated with expression of Shh in tumors with lymph node metastases. Phospho-Akt and Shh had a weak correlation with regard to the level of expression (r = 0.126), but the correlations were not statistically significant (P = 0.183). Of greater importance, the level of expression of Shh also showed a strong correlation with that of E-cadherin or MMP-9 (E-cadherin: P < 0.001, r = −0.447; MMP-9: P < 0.001, r = 0.515; Table 2 and Supplementary Fig. S6A). Finally, to further evaluate the role of Shh signaling with respect to lymph node metastasis and lymphangiogenesis in human gastric tumor specimens, we explored the correlation of expression of Shh with lymphangiogenesis in 174 gastric cancer specimens. To quantitatively analyze the degree of tumor-associated lymphangiogenesis, tumor LVD was determined by quantification of D2-40 (a specific antibody for podoplanin-positive vessels; Supplementary Fig. S6B). Increased LVD was significantly associated with lymph node metastasis (P < 0.001) and tumor-node-metastasis (TNM) staging (P = 0.032, Supplementary Table S4). Importantly, higher rates of LVD were observed in patients in the Shh-positive groups, whereas low rates of LVD were found in patients in the Shh-negative groups (P < 0.001; Table 3). Furthermore, Shh expressions were also significantly correlated with the presence of lymphatic vessel invasion (LVI, P = 0.014). We also found that the phospho-Akt-positive group had a higher density of tumor lymphatic vessels compared with the phospho-Akt-negative group (P = 0.051). Taken together, these data suggested that the Shh signaling pathway regulates lymphangiogenesis via a mechanism involving the PI3K/Akt pathway.

Table 2.

Relationship between expression of Shh, E-cadherin, and MMP-9 in lymph node metastatic gastric cancer

Number of casesShhP
PositiveNegative
E-cadherin 
Positive  28 10 (35.7) 18 (64.3) <0.001 
Negative  100 73 (73) 27 (27)  
MMP-9 
Positive  97 68 (70.1) 29 (29.9) <0.001 
Negative  31 15 (48.4) 16 (51.6)  
Number of casesShhP
PositiveNegative
E-cadherin 
Positive  28 10 (35.7) 18 (64.3) <0.001 
Negative  100 73 (73) 27 (27)  
MMP-9 
Positive  97 68 (70.1) 29 (29.9) <0.001 
Negative  31 15 (48.4) 16 (51.6)  

NOTE: The values in parenthesis are expressed in percentage.

Table 3.

Relationship between expression of Shh and LVD or LVI in lymph node metastatic gastric cancer

Number of casesShh
PositiveNegativeP
LVD 
 Positive 98 63 (64.3) 35 (33.7) <0.001 
 Negative 76 42 (55.2) 34 (44.8)  
LVI 
 Positive 75 50 (66.7) 26 (33.3) 0.014 
 Negative 99 56 (56.6) 43 (43.4)  
Number of casesShh
PositiveNegativeP
LVD 
 Positive 98 63 (64.3) 35 (33.7) <0.001 
 Negative 76 42 (55.2) 34 (44.8)  
LVI 
 Positive 75 50 (66.7) 26 (33.3) 0.014 
 Negative 99 56 (56.6) 43 (43.4)  

NOTE: The values in parenthesis are expressed in percentage.

Abbreviation: LVI, lymphatic vessel invasion.

Abundant clinical and pathologic evidence indicates that Shh signaling is involved in tumorigenesis (18–20). Although some clinical studies have suggested that Shh expression in tumors is correlated with aggressiveness of gastric cancer (19, 21), none of these studies clearly indicate the exact role of Shh signaling in the metastatic progression of gastric cancer and its underlying mechanism. In this study, we investigated the functional role of Shh signaling in gastric cancer progression using multiple experiments, both in vitro and in vivo. The results of metastatic examination in vitro and in vivo showed that Shh functions as an inducer of gastric cancer metastasis and is correlated with clinical stage, lymph node metastasis, and prognosis in patients with gastric cancer.

The molecular mechanism by which Shh signaling regulates cell motility and invasiveness is an intriguing question. Studies using constitutively activated and dominant negative Akt mutants have shown that PI3K regulation of Shh signaling is mediated through Akt signaling (15). Therefore, this activated Akt is suggested to be involved in Shh signal transduction. Furthermore, Akt activation is also best understood in the context of a major cascade stimulating EMT-related pathway, cell migration, and invasion in various human cancers. Our results showed that during migration, Akt was strongly phosphorylated at Ser473 by Shh, and the phosphorylation of this serine residue has previously been found to be involved in the motility and invasiveness of tumor cells (22, 23). Furthermore, according to our finding, Akt modulates Shh signaling through the enhancement of Gli-dependent transcriptional activity. This agrees with the finding that PI3K/Akt regulates Gli2 transcriptional activity by PKA and GSK-3β phosphorylation in LIGHT cells (15). Although the role of PI3K/Akt in modulating the Shh pathway varies between cell types, Shh-induced modification of PKA and GSK-3β through AKT might be critical for Shh-mediated EMT and metastasis in gastric cancer. However, the precise mechanisms controlling Shh/Gli-mediated EMT and metastasis through PI3K/Akt signaling need to be elucidated in detail. In addition, the mechanism whereby Shh signaling induces PI3K/Akt activities is also yet to be defined. Interestingly, elevation of Akt is a response to the loss of thePTEN, which occurs frequently in metastatic gastric cancer (24–26). Therefore, downregulation of PTEN mediated by Shh might contribute to enhance the activity of Akt and thereby further promote tumor progression. Although PI3K/Akt activities are essential for ligand-dependent Shh signaling in metastatic function of gastric cancer cell, activation may not be sufficient to drive Gli activation. Recent studies have suggested that oncogenic K-RAS is directly involved in activation of the Shh pathway (20, 27, 28). Activating RAS mutations have been frequently detected in malignant cells in many gastric cancer patients. Therefore, oncogenic RAS, through the RAF/MEK/MAPK pathway, may play an important role in Shh-mediated metastatic potential.

Lymph node metastasis is a common occurrence in cancer, and lymphatic vessels have been reported as an important route contributing to metastasis of solid tumors (12–14). Our data showed that most patients without lymph node metastasis were in the Shh-negative group, whereas patients with lymph node involvement were Shh-positive group. Despite its clinical relevance, surprisingly little is known about the mechanisms of Shh leading to lymphatic metastasis in gastric cancer. Several studies have indicated that one possible molecular mechanism promoting metastasis to regional lymph nodes is tumor-induced lymphangiogenesis. Our findings indicate Shh signaling, via a mechanism involving the PI3K/Akt pathway, promotes tumor lymphangiogenesis and, consequently, lymph node metastasis, which is in good agreement with recent findings (5). However, it is not clear whether or not Shh signaling plays a role in the formation of the lymphatic vessels and whether or not Shh-mediated lymphangiogenesis is accompanied by enhanced lymphatic metastasis. This problem may be solved with the establishment of a new model of lymphangiogenesis being able to capture the entire network of lymphatics and spread of cancer cells from the site of inoculation.

Our observations improve our understanding of the mechanism by which Shh signaling activation occurs as it relates to the metastatic behavior of gastric cancer. Investigating the expression of Shh and its related proteins in gastric cancer cells may be helpful for determining the therapeutic strategy and antagonists of this signaling may be useful in the clinical setting to suppress the spread of tumors through the lymphatic system.

No potential conflicts of interest were disclosed.

The authors thank Dr. Jae Hun Cheong for providing Gli1 promoter luciferase constructs.

This work was supported by the Korea Science and Engineering Foundation (KOSEF) grant funded by the Korea government (MEST; no. 2010-0000637), the Brain Korea 21 Project of the Ministry of Education and Human Resources Development, Republic of Korea, a Korea University Grant, and a grant of the Korea Healthcare technology R&D Project, Ministry for Health, Welfare & Family Affairs, Republic of Korea (A084953).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1.
Hooper
JE
,
Scott
MP
. 
Communicating with hedgehogs
.
Nat Rev Mol Cell Biol
2005
;
6
:
306
17
.
2.
Ingham
PW
,
McMahon
AP
. 
Hedgehog signaling in animal development: paradigms and principles
.
Genes Dev
2001
;
15
:
3059
87
.
3.
Asai
J
,
Takenaka
H
,
Kusano
KF
,
Ii
M
,
Luedemann
C
,
Curry
C
, et al
Topical sonic hedgehog gene therapy accelerates wound healing in diabetes by enhancing endothelial progenitor cell-mediated microvascular remodeling
.
Circulation
2006
;
113
:
2413
24
.
4.
Hochman
E
,
Castiel
A
,
Jacob-Hirsch
J
,
Amariglio
N
,
Izraeli
S
. 
Molecular pathways regulating pro-migratory effects of hedgehog signaling
.
J Biol Chem
2006
;
281
:
33860
70
.
5.
Bailey
JM
,
Mohr
AM
,
Hollingsworth
MA
. 
Sonic hedgehog paracrine signaling regulates metastasis and lymphangiogenesis in pancreatic cancer
.
Oncogene
2009
;
28
:
3513
25
.
6.
Feldmann
G
,
Dhara
S
,
Fendrich
V
,
Bedja
D
,
Beaty
R
,
Mullendore
M
, et al
Blockade of hedgehog signaling inhibits pancreatic cancer invasion and metastases: a new paradigm for combination therapy in solid cancers
.
Cancer Res
2007
;
67
:
2187
96
.
7.
Park
JG
,
Yang
HK
,
Kim
WH
,
Chung
JK
,
Kang
MS
,
Lee
JH
, et al
Establishment and characterization of human gastric carcinoma cell lines
.
Int J Cancer
1997
;
70
:
443
9
.
8.
Kitano
E
,
Kitamura
H
. 
Synthesis of the third component of complement (C3) by human gastric cancer-derived cell lines
.
Clin Exp Immunol
1993
;
94
:
273
8
.
9.
Yoo
YA
,
Kang
MH
,
Kim
JS
,
Oh
SC
. 
Sonic hedgehog signaling promotes motility and invasiveness of gastric cancer cells through TGF-beta-mediated activation of the ALK5-Smad 3 pathway
.
Carcinogenesis
2008
;
29
:
480
90
.
10.
Kang
MH
,
Kim
JS
,
Seo
JE
,
Oh
SC
,
Yoo
YA
. 
BMP2 accelerates the motility and invasiveness of gastric cancer cells via activation of the phosphatidylinositol 3-kinase (PI3K)/Akt pathway
.
Exp Cell Res
2010
;
316
:
24
37
.
11.
Huber
MA
,
Kraut
N
,
Beug
H
. 
Molecular requirements for epithelial-mesenchymal transition during tumor progression
.
Curr Opin Cell Biol
2005
;
17
:
548
58
.
12.
Alitalo
K
,
Tammela
T
,
Petrova
TV
. 
Lymphangiogenesis in development and human disease
.
Nature
2005
;
438
:
946
53
.
13.
Achen
MG
,
Stacker
SA
. 
Tumor lymphangiogenesis and metastatic spread-new players begin to emerge
.
Int J Cancer
2006
;
119
:
1755
60
.
14.
Pepper
MS
,
Tille
JC
,
Nisato
R
,
Skobe
M
. 
Lymphangiogenesis and tumor metastasis
.
Cell Tissue Res
2003
;
314
:
167
77
.
15.
Riobo
NA
,
Lu
K
,
Ai
X
,
Haines
GM
,
Emerson
CP
 Jr
. 
Phosphoinositide 3-kinase and Akt are essential for sonic hedgehog signaling
.
Proc Natl Acad Sci U S A
2006
;
103
:
4505
10
.
16.
Hahn
H
,
Wojnowski
L
,
Specht
K
,
Kappler
R
,
Calzada-Wack
J
,
Potter
D
, et al
Patched target Igf2 is indispensable for the formation of medulloblastoma and rhabdomyosarcoma
.
J Biol Chem
2000
;
275
:
28341
4
.
17.
Rao
G
,
Pedone
CA
,
Del Valle
L
,
Reiss
K
,
Holland
EC
,
Fults
DW
. 
Sonic hedgehog and insulin-like growth factor signaling synergize to induce medulloblastoma formation from nestin-expressing neural progenitors in mice
.
Oncogene
2004
;
23
:
6156
62
.
18.
Berman
DM
,
Karhadkar
SS
,
Maitra
A
,
Montes De Oca
R
,
Gerstenblith
MR
,
Briggs
K
, et al
Widespread requirement for Hedgehog ligand stimulation in growth of digestive tract tumours
.
Nature
2003
;
425
:
846
51
.
19.
Ma
X
,
Chen
K
,
Huang
S
,
Zhang
X
,
Adegboyega
PA
,
Evers
BM
, et al
Frequent activation of the hedgehog pathway in advanced gastric adenocarcinomas
.
Carcinogenesis
2005
;
26
:
1698
705
.
20.
Xie
K
,
Abbruzzese
JL
. 
Developmental biology informs cancer: the emerging role of the hedgehog signaling pathway in upper gastrointestinal cancers
.
Cancer Cell
2003
;
4
:
245
7
.
21.
Wang
LH
,
Choi
YL
,
Hua
XY
,
Shin
YK
,
Song
YJ
,
Youn
SJ
, et al
Increased expression of sonic hedgehog and altered methylation of its promoter region in gastric cancer and its related lesions
.
Mod Pathol
2006
;
19
:
675
83
.
22.
Suzuki
A
,
Lu
J
,
Kusakai
G
,
Kishimoto
A
,
Ogura
T
,
Esumi
H
. 
ARK5 is a tumor invasion-associated factor downstream of Akt signaling
.
Mol Cell Biol
2004
;
24
:
3526
35
.
23.
Zi
X
,
Guo
Y
,
Simoneau
AR
,
Hope
C
,
Xie
J
,
Holcombe
RF
, et al
Expression of Frzb/secreted Frizzled-related protein 3, a secreted Wnt antagonist, in human androgen-independent prostate cancer PC-3 cells suppresses tumor growth and cellular invasiveness
.
Cancer Res
2005
;
65
:
9762
70
.
24.
Sansal
I
,
Sellers
WR
. 
The biology and clinical relevance of the PTEN tumor suppressor pathway
.
J Clin Oncol
2004
;
22
:
2954
63
.
25.
Yang
L
,
Kuang
LG
,
Zheng
HC
,
Li
JY
,
Wu
DY
,
Zhang
SM
, et al
PTEN encoding product: a marker for tumorigenesis and progression of gastric carcinoma
.
World J Gastroenterol
2003
;
9
:
35
9
.
26.
Kang
YH
,
Lee
HS
,
Kim
WH
. 
Promoter methylation and silencing of PTEN in gastric carcinoma
.
Lab Invest
2002
;
82
:
285
91
.
27.
Thayer
SP
,
di Magliano
MP
,
Heiser
PW
,
Nielsen
CM
,
Roberts
DJ
,
Lauwers
GY
, et al
Hedgehog is an early and late mediator of pancreatic cancer tumorigenesis
.
Nature
2003
;
425
:
851
6
.
28.
Xie
J
,
Aszterbaum
M
,
Zhang
X
,
Bonifas
JM
,
Zachary
C
,
Epstein
E
, et al
A role of PDGFRalpha in basal cell carcinoma proliferation
.
Proc Natl Acad Sci U S A
2001
;
98
:
9255
9
.