Hedgehog signaling pathway mediates invasion and metastasis of hepatocellular carcinoma via ERK pathway.

AIM
To investigate the role of Hedgehog (Hh) signaling pathway in the invasion and metastasis of human hepatocellular carcinoma (HCC).


METHODS
Eighty six HCC tissues samples and HCC cell line Bel-7402 were examined. The protein expression of sonic hedgehog (Shh), nuclear glioma-associated oncogene-1 (Gli1), MMP-9 and p-ERK1/2 in HCC was analyzed using immunohistochemistry and Western blot analysis. Boyden chamber assay and wound-healing assay were used to quantify the invasion and metastasis of Bel-7402 cells.


RESULTS
In 86 HCC tissue samples, the positive ratio of Shh and nucleus Gli1 was 67.44% (58/86) and 60.47% (52/86), respectively; the expression of nucleus Gli1 was correlated with the tumor pathological grade (P=0.034), and with the ability of the tumor to invade and metastasize (P=0.001); the expression of nucleus Gli1 was also correlated with p-ERK1/2 (P=0.031) and with MMP-9 (P=0.034). Neither Shh, nor nucleus Gli1 was observed in normal liver tissue. KAAD-cyclopamine (KAAD-cyc), a specific inhibitor of the Hh pathway, at the concentrations of 1 and 4 μmol/L inhibited the invasion and migration of Bel-7402 cells and decreased the expression of Gli1 in nucleus and MMP-9, p-ERK1/2 proteins in Bel-7402 cells. On the other hand, Shh, a ligand of the Hh pathway, at the concentration of 0.5 μg/mL produced opposite effects. The MAPK pathway inhibitors U0126 and PD98059 at the concentrations of 5 and 10 μmol/L inhibited invasion and metastasis of Bel-7402 cells induced by Shh, and decreased the expression of p-ERK1/2 and MMP-9. However, U0126 and PD98059 had no effect on the expression of Gli1.


CONCLUSION
Hh signaling pathway mediates invasion and metastasis of human HCC by up-regulating the protein expression of MMP-9 via ERK pathway.


Introduction
Hepatocellular carcinoma (HCC) is the common primary cancer with a multifaceted molecular pathogenesis. HCC has become the fifth most prevalent malignancy worldwide and the third leading cause of cancer-related death, most importantly, the incidence of HCC is increasing [1,2] , and 82% of cases are in developing countries, with 55% in China alone [3] . Invasion and metastasis are two fundamental properties, which determine the prognosis of the HCC patients [4,5] . Many signaling pathways are thought to be involved in the development and invasion of HCC, including the MAPK pathway [6,7] , phosphatidylinositol-3 kinase (PI3K)/AKT/mammalian target of rapamycin (mTOR) pathway [8][9][10] , the wnt/beta-catenin pathway [9,11] , hepatocyte growth factor/c-MET pathway [12,13] , hedgehog (Hh) signaling pathway, and so on.
Hh signaling pathway is a highly conserved system, which plays a crucial role in tissue patterning, cell differentiation and proliferation [14] . Hedgehog, including sonic hedgehog (Shh), Indian hedgehog (Ihh), and desert hedgehog (Dhh), appear to bind to a transmembrane receptor protein, known as Patched (Ptc), which in the absence of Shh exerts an inhibitory effect on the seven transmembrane receptor smoothened (Smo). Binding of Shh to Ptc alleviates the inhibitory effect of Ptc on Smo. Once activated, Smo induces a complex series of intracellular reactions that targets the glioma-associated oncogenes (Gli) transcription factor families, the zincfinger transcription effectors [15,16] . At least three members (Gli1, Gli2, and Gli3) www.nature.com/aps Lu JT et al Acta Pharmacologica Sinica npg of nuclear proteins Gli families have been identified in mammalian tissues [14,17] . Gli1 is a transcriptional activator [18] , which induces the expression of numerous target genes that regulates proliferation, differentiation, and extracellular matrix interactions [19,20] . Increasing evidence has demonstrated that the Hh signaling pathway plays an important role in multiple tumor types, for example, basal cell carcinoma [21,22] , pancreatic cancer [23] , colon carcinoma [24] , gastric cancer [25] , small cell lung cancer [26] , prostate cancer [27][28][29] , breast cancer [30,31] , early childhood hepatoblastoma [32] , and esophageal cancers [33] .
Recent studies have revealed that the Hh signaling pathway is abnormally activated in human HCC [34][35][36] , and this pathway is thought to participate in the development of HCC [37][38][39] . Moreover, activation of the Hh pathway is correlated closely to invasion and metastasis of HCC [36,40] . Activated markers of the Hh signaling pathway such as Gli1 is significantly upregulated in the HCC tumor tissues [41] , and Gli inhibition can also suppress HCC tumor growth and metastases in vivo and in vitro [42] . However, the mechanism by which Hh signaling pathway is involved in HCC development is still unclear.
Hh signaling pathway can affect MAPK/ERK phosphorylation [43] and lead to carcinogenesis [44] . For example, it also can activate ERK1/2 in the breast epithelial cell [45] . However, other studies have revealed that in cancer the MAPK signaling pathway regulates Hh signaling, specifically Gli activity and expression [46] . Therefore, what cellular mediators are involved in the crosstalk between the Hh signaling pathway and the MAPK signaling pathway in HCC? Now, there is no report about it.
Matrix metalloproteinase-9 (MMP-9 or gelatinase-B) is mostly associated with tumor migration, invasion and metastasis for various human cancers [10,47,48] . The Hh signaling pathway up-regulates cell migration and invasion in human gliomas and in pancreatic cancer by increasing the expressions of MMP-9 [49,50] . In addition, Smo and MMP-9 were overexpressed and associated with invasion and metastasis in HCC tissues [51] . Does the Hh signaling pathway mediate the migration and invasion of HCC by increasing the expressions of MMP-9? What is the correlation between MMP-9 and ERK pathway? However, they are incompletely understood in HCC.
In the present study, we investigated the mechanisms of the Hh signaling pathway in invasion and metastasis of HCC, specifically focused on the correlation between the Hh signaling pathway and the MAPK signaling pathway. Further we determined whether Hh signaling pathway involved in human HCC invasion and metastasis by up-regulating the expression of MMP-9 through ERK pathway.

Materials and methods
Patients and specimens A total of 86 HCC patients, who had undergone liver resection without preoperative treatment at the First Affiliated Hospital of Anhui Medical University between June 2008 and December 2010, were examined. All tumor specimens were patho-logically diagnosed as HCC. Prior written informed consent was obtained from all patients according to the World Medical Association Declaration of Helsinki, and the study received ethics board approval from the Affiliated Hospital of Anhui Medical University. The age of patients ranged from 33 to 75 years, with an average age of 48.1 years. There were 68 males and 18 females, 50 I-II type-and 36 III-IV type-differentiated HCC, 37 patients with invasion and/or metastasis (Table 1).

Cell line and culture conditions
Human HCC cell line Bel-7402, obtained from the Shanghai Institute of Cell Biology, Chinese Academy of Sciences, was grown in DMEM supplemented with 10% heat-inactivated fetal bovine serum (FBS, Gbico, USA), 100 μg/mL streptomycin and 100 U/mL penicillin in a humidified atmosphere containing 5% CO 2 at 37 °C. Cells during exponential growth phase were used in the experiments. Cells were treated with KAAD-cyc, the special antagonist of Smo, Shh, U0126, or PD98059, control medium contained DMSO alone. In the invasion and metastasis experiments, cells were cultured in a serum-free medium.

Boyden chamber invasion assay
The effect of KAAD-cyc, Shh, U0126, and PD98059 on HCC Bel-7402 was determined using Boyden chamber assay. Briefly, 24-well transwell units with polycarbonate membrane filters (8 μm pore size, Costar, USA) were coated with 100 μL matrigel (25 μg in 100 μL PBS, Becton Dickinson), dried in a laminar hood overnight, and reconstituted in 100 µL, washed with phosphate-buffered saline (PBS) at 37 °C for 2 h, then PBS was discharged. Bel-7402 cells were resuspended in DMEM with 0.5% BSA (5×10 4 cells/200 μL) in the presence or absence of KAAD-cyc, Shh, U0126, or PD98059, which was added to the upper side of the invasion chamber. DMEM (500 μL) with 2.5% FBS as chemoattractant was added to the lower chamber. After 24 h, filter inserts were removed from the wells, the cells on the upper surface of the filter were wiped off using cotton swabs. The cells that penetrated to the lower surface were fixed with 4% paraformaldehyde, stained with 0.1% crystal violet in 20% ethanol, and counted in five randomly selected fields under phase contrast microscope. The invasion cells were monitored by photographing at 400× magnification with Olympus Microscope. The assay was performed in triplicate.
Wound-healing assay Cell migration was examined using the wound-healing assay. Briefly, Bel-7402 cells were cultured to about 80%-90% confluence in a 6-well plate at 37 °C and 5% CO 2 . A wound about 1 mm width was created by scratching cells with a sterile 100 μL micropipette tip. Cells were washed with PBS (pH 6.8) three times to remove floating cells, then 1 mL serum-free DMEM was added. A computer-based microscopy imaging system was used to determine wound healing at 0 h with a microscope at 200× magnification. Then 1 mL serum-free DMEM was added with different concentrations of KAAD-cyc, Shh, U0126, or PD98059. After 24 h, photoes of the wound were taken under 200× magnification. The values of wound-healing were assessed by measuring the pixel of wound area by Photoshop 7.01 software. The experiments were performed in triplicate.

Statistical analysis
Statistical analyses were performed using the SPSS 11.0 software program (SPSS Software Products, Chicago, IL, USA). All data were presented as number or mean±standard deviation (SD). Associations between protein expression and clinicopathologic variables were analyzed by ChiSquare Test. Statistical analysis among more groups was performed by oneway analysis of variance (ANOVA). The Spearman coefficient of correlation was used to examine the correlation. Statistical significance of differences were accepted at P<0.05.

Results
Overexpression of Shh, Gli1, p-ERK1/2, and MMP-9 in HCC liver tissues with invasion and metastasis compared with nonmetastasis HCC liver tissue We detected Shh, Gli1 MMP-9, and p-ERK1/2 expressions in 86 cases of HCC liver tissues with or without invasion and metastasis by IHC staining. We further evaluated the relationships between Gli1 and p-ERK1/2, Gli1 and MMP-9 on invasion and metastasis of HCC. The results indicated Shh, Gli1, p-ERK1/2, and MMP-9 expressions had no notable relationship with age and tumor diameter. However, Gli1, p-ERK1/2, and MMP-9 expressions had a significant correlation with the pathological grade and metastasis of the tumor sample (Table 1). Positive expressions of Shh, p-ERK1/2, and MMP-9 were remarkably stronger in HCC liver tissues with metastasis than in non-metastasis HCC liver tissues. A significant difference was observed in expression of Gli1 in the nucleus between HCC tissues with metastasis and non-metastatic HCC liver tissue (91.89% vs 36.74%, P<0.01, Figure 1). Interestingly, www.nature.com/aps Lu JT et al Acta Pharmacologica Sinica npg expression of Gli1 was also notably correlated to expressions of MMP-9 and p-ERK1/2 (P<0.01, Table 1). Those results suggested that Hh signal pathway mediated invasion and metastasis of human HCC by up-regulating the protein expression MMP-9 and p-ERK1/2.

KAAD-cyc inhibited the invasion and migration of Bel-7402 cells
Bel-7402 cell invasion and motility were investigated using a Boyden chamber invasion assay and wound-healing assay, respectively. KAAD-cyc is a specific inhibitor of the Hh signaling pathway and was utilized to determine the effect of this pathway on invasion and metastasis in Bel-7402 cells.
The results indicated KAAD-cyc notably inhibited migration of Bel-7402 cells on the surface of the tissue culture plate, significantly decreased area of wound-healing by 45.87% at most compared with controls in the wound-healing assay.  Meanwhile, KAAD-cyc significantly decreased the numbers of cells to the lower chamber when the cells were treated with 1 µmol/L and 4 µmol/L of KAAD-cyc for 24 h in Boyden chamber invasion assay, inhibitory rates were 43.50%±15.41% and 56.36%±15.17%, respectively. Collectively, these data demonstrated that KAAD-cyc could suppress the invasion and metastasis of Bel-7402 cells (Figure 2).

Shh increased invasion and migration of Bel-7402 cells
As illustrated in Figure 3, Shh notably improved the migration of Bel-7402 cells, significantly increased area of woundhealing by greater than 41.63%. At the same time, Shh could significantly increase the number of migrating Bel-7402 cells in the lower chamber when the cells were treated with 0.13 µg/mL and 0.5 µg/mL Shh for 24 h, incremental rates were 49.99%±14.04% with 0.13 µg/mL Shh and 69.28%±20.29% with 0.5 µg/mL Shh in invasion assays. Those data demonstrated  Effects of KAAD-cyc on the expression of Gli1, p-ERK1/2, and MMP-9 proteins in Bel-7402 cells Bel-7402 cells were cultured in serum-free media containing 1 and 4 μmol/L KAAD-cyc for 24 h. The expression of Gli1, p-ERK1/2, and MMP-9 proteins were determined by Western blot analysis. The results indicated that KAAD-cyc dramatically inhibited the expression of Gli1, p-ERK1/2, and MMP-9 proteins in Bel-7402 compared with vehicle control (Figure 5A and 5B).
Effects of Shh on the expression of Gli1, p-ERK1/2, and MMP-9 proteins in Bel-7402 cells Bel-7402 cells were cultured in serum-free media containing 0.13 and 0.50 μg/mL Shh for 24 h. The expression of Gli1, p-ERK1/2, and MMP-9 proteins were determined by Western blot analysis. The results indicated that Shh dramatically increased the expression of Gli1, p-ERK1/2, and MMP-9 in

Discussion
Hepatocellular carcinoma (HCC) is one of the most malignant cancers especially in Asian countries, and its poor prognosis is mainly due to metastasis after excision [52] . Currently there are no available effective treatment [53,54] . Studies of the underlying molecular mechanism of HCC metastasis provide potential to identify new therapeutic targets. Previous evidence has demonstrated that the Hh signaling plays an important role in multiple tumor types. This signaling pathway is involved in and participates development, invasion and metastasis of HCC [34][35][36][37][38][39][40] . In this study, we further detected expression of Shh and Gli1 in metastasis and non-metastasis HCC liver tissues, simultaneously we used KAAD-cyc, a specific inhibitor of Hh pathway and Shh, a ligand of the Hh pathway in invasion and metastasis assays of human HCC cell line, and we confirmed the results. However, to date no report has been presented to identify a potential mechanism. Therefore we investigated the mechanisms of the Hh signaling pathway in HCC invasion and metastasis. Gli1, a transcription factor activated in the Hh pathway, significantly enhanced tumor growth and metastases of other cancers through the activation of ERK1/2 [42,45] .
Overexpression of MMP-9 is also a key factor for tumor invasion and metastasis [47,48,55] . Moreover, the Hh signaling pathway may enhance migration and invasion of cancer cells by increasing the expression of MMP-9 [49,50] .
In the present study, we found that expressions of Shh, MMP-9, and p-ERK1/2 were remarkably stronger in HCC samples with metastasis than in non-metastasis HCC liver samples. Moreover, there was a significant difference in expression of Gli1 in nuclei of cells in HCC tissue samples with metastasis compared to and HCC samples with no metastasis. Moreover, expression of Gli1 was also positively correlated to expressions of both MMP-9 and p-ERK1/2. These data suggested that the Hh signal pathway may be involved in human HCC invasion and metastasis by up-regulating the protein expression MMP-9 and p-ERK1/2.
Since the Hh signaling pathway has been associated with MAPK/ERK pathway in different cancer [43][44][45][46] , furthermore our results indicate that the pathway is also notably correlated to expressions of p-ERK1/2 in HCC tissues. To confirm these results, we treated a HCC cell line, Bel-7402 cells, with KAAD-cyc, Shh, U0126, or PD98059 in vitro. Our results demonstrated that KAAD-cyc inhibited the invasion and migration of Bel-7402 cells and decreased the expression of Gli1, p-ERK1/2 proteins in Bel-7402, but Shh increased the expression of Gli1, p-ERK1/2 proteins. U0126 and PD98059 inhibited the invasion and metastasis of Bel-7402 cells induced by Shh, decreased the expression of Gli1 proteins, but they had no effect on the expression of Gli1. These indicate that the Hh signal pathway is involved in human HCC invasion and metastasis by up-regulating the expression of MAPK/ ERK pathway, instead of the MAPK/ERK pathway inducing expression of Gli1.
Matrix metalloproteinases (MMPs) have long been associ-ated with cancer cell invasion and metastasis. MMPs are proteolytic enzymes, their basic mechanism of action is to degrade proteins in extracellular matrix. Activation of MMPs has been detected in almost all type of human cancers, which is closely correlated to advanced tumor stage, increasing tumor invasion and metastasis. Onishi et al reported that the Hh signaling pathway up-regulated cell migration and invasion in human cancers by increasing expressions of MMP-9 [50] . However, the relationship among the Hh signaling pathway, MAPK/ERK pathway and MMP-9 remains unclear. Our results indicate that the Hh pathway is also correlated with expressions of p-ERK1/2 and MMP-9 in HCC tissues. In Bel-7402 cellular assays, KAAD-cyc inhibited expression of MMP-9 proteins, and Shh up-regulated expression of MMP-9 proteins. Both U0126 and PD98059 were able to inhibit the expression of MMP-9 elevated by Shh. Therefore, we concluded that the Hh signaling pathway mediated the protein expression MMP-9 through MAPK/ERK pathway. Most importantly, we have confirmed that the Hh signal pathway is involved in human HCC invasion and metastasis. We hypothesize and also deduce that the Hh signaling pathway activates the ERK pathway, subsequently, the ERK pathway up-regulates the protein expression of MMP-9, thereby mediating human HCC invasion and metastasis. These data may assist in identifying novel diagnostic markers and therapeutic targets for the treatment of highly aggressive HCC.