Original Article | Published:

ACP5, a direct transcriptional target of FoxM1, promotes tumor metastasis and indicates poor prognosis in hepatocellular carcinoma

Oncogene volume 33, pages 13951406 (13 March 2014) | Download Citation

Abstract

Tartrate-resistant acid phosphatase 5 (ACP5), which is essential for bone resorption and osteoclast differentiation, promotes cell motility through the modulation of focal adhesion kinase phosphorylation. However, whether ACP5 contributes to the metastasis and progression of hepatocellular carcinoma (HCC) remains unknown. In this paper, a complementary DNA microarray, serial deletion, site-directed mutagenesis and a chromatin immunoprecipitation assays confirmed that ACP5 is a direct transcriptional target of Forkhead box M1 (FoxM1). ACP5 expression was markedly higher in HCC tissues compared with adjacent noncancerous tissues. ACP5 overexpression was correlated with microvascular invasion, poor differentiation and higher tumor-node-metastasis stage. HCC patients with positive ACP5 expression had poorer prognoses than those with negative ACP5 expression. A multivariate analysis revealed that ACP5 expression was an independent and significant risk factor for disease recurrence and reduced-patient survival following curative resection. Transwell assays and an orthotopic metastatic model showed that the upregulation of ACP5 promoted HCC invasion and lung metastasis, whereas ACP5 knockdown inhibited these processes. The knockdown of ACP5 significantly attenuated FoxM1-enhanced invasion and lung metastasis. Immunohistochemistry revealed that ACP5 expression was positively correlated with FoxM1 expression in human HCC tissues, and their coexpression was associated with poor prognoses. In summary, ACP5 is a direct transcriptional and functional target of FoxM1. This novel FoxM1/ACP5 signaling pathway promotes HCC metastasis and may be a candidate biomarker for prognosis and a target for new therapies.

Introduction

Hepatocellular carcinoma (HCC) is the fifth most common cancer and the third most common cause of cancer mortality worldwide.1 Metastasis is the major reason for the high mortality of HCC patients after curative resection.2, 3 Nonetheless, the molecular mechanisms underlying HCC metastasis remain largely unclear.

Forkhead box M1 (FoxM1), which is a member of the Fox transcription factor family, is a master regulator of tumor metastasis.4 FoxM1 induces an epithelial-mesenchymal-like transition phenotype and increases cell migration and invasion by transactivating MMP-2, MMP-9, JNK1 and Caveolin expression.5, 6, 7, 8, 9 FoxM1 also induces a pre-metastatic niche at the distal organ of metastasis by transactivating lysyl oxidase and lysyl oxidase-like 2 expressions.10 Interestingly, several recent studies indicated that FoxM1 has an essential role in the development of HCC. FoxM1 is necessary for hepatocyte DNA replication and mitosis during mouse liver regeneration.11, 12 Conditionally deleted FoxM1-mouse hepatocytes failed to proliferate and were highly resistant to the development of HCC in response to diethylnitrosamine.13 A p19ARF peptide that inhibited FoxM1 transcriptional activity efficiently diminished HCC proliferation and induced apoptosis of HCC cells in FoxM1 transgenic mice.14 Furthermore, FoxM1 overexpression has been associated with the poor prognoses of HCC patients.15 Another study reported that FoxM1 overexpression predicted a worse clinical outcome in HCC patients following orthotopic liver transplantation.16 In a previous study, we found that FoxM1 promoted HCC metastasis by transactivating MMP-7, RhoC and ROCK1 expression. We also found that FoxM1 expression was an independent risk factor for recurrence and the reduced survival of HCC patients following curative resection.17, 18 These studies strongly suggest that FoxM1 contributes to the metastasis and malignant progression of HCC. However, the molecular mechanism underlying the promotion of HCC metastasis by FoxM1 remains unclear.

To further delineate the manner by which FoxM1 promotes the metastases of HCC cells, we conducted a detailed comparison of gene expression in SMMC7721-FoxM1 cells (stable transfection of FoxM1) and SMMC7721-control cells using a complementary DNA microarray. In this analysis, we focused on genes that are involved in metastasis. Of particular interest was tartrate-resistant acid phosphatase 5 (ACP5), which was upregulated by 12.19-fold in response to FoxM1 overexpression (Table 1, Supplementary Figure S1). ACP5 is a metalloprotein enzyme that belongs to the acid phosphatase family and is known to be expressed by osteoclasts. Furthermore, it is a classic marker for bone resorption and osteoclast differentiation.19 ACP5 expression has been shown to be significantly upregulated in breast and ovarian cancers and melanoma20, 21 and is a useful serum marker for extensive bone metastasis.22, 23, 24, 25, 26 Recent studies have reported that ACP5 promotes the invasion and distal metastasis of melanoma and breast cancer cells through the modulation of focal adhesion kinase (FAK) phosphorylation. Additionally, ACP5 overexpression is indicative of the poor prognoses of human melanoma patients.27 These studies suggest that ACP5 may have an important role in promoting tumor metastasis. However, its role in HCC metastasis remains unknown.

Table 1: Metastatic-related genes upregulated after FoxM1 overexpression in SMMC7721 cells

To date, no studies have reported the clinicopathologic significance of ACP5 in HCC. In the present study, we provide the first evidence that ACP5, which is a direct transcriptional target of FoxM1, promotes HCC metastasis and is associated with the poor prognoses of human HCC patients following curative resection. The knockdown of ACP5 attenuates FoxM1-enhanced HCC invasion and metastasis. Additionally, ACP5 expression is positively correlated with FoxM1 expression in human HCC tissues. Thus, this FoxM1/ACP5 signaling pathway may have an important role in HCC metastasis.

Results

FoxM1 upregulates ACP5 expression in human HCC cells

To determine the manner by which FoxM1 promotes invasion and metastasis in HCC cells, we compared the gene expression patterns in SMMC7721-FoxM1 cells and SMMC7721-control cells, focusing on genes that are involved in metastasis. FoxM1 overexpression substantially upregulated the expression of a number of metastasis-related genes, including CA9, ACP5, ANGPTL4, IGFBP3, CXCR1, MMP24 and so on (Table 1). The changes in expression levels of these downstream targets were further validated by real-time PCR in four different cell lines (Supplementary Figures S1 and 9).

Of particular interest was ACP5, which was upregulated by 12.19-fold in response to FoxM1 overexpression (Table 1). Considering the critical role of ACP5 in metastasis, we wanted to determine whether it was involved in FoxM1-mediated HCC metastasis. Real-time PCR and western blot assays showed that FoxM1 upregulated ACP5 expression in Huh7 and SMMC7721 cells (low metastatic potential). However, the knockdown of FoxM1 expression decreased ACP5 expression in HCCLM3 and MHCC97H cells (high metastatic potential) (Figure 1a).

Figure 1
Figure 1

ACP5 is a direct transcriptional target of FoxM1. (a) FoxM1 upregulates ACP5 expression. Following the infection of SMMC7721 and HCCLM3 cells with LV-FoxM1 or LV-shFoxM1, ACP5 mRNA and protein levels were detected using real-time PCR and western blotting, respectively. (b) FoxM1 transactivates ACP5 promoter activities. The ACP5 promoter luciferase construct (−2076/+33)ACP5 was co-transfected with pCMV-FoxM1, and promoter activities were detected using a luciferase reporter assay. (c) A sequence analysis of the 5′-flanking region of the human ACP5 gene revealed four putative FoxM1-binding sites, three of which overlapped. The transcription start site is also indicated. (d) Deletion and selective mutation analyses identified FoxM1-responsive regions in the ACP5 promoter. Serially truncated and mutated ACP5 promoter constructs were co-transfected with pCMV-FoxM1, and relative luciferase activities were determined. The schematic constructs are shown (left), and the bar graphs present the relative levels of luciferase activity in each of the samples (right). (e) A chromatin immunoprecipitation assay demonstrated the direct binding of FoxM1 to the ACP5 promoter in HCC cells. Real-time PCR was performed to detect the amounts of immunoprecipitated products. (f) FoxM1 directly binds to the ACP5 promoter in HCC tissues. The hepatocytes were separated from the liver tissues of HCC patients and healthy controls (HC). The cells were cross-linked and the chromatin were immunoprecipitated by anti-FoxM1 or control antibody, respectively. *P<0.05.

ACP5 is a direct transcriptional target of FoxM1

To determine whether FoxM1-induced ACP5 expression is mediated through the transactivation of the ACP5 promoter, Huh7 and SMMC7721 cells were co-transfected with a reporter plasmid containing the promoter of the human ACP5 gene (−2076/+33)ACP5 and pCMV-FoxM1 or pCMV-Taq. Luciferase reporter assays showed that the ACP5 promoter activity was significantly enhanced in the cells transfected with pCMV-FoxM1 (Figure 1b).

Previous studies have reported that the FoxM1-consensus binding sequence is 5′-A(C/T)AAA(C/T)AA-3′ and 5′-TAATCA-3′.28 We found four putative FoxM1-binding sites in the ACP5 promoter, three of which overlapped (Figure 1c). To define the roles of the cis-regulatory elements of the ACP5 promoter in response to FoxM1 regulation, a series of ACP5 promoter truncation mutants were generated. The reporter constructs containing serial 5′ deletions of the ACP5 promoter (−2076/+33)ACP5, (−893/+33)ACP5, (−573/+33)ACP5 and (−221/+33)ACP5) were co-transfected with pCMV-FoxM1 into SMMC7721 cells. The luciferase reporter assay showed that a deletion from nt-2076 to nt-573 had no effect on FoxM1-induced ACP5 promoter activity. The extension of this deletion from nt-573 to nt-221 significantly decreased the FoxM1-induced ACP5 promoter activity (Figure 1d), indicating that the sequence that is located between nt-573 and nt-221 is critical for the activation of the ACP5 promoter as regulated by FoxM1. The three overlapping FoxM1-binding sites are located in this region, and thus, site-directed mutagenesis was used to mutate them. Luciferase reporter assays showed that mutation of these FoxM1-binding sites significantly reduced the FoxM1-induced transactivation of the ACP5 promoter (Figure 1d).

To determine whether FoxM1 is recruited to the ACP5 promoter, chromatin immunoprecipitation assays were performed using chromatin that had been prepared from the cells that were transfected with pCMV-FoxM1 or pCMV-tag. The 99-bp DNA fragment containing the second FoxM1-binding site in the ACP5 promoter was pulled down by the FoxM1 antibody but not by the negative control antibody (Figure 1e). To determine whether FoxM1 binds to the ACP5 promoter in the physiological level, three normal liver tissues (health control) and three HCC tissues were collected. A chromatin immunoprecipitation assay showed that the FoxM1-binding activity to the ACP5 promoter was much higher in HCC tissues than in healthy control (Figure 1f). Taken together, these data suggest that ACP5 is a direct transcriptional target of FoxM1.

ACP5 is significantly upregulated in human HCC tissues and indicates poor prognosis

To explore the potential role of ACP5 in determining the clinical outcomes of the HCC patients, we assessed its expression in a tissue microarray of 406 HCC patients. The immunohistochemistry results showed that ACP5 was primarily localized to the cytoplasm (Figure 2a, Supplementary Figure S10). Positive ACP5 expression was found in 223 out of 406 (54.9%) primary HCC samples compared with only 85 out of 406 (20.9%) in adjacent nontumor tissues (Figure 2b, P<0.001). To investigate the role of ACP5 in HCC metastasis, ACP5 expression was compared in primary and metastatic HCCs using an immunohistochemical assay in 20 pairs of HCC specimens. Metastatic HCC tumors (n=20), included nine lymph node, seven peritoneal, one kidney, one adrenal cortex and two bone metastases. Twelve of them had positive expression of ACP5 (Supplementary Figure S12). Overall, 12 pairs of HCCs (60%) showed higher levels of ACP5 expression in metastatic lesions compared with the corresponding primary tumor samples (Figure 2c). Real-time PCR analysis showed that ACP5 mRNA levels were much higher in metastatic HCC tissues than that in primary HCC tissues (Figure 2d). In addition, ACP5 mRNA expression was much higher in primary HCC tissues from patients who develop metastasis than that in primary HCC tissues from patients who did not develop metastasis (Supplementary Figure S11). ACP5 expression was detected in several HCC cell lines with varying metastatic capabilities. ACP5 mRNA and protein levels were increased in the poorly metastatic HCC cells compared with the normal liver cells and were greatest in the highly metastatic HCC cells (Figure 2e).

Figure 2
Figure 2

ACP5 is significantly upregulated in human HCC tissues and indicates poor prognosis. (a) Immunohistochemistry analysis of ACP5 expression in 406 paired HCC tissues. (b) Statistical analysis of ACP5 expression in HCCs. (c) Representative ACP5 expression in primary and metastatic HCC tissues detected by immunohistochemical methods. (d) Real-time PCR analysis of ACP5 expression in normal livers (n=10) and 20 paired primary HCCs and metastatic HCC tissues. (e) Real-time PCR and western blot analyses of ACP5 expression in different human HCC cell lines. (f) Kaplan–Meier analysis of the correlation between ACP5 expression and the recurrence or OS of HCC patients. *P<0.05.

The correlation regression analysis indicated that the overexpression of ACP5 was significantly correlated with maximal tumor size, loss of tumor encapsulation, microvascular invasion, poor tumor differentiation and a higher tumor-node-metastasis (TNM) stage (Table 2). The Kaplan–Meier analysis showed that patients with positive ACP5 expression had shorter overall survival (OS) and higher recurrence rates than those with negative ACP5 expression (Figure 2f). A univariate analysis revealed that the tumor number, maximal tumor size, tumor encapsulation, microvascular invasion, tumor differentiation, TNM stage and ACP5 expression levels were statistically correlated with both recurrence and survival. These individual parameters were further subjected to a multivariate Cox proportional hazards model, which indicated that ACP5 expression was an independent and significant factor for recurrence (P=0.014) and reduced survival (P=0.037) (Table 3). Taken together, these studies suggest that ACP5 overexpression is indicative of poor prognoses of human HCC patients following curative resection.

Table 2: Correlation between ACP5 expression and clinicopathological characteristics of 406 HCCs
Table 3: Univerate and multivariate analysis of factors associated with survival and recurrence of 406 HCCs

ACP5 promotes HCC cell invasion in vitro and lung metastasis in vivo

It has been reported that ACP5 promotes the invasion and lung metastasis of human melanoma cells.27 To determine the role of ACP5 in the migration and invasion of HCC cells, we established two stable cell lines (denoted as SMMC7721-ACP5 and HCCLM3-shACP5) after lentiviral infection with LV-ACP5 or LV-shACP5. Both the upregulation and knockdown of ACP5 expression were confirmed by western blot. Three target sites were selected for the knockdown of ACP5 expression. Target site 3 (used in this study) was the most effective and was therefore chosen for further study (Figure 3a). Transwell assays showed that the upregulation of ACP5 expression significantly enhanced the migration and invasion of the SMMC7721 cells (low metastatic potential). Conversely, the inhibition of ACP5 expression in the HCCLM3 cells (high metastatic potential) dramatically decreased cell migration and invasion (Figure 3b).

Figure 3
Figure 3

ACP5 promotes HCC cell invasion in vitro and metastasis in vivo. (a) After cells were infected with LV-ACP5 or LV-shACP5-1,-2 and -3, ACP5 protein expression was detected by western blot. (b) Upregulation of ACP5 expression enhanced SMMC7721 (low metastatic potential) cell migration and invasion in vitro, whereas inhibition of ACP5 expression decreased HCCLM3 (high metastatic potential) cell migration and invasion. (cg) In vivo metastasis assays. The above four cell lines were transplanted into the livers of nude mice. (c) Representative bioluminescent imaging of the different groups is shown at 10 weeks following orthotopic implantation. (d) The incidences of lung metastases in the different groups of nude mice. (e) The numbers of lung metastatic foci in each group. (f) The OS times of the nude mice in the different groups. (g) The representative HE staining of lung tissues from the different groups is shown. *P<0.05.

The above four stable cell lines were transplanted into the livers of nude mice. Tumor metastases in the mice were monitored by an imaging system that detected the luciferase signal. The representative bioluminescent imaging of the different groups is shown in Figure 3c. Histological analyses (Figure 3g) confirmed that the incidence of lung metastases in the SMMC7721-ACP5 group was dramatically increased compared with the control group (60% versus 6.7%), indicating that ACP5 overexpression promoted HCC metastasis. In the HCCLM3-shcontrol group, the entire population developed lung metastases; however, only seven mice developed them in the HCCLM3-shACP5 group (100% versus 46.7%), indicating that ACP5 inhibition suppressed HCC metastasis (Figure 3d). The number of metastatic lung nodules in the SMMC7721-ACP5 group was increased compared with the SMMC7721-control group. However, the number of metastatic lung nodules in the HCCLM3-shACP5 group was significantly reduced compared with the HCCLM3-shcontrol group (Figure 3e). Furthermore, the SMMC7721-ACP5 group had a shorter OS time compared with the SMMC7721-control group. In contrast, the HCCLM3-shACP5 group displayed a longer OS time compared with the HCCLM3-shcontrol group (Figure 3f). Thus, these studies indicate that ACP5 promotes HCC invasion and metastasis.

Knockdown of ACP5 significantly decreased FoxM1-mediated HCC metastasis

Considering that ACP5, which promoted HCC invasion and metastasis, is a direct transcriptional target of FoxM1, we investigated whether ACP5 is involved in FoxM1-mediated HCC metastasis. The lentivirus LV-shACP5 was used to knockdown ACP5 expression in SMMC7721-FoxM1 cells that had been stably transfected with FoxM117 (Figure 4a). Transwell assays showed that the inhibition of ACP5 expression decreased FoxM1-enhanced cell migration and invasion (Figure 4b).

Figure 4
Figure 4

ACP5 knockdown significantly attenuates FoxM1-enhanced HCC invasion and metastasis. (a) After cells were infected with LV-shACP5, the protein expression of FoxM1 and ACP5 was detected by western blot. (b) ACP5 inhibition decreased FoxM1-enhanced cell migration and invasion in vitro. Following the infection of the SMMC7721-FoxM1 cells (stably transfected with FoxM1) with the lentivirus LV-shACP5, the cell migration and invasion capacities were assessed using transwell assays. (cg) In vivo metastatic assay. (c) Cell lines were transplanted into the livers of nude mice. Representative bioluminescent imaging of the different groups is shown at 10 weeks following orthotopic implantation. (d) The incidences of lung metastases in the different groups of nude mice. (e) The numbers of lung metastatic foci in each group. (f) The OS times of the nude mice in the different groups. (g) The representative HE staining of the lungs from the different groups is shown. *P<0.05.

In vivo metastatic assays confirmed that 10 cases developed lung metastases in the control group (SMMC7721-FoxM1+LV-shcontrol). However, in the ACP5-inhibition group (SMMC7721-FoxM1+LV-shACP5), there were only four cases with lung metastases (Figures 4c), as confirmed by both bioluminescent imaging and histological analyses. The number of metastatic lung nodules in the ACP5-inhibition group was significantly reduced compared with the control group (Figure 4e). Moreover, the ACP5-inhibition group had a longer OS time compared with the control group (Figure 4f). Taken together, these studies suggest that the knockdown of ACP5 significantly attenuates FoxM1-enhanced invasion and metastasis.

ACP5 rescues FoxM1-mediated invasion and metastasis upon knockdown of FoxM1

To further investigate the role of ACP5 in FoxM1-mediated HCC invasion and metastasis, the lentivirus LV-ACP5 was used to upregulate ACP5 expression in HCCLM3-shFoxM1 cells (stable knockdown of FoxM1 protein).17 The protein expression of FoxM1 and ACP5 were shown in Supplementary Figure S7A. Transwell assays showed that the stable knockdown of FoxM1 in HCCLM3 cells led to a dramatic decrease in migration and invasion abilities, which was partially rescued by the overexpression of ACP5 (Supplementary Figure S7B).

In vivo metastatic assays confirmed that nine cases developed lung metastases in the ACP5-overexpression group (HCCLM3-shFoxM1+LV-ACP5). However, in the control group (HCCLM3-shFoxM1+LV-control), there were only four cases with lung metastases (Supplementary Figure S7C, D and G), as confirmed by both bioluminescent imaging and histological analyses. The number of metastatic lung nodules in the ACP5-overexpression group was significantly increased compared with the control group (Supplementary Figure S7E). Moreover, the ACP5-overexpression group had a OS time compared with the control group (Supplementary Figure S7F). Taken together, these studies suggest that the overexpression of ACP5 rescues FoxM1-mediated invasion and metastasis upon knockdown of FoxM1.

ACP5 expression is positively correlated with FoxM1 expression in human HCC tissues and their coexpression indicates poor prognosis

We further evaluated the possible association between FoxM1 and ACP5 expression in human HCC tissues. Immunohistochemistry assays revealed that there was a significant positive correlation between FoxM1 and ACP5 expression in 406 human HCC tissues (P=0.026, Figures 5a and b).

Figure 5
Figure 5

ACP5 expression is positively correlated with FoxM1 expression, and their coexpression indicates poor prognosis. (a) An immunohistochemical analysis of FoxM1 and ACP5 expression in HCC tissues. (b) The association between FoxM1 and ACP5 expression. (c) The Kaplan–Meier analysis of concurrent FoxM1 and ACP5 expression with recurrence and OS. (d) The correlation of ACP5 expression with recurrence and OS among patients with FoxM1-positive tumors. (e) The correlation of ACP5 expression with recurrence and OS among patients with FoxM1-negative tumors.

Because the staining analyses revealed a statistically significant positive correlation between FoxM1 and ACP5 expression, we set out to detect whether the impact of FoxM1 on the prognoses of HCC patients was influenced by their ACP5 statuses. The patients were subsequently divided into four groups according to their combined expression levels of FoxM1 and ACP5. The Kaplan–Meier analysis revealed statistically distinct recurrence rates and survival times among the four subgroups. Among them, the patients with the positive coexpression of FoxM1 and ACP5 showed the highest recurrence rates and lowest OS times (Figures 5c and d). In addition, in patients with both FoxM1-positive and FoxM1-negative tumors that tested positive for ACP5 expression, shorter recurrence rates and survival times were observed compared with patients with ACP5-negative tumors (Figures 5e and f). Taken together, these results suggest that FoxM1 and ACP5 coexpression are indicative of poor prognoses of HCC patients.

Discussion

Tumor recurrence and metastases are the main cause of deaths in patients with HCC.2, 3 Therefore, it is extremely important to explore the molecular events underlying the development of metastasis in HCC. In the present study, we showed that ACP5 was significantly upregulated in human HCC tissues compared with adjacent noncancerous tissues. ACP5 overexpression was correlated with increased tumor size, loss of tumor encapsulation, microvascular invasion, malignant differentiation and a higher TNM stage. In addition, HCC patients showing positive ACP5 expression had poorer prognoses than those showing negative ACP5 expression. Furthermore, a multivariate analysis revealed that ACP5 was an independent risk factor for recurrence and the decreased survival of human HCC patients following curative resection. These clinical data strongly suggest that ACP5 contributes to the malignant progression of HCC and may be a biomarker for prognosis.

Previous studies have indicated that ACP5 promotes the invasion and distal metastases of melanoma and breast cancer cells.27 Furthermore, ACP5 is a marker of bone metastases in breast cancer.23, 25 However, the role of ACP5 in HCC metastasis remains unknown. This study has indicated the presence of a close association between ACP5 expression and HCC metastasis. First, the levels of ACP5 protein and mRNA correlated with the metastatic potentials of the HCC cell lines that were examined. Second, ACP5 overexpression was positively correlated with several metastatic characteristics, including loss of tumor encapsulation, microvascular invasion and a higher TNM stage. Third, the upregulation of ACP5 significantly promoted the invasion and lung metastases of HCC cells. Conversely, ACP5 knockdown decreased the invasion and metastases.

In addition, we also investigated the role of FoxM1 and ACP5 in melanoma. We found that both FoxM1 and ACP5 expression was significantly higher in primary melanoma and metastatic melanoma tissues than in benign nevi (Supplementary Figure S3A). FoxM1 expression was positively correlated with ACP5 expression in primary melanoma tissues (Supplementary Figure S3B). ACP5 protein levels increased progressively from immortalized melanocyte to melanoma cells with low metastatic potential and finally to melanoma cells with high metastatic potential (Supplementary Figure S4). FoxM1 also upregulated ACP5 expression in melanoma cells (Supplementary Figure S5). Furthermore, ACP5 knockdown significantly reduced FoxM1-enhanced melanoma cell migration and invasion (Supplementary Figure S6). Taken together, these studies suggested that FoxM1/ACP5 signaling pathway may have a role in melanoma progression.

FAK is an important mediator of cancer cell invasion and metastasis.29 A previous study reported that overexpression of ACP5 in melanoma cells led to a reproducible decrease in FAK auto-phosphorylation at Tyr397, which in turn promoted melanoma cell invasion.30 Auto-phosphorylation of FAK on Tyr397 creates a high-affinity binding site for the SRC. The FAK–SRC signaling complex acts to recruit and phosphorylate a number of signaling proteins and is involved in adhesion regulation and invasive phenotype.31 In the revised manuscript, we also found that ACP5 overexpression did not affect the total level of FAK, but decreased the phosphorylation level of FAK on Tyr397. Knockdown of ACP5 in FoxM1-overexpressing cells dramatically increased the phosphorylation level of FAK on Tyr397 (Supplementary Figure S2). These studies suggested that ACP5 might promote HCC invasion through inhibition of FAK auto-phosphorylation at Tyr397. Although we found that ACP5 is also upregulated in metastatic HCC tissues, whether or not ACP5 is involved in the establishment of a pre-metastatic niche remains unknown.

To date, the mechanisms that are responsible for ACP5 overexpression in malignancies have been unknown. In this study, we provided both experimental and clinical evidence to confirm that FoxM1 regulates ACP5 overexpression. Furthermore, we showed that ACP5 is involved in FoxM1-mediated HCC metastasis. The luciferase reporter assays showed that FoxM1 transactivates ACP5 promoter activity. Serial deletion, site-directed mutagenesis and chromatin immunoprecipitation assays further confirmed that FoxM1 binds directly to the ACP5 promoter, indicating that ACP5 is a direct transcriptional target of FoxM1. Additionally, the inhibition of ACP5 expression significantly attenuated FoxM1-enhanced cell invasion in vitro and lung metastasis in vivo, indicating that ACP5 is essential for FoxM1-mediated HCC metastasis. Finally, ACP5 expression is positively correlated with FoxM1 expression in human HCC tissues, and their positive coexpression is indicative of a poor prognosis. Thus, the FoxM1/ACP5 signaling pathway may have an important role in HCC metastasis.

In conclusion, these clinical and mechanistic findings suggest that ACP5 is a direct transcriptional and functional target of FoxM1. The FoxM1/ACP5 signaling pathway promotes HCC metastasis and may be a candidate biomarker for prognosis and a target for new therapies.

Materials and methods

Cell culture

Human HCC cells (HepG2, Huh7) were purchased from the American Type Culture Collection. Human HCC cells (SMMC7721, MHCC97L, MHCC97H and HCCLM3) were kindly provided by Dr Tang ZY (Liver Cancer Institute, Zhongshan Hospital, Fudan University, Shanghai, China). The SMMC7721 cell line is an HCC cell line with low metastatic potential. MHCC97L, MHCC97H and HCCLM3 cells are stepwise potentially metastatic cell lines with the same genetic background but different lung metastatic potentials.32 Immortalized liver cell lines (Chang liver and HL-7702) were purchased from the Institute of Biochemistry and Cell Biology, Chinese Academy of Science, Shanghai, China. All of the cells were cultured in Dulbecco’s modified eagle medium at 37 °C in a 5% CO2 incubator. The medium was supplemented with 10% fetal bovine serum, 100 μg/ml penicillin and 100 μg/ml streptomycin.

Plasmid construction

Plasmid construction was performed according to the standard procedures as outlined in our previous study.33 All of the primers are shown in Supplementary Table S1. The ACP5 promoter construct, (−2076/+33)ACP5, was generated from human genomic DNA. This construct corresponded to the sequence from −2076 to +33 (relative to the transcriptional start site) of the 5′-flanking region of the human ACP5 gene. It was generated with forward and reverse primers incorporating KpnI and XhoI sites at the 5′ and 3′-ends, respectively. The polymerase chain reaction (PCR) product was cloned into the KpnI and XhoI sites of the pGL3-Basic vector (Promega, Madison WI, USA). The 5′-flanking deletion constructs of the ACP5 promoter, ((−893/+33)ACP5, (−573/+33)ACP5 and (−221/+33)ACP5), were similarly generated with the (−2076/+33)ACP5 construct as the template. The FoxM1-binding sites in the ACP5 promoter were mutated using a QuikChange II Site-Directed Mutagenesis Kit (Strategene, La Jolla, CA, USA). The constructs were confirmed by DNA sequencing.

Patients and follow-up

A total of 406 adult patients with HCC who underwent curative resection between 1999 and 2001 at the Xijing Hospital of Fourth Military Medical University (Xi’an, China) were enrolled in this study. Ethical approval was obtained from the research ethics committee of Xijing Hospital and a written, informed consent was obtained from each patient. A preoperative clinical diagnosis of HCC was required to meet the diagnostic criteria of the American Association for the Study of Liver Diseases. Briefly, the inclusion criteria are as follows: (a) distinctive pathologic diagnosis, (b) no preoperative anti-cancer treatment or distant metastases, (c) curative liver resection and (d) complete clinical-pathologic and follow-up data. The clinical characteristics of the patients are listed in Table 2. The grading of the differentiation statuses was performed according to the method of Edmondson and Steiner.34 The pTNM classification for HCC was based on The American Joint Committee on Cancer/International Union Against Cancer staging system (6th edition, 2002).

In addition, 40 pairs of frozen fresh tumor liver tissues and peripheral nontumor tissues were collected after surgical resection and stored in liquid nitrogen. These tissue pairs were used to detect the mRNA expression of ACP5. A total of 20 pairs of primary and matched metastatic HCC specimens were collected from archived paraffin-embedded tissues. Metastatic tumors included nine lymph node, seven peritoneal, one kidney, one adrenal cortex and two bone metastases.

The follow-up data were summarized at the end of December 2009 with a median follow-up of 52.9 months (range 4–96 months). The patients were evaluated every 2-3 months during the first 2 years and every 3–6 months thereafter. All of the follow-up examinations were performed by physicians who were unaware of the study. During each check-up, the patients were monitored for tumor recurrence by assaying serum-AFP levels and performing abdominal ultrasound examinations. A computed tomography and/or magnetic resonance imaging examination was performed every 3–6 months together with a chest radiographic examination. The diagnostic criteria for HCC recurrence were the same as the preoperative criteria. The time to recurrence and OS were considered to be the primary end points. The time to recurrence was calculated from the date of resection to the date of tumor recurrence diagnosis. The OS was calculated from the date of resection to the date of death or last follow-up.

Construction of tissue microarrays and immunohistochemistry

HCC samples and the corresponding adjacent liver tissues (n=406) were used for the construction of a tissue microarray (Shanghai Biochip Co, Ltd, Shanghai, China). The tissue microarray was stained for ACP5 (Abcam, Cambridge, UK, no. 49507) expression. The array was scored independently by two pathologists for both the staining intensity and the extent of ACP5 expression across the section. Immunohistochemical staining was performed using the Dako Envision Plus System (Dako, Carpinteria, CA, USA) according to the manufacturer’s instructions.

The final scores for each sample (negative or positive) were assessed by summarizing the results of the intensities and extents of staining. The intensity of staining was scored as 0 (negative), 1 (weak) or 2 (strong). The extent of staining was based on the percentage of positive tumor cells: 0 (negative), 1 (1–25%), 2 (26–50%), 3 (51–75%) and 4 (76–100%). Therefore, each case was ultimately considered negative if the final score was 0 to 1 (−) or 2 to 3 (±) and positive if the final score was 4 to 5 (+) or 6 to 7 (+ +).

Construction of lentivirus and establishment of stable ACP5 overexpressing cells and ACP5 knockdown cells

On the basis of the ACP5 sequence (NM_001611), three short hairpin (sh) RNAs were designed using the small interferingRNA Target Finder (InvivoGene, San Diego, CA, USA): shACP5-1, 5′-GCCCAGATTGCATACTCTAAG-3′; shACP5-2, 5′-GATTGCATACTCTAAGATCTC-3′; shACP5-3, 5′- GGCTATCTGCGCTTCCACTAT-3′. Lentiviral vectors encoding shRNAs were generated using PLKO.1-TRC (Addgene, Cambridge, MA, USA) and were designated as LV-shACP5-1, LV-shACP5-2, LV-shACP5-3 and LV-shcontrol. Lentiviral vectors encoding the human ACP5 gene were constructed in FUW-teto (Addgene) and designated as LV-ACP5. An empty vector was used as the negative control and was designated as LV-control.

The lentiviral vectors were transfected into the HCC cells with a multiplicity of infection ranging from 30 to 50 in the presence of polybrene (6 μg/ml). At 48 h after transfection, 2.5 μg/ml puromycin (Origene, Rockville, MD, USA) was added, and the cells were incubated for 2 weeks to select for transfected cells. The pooled populations of knockdown and overexpressing cells, which were obtained at 2 weeks following drug selection without subcloning, were subjected to both in vitro and in vivo experiments.

Real-time PCR

Total RNA was extracted using the TRIzol Reagent (Invitrogen, Carlsbad, CA, USA) and reverse transcription was performed using the Advantage RT-for-PCR Kit (Takara, Otsu, Japan) according to the manufacturer’s instructions. For the real-time PCR analysis, aliquots of double-stranded complementary DNA were amplified using a SYBR Green PCR Kit (Applied Biosystems, Foster City, CA, USA). The cycling parameters were as follows: 95 °C for 15 s, 55 °C for 15 s and 72 °C for 15 s for 45 cycles. A melting curve analysis was then performed. The Ct was measured during the exponential amplification phase, and the amplification plots were analyzed using the SDS 1.9.1 software (Applied Biosystems). For the cell lines, the relative expression level (defined as fold change) of the target gene was determined by the following equation: 2–ΔΔCtCtCttarget−ΔCtGAPDH;ΔΔCtCtexpressing vector−ΔCtcontrol vector). The expression level was normalized to the fold change that was detected in the corresponding control cells, which was defined as 1.0. For the clinical tissue samples, the fold change of the target gene was determined by the following equation: 2–ΔΔCt (ΔΔCtCttumor−ΔCtnontumor). This value was normalized to the average fold change in the normal liver tissues, which was defined as 1.0. All of the reactions were performed in duplicate. The primer sequences are listed in Supplementary Table S1.

In vivo metastatic model and bioluminescent imaging

BALB/C nude mice (5-weeks-old) were housed under standard conditions and cared for according to the institutional guidelines for animal care. All animal experiments were approved by the Committee on the Use of Live Animals in Teaching and Research, Fourth Military Medical University.

For the in vivo metastasis assays, 4 × 106 cells in 100 μl of phosphate-buffered saline were injected subcutaneously into the flanks of nude mice. After 4 weeks, the subcutaneous tumors were resected and diced into 1-mm3 cubes, which were then implanted into the left lobes of the livers of the nude mice (15 for each group). For the in vivo tracking, different group of cells were infected with firefly luciferase. The in vivo tumor formation and metastases were imaged by bioluminescence. D-luciferin (Xenogen, Hopkinton, MA, USA) at 100 mg/kg was injected intraperitoneally into the mice and bioluminescence was detected using an IVIS 100 Imaging System (Xenogen). After acquiring photographic images of each mouse, luminescent images were captured using various (1–60 s) exposure times. The resulting grayscale photographic and pseudocolor luminescent images were automatically superimposed using the IVIS Living Image (Xenogen) software, which was performed to facilitate matching of the observed luciferase signal with its location on the mouse. The survival of the mice was recorded daily. After 10 weeks, the mice were killed and their lungs were dissected and prepared for standard histological examination.

Luciferase reporter assay

Luciferase activity was detected using the Dual Luciferase Assay (Promega) according to the manufacturer’s instructions. The transfected cells were lysed in the culture dishes containing a lysis buffer, and the resulting lysates were centrifuged at maximum speed for 1 min in an Eppendorf microcentrifuge. Relative luciferase activity was determined using a Modulus TD20/20 Luminometer (Turner Biosystems, Sunnyvale, CA, USA), and the transfection efficiencies were normalized according to the Renilla activity.

Chromatin immunoprecipitation assay

Cells that were transfected with the appropriate plasmids were cross-linked in 1% formaldehyde at 37 °C for 10 min. After washing with phosphate-buffered saline, the cells were resuspended in 300 μl of lysis buffer (50 mM Tri (pH 8.1), 10 mM EDTA, 1% SDS and 1 mM phenylmethylsulfonyl fluoride). The DNA was sheared to small fragments by sonication. The supernatants were precleared using a herring sperm DNA/protein G-sepharose slurry (Sigma-Aldrich, St Louis, MD, USA). The recovered supernatants were incubated with specific antibodies or an isotype control immunoglobulin G for 2 h in the presence of herring sperm DNA and protein G-sepharose beads. The immunoprecipitated DNA was retrieved from the beads with 1% SDS and a 1.1 M NaHCO3 solution at 65 °C for 6 h. The DNA was then purified using a PCR Purification Kit (Qiagen, Hilden, Germany). The primers are shown in Supplementary Table S1.

Statistical analysis

Comparisons of the quantitative data between the two groups were analyzed using the student’s t-test. The categorical data were analyzed using the Fisher’s exact test. The cumulative recurrence and survival rates were determined using the Kaplan–Meier method and log-rank test. The Cox proportional hazards model was used to determine the independent factors influencing survival and recurrence based on the variables that had been selected from the univariate analysis. A value of P<0.05 was considered to be statistically significant. All of the analyses were performed using the SPSS software version 11.0 (SPSS, Chicago, IL, USA).

References

  1. 1.

    , . Hepatocellular carcinoma: a global view. Nat Rev Gastroenterol Hepatol 2010; 7: 448–458.

  2. 2.

    , , , . Diagnosis and treatment of hepatocellular carcinoma. Gastroenterology 2008; 134: 1752–1763.

  3. 3.

    . Hepatocellular carcinoma. N Engl J Med 2011; 365: 1118–1127.

  4. 4.

    , . FoxM1: a master regulator of tumor metastasis. Cancer Res 2011; 71: 4329–4333.

  5. 5.

    , , , , , et al. Over-expression of FoxM1 leads to epithelial-mesenchymal transition and cancer stem cell phenotype in pancreatic cancer cells. J Cell Biochem 2011; 112: 2296–2306.

  6. 6.

    , , , , , et al. Aberrant FoxM1B expression increases matrix metalloproteinase-2 transcription and enhances the invasion of glioma cells. Oncogene 2007; 26: 6212–6219.

  7. 7.

    , , , , , et al. FoxM1 regulates transcription of JNK1 to promote the G1/S transition and tumor cell invasiveness. J Biol Chem 2008; 283: 20770–20778.

  8. 8.

    , , , , . Down-regulation of Forkhead Box M1 transcription factor leads to the inhibition of invasion and angiogenesis of pancreatic cancer cells. Cancer Res 2007; 67: 8293–8300.

  9. 9.

    , , , , , et al. A novel FoxM1-caveolin signaling pathway promotes pancreatic cancer invasion and metastasis. Cancer Res 2012; 72: 655–665.

  10. 10.

    , , , , , et al. Deregulation of FoxM1b leads to tumour metastasis. EMBO Mol Med 2011; 3: 21–34.

  11. 11.

    , , , , , . Increased levels of forkhead box M1B transcription factor in transgenic mouse hepatocytes prevent age-related proliferation defects in regenerating liver. Proc Natl Acad Sci USA 2001; 98: 11468–11473.

  12. 12.

    , , , . The Forkhead Box m1b transcription factor is essential for hepatocyte DNA replication and mitosis during mouse liver regeneration. Proc Natl Acad Sci USA 2002; 99: 16881–16886.

  13. 13.

    , , , , , et al. Foxm1b transcription factor is essential for development of hepatocellular carcinomas and is negatively regulated by the p19ARF tumor suppressor. Genes Dev 2004; 18: 830–850.

  14. 14.

    , , , , , et al. A cell-penetrating ARF peptide inhibitor of FoxM1 in mouse hepatocellular carcinoma treatment. J Clin Invest 2007; 117: 99–111.

  15. 15.

    , , , , , et al. Forkhead box M1B is a determinant of rat susceptibility to hepatocarcinogenesis and sustains ERK activity in human HCC. Gut 2009; 58: 679–687.

  16. 16.

    , , , , , et al. FOXM1 expression predicts the prognosis in hepatocellular carcinoma patients after orthotopic liver transplantation combined with the Milan criteria. Cancer Lett 2011; 306: 214–222.

  17. 17.

    , , , , , et al. Upregulated FoxM1 expression induced by hepatitis B virus X protein promotes tumor metastasis and indicates poor prognosis in hepatitis B virus-related hepatocellular carcinoma. J Hepatol 2012; 57: 600–612.

  18. 18.

    , , , , , et al. The TNF-alpha/ROS/HIF-1-induced upregulation of FoxMI expression promotes HCC proliferation and resistance to apoptosis. Carcinogenesis 2012; 33: 2250–2259.

  19. 19.

    , , , , , . Serum tartrate-resistant acid phosphatase 5b is a specific and sensitive marker of bone resorption. Clin Chem 2001; 47: 597–600.

  20. 20.

    , , . Human breast cancer cell lines and tissues express tartrate-resistant acid phosphatase (TRAP). Cell Biol Int 2007; 31: 191–195.

  21. 21.

    , , , , , et al. Increased tartrate-resistant acid phosphatase (TRAP) expression in malignant breast, ovarian and melanoma tissue: an investigational study. BMC Cancer 2006; 6: 199.

  22. 22.

    , , , . Evaluation of tartrate-resistant acid phosphatase (TRAP) 5b as serum marker of bone metastases in human breast cancer. Anticancer Res 2003; 23: 1011–1015.

  23. 23.

    , , , , , et al. Tartrate-resistant acid phosphatase 5b is a useful serum marker for extensive bone metastasis in breast cancer patients. Clin Cancer Res 2005; 11: 544–550.

  24. 24.

    , , , , . Tartrate-resistant acid phosphatase as a marker of bone metastases in patients with breast cancer and prostate cancer. Bull Exp Biol Med 2004; 138: 77–79.

  25. 25.

    , , , , , et al. Serum tartrate-resistant acid phosphatase 5b activity as a prognostic marker of survival in breast cancer with bone metastasis. BMC Cancer 2010; 10: 158.

  26. 26.

    , , , , , et al. Differential expression of tartrate-resistant acid phosphatase isoforms 5a and 5b by tumor and stromal cells in human metastatic bone disease. Clin Exp Metastasis 2011; 28: 65–73.

  27. 27.

    , , , , , et al. Proinvasion metastasis drivers in early-stage melanoma are oncogenes. Cancer Cell 2011; 20: 92–103.

  28. 28.

    , , . FoxM1: at the crossroads of ageing and cancer. Biochim Biophys Acta 2007; 1775: 92–102.

  29. 29.

    , , , , , . The role of focal-adhesion kinase in cancer—a new therapeutic opportunity. Nat Rev Cancer 2005; 5: 505–515.

  30. 30.

    , , , , , et al. Proinvasion metastasis drivers in early-stage melanoma are oncogenes. Cancer Cell 2011; 20: 92–103.

  31. 31.

    , , . Focal adhesion kinase: in command and control of cell motility. Nat Rev Mol Cell Biol 2005; 6: 56–68.

  32. 32.

    , , . Hepatocellular carcinoma: insight from animal models. Nat Rev Gastroenterol Hepatol 2012; 9: 32–43.

  33. 33.

    , , , , , et al. Upregulation of IL-23 expression in patients with chronic hepatitis B is mediated by the HBx/ERK/NF-kappaB pathway. J Immunol 2012; 188: 753–764.

  34. 34.

    , . Primary carcinoma of the liver: a study of 100 cases among 48,900 necropsies. Cancer 1954; 7: 462–503.

Download references

Acknowledgements

This study was supported by combined grants from the National Natural Science Foundation of China (No. 81272652, No. 81172290, No. 91129723, No.81090270, No.81090273 and No.81120108005), the National Key and Basic Research Development Program of China (No. 2010CB529302 and 2010CB529306), the National Municipal Science and Technology Project (2009ZX09103-667 and 2009ZX09301-009-RC06) and the Chinese Postdoctoral Science Foundation (No. 20100471776 and No.201104757).

Author information

Author notes

    • L Xia
    •  & W Huang

    These authors contributed equally to this work.

Affiliations

  1. State Key Laboratory of Cancer Biology and Xijing Hospital of Digestive Diseases, Fourth Military Medical University, Xi’an, Shaanxi Province, People’s Republic of China

    • L Xia
    • , W Huang
    • , Z Chen
    • , L Zhang
    • , H Hu
    • , J Liu
    • , G Tang
    • , J Dou
    • , S Sha
    • , B Xu
    • , C Liu
    • , J Ma
    • , S Zhang
    • , M Li
    • , D Fan
    • , Y Nie
    •  & K Wu
  2. Department of Gastroenterology, Tongji Hospital of Tongji Medical College, Huazhong University of Science and Technology, Wuhan, Hubei Province, People’s Republic of China

    • D Tian
  3. University of British Columbia, Vancouver, British Columbia, Canada

    • Y Li

Authors

  1. Search for L Xia in:

  2. Search for W Huang in:

  3. Search for D Tian in:

  4. Search for Z Chen in:

  5. Search for L Zhang in:

  6. Search for Y Li in:

  7. Search for H Hu in:

  8. Search for J Liu in:

  9. Search for Z Chen in:

  10. Search for G Tang in:

  11. Search for J Dou in:

  12. Search for S Sha in:

  13. Search for B Xu in:

  14. Search for C Liu in:

  15. Search for J Ma in:

  16. Search for S Zhang in:

  17. Search for M Li in:

  18. Search for D Fan in:

  19. Search for Y Nie in:

  20. Search for K Wu in:

Competing interests

The authors declare no conflict of interest.

Corresponding authors

Correspondence to Y Nie or K Wu.

Supplementary information

About this article

Publication history

Received

Revised

Accepted

Published

DOI

https://doi.org/10.1038/onc.2013.90

Supplementary Information accompanies this paper on the Oncogene website (http://www.nature.com/onc)

Further reading