Hepatocyte growth factor (HGF) activator inhibitor type 2 (HAI-2/SPINT2) encodes Kunitz-type protease inhibitor that regulates HGF activity. Inspection of the human HAI-2/SPINT2 locus uncovered a large and dense CpG island within the 5′ region of this gene. Analysis of cultured human gastric tumor lines indicated that HAI-2/SPINT2 expression is either undetectable or in low abundance in several lines; however, enhanced gene expression was measured in cells cultured on the DNA demethylating agent 5-aza-2′-deoxycytidine. Bisulfite DNA sequencing confirmed the densely methylated HAI-2/SPINT2 promoter region. Forced expression of HAI-2/SPINT2 induced cell apoptosis, suppressed anchorage independent growth in vitro and tumor growth in vivo. We investigated HAI-2/SPINT2 aberrant methylation in patients with gastric cancer. The HAI-2/SPINT2 methylation was found preferentially in cancerous tissues (30 of 40, 75%) compared with nontumor tissues (no methylation was detected), indicating that this aberrant characteristic is common in gastric malignancies. In conclusion, epigenetic inactivation of HAI-2/SPINT2 is a common event contributing to gastric carcinogenesis and may be a potential biomarker for gastric cancer.
HAI-2/SPINT2 is an endogenous inhibitor of hepatocyte growth factor (HGF) activator (HGFA).1 HGFA is an enzyme that transforms the inactive, single-chain proform of HGF to its active heterodimeric form, which can initiate MET signaling by binding to MET receptor.2, 3 HAI-2/SPINT2 can inhibit the HGF/MET pathway by inhibiting HGFA.4 It has recently been shown that HAI-2/SPINT2 mRNA is downregulated or undetectable in several human cancers including gliomas,5 colorectal cancer6 and liver cancer. It was reported that expression of HAI-2/SPINT2 was inversely correlated with the histologic grade of gliomas and reexpression of HAI-2/SPINT2 in cell lines inhibited the cellular fibrinolitic activity and Matrigel invasion in vitro.5 In addition, downregulation of HAI-2/SPINT2 was also linked with poor prognosis in breast cancer.7 HGF/MET pathway has been implicated in the pathogenesis of medulloblastoma,4 gastric,8 brain cancer9 and hepatocellular carcinoma.10 Both c-MET and HGF have been shown to be deregulated and to correlate with poor prognosis in a number of major human cancers.11 Overexpression of c-MET is often observed in gastric carcinoma.8, 12, 13 In support of c-MET activation implicating in the pathogenesis of human cancers, introduction of c-Met and HGF into cells conferred the properties of motility, invasiveness and tumorgenicity to the transformed cells.14 Conversely, the inhibition of c-MET with a variety of receptor antagonists inhibited the motility, invasiveness and tumorgenicity of human tumor cell.15–17
Gastric cancer is the second most frequent cause of death from cancer in the world. The molecular mechanisms underlying the gastric carcinogenesis have not been fully understood, although several factors, including low intake of vegetables and fruit, Helicobacter pylori infection, have been found to be related with gastric carcinogenesis. In recent years, epigenetic silencing of tumor-related genes due to CpG island hypermethylation has been regarded as one of the most important alternations in gastric cancer.18, 19 The development and progression of gastric cancer involves a number of epigenetic alterations of tumor suppressor and tumor-related genes, such as APC,20p16,21E-cadherinand RASSF2A.23 Hypermethylation can occur in both neoplastic and non-neoplastic gastric epithelia and, therefore, is regarded as an early event in gastric carcinogenesis.24 Promoter methylation and transcriptional silencing of HAI-2/SPINT2 has been described previously in hepatocellular carcinoma,10 medulloblastoma,4 gliomas25 and renal carcinoma.26, 27 These findings prompted us to examine whether methylation-mediated epigenetic silencing of HAI-2/SPINT2 was involved in gastric carcinogenesis.
Material and Methods
Cell lines and tissue specimens
The SV40-transformed immortal human gastric mucosal epithelial cell line (GES-1) and four gastric cancer cell lines (AGS, SGC-7901, MKN-45 and BCG-823) were all preserved in our laboratory and maintained in RPMI 1640 with 10% fetal bovine serum. Cells were treated with 5 μM/L 5′-aza-2′-deoxycytidine (DAC; Sigma) for 48 h. Forty gastric cancer samples were obtained from the Department of Gastroenterology, Rui-jin Hospital. Samples of adjacent nontumor tissues from 19 patients were also collected. The clinicopathologic characteristics were analyzed according to tumor size, histologic grading and presence of nodal metastasis. Both tumor and adjacent nontumor tissues were sampled respectively, with approximate 1 cm3 size of each specimen and were proved by pathologic examination.
DNA methylation analysis of the HAI-2/SPINT2 gene
Genomic DNA (2 μg) was modified with sodium bisulfite using EpiTect Bisulfite kit (Qiagen, Germany). Methylation status was analyzed by bisulfite genomic sequencing of the CpG islands. The fragment covering 85 CpG sites from HAI-2/SPINT2 promoter region was amplified from bisulfite-modified DNA. The primers used were 5′-AAGGGAAGGGTGGTAGGTG-3′ (sense) and 5′-CCAATTCTCCCTACTCAAACC-3′ (antisense). Amplified products were cloned into pMD18-T simple vector (Takara, Japan); 5 independent clones were sequenced.
Semiquantitative polymerase chain reaction
Total RNA was extracted from gastric cancer or normal tissues using TRIZOL (Invitrogen, USA). Reverse transcription was carried out using Moloney Murine leukemia virus reverse transcriptase (Promega, Germany). The polymerase chain reaction (PCR) was performed using SYBER green kit following the manufacture's protocol (Takara, Japan). Primers used in RT-PCR: HAI-2/SPINT2-F: 5′-AGGTGGTGGTACAATGTCA CT-3′; HAI-2/SPINT2-R: 5′-GGGACAGAGGAATCCGCTG-3′. Actin-F: 5′-AGATGTGGATCAGCAAGCGGAGT-3′; Actin-R: 5′-GCAATCAAAGTCCTCGGCCACATT-3′. A 7900HT Fast Real-time PCR System (Applied Biosystems, USA) was used for testing. Following the protocol of the manufacturer, the amount of HAI-2/SPINT2 expression, normalized to a human actin endogenous reference is given by: 2▵▵Ct. Real-time RT-PCR was repeated at least three times for each specimen, and mean was obtained.
To study HAI-2/SPINT2 promoter hypermethylation, we also performed methylation-specific PCR (MSP) as described previously. For the HAI-2/SPINT2 promoter methylation study, we used two sets of primers that could amplify the modified DNA of either the methylated or unmethylated alleles separately. For the methylated allele, the sense primer was 5′-GTTTTGGCGATTTTCGCGC-3′, and the antisense primer was 5′-CCGACCTTCTCGACGCG-3′. For the unmethylated allele, the sense primer was 5′-GTTTTGGTGATTTTTGTGT-3′, and the antisense primer was 5′-CCAACCTTCTCAACACA-3′. The annealing temperatures for the methylated and unmethylated DNA were 56 and 53°C, respectively, for 30 s each. Hot-start PCR with a total cycle number of 35 was used in all MSP DNA amplifications. Denaturation and extension cycles were maintained for 45 and 30 s, respectively.
Construction of plasmids and stable cell line generation
For construction of pCMV4-flag/SPINT2, the HAI-2/SPINT2 cDNA was generated by reverse transcription PCR using SPINT2 forward primer (5′-AAAGCCTTATGGCGCAGCT GTGCGGGCTG-3′) and reverse primer (5′-GGATCCTTTC ACAGGACATATGTGTTCT-3′). The sequence was confirmed by DNA sequencing and ligated into the HindIII and BamHI sites of pCMV-flag vector (Invitrogen, USA). For transfection experiments, MKN45 cells were plated into 6-well plates 24 h before transfection. The cells were transfected with 5 μg/well of empty pCMV-flag or pCMV4-flag/SPINT2 using Superfect (Qiagen, Germany) according to manufacturer's instructions. For 48 h after transfection, the cells were passaged at 1:5 and cultured in medium supplemented with G418 at 500 μg/ml for 4 weeks. One stable MKN45 clone reexpressing HAI-2/SPINT2 (MKN45/SPINT2) was selected for further study, with HAI-2/SPINT2 expression verified by western blot. As a control group, cells stably transfected with an empty vector pCMV-flag were also established (MKN45/vector).
Cells were harvested and samples (20 μg) of the cell lysate were subjected to 10% SDS-PAGE gel electrophoresis, after which the resolved proteins were transferred to nitrocellulose membranes (Amersham Biosciences, UK). The membranes were then blocked with 5% nonfat milk and 0.1% Tween 20 in Tris-buffered saline and probed with anti-flag antibody (Sigma, USA), after which the blots were visualized using enhanced chemiluminescence (Amersham, UK).
Cell proliferation analysis
Approximately 5 × 104 cells were seeded in 12-well plates (Corning, USA), cells counts and viability (trypan blue dye exclusion) were determined daily, from day 1 to day 7 of culture. Experiment was performed in triplicates.
Evaluation of apoptosis
Apoptosis was detected by flow cytometric analysis of Annexin V staining. Annexin V-FITC (Fluorescein Isothiocyanate) versus PI (Propidium Iodide) assay was performed as previously reported. Briefly, adherent cells were harvested and suspended in the Annexin-binding buffer (1 × 106 cells/ml). Thereafter, cells were incubated with Annexin V-FITC and PI for 15 min at room temperature in the dark and immediately analyzed by flow cytometry. The data are presented as bi-parametric dot plots showing Annexin V-FITC green fluorescence versus PI red fluorescence.
Anchorage-independent growth and mouse xenograft model
For anchorage-independent growth assay, 2 × 104 cells were plated in 0.3% low melting point agar/growth medium onto 6-cm dishes with a 0.6% agar underlay. After 4 weeks, the number of colonies was determined. Six-week-old female athymic nude mice were used for MKN45 tumor xenografts. The tumors were established by subcutaneous injection of 1 × 106 MKN45/vector, MKN45/SPINT2 into nude mice. Ten mice were included in each group in all experiments.
Pearson χ2 and one ANOVA were used for statistical analysis of group differences. Pearson χ2 were performed to evaluate the significance of the differences between the frequencies of HAI-2/SPINT2 promoter hypermethylation status of the various tissue categories and comparisons with clinical characteristics, and p values less than 0.05 was considered significant.
Bisulfite sequencing analysis
Recently, promoter methylation of HAI-2/SPINT2 gene and the resultant gene suppression have been shown in medulloblastoma,4 renal carcinoma26 and hepatocellular carcinoma,10 these findings prompted us to examine whether such methylation is also present in gastric cancer. We determined the HAI-2/SPINT2 CpG island methylation status in GES-1 and gastric cancer cell lines by bisulfite genomic sequencing. The area of the CpG-rich region around the transcription initiation site of HAI-2/SPINT2 gene between the nucleotides −239 and +521, which spanned 85 CpG sites, was sequenced. The PCR products were cloned into a plasmid vector, and five independent clones were sequenced. As shown in Figure 1, most CpG dinucleotides were methylated in gastric cancer cell lines, whereas aberrant methylation was absent in GES-1 cells.
Association of the HAI-2/SPINT2 promoter methylation with transcriptional gene silencing
To elucidate whether the aberrant methylation of HAI-2/SPINT2 is associated with loss of HAI-2/SPINT2, we examined the HAI-2/SPINT2 mRNA expression in gastric cancer cell lines using real-time RT-PCR. Abundant levels of HAI-2/SPINT2 mRNA were observed in normal gastric epithelia cell line (GES-1), in contrast, HAI-2/SPINT2 expression was significantly reduced in a majority of gastric cell lines (Fig. 2a and 2b).
To confirm that this loss of expression was because of the HAI-2/SPINT2 promoter methylation, gastric cancer cell lines were treated with DAC (an inhibitor of the methylase enzyme, which can reactivate mRNA expression suppressed by methylation) for 48 h and performed RT-PCR to detect HAI-2/SPINT2 expression (Fig. 2a). After treatment with DAC, we found the levels of HAI-2/SPINT2 mRNA were induced in these gastric cancer cell lines, suggesting that the HAI-2/SPINT2 gene silencing is accounted for by hypermethylation. The expression levels of HAI-2/SPINT2 were quantified by real-time PRC (Fig. 2b). The demethylation of HAI-2/SPINT2 by DAC in these gastric cancer cells was verified by use of MSP (Fig. 2c).
Expression and methylation analysis of HAI-2/SPINT2 in primary gastric cancers
HAI-2/SPINT2 mRNA expression levels in each of the 40 gastric cancer tissues were assessed relative to the normal tissues using real-time RT-PCR analysis. The results showed that HAI-2/SPINT2 was significantly downregulated in tumor tissues compared with normal tissues (p = 0.004; Fig. 3a). To determine whether the hypermethylation of CpG islands in primary gastric cancer samples was correlated with loss of HAI-2/SPINT2 expression, we also performed MSP analysis. The result of MSP revealed that the HAI-2/SPINT2 was methylated in 30 of 40 primary gastric cancers (75%) tested, no methylation was detected in normal gastric samples. To confirm the methylation status of HAI-2/SPINT2 found by MSP, the samples positive for hypermethylation of HAI-2/SPINT2 were subjected to bisulfite sequencing. Bisulfite sequencing showed good concordance with MSP. The samples were divided into three different groups, depending on percentage of methylation on methylated CpG sites of a total of 85 CpGs, as <30%, 30–70%, and >70%.28 Correlation of the promoter methylation with HAI-2/SPINT2 expression showed a significantly lower expression level in the tumors with >70% and 30–70% methylation compared with the tumors with <30% methylation (Fig. 3b). Representative examples of the gel analysis of MSP and bisulfite DNA sequencing were shown in Figure 4a and 4b. In total, these data suggested that the frequent mechanism of transcriptional downregulation of HAI-2/SPINT2 in gastric cancers was HAI-2/SPINT2 promoter hypermethylation.
The association between HAI-2/SPINT2 methylation and the clinicopathologic features of patients are listed in Table 1. The prevalence of HAI-2/SPINT2 methylation was significantly different between normal and cancerous samples (p < 0.001). There was no significant correlation between hypermethylation of HAI-2/SPINT2 and unfavorable variables, such as tumor size (P = 0.141), but not the type of cell differentiation (P = 0.011) and metastasis (P = 0.027). There was no obvious difference between male and female patients in this study (data not shown).
Table 1. Clinical characteristics of gastric cancer patients according to hypermethylation status of HAI-2/SPINT2
We also determined the correlation between H. pylori infection and HAI-2/SPINT2 methylation. H. pylori infection was detected in 57% of gastric cancer tissues with methylation in HAI-2/SPINT2 gene and 40% of gastric cancer tissues without HAI-2/SPINT2 methylation. There was no association between promoter hypermethylation of HAI-2/SPINT2 gene and the status of H. pylori infection (p = 0.557).
HAI-2/SPINT2 inhibits cell proliferation and induces cell apoptosis in vitro
The frequent silencing of HAI-2/SPINT2 by methylation in gastric cancer but not in adjacent gastric mucosa suggests a potential tumor suppressor role of this gene. To test this speculation, we examined the effect of HAI-2/SPINT2 transfection on cell proliferation and apoptosis of gastric cancer cells. We established MKN45 stable transfectants with empty vector and HAI-2/SPINT2 expression plasmid. The expression of HAI-2/SPINT2 in MKN45/SPINT2 cells was confirmed by Western analysis (data not shown). Compared with the control, the proliferation in the HAI-2/SPINT2-expressing MKN45 cells was inhibited (Fig. 5a; p < 0.05). The frequent silencing of HAI-2/SPINT2 by methylation in gastric cancer but not in nontumor mucosa suggests a potential tumor suppressor role of this gene; 27 ± 4% of MKN45/SPINT2 at 24 h and 39 ± 6% at 48 h were stained positively with Annexin V as determined by FCM, whereas there were no significant changes in control or MKN45/vector cells (Fig. 5b; p < 0.05).
HAI-2/SPINT2 reexpression results in growth inhibition of colony assays and nude mice
We further showed the growth inhibitory features of HAI-2/SPINT2 reintroduction in colony formation assay and nude models. We observed that HAI-2/SPINT2 reexpression revealed tumor suppressor activity, there was lower colony formation compared with control vector (p < 0.05; Fig. 6a). We examined the ability of MKN45/SPINT2 cells to form tumors compared with MKN45/vector. All mice injected with MKN45/vector cells and none of 10 mice received injection of MKN45/SPINT2 cells developed palpable tumors (Fig. 6b).
HAI-2/SPINT2 is a novel member of the Kunitz family of serine protease inhibitors. HGF activator is a serine protease responsible for proteolytic activation of HGF. HGF, also known as the scatter factor, is considered to play a pivotal role in carcinogenesis due to its ability to trigger proliferation and motility of tumor cells. HGF is also a powerful inducer of angiogenesis,9, 29 a process required for tumor growth as well. The biological responses to HGF are mediated through its cellular receptor c-MET.
The majority of gastric adenocarcinomas have demonstrated overexpression of the c-MET protein, such as MKN45 gastric cancer cell line.12, 13, 30 The HGF/MET pathway plays a pivotal role in gastric tumor growth, invasion, metastasis and angiogenesis and correlates with poor survival of patients. c-MET can be activated aberrantly through several mechanisms, including mutation31 and gene amplification,8 formation of autocrine loops.14, 32
HAI-2/SPINT2 has a typical CpG island around its transcription start site (Fig. 1). We first determined the HAI-2/SPINT2 CpG island methylation status of a panel of four human gastric cancer cell lines and one normal control GES-1 by bisulfite genomic sequencing of multiple clones. Strikingly, HAI-2/SPINT2 CpG island hypermethylation was found in all gastric cancer cell lines tested, whereas aberrant methylation was absent in GES-1 cells. We examined a further link between HAI-2/SPINT2 CpG island hypermethylation and its gene silencing by the treatment of these cancer cell lines with DAC (a DNA demethylating agent). After the treatment of DAC, the expression of HAI-2/SPINT2 mRNA was restored.
We extended our HAI-2/SPINT2 CpG island hypermethylation analysis to 40 patients with primary gastric cancer by use of MSP. We observed that HAI-2/SPINT2 CpG island hypermethylation was a common event in gastric cancer tissues (30 of 40, 75%), whereas no methylation was detected in adjacent nontumor tissues. Moreover, the degree of methylation correlated inversely with HAI-2/SPINT2 expression. It has recently been shown that HAI-2/SPINT2 mRNA is downregulated or undetectable in several human cancers and cancer cell lines. For example, HAI-2/SPINT2 mRNA was abundantly expressed in normal kidney but was down-regulated in advanced stage renal cell carcinoma.26 Although the expression of HAI mRNA was conserved in colorectal cancers, the levels of expression were decreased in the adenocarcinoma tissues compared with the normal counterparts. There was a tendency toward an inverse correlation, albeit not well defined, between the amounts of HAI mRNA and the tumor progression. Immunohistochemical study indicated that HAI protein is present predominantly on the surface of epithelial cells of the colon and the immunoreactivity was decreased in the adenocarcinoma cells.6 Similarly, we found that HAI-2/SPINT2 mRNA was downregulated in gastric cancer tissues compared with normal tissue, which was kept in line with its role of tumor suppressor gene. Although promoter hypermethylation of HAI-2/SPINT2 is significantly linked to cell differentiation and metastasis, there was no significant correlation with the other parameters, such as gender, tumor size and H. pylori infection.
From a functional standpoint, we next wanted to examine whether epigenetic inactivation of HAI-2/SPINT2 inhibited growth suppression in gastric cancer cells. We established MKN45 cells stably overexpressing HAI-2/SPINT2. MKN45/SPINT2 cells underwent induced apoptosis when compared with the control cells. We further observed that reintroduction of HAI-2/SPINT2 reduced the anchorage-independent growth ability of MKN45 in vitro. All mice injected with MKN45/SPINT2 cells remained tumor free for an additional 1 month, whereas mice received injection of control vector-transfected cells developed palpable tumors within 1 month, suggesting that HAI-2/SPINT2 reexpression could have reduced tumorigenicity in nude mice xenografts.
In addition, HAI-2/SPINT2 is also known as placental bikunin. Overexpression of bikunin in an ovarian cancer cell line suppressed invasion,33, 34 and bikunin has been investigated as a treatment for ovarian carcinoma.35 For incidence, once-daily oral bikunin reduced tumor in a nude mouse model and in human ovarian cancer.36–39 HGF/MET participate in all stages of malignant progression and become one of leading candidates for targeted cancer therapies.40 In ovarian cancer, HAI-2/SPINT2 was identified as an inhibitor of tumor invasion and metastasis. Stable reexpression of HAI-2/SPINT2 in renal cancer cells reduced proliferative capacity, anchorage-independent growth and cell motility in vitro.4 HAI-2/SPINT2 as a novel serine protease inhibitor may be a novel therapeutic approach for treating gastric cancer.30 HAI-2/SPINT2 may inhibit the activation of HGF in vivo via its potent inhibitory activity against HGFA, which in turn suppresses the HGF-induced signaling transduction that could have an important role in the progression of gastric cancer. We attempted to find out the exact mechanism by which HAI-2/SPINT2 inhibits the HGF-induced signaling. We detected the activity of pERK and pAKT in MKN-45/vector and MKN-45/SPINT2 cells with or without the presence of exogenous HGF.9 The activity of pERK and pAKT, the important downstream factors of MET, did not alter (Supporting Information Fig. 1). Certainly, there are remains to be done to explore the exact factors in HGF/MET pathway inhibited by HAI-2/SPINT2, which contribute to the tumor suppression of HAI-2/SPINT2.
In conclusion, we found that epigenetic inactivation of HAI-2/SPINT2 is a common event in gastric cancer and may be a potential biomarker for gastric cancer. We further demonstrated that HAI-2/SPINT2 could act as a functional tumor suppressor gene in gastric cancer.