Background
All the five secreted glycoproteins in Secreted Frizzled-Related Protein (SFRP) family contain a cysteine-rich domain (CRD) homologous to that of Wnt receptor Frizzled proteins. It is noted that via the CRD, they compete with the Frizzled proteins for Wnt binding, interfering with Wnt signaling. Downregulation of them can lead to aberrant activation of the Wnt pathway and induce tumorigenesis. SFRP1 is particularly well known for its frequent downregulation in human tumors, mainly due to hypermethylation of
SFRP1 promoter [
1‐
3]. Other members of this family, such as SFRP2, SFRP4 and SFRP5 [
4,
5], were also found to exert their roles in human cancers. These studies have labeled SFRPs as the candidate tumor suppressors in carcinogenesis.
It is also noted that matrix metalloproteinases (MMPs) are involved in a variety of physiological and pathological conditions, among which, their link with cancer has been most extensively studied. As a family of zinc-dependent endopeptidases, MMPs participate in the degradation of extracellular matrix components. In addition, their specific substrates also include latent growth factors, growth factor binding proteins, chemokines and cell adhesion molecules. Naturally, MMPs are known not only for their roles in invasion and metastasis of tumor, but also in growth and angiogenesis of it. Overexpression of MMPs has been observed in a series of cancers, with that of MMP-2[
6,
7], MMP-7[
8,
9], MMP-9[
10] and MT1-MMP[
11], in particular, typically present in gastric cancer.
Our initial finding showed that SFRP5 expression is significantly downregulated in gastric cancer, while the subsequent quantitative real-time PCR analysis revealed that MMP-2, MMP-7, MMP-9 and MT1-MMP are all upregulated in it. Such a finding that these MMPs are the targets of Wnt signaling [
12‐
15] led to our initial hypothesis: there may be a link between downregulation of SFRP5 and overexpression of these MMPs in gastric cancer. Another hypothesis was that the upregulation of these MMPs may be related to aberrant activation of Wnt signaling pathway due to SFRP5 downregulation.
Methods
Cell lines and tissue specimens
The human gastric cancer cell lines AGS, NCI-N87 and KatoIII were obtained from the American Type Culture Collection (Rockville, MD, USA); and MKN1, MKN28, MKN45, SGC-7901 and HGC-27 were obtained from Keygen Biotech. Co. (Nanjing, China). These cell lines were cultured in RPMI 1640 supplemented with 10% fetal bovine serum, at 37°C in a humid incubator with 5% CO2. 32 primary gastric cancer specimens and their corresponding matched cancer adjacent normal tissue specimens were obtained from patients under operation at Shengjing hospital, Chinese Medical University. Informed consent was obtained from all patients before collection of the specimens, which were frozen in liquid nitrogen immediately after surgical removal. Haematoxylin- and eosin-staining sections were prepared for assessment of the percentage of tumor cells, and only specimens with >70% tumor cells were selected for analysis. This study was carried out with the approval of the ethical committee of China Medical University.
RT-PCR and real-time reverse-transcription PCR
Total RNA was isolated from tissues and cell lines by Trizol (Takara, Dalian, China) according to the protocol supplied by the manufacturers. cDNA was synthesized from 1 μg RNA using random 9 primer and AMV reverse transcriptase. RT-PCR was performed using RNA PCR 3.0 Kit (Takara, Dalian, China). Real-time PCR was performed using the LightCycler system together with the LightCycler DNA Master SYBR Green I Kit (LightCycler, Roche Diagnostics). The housekeeping gene glyceraldehyde-3-phosphate dehydrogenase (
GAPDH) was used as an internal control. Gene expression was quantified by the comparative CT method, normalizing CT values to
GAPDH and calculating relative expression values. Primer sequences for
SFRP5,
MMP-2,
MMP-7,
MMP-9,
MT1-MMP and
GAPDH are shown in table
1.
Table 1
Primer sequences for RT-PCR
MMP-2 | Sense | 5'-GGATGATGCCTTTGCTCG-3' | 487 bp | 55°C |
| Anti-sense | 5'-ATAGGATGTGCCCTGGAA-3' | | |
MMP-7 | Sense | 5'-CTTCAGGCAGAACATCCA-3' | 220 bp | 49°C |
| Anti-sense | 5'-ATTTATTGACATCTACCCAC-3' | | |
MMP-9 | Sense | 5'-AGGACGGCAATGCTGATC-3' | 127 bp | 58°C |
| Anti-sense | 5'-TCGTAGTTGGCGGTGGTG-3' | | |
MT1-MMP | Sense | 5'-CCTGCGTCCATCAACACT-3' | 481 bp | 57°C |
| Anti-sense | 5'-TCACCTCCGTCTCCTCCTC-3' | | |
SFRP5 | Sense | 5'-AGGGTAAGGGAAAGGTGGAG-3' | 379 bp | 60°C |
| Anti-sense | 5'-GAGAAGCAGAGGCTGAGGAA-3' | | |
GAPDH | Sense | 5'-GGGAAACTGTGGCGTGAT-3' | 309 bp | 56°C |
| Anti-sense | 5'-AAAGGTGGAGGAGTGGGT-3' | | |
Methylation-specific PCR and DNA demethylation
Genomic DNA was extracted from tissues and cell lines by a standard phenol/chloroform extraction and ethanol precipitation procedure. Methylation of
SFRP5 was detected by Genmed MSP Kit (Genmed, Shanghai, China). The procedure was performed according to the manufacturer's instructions. Primers for
SFRP5 methylated sequence and unmethylated sequence were described in reference [
4]. Cell lines with
SFRP5 methylation were demethylated by 5-Aza-2'-deoxycytidine (DAC) (2 μM, Sigma, St Louis, MO, USA). Cells were seeded at a density of 3 × 10
4 cells/cm
2 in a 24 well plate on day 0, and exposed to DAC on day 1, 2, and 3. After each treatment, the cells were cultured in fresh medium. Control cells were incubated without the addition of DAC. Cells were harvested on day 4 for experiment.
Western blotting
Standard protocol was used. In brief, a total of 10 μg protein of each sample was run on a 12% sodium dodecyl sulfate (SDS)/acrylamide gel. The proteins on acrylamide gel were transferred to a nylon membrane, which was blocked overnight (4°C in PBS with 0.1% Tween and 10% milk powder). Polyclonal antibodies for SFRP5 (Santa Cruz, CA, America), MMP-2, MMP-7, MMP-9 and MT1-MMP (Abcam, Cambridge, United Kingdom), and the corresponding secondary antibodies (Santa Cruz, CA, America) were applied before immunoblotting. The human gene β-actin (Santa Cruz, CA, America) was used as an internal control.
RNA interference
SFRP5 RNAi plasmid and nonsilencing control RNAi plasmid were purchased from Takala Biotech. Co. (Dalian, China). The target sequence of SFRP5 siRNA was 5'-AAGAAGAATAAGGAGATGAAGTT-3', corresponding to 1022 to 1044 of the human SFRP5 cDNA sequence (GenBank: NM 003015). Cells were seeded into a 24-well plate at a density of 2 × 105. On the following day cells were transfected with SFRP5 siRNA or Control siRNA using Lipofectamine 2000 (Invitrogen, Paisley, United Kingdom) according to the manufacturer's instructions.
Construction of expression plasmids and transient transfection
The full-length pcDNA3.1 (Invitrogen, Paisley, United Kingdom) SFRP5 vector was made by cloning of the full-length PCR product of SFRP5 with PFU DNA polymerase (Invitrogen, Paisley, United Kingdom). All the plasmid sequences were confirmed by DNA sequencing. For transient transfection experiments, cells were plated in a 24-well plate at a density of 2 × 105 24 hours before transfection. Lipofectamine 2000 (Invitrogen, Paisley, United Kingdom) was used to perform transfection with 2.0 μg pcDNA3.1 SFRP5 vector or 2.0 μg pcDNA3.1 empty vector (as control) according to the manufacturer's protocol.
Statistical analysis
χ2 test was used to compare the methylation frequency between tumors and tumor adajcent normal tissues. The correlation between methylation and mRNA expression in primary gastric cancer was analyzed by Fisher's exact test. Mann-Whitney U-test was used to compare mRNA expression between tumors and tumor adjcent normal tissues. mRNA expression in gastric cell lines was compared using Student's t-test or one way ANOVA. Statistical analysis was carried out using SPSS version 13.0 (SPSS, Chicago, IL, USA). Difference was considered significant when P-value was < 0.05.
Discussion
Lots of studies have shown that aberrant Wnt signaling pathway is involved in carcinogenesis, in fact, alterations of downstream components in Wnt pathway have been found in a series of human cancers. A case in point is the association of mutations of
APC with most of human colon tumors [
16]. In addition,
β-catenin, another downstream component in Wnt pathway, is also frequently mutated in intestinal cancers [
16]. These mutations are known to have led to the decreased degradation of β-catenin before it is translocated and accumulated in the nucleus. There, β-catenin interacts with T-cell factor/lymphocyte enhancer factor (TCF/LEF) to promote the transcription of target genes in Wnt pathway. However, unlike colon cancer, mutations in
APC and
β-catenin are quite rare in gastric cancer, despite the occassional β-catenin accumulation in it [
17].
Interestingly, within Wnt pathway, the role of alterations of upstream components in gastric carcinogenesis has been drawing attention. The initial focus is the obvious overexpression in certain members of Wnt family, such as Wnt5a[
18], Wnt8b[
19], Wnt10a[
20], and Wnt receptor Frizzled 7[
21]. Another spotlight is the hypermethylation and downregulation of Wnt inhibitor SFRP1 in our recent group report[
22]. In addition, diminished expression of SFRP2[
23] and DKKs[
24] (another Wnt inhibitor family) was also observed. Both overexpression of Wnts or their receptor Frizzled proteins, and downregulation of Wnt inhibitors will lead to aberrant activation of Wnt pathway and increase the transcription of Wnt target genes, such as
c-myc[
25],
cyclin D1[
26] and some
MMPs, which are involved in initiation and progression of tumor. The focus is then shifted onto the significant downregulation (>10-fold) of SFRP5, another member of SFRP family, in 75% of the total gastric cancer specimens, and no detectable expression of it in 5 out of the total 8 gastric cell lines tested. MSP analysis indicated a connection between the absence of SFRP5 and its methylation in all these five cell lines and 21 out of the 24 SFRP5-downregulated gastric cancer specimens, for there is a resumption of it with these cell lines demethylated by DAC. Taken together, these results suggest a strong link between the common SFRP5 downregulation and its methylation in gastric cancer.
The assumption of a link between downregulation of SFRP5 and upregulation of MMPS in gastric cancer is based on the following findings. First, as a putative tumor suppressor gene, SFRP5 is downregulated in a series of human cancers, including gastric cancer; and in most cases, it can lead to aberrant activation of Wnt signal pathway. One recent study reported that
cyclinD1 and
c-myc, both known as Wnt targer genes, were significantly downregulated, when the expression of SFRP5 was restored by demethylation agent DAC after it was methylated in breast cancer cell lines. Second, some MMPs, such as MMP-2, MMP-7, MMP-9 and MT1-MMP, are frequenly overexpressed in gastric cancers, primarily in both tumor cells and stromal cells (except MMP-7, which is only expressed in tumor cells). And levels of MMP-2 and MMP-9 are also found elevated in plasma of gastric cancer patients. Thus it is obvious that the overexpression of these MMPs plays important roles in initiation and progression of gastric cancer. Third, other studies have shown Wnt signals can upregulate many members of MMPs family directly or indirectly, including MMP-1[
27], MMP-2[
12,
13], MMP-3[
28], MMP-7[
14], MMP-9[
12,
13], MMP-13[
13], MMP-26[
29], MT1-MMP[
15] and MT3-MMP[
30].
In this study, real-time PCR was used to determine mRNA levels of MMP-2, MMP-7, MMP-9 and MT1-MMP in gastric cancers. We found these four MMPs were all upregulated (from 1.4 folds to 4.2 folds). Next we analyzed the correlation between downregualtion of SFRP5 and upregulation of these four MMPs. The first line of evidence comes from the expression status of SFRP5 in primary gastric cancers. Levels of MMP-7 and MT1-MMP were found significantly high in the SFRP5-negative group. Second, we observed that cell lines AGS and NCI-N87, with no SFRP5 expression due to promoter hypermethyaltion, presented high levels of MMP-7 and MT1-MMP; while cell line MKN1 with strong expression of SFRP5 presented relatively low levels of them. Third, when we restored SFRP5 expression with demethylation agent DAC in AGS and NCI-N87, or enhanced SFRP5 expression with SFRP5 transfection in KatoIII, both MMP-7 expression and MT1-MMP expression are found to have reduced significantly in these cell lines. Finally, to further investigate the action of SFRP5 downregulation on expression of MMPs, we abrogated SFRP5 expression in MKN1 with SFRP5 RNAi. Just as presumed, both MMP-7 expression and MT1-MMP expression were significantly elevated. Taken together, all the lines of evidence suggest a prevalence of SFRP5 downregulation in gastric cancer, which is responsible for the upregulation of the expression of MMP-7 and MT1-MMP.
The mechanism involved in the upregulation of MMP-7 and MT1-MMP by SFRP5 downregulation may be presumably related to aberrant activation of Wnt signals. Now it is well accepted that aberrant activation of Wnt signaling pathway may induce carcinogenesis, and that these Wnt-related tumors often have increased invasiveness and metastasis, which is linked to some members in MMP family. Studies have shown some putative TCF/LEF binding sites lie in the promoters of
MMP-7[
31],
MT1-MMP[
32],
MMP-2 and
MMP-9[
12], and via these binding sites, Wnt/β-catenin can directly upregulate these
MMPs. For instance, it has been found that MMP-7 and MT1-MMP are upregulated by β-catenin/LEF-TCF compound in intestinal tumors [
31,
32]. Other researches showed Wnt signals can upregulate MMP-2 and MMP-9 expression in T cells and regulate T cell transmigration [
12]. In addition, Wnt/β-catenin can also upregulate MMP indirectly. Lowy et al observed MT3-MMP was upregulated in gastric cancer, however, no functional TCF/LEF binding site was found in the promoter of MT3-MMP[
30]. Taken together, it is clear that canonical Wnt signal pathway is involved in the upregulation of these MMPs.
Our results did not show MMP-2 and MMP-9 were upregulated by SFRP5 downregualtion, though they were also overexpressed in primary gastric cancers, and can be upregulated in T cells by Wnt signals [
12]. And in another study, breast cancers induced by Wnt-1 overexpressed a series of MMPs, including MMP-2 and MMP-9, except MMP-7 [
13], a long held Wnt target gene in intestinal tumors. These studies suggest that Wnt target genes may be tissue-specific, though it is true that the regulation of MMP expression is complex and it is certain that Wnt signaling is not alone among numerous factors, such as TNF-α [
33], H. pylori [
34] in regulating MMP expression in gastric cancer.
It should be noted that downregulation of SFRPs seems to be involved in carcinogenesis via both canonical Wnt pathway and noncanonical Wnt pathway. One recent study has demonstrated SFRP1 can inhibit the canonical Wnt/β-catenin pathway in breast cancer cells[
35]. Interestingly, in β-catenin-deficient human mesothelioma cell lines, SFRP4 can still inhibit cell growth and promote apoptosis[
36]. Another studies showed noncanonical Wnt pathway is also involved in the regulation of MMPs. For instance, Wnt5a has been found to upregulate MMP1 in endothelial cells[
27] and MMP-3 in mouse mammary cells[
30]. Therefore up to now, we can not rule out the possibility that SFRP5 downregulation upregulates MMP-7 and MT1-MMP via noncanonical Wnt pathway.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
ZC designed the study, carried out PCR analysis, analyzed and interpreted the data, and drafted the manuscript. BM performed cell transfection. ZN was engaged in drafting the manuscript and in statistical analysis. WW performed Western analysis. All authors read and approved the final manuscript.