Fax: (011) 86-20-85227039
Mammary serine protease inhibitor inhibits epithelial growth factor-induced epithelial-mesenchymal transition of esophageal carcinoma cells
Article first published online: 17 DEC 2008
Copyright © 2008 American Cancer Society
Volume 115, Issue 1, pages 36–48, 1 January 2009
How to Cite
Cai, Z., Zhou, Y., Lei, T., Chiu, J.-F. and He, Q.-Y. (2009), Mammary serine protease inhibitor inhibits epithelial growth factor-induced epithelial-mesenchymal transition of esophageal carcinoma cells. Cancer, 115: 36–48. doi: 10.1002/cncr.23991
- Issue published online: 29 DEC 2008
- Article first published online: 17 DEC 2008
- Manuscript Accepted: 1 AUG 2008
- Manuscript Revised: 15 JUL 2008
- Manuscript Received: 13 MAY 2008
- epithelial-mesenchymal transition;
- esophageal squamous cell carcinoma;
- mammary serine protease inhibitor;
By using proteomic technology, the authors previously observed the substantial down-regulation of mammary serine protease inhibitor (maspin) in esophageal squamous cell carcinoma and metastases. In the current study, they examined the effects of maspin re-expression in a maspin-null esophageal cancer cell line EC109 and also investigated the underlying mechanism.
A cell line with stable maspin expression was established. An epithelial growth factor (EGF)-induced epithelial-mesenchymal transition (EMT) model was used to mimic some aspects of the metastatic process in vitro. The effects of maspin reintroduction on EGF-induced EMT and cell growth characteristics were evaluated. Comparative proteomic analysis of transfected cells versus parental cells was then performed to explore the potential mechanism.
The introduction of maspin into EC109 cells was able to inhibit EGF-induced EMT and altered cell growth characteristics, including the serum dependence, proliferative response to EGF stimulation, and colony formation ability in soft agar, indicating a conversion from a malignant phenotype to a benign phenotype. Proteomic analysis revealed a significant down-regulation of a group of glycolytic enzymes in maspin-transfected cells. In addition, maspin-transfected cells expressed much lower levels of hypoxia-inducible factor 1α than parental cells or empty vector transfected cells.
Maspin exhibited a metastasis-suppressive effect, which may be a consequence of the reversal of the malignant phenotype of EC109 cells. The switch of cellular metabolic phenotype to low glycolysis by the gain of maspin function may play a key role in the process. This finding provides additional evidence of the tumor metastasis-suppressive activity of maspin and may indicate a new direction for future studies of the mechanism of maspin. Cancer 2009. © 2008 American Cancer Society.
Widely distributed in human tissue, mammary serine protease inhibitor (maspin) is a member of the serine protease inhibitor superfamily.1 Sequence comparison categorizes maspin to the ov-serpin subfamily (clade B). However, to our knowledge, there currently are no known target proteases for maspin, and structural features indicate that maspin is a noninhibitory serpin.2 Therefore, maspin was investigated to determine whether it functions in a manner similar to that of traditional serine protease inhibitors. Together with several other clade B members, including squamous cell carcinoma (SCC) antigens 1 and 2 and plasminogen activator inhibitor type 2, maspin is localized at chromatin 18q21.3, a region that often undergoes loss of heterozygosity in some forms of human cancers,3 suggesting its potential role in carcinogenesis. In 1994, maspin initially was identified by Zou et al as a putative tumor-suppressor gene in breast carcinoma.4 Since then, maspin has been reported sequentially as inversely related to cancer progression and metastasis in various cancers, such as prostate cancer, lung cancer, and oral squamous cell carcinoma. Although there were some paradoxical findings in ovarian and pancreatic cancers, the bulk of accumulated evidence from in vivo and in vitro experiments ascertained the multifaceted roles of maspin as an anti-invasive, antiangiogenic, and proapoptotic molecule involved in tumor suppression.5
It has been demonstrated that the expression of maspin is regulated by wild-type p53, hormonal factors, nitric oxide, manganese superoxide dismutase (MnSOD), and some DNA damage agents; and it was believed that the up-regulation of maspin was 1 of the mechanisms used by these molecules to perform their tumor-suppressive functions.6, 7 However, the precise mechanism by which maspin exerts its unique metastasis-suppressive property remains unknown. Important progress recently was made by Luo et al, who demonstrated that the activation and translocation of IκB kinase α, a key player in the nuclear factor (NF) κB signaling pathway, from cytoplasm to nucleus was required for the repression of maspin transcription in prostate epithelial cells, eventually promoting malignant prostatic epithelial cells to a metastatic fate.8, 9 In addition, previous studies revealed several putative mechanisms that affect extracellular matrix (ECM) adhesion, cell motility, proteolytic ability, apoptosis, and angiogenesis by up-regulating cell surface adhesion molecule integrins, reducing the activity of ras-related C3 botulinum toxin (PAK1) (Rac1) and Rac1 effector p21-activated kinase 1 substrate, decreasing urokinase-type plasminogen activator (uPA)/uPA receptor system activity, and activating the caspase cascade.10 The identification of maspin-interacting proteins, such as glutathione S-transferase and interferon-regulatory factor 6, also provides some insights into the molecular basis of maspin's functions.11, 12 These findings indicate that maspin may perform its multiple functions through the cooperation of various pathways, even in the same cells.
Investigations on maspin expression in esophageal SCC (ESCC) are sparse, except for an immunohistochemistry study by Wang et al.13 However, several studies concerning maspin expression in oral SCC reported that the expression of maspin was associated significantly with the grade of tumor differentiation and also with lymph node metastasis.14 Because both the esophagus and the oral cavity are lined by stratified squamous epithelium and are closely related in the digestive tract, it was believed that ESCC and oral SCC were subjected to similar mechanisms in terms of cellular malignant transformation and progression. Our previous comparative proteomic study demonstrated that maspin was down-regulated in esophageal cancer cells compared with immortalized epithelial cells. This change was confirmed further in a set of clinical specimens. In addition, in a xenograft model, we also observed significantly reduced maspin expression in metastasis-derived transplants compared with those derived from primary tumors (unpublished data). These findings indicate that maspin plays a tumor-and metastasis-suppressive role in ESCC similar to that observed in breast, prostate, and oral cancers. To gain further insights into the role of maspin in ESCC progression, we investigated the effect of maspin expression with an epithelial-mesenchymal transition (EMT) model by reintroducing maspin into a maspin-null esophageal cancer line, EC109 (Fig. 1). Our current collective data suggest that maspin reintroduction is capable of reversing cells to a low malignant phenotype and inhibiting epithelial growth factor (EGF)-induced EMT in EC109 cells.
MATERIALS AND METHODS
Cell Culture and Treatment
Human ESCC EC109 cells were maintained in Dulbecco Modified Eagle Medium (DMEM) supplemented with 10% fetal bovine serum (FBS), 2 mM L-glutamine, 100 U/mL penicillin, and 100 μg/mL streptomycin. Recombinant human EGF was prepared in 1% bovine serum albumin at 2 μg/mL and diluted in serum-free medium before use. For EGF treatment, cells were serum starved for 12 hours; then, EGF was added to cultures at a final concentration of 50 ng/mL in serum-free medium. Several other cell lines of esophageal origin also were used in the current experiments, including 2 human immortalized esophageal epithelial cell lines NE1 and NE3, which were maintained routinely in keratin serum-free medium supplemented with bovine pituitary extract (30 μg/mL) and EGF (0.2 ng/mL); the ESCC cell lines HKESC-1, HKESC-2, and SLMT-1, which were developed from patients with ESCC patients15–17 and maintained in MEM-α medium with 10% FBS; and the ESCC cell lines EC1 and EC18, which were cultured in DMEM supplemented with 10% FBS. All cell lines were maintained in a 5% carbon dioxide humidified incubator at 37°C.
Establishment of Maspin Transfectants
Full-length maspin complementary DNA was amplified from immortalized esophageal NE1 cells by using the following primer set: forward, atcaagcttcggatgcccctgcaactagcaaattc; reverse, gggattgaattcttaaggagaacafaatttgcc. The polymerase chain reaction (PCR) product was then cloned into the pEGFP-C1 vector (Clontech, Palo Alto, Calif) by using HindIII-EcoRI site. Maspin-null EC109 cells were transfected with either pEGFP-C1 or maspin/pEGFP-C1 plasmid DNA, which encode green fluorescent protein (GFP) alone or GFP-maspin fusion protein, respectively, by using Lipofectamine 2000 (Invitrogen). Stable transfectants were selected with 0.8 mg/mL Geneticin. Positive clones were maintained in the presence of 0.4 mg/mL of Geneticin.
Cell Growth Assay
Cell growth was measured by using a colorimetric naphthol blue black (NBB) staining protocol as described previously.18 Briefly, 5 × 103 cells were seeded into 96-well plates and treated for the indicated times. At the end of the experiment, the medium was removed, and the cells were fixed and stained in NBB solution. After washing, the dyes were dissolved in 50 mM NaOH, and the optic densities (OD) were measured at 595 nanometers (nm). Cell doubling times were calculated using a standard curve based on OD 595-nm values of known cell numbers ranging from 104 to 2 × 105.
Cell Migration Assay
The migration assay was performed according to the method of Brooks and Schumacher19 using Boyden chambers equipped with 8-μm porosity polyvinylpyrrolidone-free polycarbonate filters. Next, 5 × 104 cells in 0.5 mL DMEM with 1% serum were seeded into the upper chamber of transwells, and 0.75 mL DMEM containing 50 ng/mL EGF was added to the lower chamber as a chemoattractant. Duplicate wells without EGF addition served as controls. After 48 hours, the inserts were transferred to a clean companion plate, and the cells were fixed with 4% formalin for 30 minutes at 4°C. Cells on the upper surface of the membrane were then removed by scraping with sterile cotton swabs, and the cells remaining on the lower surface of the insert were stained with hematoxylin. The migrated cells were counted under a microscope at ×200 magnification. The cells from 4 different fields in each well were counted with 3 wells per treatment.
Cell Invasion Assay
The invasion assay was conducted as described for the migration assay with several differences. The transwell inserts were precoated with a thin layer of growth factor-reduced Matrigel solution (200 μg/mL). Duplicate wells without the Matrigel overlay served as controls. Cells that invaded into the lower chamber were counted at 48 hours after seeding. The percentage of invasion was calculated by dividing the total number of invading cells by the total number of cells that migrated through the control membrane.
Anchorage-independent Growth Assay
Anchorage-independent cell growth was examined by using a soft agar assay. The assay was done in 6-well plates with a base layer containing 0.6% agar in DMEM and 10% FBS. Then, 3 × 103 cells in 0.3% agar containing 10% FBS were embedded as a second layer. The plates were incubated at 37°C for 10 days, and the number of colonies that consisted of >10 cells was counted.
Two-dimensional Gel Electrophoresis (2DE) Image Analysis and Protein Identification
2DE were performed with 13 cm, pH 3-10, nonlinear, immobilized pH gradient gel strips by following a previously described protocol.18 The gels were silver stained, and the images were acquired at 300 dots-per-inch resolution and analyzed with ImageMaster 2D Elite software (version 4.01; GE Healthcare, Milwaukee, Wis). Spots that exhibited a consistent and significant difference in normalized volume (±2 fold) between compared samples in 3 separate experiments were selected for identification by tandem mass spectrometry (MS/MS).
Matrix-assisted laser desorption/ionization-time of flight (MALDI-TOF) and data-dependent MS/MS analyses were preformed on a Voyager 4700 mass spectrometer (Applied Biosystems, Foster City, Calif). MALDI-TOF mass spectra were acquired in reflector positive ion mode with an average 1500 laser shots per spectrum. MS/MS fragmentation spectra were acquired in a data-dependent fashion based on the MALDI-TOF peptide mass map for each protein, and the 10 most abundant ions present were selected in each sample (excluding trypsin-autolytic peptides and other known background ions).
Protein identification was performed by searching NCBInf protein database using automated GPS Explorer 2.0 software (Applied Biosystems) with the following criteria: 1) a maximum of 1 missed cleavage; 2) stated variable modifications were monoisotopic; 3) carbamidomethyl cysteine was selected as a fixed modification, and partial oxidation of methionine residues was considered; 4) mass tolerance was kept at 0.2 Da; and 5) taxonomy searched was limited to Homo sapiens.
Western Blot Analysis
Equal amounts of proteins were loaded on 10% or 12.5% sodium dodecylsulfate-polyacrylamide gel electrophoresis gels. After electrophoresis, the proteins were transferred electrophoretically to a polyvinylidene fluoride membrane. Membranes were incubated with 5% nonfat dry milk in Tris-buffered saline and 0.1% Tween-20 for 2 hours; then, they were probed with various primary antibodies for 2 hours at room temperature with gentle rocking. After incubation in a peroxidase-conjugated secondary antibody, the immunoblots were detected using an enhanced chemiluminescence kit.
The data represent the mean ± standard deviation from 3 independent experiments. Statistical analyses were performed by using 2-tailed Student t tests at a significance level of P < .05.
Maspin Reconstitution Inhibited the Epidermal Growth Factor-induced Epithelial-mesenchymal Transition of EC109 Cells
In our previous study, we demonstrated that EGF was able to induce EC109 cells to precede EMT (unpublished data), an in vitro process that some currently believe is correlated with tumor progression. To investigate the role(s) played by maspin in the process of esophageal cancer progression, we examined the effect of maspin expression with this established EMT model. A maspin expression vector that contained an enhanced GFP gene was constructed and transfected into the esophageal cancerous cell line EC109, which originally had no maspin expression (Fig. 1A), to generate maspin-stable transfectants (Fig. 1B). Figure 2A shows that both empty vector (109-Neo) and maspin transfection (Mas-8D) had no effect on the morphology of EC109 cells (0 hours). With exposure to EGF, the morphology of parental EC109 and 109-Neo cells switched from an epithelioid to spindle-like shape and demonstrated significant scattering (48 hours). In contrast, maspin-transfected cells remained tightly clustered, although a few marginal cells demonstrated an extension trend from the edge of cell sheets. An in vitro Boyden chamber assay revealed that the maspin-transfected Mas-8D cells had much lower migration, were less invasive than the parental and vector-transfected cells, and had no significant response to EGF stimulation (Fig. 2B). In agreement with these results, Western blot analysis demonstrated the reciprocal expression of pancytokeratin and vimentin, the lineage marker proteins for epithelial and mesenchymal cells, respectively, in the EC109 cells when exposed to EGF. With increased time of EGF stimulation, pancytokeratin tended to decrease its expression, but vimentin exhibited increased expression in EC109 and 109-Neo cells, whereas these 2 marker proteins maintained steady-state levels in Mas-8D cells (Fig. 2C). These observations indicate that maspin reconstitution in EC109 cells prevented the EGF-induced invasive phenotype transition.
Maspin Reintroduction Altered the Growth Characteristics of EC109 Cells
Although the published data suggested that the maspin re-expression did not affect the proliferation of transfected cells,20 we observed that the maspin transfectants grew much more slowly in serum-free medium than parental EC109 cells and empty vector-transfected 109-Neo cells. To clarify the effect of maspin expression on cell proliferation, we calculated the doubling time of EC109 cells, 109-Neo cells, and 2 stable maspin expression clones. Normally, the doubling time of maspin clones (Mas-8D, 35.7 ± 3.6 hours; Mas-1N, 36.2 ± 2.9 hours) was slightly longer than for EC109 cells (31.2 ± 1.4 hours) and 109-Neo cells (33.2 ± 2.1 hours), with no statistically significant difference (P > .05). Under the serum-free culture condition, however, the proliferation of the maspin transfectants was retarded markedly, but no significant effect on the growth of EC109 or 109-Neo cells was observed (Fig. 3A). Instead, the EC109 and 109-Neo cells appeared to adapt to the serum-free condition completely after 48 hours in culture. It is noteworthy that, when cells were serum starved for 12 hours and then exposed to EGF, as indicated in Figure 3B, maspin-transfected cells exhibited higher proliferative activity than parental EC109 and 109-Neo cells. These disparities in growth characteristics (Table 1) suggest distinct biochemical differences between the cells with and without maspin expression.
|Doubling time, h||31.2 ± 1.4||33.2 ± 2.1||35.7 ± 3.6||36.2 ± 2.9|
|Proliferation rate (response to EGF)||Moderate||Moderate||High||High|
|No. of colonies in soft agar||56.7 ± 3.2||58.7 ± 2.3||7.3 ± 2.5||10.3 ± 4|
Maspin-stable Expression Inhibited Anchorage-Independent Growth
It is believed that the ability to grow in an anchorage-independent manner is correlated with the degree of tumor malignancy; therefore, it is used to evaluate tumorigenicity. We observed that maspin-transfected Mas-8D and Mas-1N cells had decreased anchorage-independent growth, as determined by their ability to grow as colonies in soft agar (Fig. 3C). The number of colonies formed by maspin transfectants was significantly fewer than either parental EC109 cells or 109-Neo-transfected cells (P < .001).
Maspin Transfection Induced Protein Alterations in EC109 Cells
The results described above indicate that maspin reconstitution may convert the EC109 cells to a lower malignant phenotype. To gain insight into the molecular details of this change, we investigated the protein expression profiles of EC109, 109-Neo, and Mas-8D cells by using 2DE coupled with MS/MS. Proteins with levels that changed >2-fold in Mas-8D cells, compared with EC109 and 109-Neo cells, were subjected to MS/MS analysis. Thirteen protein spots were identified that belonged to 9 proteins, including glycolytic enzymes enolase 1, phosphoglycerate kinase 1 (PGK-1), triosephosphate isomerase (TPI), and glyceradehyde-3-phosphate dehydrogenase (GAPDH); the stress proteins heat-shock protein 27 (hsp27), hsp70, and manganese superoxide dismutase (MnSOD); and the oncoproteins DJ-1 and stathmin (Fig. 4). Detailed information of the identified proteins is summarized in Table 2. The altered expression levels of several identified proteins were confirmed further by Western blot analysis (Fig. 4C).
|Spot No.||Protein Name||Accession No.||Mr/pI||Mas-8D vs EC109||Mas-8D vs109-Neo||Protein Functions|
|560||PGK-1||gi|4505763||44,586.1/8.3||−2.32 ± 0.23||−2.17 ± 0.28||Carbonate metabolism, glycolysis|
|363||PGK-1||gi|4505763||44,586.1/8.3||−1.97 ± 0.13||−2.52 ± 0.41|
|446||Enolase 1||gi|4503571||47,139.3/7.01||−3.23 ± 0.57||−2.95 ± 0.25|
|2601||Enolase 1||gi|4503571||47,139.3/7.01||−2.93 ± 0.21||−3.16 ± 0.84|
|673||TPI-1||gi|4507645||26,652.7/6.45||−3.56 ± 0.62||−2.81 ± 0.56|
|693||TPI||gi|136066||26,608.8/7.1||−2.46 ± 0.54||−2.58 ± 0.27|
|522||GAPDH||gi|7669492||36,030.4/8.57||−3.66 ± 0.73||−3.46 ± 0.30|
|533||GAPDH||gi|7669492||36,030.4/8.57||−3.03 ± 1.08||−2.92 ± 0.74|
|626||Hsp27||gi|662841||22,313.3/7.83||−3.42 ± 0.40||−2.67 ± 0.24||Chaperon, stress-related antiapoptosis|
|877||Hsp70||gi|17511780||70,009/5.48||−2.12 ± 0.55||−2.52 ± 0.44|
|763||MnSOD||gi|110590807||22,190.2/6.86||+2.35 ± 0.44||+2.19 ± 0.32||Antioxidative, tumor suppressive|
|876||Stathmin 1||gi|5031851||17,291.9/5.76||−2.23 ± 0.30||−2.09 ± 0.11||Oncoprotein, cell proliferation, motility|
|1006||DJ-1 protein||gi|31543380||19,878.5/6.329||−3.08 ± 0.61||−2.54 ± 0.32||Oncoprotien, antiapoptosis|
Maspin Reconstitution Altered Metabolic Phenotype of EC109 Cells
Comparative proteomic analysis indicated that the expression levels of a set of glycolytic enzymes in maspin transfectants were much lower than the levels in parental cells and vector-transfected cells, suggesting that changes in cellular metabolism occurred because of the introduction of maspin. This finding prompted us to examine the expression of hypoxia-inducible factor 1α (HIF1α) in these cell lines, because it is believed that HIF1α is the central transcription factor in metabolic regulation.21 Western blot assay revealed a significant difference between cells with and without maspin (Fig. 5A). Parental EC109 cells and vector transfectants expressed high levels of HIF1α protein under normoxia, suggesting the constitutive expression of HIF1α protein in these cells. All of the stable maspin-transfected cell lines demonstrated much lower HIF1α expression, indicating the possibility that a transition in metabolic phenotype of these cells had been induced by maspin reconstitution. Next, we examined the effect on cell growth and survival when the glycolytic pathway was inhibited by 2-deoxy-D-glucose (2DG). The cells were treated with 2DG at different concentrations for 48 hours; then, the growth was determined. We observed that treatment with low doses of 2DG (2 mM and 4 mM) reduced the proliferation more profoundly in Mas-1N and Mas-8D cells than in EC109 and 109-Neo cells (Fig. 5B). However, a high dose of 2DG (10 mM) appeared to be more cytotoxic to parental EC109 and 109-Neo cells than to the maspin transfectants (P < .01).
In this study, we initially attempted to evaluate the effect of maspin on esophageal cancer metastasis by using an established EGF-induced EMT model because it is believed generally that EMT is a major mechanism for carcinoma progression and metastasis.22 Our data on cell morphology, cell migration and invasion, and the expression of lineage marker proteins clearly indicated that maspin was able to inhibit the transition of EC109 cells from epithelial to mesenchymal cells induced by EGF (Fig. 2), suggesting that the maspin expression level is correlated negatively with the metastatic behaviors of malignant esophageal cells.
Coincidentally, we noticed that the maspin-transfected cells grew much more slowly than the parental cells or the vector transfectants when maintained in serum-free medium, although this effect was not prominent under standard culture conditions. A cell proliferation assay confirmed this difference. Serum deprivation markedly retarded the growth of maspin-expressing cells but had no impact on parental or empty vector-transfected cells (Fig. 3A). This phenomenon suggests the possibility that maspin re-expression in EC109 cells may convert the cell phenotype toward benign, because serum-independent proliferation generally is accepted as a hallmark of tumor cells but not normal cells. Indeed, this speculation is supported further by the observation that significantly fewer colonies were formed by maspin transfectants in soft agar than by parental or vector-transfected cells (Fig. 3C).
We sequentially investigated the influence of EGF on cell growth under serum-free conditions. Maspin-transfected cells exhibited an obvious proliferation in response to EGF stimulation, whereas parental EC109 and 109-Neo cells exhibited moderate proliferation with a concomitant transition to mesenchymal cell type (Figs. 2 and 3B). The cell-specific differential effects of EGF have been reported extensively.23, 24 For example, Yeudall et al studied a pair of cell lines (HN4 and HN12) derived from a primary tongue squamous carcinoma and a lymph node metastasis, respectively, and demonstrated that EGF treatment at the same concentration produced completely different effects on the 2 cell lines.23 Primary tumor-derived HN4 cells proliferated, whereas the growth of metastasis-derived HN12 cells was repressed compared with untreated controls in the presence of high concentration of EGF. Moreover, HN12 cells demonstrated enhanced invasion in response to EGF, whereas HN4 cells did not. On the basis of these reports, we speculate that the observed differential responses to EGF stimulation between maspin-transfected and maspin-untransfected cells may be the consequence of altered cellular phenotype. This provides an additional instance to support the notion that maspin reintroduction may have caused certain key biochemical differences in EC109 cells, consequently resulting in the conversion of cell phenotype.
We decided to use the proteomic approach to identify protein alterations that are supposed to result, directly or indirectly, from maspin transfection to seek some useful hints concerning how to delineate the mechanism(s) of the maspin-related conversion. Several proteins, including hsp27, hsp70, oncoproteins DJ-1 and stathmin 1, and MnSOD, were identified that had affirmative alterations in expression because of maspin transfection (Table 2). Numerous studies have reported the close association between the expression levels of these proteins and cellular proliferation, apoptosis, malignant transformation, and tumor progression. Usually, these proteins, including hsp27, hsp70, DJ-1, and stathmin, have high basal levels in malignant cells compared with normal cells. Their over-expression has been reported previously in various malignancies and is regarded as an indicator of invasive phenotype and/or high resistance to apoptosis.25-30 In the current study, we observed that maspin transfection led to a marked reduction in their expression.
MnSOD was the only protein we identified that was up-regulated in maspin transfectants. This protein is supposed to possess tumor-suppressor activity. It was demonstrated that increased MnSOD activity was able to reverse part of the malignant phenotype in simian virus (SV) 40-transformed human fibroblast cells.31 Two earlier articles reported that forced expression of MnSOD caused the accumulation of maspin messenger RNA (mRNA) in both breast and prostate cancers, which also was postulated as 1 of the mechanisms for the antitumor effect of MnSOD.32, 33 In the current study, we demonstrated the reverse: that the MnSOD expression was up-regulated by the introduction of maspin, suggesting a probable bidirectional interaction between these 2 proteins and the involvement of maspin in the regulation of cellular redox homeostasis. Collectively, these findings in protein expression alterations induced by forced maspin expression indicate the capacity of maspin to normalize the expression of some oncoproteins and stress proteins that possibly were involved in the process of malignant progression.
Most intriguingly, a group of glycolytic enzymes, including enolase 1, PGK-1, TPI, and GAPDH, was identified that had decreased expression in maspin transfectants, indicating lower glycolytic potential in these cells. Increased aerobic glycolysis is observed exclusively in cancers, particularly in invasive cancers.34 A correlation between a high glycolytic rate and tumor aggressiveness has been verified repeatedly.35 Gatenby and Gillies proposed that the acquisition of the glycolytic phenotype may be an evolutionary consequence in malignant cells of the adaptation to hypoxia/anoxia during the development of invasive cancer.36 Moreover, this biochemical characteristic could be maintained by tumor cells (even in the presence of normoxia) or cultured in vitro and, thus, was considered an essential component of the malignant phenotype and a hallmark of invasive cancers.36 Persistently increased activity of the glycolytic pathway is necessary particularly for invasive tumor growth, not only because the glycolytic phenotype helps cells to survive in periodic or sustained hypoxia, but also because the local acidic environment derived from increased glycolysis facilitates the degradation of the ECM, thereby promoting invasion and metastasis.36
To our knowledge, currently, there is no report concerning the direct modulation of glucose metabolism by maspin. However, Amir et al reported the induction in mRNA transcripts of HIF1α by hypoxia in highly invasive MDA-MB-231 cells, but not in less invasive, maspin-stable transfectants, suggesting that maspin can subdue hypoxia-stimulated elevation in HIF1α.37 Transcription factor HIF1α is a master regulator of glycolytic energy metabolism. It directly modulates the transcription of nearly all of the core enzymes of glycolysis and also can initiate a pleiotropic response that includes apoptosis, angiogenesis, stress response, and erythropoeisis.21 Research by Seagroves et al on a pair of cell lines with and without HIF1α expression demonstrated that loss of HIF1α resulted in a decreased glycolytic rate, which then affected the growth characteristics of cells.38 Those authors also compared the global protein expression patterns of these 2 cell lines and identified 3 glycolytic enzymes, PGK-1, GAPDH, and TPI, significantly different in expression, which they believed were HIF1α-dependent variations.38 These protein alterations are similar to our findings in maspin-induced proteome alteration in the current study. Thus, it was reasonable for us to speculate that maspin reintroduction to EC109 cells would elicit a metabolic switch from high glycolysis to relatively normal physiologic levels, most likely by alleviating HIF1α levels. Western blot results, which demonstrated HIF1α express at much lower levels in all maspin-stable transfectant clones than in EC109 and 109-Neo cells, sustained this hypothesis (Fig. 5A).
We also investigated the effects of the glycolytic inhibitor 2DG on cell growth and survival. All of the examined cells exhibited significant growth inhibition with 2DG in a dose-dependent manner. However, EC109 and 109-Neo cells appeared to be less sensitive to low-dose 2DG (2 mM and 4 mM) but more susceptible to high-dose 2DG (10 mM) compared with Mas-1N and Mas-8D cells (Fig. 5B). 2DG, an analog of glucose, competes with glucose for transporting into cells, and then is phosphorylated to 2-deoxyglucose phosphate, which cannot be metabolized further, to produce adenosine-5′-triphosphate. When cells are exposed to the low dose of 2DG, the higher level of glycolytic enzymes expressed in parental EC109 cells and vector transfectants may help the cells to take up and use the glucose more effectively; thus, these cells appear to be more resistant to 2DG than maspin transfectants, which have lower glycolytic enzyme expression (Table 2). In addition, a high dose of 2DG may lead to severe blockage of glycolysis, resulting in more cytotoxicity to EC109 and 109-Neo cells, because the hypoxic tumor cells rely primarily on glucose as an energy source. These data provide further evidence for our speculation that maspin reintroduction induced a glycolytic switch in EC109 cells.
In summary, the current observations demonstrate that maspin re-expression antagonized EGF-induced EMT in esophageal cancer EC109 cells, indicating an important metastasis-suppressive role of maspin and a promising prospect for maspin as a valuable therapeutic drug in ESCC. The metastasis-suppressive effect may be a consequence of a maspin-induced reversal of the malignant phenotype of EC109 cells. Taken together with the proteomic results and the findings of others,38 we hypothesize that maspin transfection stimulates the reversion of malignant cells toward a relatively normal phenotype after a switch of metabolic phenotype to low glycolysis through disruption of the HIF1α pathway. These data provide additional evidence and indicate a new direction for future studies of the mechanism of maspin.
Conflict of Interest Disclosures
Supported in part by Guangzhou 2007 Municipal Programs of Science and Technology and Chang-Jiang Scholars Program 2007 (to Q.-Y.H.) and Hong Kong Research Grants Council Grants HKU 7395 of 03M (to J.-F.C.).
- 13Expression of tumor suppressor gene maspin in esophageal squamous cell carcinoma. J Henan Med Univ. 1997; 34: 48-49., , , , .
- 19Brooks SA, Schumacher U, eds. Metastasis Research Protocols, Vols 1 and 2. Totowa, NJ: Humana Press; 2001.