Transcription Factor Sp1 Expression Is a Significant Predictor of Survival in Human Gastric Cancer

Vol. 9, 6371–6380, December 15, 2003 Clinical Cancer Research 6371

Liwei Wang,1 Daoyan Wei,1 Suyun Huang,2,4 angiogenesis, its disregulated expression and activation may Zhihai Peng,5 Xiangdong Le,1 Tsung Teh Wu,3 play important roles in gastric cancer development and James Yao,1 Jaffer Ajani,1 and Keping Xie1,4 progression. Departments of 1Gastrointestinal Medical Oncology, 2Neurosurgery, 3Pathology, and 4Cancer Biology, The University of Texas M. D. Anderson Cancer Center, Houston, Texas, and 5Department of INTRODUCTION General Surgery, Shanghai Jiaotong University Affiliated First Although the incidence of gastric cancer declined in the People’s Hospital, Shanghai, People’s Republic of China West from the 1940s to the 1980s, it remains a major public health problem throughout the world (1). In Asia and parts of South America in particular, it is the most common epithelial ABSTRACT malignancy and leading cause of cancer-related deaths. In fact, The transcription factor Sp1 regulates the expression of gastric cancer remains the second most frequently diagnosed multiple genes. However, its expression and role in human malignancy worldwide and cause of 12% of all cancer-related tumor development and progression remain unclear. Using deaths each year (1, 2). Advances in treatment of this disease are immunohistochemistry, we investigated Sp1 expression pat- likely to come from a fuller understanding of its biology and terns in 86 cases of human gastric cancer having various behavior. clinicopathologic characteristics, 57 normal gastric tissue The aggressive nature of human metastatic gastric carci- specimens, and 53 lymph node metastases. We found that noma is related to mutations of various oncogenes and tumor Sp1 protein was expressed predominantly in the nuclei of suppressor genes (3–6), and abnormalities in several growth cells located in the mucous neck region, whereas Sp1 expression was not detected either in the cells located toward the factors and their receptors (4, 5). These abnormalities affect the gastric pit (foveolar differentiation) or cells of the glandular downstream signal transduction pathways involved in the con- epithelium (glandular differentiation). In sharp contrast, trol of cell growth and differentiation. Specifically, these per- strong Sp1 expression was detected in tumor cells, whereas turbations confer a tremendous survival and growth advantage no or very weak Sp1 expression was detected in stromal cells to gastric cancer cells (4–6). Previous studies indicated the role and normal glandular cells surrounding or within the tu- of several growth factor families in gastric cancer development mors. and progression, including basic fibroblast growth factor, trans- We also evaluated the effect of Sp1 expression on the forming growth factor, nitric oxide synthase, matrix metallopro- survival of patients who have undergone surgical resection. teinase-2, epidermal growth factor receptor, insulin-like growth The median survival duration in patients who had a tumor factor receptor, interleukin 8, and vascular endothelial growth with negative, weak, and strong Sp1 expression was 43, 37, factor (VEGF; Refs. 7–11). However, the molecular mecha- Next, Sp1 expres- nisms behind the abnormal expression of multiple molecules .(0.0075 ؍ and 8 months, respectively (P sion, stage, completeness of resection, age, and sex were remain unclear. Our recent study of human pancreatic cancer entered into a Cox proportional hazard model. In multiva- suggested that abnormal Sp1 activation augments the angio- -and stage (P < 0.001) were genic and metastatic capacity of tumor cells through overex (0.003 ؍ riate analysis, Sp1 (P independently prognostic of survival. Therefore, normal and pression of multiple Sp1 downstream genes, including the key malignant gastric tissues have unique Sp1 expression pat- angiogenic factor VEGF (12). terns. Given the importance of Sp1 in the expression of Sp1 is a sequence-specific DNA-binding protein that is multiple molecules key to tumor cell survival, growth, and important in the transcription of many cellular and viral genes that contain GC boxes in their promoter (13). Additional transcription factors (Sp2, Sp3, and Sp4) that are similar to Sp1 in their structural and transcriptional properties have been cloned, thus forming a Sp1 multigene family (14–20). Because Sp1 is Received 4/1/03; revised 8/26/03; accepted 9/8/03. Grant support: Research Scholar Grant CSM-106640 from American an essential transcription factor for many genes that are key to Cancer Society, and Grant 1R01-CA093829 and Cancer Center Support the regulation of multiple aspects of tumor cell survival, growth, Core Grant CA 16672-23 from the National Cancer Institute, NIH and angiogenesis, abnormal Sp1 expression and activation may (K. X.). contribute to gastric cancer development and progression. For The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked example, Sp1 mRNA and Sp1 and Sp3 DNA-binding activity advertisement in accordance with 18 U.S.C. Section 1734 solely to are increased in epithelial tumors when compared with that in indicate this fact. papillomas in skin (21). Interestingly, interference of Sp1 func- Requests for reprints: Keping Xie, Department of Gastrointestinal Med- tion has been shown to inhibit cell growth (22). Consistent with ical Oncology, Unit 426, The University of Texas M. D. Anderson Cancer Center, 1515 Holcombe Boulevard, Houston, TX 77030. Phone: (713) 792- these findings, we have shown that Sp1 is constitutively over- 2013; Fax: (713) 745-1163; E-mail: [email protected]. activated in human pancreatic cancer (12). However, it is un-

known whether and how Sp1 signaling pathways contribute to each with distilled water and PBS. Afterward, the slides were tumor development and progression. mounted using Universal Mount (Research Genetics) and ex- In the present study, we examined Sp1 expression in hu- amined using a bright-field microscope. A positive reaction was man gastric cancer. We found that elevated Sp1 activation indicated by a reddish-brown precipitate in the nuclei (23, 24). occurred in human gastric cancer and was inversely correlated Besides clone PEP2, other specific antihuman Sp1 antibodies

with patient survival, suggesting that abnormally activated Sp1 were used, including a polyclonal rabbit antihuman Sp1 NH2 expression may represent a potential molecular marker for poor terminus peptide antibody (clone H-225; Santa Cruz Biotech- prognosis and directly contribute to gastric cancer development nology) and monoclonal mouse antihuman Sp1 amino acid and progression. 520–538 peptide antibody (clone 1C6; Santa Cruz Biotechnol- ogy). All three of the antibodies were used on formalin-fixed, MATERIALS AND METHODS paraffin-embedded tumor sections. A positive reaction was indicated by a reddish-brown precipitate in the nuclei (23, 24). Human Tissue Specimens and Patient Information. Depending on the percentage of positive cells and staining We used human gastric cancer tissue specimens preserved in the Gastric Cancer Tissue Bank, and information about the patients intensity, Sp1 staining was classified into three groups: nega- was obtained from what was recorded in its comprehensive tive, weak positive, and strong positive. Specifically, the per- database at The University of Texas M. D. Anderson Cancer centage of positive cells was divided into five grades (percent- Ͻ Center. Primary gastric cancer in these patients was diagnosed age scores): 10% (0), 10–25% (1), 25–50% (2), 50–75 (3), and treated at M. D. Anderson Cancer Center from 1985 to and Ͼ75% (4). The intensity of staining was divided into four 1998. The patients had well-documented clinical histories and grades (intensity scores): no staining (0), light brown (1), brown follow-up information. None of them underwent preoperative (2), and dark brown (3). Sp1 staining positivity was determined chemotherapy and/or radiation therapy. Eighty-six patients with by the formula: overall scores ϭ percentage score ϫ intensity primary gastric cancer were randomly selected for this study. score. The overall score of Յ3 was defined as negative, of Ͼ3 All of the patients had undergone gastrectomy with lymph node - Յ6 as weak positive, and of Ͼ6 as strong positive. dissection. All of the patients also were observed at M. D. Electrophoretic Mobility Shift Analysis and Supershift Anderson through the end of 1999. The median follow-up Assay. Nuclear protein extracts from normal stomach and period for the 86 patients was 25.7 months. At the end of 1999, gastric cancer tissues were prepared by lysing cells in a high-salt 30 patients were still alive, whereas 56 patients had died. Fifty- buffer [20 mM HEPES (pH 7.9), 20 mM NaF, 1 mM Na3P2O7,1 three lymph node metastasis specimens and 57 normal gastric mM Na3VO4,1mM EDTA, 1 mM EGTA, 1 mM DTT, 0.5 mM tissue specimens obtained from patients without gastric cancer phenylmethylsulfonyl fluoride, 420 mM NaCl, 20% glycerol, 1 were also included in this study. ␮g/ml leupeptin, and 1 ␮g/ml aprotinin] followed by snap- Western Blot Analysis. Whole cell lysates were pre- freezing in ethanol/dry ice for 5 min and thawing on ice for 10 pared from human normal and gastric cancer tissue specimens. min; the freezing and thawing procedures were performed twice. Standard Western blotting was performed using a polyclonal The supernatant was then centrifuged and harvested. Protein rabbit antibody against human Sp1 (clone PEP2; Santa Cruz concentrations were determined using the Coomassie Plus Pro- Biotechnology, Santa Cruz, CA) and antirabbit IgG, a horserad- tein Assay kit (Pierce Biotechnology, Rockford, IL) according ish peroxidase-linked F(abЈ) fragment obtained from a donkey 2 to the manufacturer’s protocol. Protein extracts (10 ␮g) were (Amersham Life Sciences, Arlington Heights, IL). Equal protein incubated in a final 20-␮l volume of 10 mM HEPES (pH 7.9), 80 sample loading was monitored by incubating the same mem- mM NaCl, 10% glycerol, 1 mM DTT, 1 mM EDTA, and 100 brane filter with an anti-␤-actin antibody (12). The probe pro- ␮ teins were detected using the Amersham enhanced chemilumi- g/ml poly(deoxyinosinic-deoxycytidylic acid) using electro- nescence system according to the manufacturer’s instructions. phoretic mobility shift analysis with the radiolabeled double- Ј Immunohistochemistry. Sections (5-␮m thick) of for- stranded Sp1 consensus-binding motif 5 -ATTCGATCG Ј malin-fixed, paraffin-embedded tumor specimens were deparaf- GGGCGGGGCGAGC-3 (Santa Cruz Biotechnology) or as in- finized in xylene and rehydrated in graded alcohol. Antigen dicated for 20 min at room temperature. For supershift analysis, retrieval was performed with 0.05% saponin for 30 min at room the nuclear extracts were preincubated with specific antibodies temperature. Endogenous peroxidase was blocked using 3% against Sp1 (Santa Cruz Biotechnology). The protein-DNA hydrogen peroxide in PBS for 12 min. The specimens were complexes were resolved on a 4.5% nondenaturing polyacryl- incubated for 20 min at room temperature with a protein- amide gel containing 2.5% glycerol in 0.25ϫ Tris-borate EDTA blocking solution consisting of PBS (pH 7.5) containing 5% at room temperature. The results were autoradiographed (12). 2 normal horse serum and 1% normal goat serum and then incu- Statistical Analysis. The two-tailed ␹ test was per- bated at 4°C in a 1:200 dilution of rabbit polyclonal antibody formed to determine the significance of the difference between against human Sp1 (clone PEP2). The samples were then rinsed the covariates. Survival durations were calculated using the and incubated for1hatroom temperature with peroxidase- Kaplan-Meier method. The log rank test was used to compare conjugated antirabbit IgG. Next, the slides were rinsed with PBS cumulative survival of patient groups. Cox proportional hazard and incubated for 5 min with diaminobenzidine (Research Ge- model was used to provide univariate and multivariate hazard netics, Huntsville, AL). The sections were washed three times ratios for the study parameters. The level of Sp1 expression, age, with distilled water, counterstained with Mayer’s hematoxylin sex, Lauren’s histology type, stage (American Joint Committee (Biogenex Laboratories, San Ramon, CA), and washed once on Cancer), and completeness of surgical resection (R0 versus

Fig. 1 Sp1 protein expression in human normal gastric tissue. Tissue sections were prepared from formalin-fixed, paraffin-embedded specimens of normal human gastric tissue (57 cases). Immunohistochemical staining was performed using a specific anti-Sp1 antibody (clone PEP2). Of note is that the majority of the normal gastric mucous cells were negative for Sp1 in the nuclei, whereas only some of the cells in the mucous neck region were positive for Sp1. Calibration bar, 100 ␮m(A), 40 ␮m(B), and 20 ␮m(C).

R1 and R2) were included in the model. In all of the tests, a P Ͻ in the nuclei of cells localized predominantly in the mucous 0.05 was defined as statistically significant. SPSS version 11.05 neck region, whereas Sp1 expression was not detected in the (SPSS Inc., Chicago, IL) was used for analyses. cells located toward the gastric pit (foveolar differentiation) or cells of the glandular epithelium (glandular differentiation; Fig. RESULTS 1). In sharp contrast, strong Sp1 staining was seen in the Unique Sp1 Expression Pattern in Human Normal Gas- tumor-cell nuclei of various types of gastric cancer, whereas no tric and Tumor Tissue. To localize Sp1 protein expression in or very weak Sp1 expression was detected in the stromal cells both human normal gastric and tumor tissue specimens, we and normal glandular cells surrounding or within the tumors performed immunohistochemical staining of paraffin-embedded (Fig. 2). These Sp1 expression patterns were additionally con- tissue specimens using an antibody against Sp1 protein (clone firmed using three different anti-Sp1 antibodies (data not PEP2). There was either weak or negative expression in most of shown). These results suggested that Sp1 is rarely expressed in the cell nuclei in the gastric tissue specimens obtained from normal human gastric cells but commonly expressed in human patients who did not have cancer. Positive staining was observed gastric cancer cells.

Fig. 2 Sp1 protein expression in human gastric cancer tissue. Tissue sections were prepared from formalin-fixed, paraffin-embedded specimens of various types of gastric cancers (86 cases): A, diffuse; B, papillary; C, tubular; D, tubular with adjacent normal; E, mucinous; and F, signet ring. Immunohistochemical staining was performed using a specific anti-Sp1 antibody (clone PEP2). Of note is that the majority of the tumor cells were positive for Sp1 in the nuclei, whereas the majority of the normal gland (A), stromal (C), and adjacent tissue cells (D), and infiltration lymphocytes were negative for Sp1. Insets in A and F are representative areas from A and F, respectively, with higher magnifications. Calibration bar, 50 ␮m (A–F), and 20 ␮m(insets).

Sp1 DNA-Binding Activity in Normal and Tumor Tis- obtained from 8 patients with known Sp1 expression as detected sues. To determine the functional level of Sp1 expression in via immunostaining. Sp1 protein expression was determined both human normal gastric and tumor tissue cells, whole cell using Western blot analysis with 10-␮g whole cell protein and nuclear protein extracts were prepared from specimens extracts (Fig. 3A), whereas Sp1 DNA-binding activity was de-

Fig. 3 Sp1 protein expression and DNA-binding activity. Whole-cell and nuclear protein extracts were prepared from 8 paired normal gastric (N) and gastric tumor tissue (T) specimens obtained from the patients with known Sp1 expression via immunostaining. A, Sp1 protein expression was determined using Western blot analysis with 10-␮g whole-cell protein extracts. Equal protein sample loading was monitored by hy- bridizing the same membrane filter with an anti-␤-actin antibody. B, Sp1 DNA-binding activity was determined via electrophoretic mobility shift analysis using 10-␮g nuclei protein extracts and a 32P-labeled oligonucleotide probe containing a consensus-binding motif for Sp1. For characterization of the Sp1 DNA-binding complex (B4), nuclear protein extracts were preincubated with a specific antibody against Sp1 (Tab) or control serum (Tns). The positions of the Sp1 and supershifted complexes were marked SS. Of note is that the level of both Sp1 protein expression and its DNA-binding activity was significantly elevated in tumor tissue compared with that in normal tissue.

termined via electrophoretic mobility shift analysis using 10-␮g yses included stage (P Ͻ 0.0001) and completeness of resection nuclear protein extracts and a 32P-labeled oligonucleotide probe (P Ͻ 0.0002). Age at diagnosis (as a continuous variable by containing a consensus-binding motif for Sp1 (Fig. 3B). It was univariate Cox proportional hazard analysis), sex, and Lauren’s apparent that cells in the tumor tissue specimens had a signifi- histology type did not have a statistically significant affect on cantly higher level of Sp1 activity than those in the normal survival. tissue specimens did, which was consistent with the level of Sp1 Next, Sp1 expression, stage, completeness of resection, protein expression determined via immunostaining. age, and sex were entered into a Cox proportional hazard model Sp1 Expression in Primary Tumor and Nodal Metasta- for multivariate analysis (Table 2). Adjust for the effect of ses of Gastric Cancer. We analyzed Sp1 expression in 86 covariates, high Sp1 expression (P ϭ 0.003) and advanced stage gastric tumor, 53 lymph node metastasis, and 57 normal human (P Ͻ 0.001) were independent predictors of poor survival. The gastric tissue specimens. Overall, lymph node metastasis spec- hazard ratio for the group with strong Sp1 expression (hazard imens exhibited the highest level of Sp1 expression (54.7% ratio, 4.5; 95% confidence interval, 1.8–11.2) was significantly strong positive; P Ͻ 0.001), and tumor tissue specimens had higher than that of the Sp1-negative group (referent). Complete- higher levels of Sp1 expression than normal tissue specimens ness of resection, age at diagnosis, sex, and Lauren’s histology (24.4% versus 0% strong positive; P Ͻ 0.001; Table 1; Fig. 4A). type did not have a statistically significant affect on survival in Sp1 Expression and Survival in Gastric Cancer Pa- multivariate analyses. tients. The median survival duration in patients who had a tumor with negative, weak, and strong Sp1 expression was 43, 37, and 8 months, respectively. The elevated Sp1 expression DISCUSSION was associated with inferior survival duration (P ϭ 0.0075; Fig. In the present study, we provided evidence of distinct Sp1 4B). Other parameters that affected survival in univariate anal- expression patterns in normal gastric and gastric tumor tissues.

Table 1 Clinical characteristics of gastric cancer patients and Sp1 expression The two-tailed ␹2 test was performed using SPSS software. Sp1 staining Case (%) Negative Weak Strong (n ϭ 86) (n ϭ 17) (n ϭ 48) (n ϭ 21) P Age Յ60 39 (45.3) 11 (28.2) 21 (53.8) 7 (17.9) Ͼ60 47 (54.7) 6 (12.8) 27 (57.4) 14 (29.8) 0.146 Sex F 30 (34.9) 8 (26.7) 15 (50.0) 7 (23.3) M 56 (65.1) 9 (16.1) 33 (58.9) 14 (25.0) 0.494 Stage I 14 (16.3) 4 (28.6) 8 (57.1) 2 (14.3) II 28 (32.6) 5 (17.9) 18 (64.3) 5 (17.9) III 30 (34.9) 5 (16.7) 15 (50.0) 10 (33.3) IV 14 (16.3) 3 (21.4) 7 (50.0) 4 (28.6) 0.737 Pathology type Papillary 12 (14.0) 2 (16.7) 6 (50.0) 4 (33.3) Tubular 28 (32.6) 7 (25.0) 16 (57.1) 5 (17.9) Diffuse 8 (9.3) 1 (12.5) 3 (37.5) 4 (50.0) Mucinous 5 (5.8) 1 (20.0) 4 (80.0) 0 (00.0) Signet ring 21 (24.4) 6 (28.6) 11 (52.4) 4 (19.0) Mixed 12 (14) 0 (00.0) 8 (66.7) 4 (33.3) 0.429 Lauren’s classification Intestinal 53 (61.6) 14 (26.4) 30 (56.6) 9 (17.0) Diffuse 33 (38.4) 3 (9.1) 18 (54.5) 12 (36.4) 0.044

Specifically, we found that Sp1 protein was expressed predom- tion of angiogenesis (36). Furthermore, Sp1 has been shown to inantly in the nuclei of cells located in the mucous neck region regulate the promoters of many apoptosis-related genes, includ- where stem cells reside, suggesting that dividing cells express ing Bcl-2 and Bcl-x (37, 38), survivin (39), and TGF-␤ and its Sp1, whereas Sp1 expression was not detected in either cells receptors (40, 41). Thus, increasing evidence indicates that Sp1 located toward the gastric pit (foveolar differentiation) or cells is involved in the regulation of apoptosis in numerous systems of the glandular epithelium (glandular differentiation), suggest- (42–45). Moreover, members of the Sp1 family regulate (pos- ing that terminally differentiated cells lost their Sp1 expression. itively or negatively) the expression of VEGF and basic fibro- In contrast, significantly elevated Sp1 activity was observed in blast growth factor, two major signaling molecules key to tumor gastric tumor specimens and inversely correlated with patient angiogenesis (12, 46–48). Accumulating evidence has also sug- survival, suggesting that Sp1 is an independent prognostic gested that Sp1 family members regulate multiple aspects of marker for gastric cancer and that the Sp1 signaling pathway angiogenesis and that this involvement appears to extend to all may play an important role in gastric cancer development and of the cell types involved in vessel formation (49, 50). Further- progression. more, evidence has shown that Sp1 also up-regulates matrix The notion that abnormal Sp1 expression and activation metalloproteinase-2 (51, 52) and urokinase-type plasminogen contribute to gastric cancer development and progression is well activator (53), which play very important roles in tumor inva- supported not only by our study of pancreatic cancer (12) but sion and metastasis. also by other lines of evidence showing that Sp1 may regulate Given the demonstrated role of the Sp1 factors in growth many aspects of cancer biology, including cell growth, survival, regulation, cell survival, angiogenesis, and metastasis, it is con- invasion, and angiogenesis. Early studies indicated that Sp1 ceivable that members of the Sp1 family have been implicated regulates multiple growth-regulated genes, arguing that Sp1 in tumor development and progression. In fact, Sp1-mediated may be important for cell-growth regulation. Direct evidence of transcription is involved in many signal transduction pathways the ability of Sp1 to regulate transcription during changes in cell linked to cancer; this role has been shown to directly impinge growth came with the demonstration that Sp1 affects serum upon transformation (54). In addition, Sp1 mRNA, and Sp1 and stimulation of quiescent cells at the rep3a promoter (25), as well Sp3 DNA-binding activity are increased in epithelial tumors as at the hamster dihydrofolate reductase (26, 27) and ornithine when compared with papillomas, indicating that increased ac- decarboxylase promoters (21). More recently, several studies tivity of these factors contributes to tumor progression in skin established the ability of Sp1 to mediate the growth induction of (21). Interestingly, a dominant-negative Sp1 (a truncated a variety of promoters, including those of the genes encoding COOH-terminal Sp1, which contains the zinc finger DNA- FGFR1, epidermal growth factor receptor, insulin-like growth binding domain but not the major transactivating domains of the factor receptor 1 (28–31), insulin-like growth factor-binding protein) has been shown to inhibit the growth of HeLa cells (22). protein 2 (32), VEGF (12, 33), thymidine kinase (34), and serum Targeting Sp1 using Sp1 decoy oligonucleotides or small inhib- response factor (35). In addition to altered growth regulation, itory RNA blocks cell cycle progression and suppresses cell tumor progression requires resistance to apoptosis and stimula- growth (55, 56). In fact, nuclear levels of Sp1 protein have been

Fig. 4 Correlation of elevated Sp1 expression with reduced survival in gastric cancer patients. A, Sp1 expression was determined in 86 gastric cancer, 53 metastatic lymph node, and 57 normal human gastric tissue specimens. Over- all, lymph node metastasis specimens exhibited the highest level of Sp1 expression (P Ͻ 0.001), and tumor tissue specimens had a much higher level of Sp1 expression than adjacent normal tissue specimens did (P Ͻ 0.001). B, Kaplan-Meier plots of overall survival in patients who had a tumor with negative, weak, or strong Sp1 expression. The survival curve for 17 patients who had a tumor with negative Sp1 expression was not statistically different from that for 48 patients who had a tumor with weak Sp1 expression (P ϭ 0.89). However, the survival curve for 21 patients who had a tumor with strong Sp1 expression was significantly worse than that for the 17 patients with negative Sp1 expression (P Ͻ 0.001) as well as that for the 48 patients with weak Sp1 expression (P Ͻ 0.001).

Table 2 Survival analyses of gastric cancer patients Median survival Multivariate hazard ratio Parameters (months) (95% confidence interval) Multivariate P Sp1 expression Negative 43 1 (Referent) 0.003 Weak 37 1.8 (0.8–4.1) Strong 8 4.5 (1.8–11.2) Stage I Not reached 1 (Referent) Ͻ0.001 II 63 1.4 (0.4–4.6) III 23 5.2 (1.7–15.6) IV 9 7.6 (2.0–28.6) Resection status R0 41 1 (Referent) 0.160 R1 and R2 10 1.9 (0.8–4.4) Lauren’s histology type Intestinal 31 1 (Referent) 0.799 Diffuse 28 .9 (0.5–1.8) Age (continuous variable) 1 (0.97–1.01) 0.609 Gender Female 48 1 (Referent) 0.661 Male 26 1.1 (0.6–2.1)

shown to correlate directly with cell proliferation and predom- VEGF expression (66). Whether oncogenes and/or tumor sup-

inate in the G1 phase of the cell cycle through a proteasome- pressor genes contribute to constitutive Sp1 activation in human dependent degradation mechanism (57). Presumably, a high cancer cells and how their status affects that activation are not level of cell proliferation results in reduced Sp1 proteolysis, fully elucidated. It appears that activated oncogenes, including whereas reduced Sp1 degradation and enhanced Sp1 DNA bind- Ras, Src, and Raf, enhance the transcription activity of Sp1 ing may additionally augment the proliferation by overexpress- through constitutive activation of the p42/p44 mitogen-activated ing Sp1-responsive genes, including ODC and cyclin D1 (57). protein kinase pathway, which has often been observed in many We have consistently shown that Sp1 is constitutively overac- human tumors. In fact, p42/p44 mitogen-activated protein ki- tivated in and directly correlates with the angiogenic potential of nase directly phosphorylates Sp1 on threonines 453 and 739, human pancreatic cancer (12). The results of the current study and increases the DNA-binding ability of Sp1, thus activating may indicate that abnormally activated Sp1 contributes directly many Sp1-regulated genes implicated in tumor growth and to gastric cancer progression. However, additional studies are progression (66, 67). clearly needed to test whether Sp1 is a reliable tumor progres- Other factors may also be involved in Sp1 overactivation. sion marker and/or effective therapeutic target for gastric can- For example, development of stressful tumor microenvironment cer. may be important, especially at advanced tumor stages. It has On the other hand, the underlying mechanism that results in been shown that Sp1 activity can be modulated by stress factors Sp1 overactivation is currently unknown. During tumor devel- such as hypoxia (68, 69) and overproduction of free radicals, opment and progression, Sp1 can be overactivated through e.g., reactive oxygen species and nitric oxide (70–72), which are several potential mechanisms. A recent study showed that Sp1 prominent in the tumor microenvironment. Several lines of protein positively regulates its own promoter (58), although evidence additionally indicated that such stress factors can ac- post-translational modification and degradation of Sp1 protein tivate the p42/p42 mitogen-activated protein kinase pathway

directly regulate its transcriptional activity (13). Aberrant Sp1 and, to a lesser extent, c-Jun NH2-terminal kinase-related sig- activity may also be due to an altered Sp1:Sp3 ratio (18), naling pathways, which may be responsible for Sp1 overactiva- methylation status of Sp1-binding sites (19), and potentially tion (66). Therefore, both genetic and epigenetic factors can inhibitory proteins that bind to Sp1/Sp3 (20). The involvement cause Sp1 overactivation, which leads to altered expression of of altered oncogenes and suppressor genes is particularly im- multiple Sp1 downstream genes and contributes to tumor portant. Tumor suppressors such as p53 (59), p73 (60), von growth and metastasis. Hippel-Landau (61), and retinoblastoma protein (19, 62) can In summary, we discovered unique Sp1 expression in nor- physically interact with Sp1, form a complex, and block Sp1 mal gastric tissue and various gastric cancer tissues. The level of binding to the promoter of Sp1-regulated genes. For example, Sp1 expression was closely related to the postoperative progno- tumor cells harboring the inactivated von Hippel-Landau gene sis of gastric cancer. This study demonstrated that abnormally are associated with increased VEGF expression and angiogen- activated Sp1 expression represents a potential risk for poor esis (63–65). Also, von Hippel-Landau-mediated repression of prognosis and directly contributes to gastric cancer development VEGF expression has convincingly been shown to be mediated and progression. Therefore, preoperative determination of Sp1 by transcriptional and post-transcriptional mechanisms (49, 62). activity may be useful in deciding the modality and extent of At the transcriptional level, von Hippel-Landau forms a com- postoperative therapy. We are currently investigating the mo- plex with the Sp1 transcription factor and inhibits Sp1-mediated lecular mechanism by which the Sp1 signaling pathway regu-

lates gastric cancer development and progression, and whether factor receptor expression through a novel recognition sequence. J. Biol. this Sp1 pathway is a potential therapeutic target for controlling Chem., 273: 19294–303, 1998. gastric cancer growth and metastasis. 19. Clark, S. J., Harrison, J., and Molloy, P. L. Sp1 binding is inhibited by (m)Cp(m)CpG methylation. Gene, 195: 67–71, 1997. ACKNOWLEDGMENTS 20. Udvadia, A. J., Rogers, K. T., Higgins, P. D., Murata, Y., Martin, K. H., Humphrey, P. A., and Horowitz, J. M. Sp-1 binds promoter We thank Judy King for expert help in the preparation of this elements regulated by the RB protein and Sp-1-mediated transcription is manuscript and Don Norwood for editorial comments. stimulated by RB coexpression. Proc. Natl. Acad. Sci. USA, 90: 3265– 3269, 1993. REFERENCES 21. Kumar, A. P., and Butler, A. P. Serum responsive gene expression 1. Parkin, D. M., Pisani, P., and Ferlay, J. Global cancer statistics. CA mediated by Sp1. Biochem. Biophys. Res. Commun., 252: 517–523, Cancer J. Clin., 49: 33–64, 1999. 1998. 2. Parkin, D. M., Pisani, P., and Ferlay, J. Estimates of the worldwide 22. Chen, F., Zhang, F., Rao, J., and Studzinski, G. P. Ectopic expres- incidence of 25 major cancers in 1990. Int. J. Cancer, 80: 827–841, sion of truncated Sp1 transcription factor prolongs the S phase and 1999. reduces the growth rate. Anticancer Res., 20: 661–667, 2000. 3. Tahara, E. Molecular aspects of invasion and metastasis of stomach 23. Shi, Q., Wang, B., Xiong, Q., Le, X., Khan, N., and Xie, K. cancer. Verh. Dtsch. Ges. Pathol., 84: 43–49, 2000. Influence of nitric oxide synthase II gene disruption on tumor growth 4. Chan, A. O., Luk, J. M., Hui, W. M., and Lam, S. K. Molecular and metastasis. Cancer Res., 60: 2645–2650, 2000. biology of gastric carcinoma: from laboratory to bedside. J. Gastroen- 24. Wang, B., Xiong, Q., Shi, Q., Le, X., and Xie, K. Genetic disruption terol. Hepatol., 14: 1150–1160, 1999. of host interferon-␥ drastically enhanced the metastasis of pancreatic 5. Xie, K., and Huang, S. Regulation of cancer metastasis by stress adenocarcinoma through impaired expression of inducible nitric oxide pathways. Clin. Exp. Metastasis, 20: 31–43, 2003. synthase. Oncogene, 20: 6930–6937, 2001. 6. Siewert, J. R., Bottcher, K., Stein, H. J., and Roder, J. D. Relevant 25. Schilling, L. J., and Farnham, P. J. The bidirectionally transcribed prognostic factors in gastric cancer: ten-year results of the German dihydrofolate reductase and rep-3a promoters are growth regulated by Gastric Cancer Study. Ann. Surg., 228: 449–461, 1998. distinct mechanisms. Cell Growth Differ., 6: 541–548, 1995. 7. Xie, K. Interleukin-8 and human cancer biology. Cytokine Growth 26. Jensen, D. E., Black, A. R., Swick, A. G., and Azizkhan, J. C. Factor Rev., 12: 375–391, 2001. Distinct roles for Sp1 and E2F sites in the growth/cell cycle regulation 8. Brown, L. F., Berse. B., Jackman, R. W., Tognazzi, K., Manseau, of the DHFR promoter. J. Cell Biochem., 67: 24–31, 1997. E. J., Senger, D. R., and Dvorak, H. F. Expression of vascular perme- 27. Noe, V., Chen, C., Alemany, C., Nicolas, M., Caragol, I. Chasin. ability factor (vascular endothelial growth factor) and its receptors in L. A., and Ciudad, C. J. Cell-growth regulation of the hamster dihydro- adenocarcinomas of the gastrointestinal tract. Cancer Res., 53: 4727– folate reductase gene promoter by transcription factor Sp1. Eur. J. Bio- 4735, 1993. chem., 249: 13–20, 1997. 9. Guo, Y. S., Beauchamp, R. D., Jin, G. F., Townsend, C. M., Jr., and 28. Zhang, X., Li, Y., Dai, C., Yang, J., Mundel, P., and Liu, Y. Sp1 and Thompson, J. C. Insulinlike growth factor-binding protein modulates the Sp3 transcription factors synergistically regulate HGF receptor gene growth response to insulinlike growth factor 1 by human gastric cancer expression in kidney. Am. J. Physiol. Renal. Physiol., 284: F82–F94, cells. Gastroenterology, 104: 1595–1604, 1993. 2003. 10. Yoshikawa, T., Tsuburaya, A., Kobayashi, O., Sairenji, M., Mo- 29. DiMario, J. X. Activation and repression of growth factor receptor tohashi, H., Yanoma, S., and Noguchi, Y. Plasma concentrations of gene transcription (Review). Int. J. Mol. Med., 10: 65–71, 2002. VEGF and bFGF in patients with gastric carcinoma. Cancer Lett., 153: 7–12, 2000. 30. Parakati, R., and DiMario, J. X. Sp1- and Sp3-mediated transcriptional regulation of the fibroblast growth factor receptor 1 gene in 11. Sier, C. F., Kubben, F. J., Ganesh, S., Heerding, M. M., Griffioen, chicken skeletal muscle cells. J. Biol. Chem., 277: 9278–85, 2002. G., Hanemaaijer, R., van Krieken, J. H., Lamers, C. B., and Verspaget, H. W. Tissue levels of matrix metalloproteinases MMP-2 and MMP-9 31. Patel, S. G., and DiMario, J. X. Two distal Sp1-binding cis-ele- are related to the overall survival of patients with gastric carcinoma. ments regulate fibroblast growth factor receptor 1 (FGFR1) gene ex- Br. J. Cancer, 74: 413–417, 1996. pression in myoblasts. Gene, 270: 171–180, 2001. 12. Shi, Q., Le, X., Peng, Z., Tang, H., Xiong, Q., Wang, B., Li, X-C., 32. Kutoh, E., Margot, J. B., and Schwander, J. Identification and Abbruzzese, J. L., and Xie, K. Constitutive Sp1 activity is essential for characterization of the putative retinoblastoma control element of the rat differential constitutive expression of vascular endothelial growth factor insulin-like growth factor binding protein-2 gene. Cancer Lett., 136: in human pancreatic adenocarcinoma. Cancer Res., 61: 4143–4154, 187–194, 1999. 2001. 33. Milanini, J., Vinals, F., Pouyssegur, J., and Pages, G. p42/p44 MAP 13. Suske, G. The Sp-family of transcription factors. Gene, 238: 291– kinase module plays a key role in the transcriptional regulation of the 300, 1999. vascular endothelial growth factor gene in fibroblasts. J. Biol. Chem., 14. Hagen, G., Dennig, J., Prei␤, A., Beato, M., and Suske, G. Func- 273: 18165–18172, 1998. tional analyses of the transcription factor Sp4 reveal properties distinct 34. Sorensen, P., and Wintersberger, E. Sp1 and NF-Y are necessary from Sp1 and Sp3. J. Biol. Chem., 270: 24989–24994, 1995. and sufficient for growth-dependent regulation of the hamster thymidine 15. Dennig, J., Beato, M. and Suske, G. An inhibitor domain in Sp3 kinase promoter. J. Biol. Chem., 274: 30943–30949, 1999. regulates its glutamine-rich activation domains. EMBO J., 15: 5659– 35. Spencer, J. A., and Misra, R. P. Expression of the SRF gene occurs 5667, 1996. through a Ras/Sp/SRF-mediated- mechanism in response to serum 16. Majello, B., De Luca, P., and Lania, L. Sp3 Is a bifunctional growth signals. Oncogene, 18: 7319–7327, 1999. transcription regulator with modular independent activation and repres- 36. Hanahan, D., and Weinberg, R. A. The hallmarks of cancer. Cell, sion domains. J. Biol. Chem., 272: 4021–4026, 1997. 100: 57–70, 2000. 17. Kwon, H-S., Kim, M-S., Edenberg, H. J., and Hur, M-W. Sp3 and 37. Grillot, D. A., Gonzalez-Garcia, M., Ekhterae, D., Duan, L., Ino- Sp4 can repress transcription by competing with sp1 for the core hara, N., Ohta, S., Seldin, M. F., and Nunez, G. Genomic organization, cis-elements on the human ADH5/FDH minimal promoter. J. Biol. promoter region analysis, and chromosome localization of the mouse Chem., 274: 20–28, 1999. bcl-x gene. J. Immunol., 158: 4750–4757, 1997. 18. Hata, Y., Duh, E., Zhang, K., Robinson, G. S., and Aiello, L. P. 38. Rebollo, A., Dumoutier. L., Renauld, J. C., Zaballos, A., Ayllon, V., Transcription factors Sp1 and Sp3 alter vascular endothelial growth and Martinez, A. C. Bcl-3 expression promotes cell survival following

interleukin-4 deprivation and is controlled by AP1 and AP1-like tran- transfected to carcinoma cells suppresses the expression of vascular scription factors. Mol. Cell. Biol., 20: 3407–3416, 2000. endothelial growth factor, transforming growth factor ␤1, and tissue 39. Li, F., and Altieri, D. C. The cancer antiapoptosis mouse survivin factor and also cell growth and invasion activities. Cancer Res., 60: gene: characterization of locus and transcriptional requirements of basal 6531–6536, 2000. and cell cycle-dependent expression. Cancer Res., 59: 3143–3151, 56. Abdelrahim, M., Samudio, I., Smith, III, R., Burghardt, R., and 1999. Safe. S. Small inhibitory RNA duplexes for Sp1 mRNA block basal and 40. Kim, S. J., Denhez, F., Kim, K. Y., Holt, J. T., Sporn, M. B., and estrogen-induced gene expression and cell cycle progression in MCF-7 Roberts, A. B. Activation of the second promoter of the transforming breast cancer cells. J. Biol. Chem., 277: 28815–28822, 2002. growth factor-␤ 1 gene by transforming growth factor-␤ 1 and phorbol 57. Grinstein, E., Jundt, F., Weinert, I., Wernet, P., and Royer, H-D. ester occurs through the same target sequences. J. Biol. Chem., 264: Sp1 as G1 cell cycle phase specific transcription factor in epithelial cells. 19373–19378, 1989. Oncogene, 21: 1485–1492, 2002. 41. Wolf, G., Hannken, T., Schroeder, R., Zahner, G., Ziyadeh, F. N., 58. Nicolas, M., Noe, V., Jensen, K. B., and Ciudad, C. J. Cloning and and Stahl, R. A. Antioxidant treatment induces transcription and expres- characterization of the 5Ј-flanking region of the human transcription ␤ sion of transforming growth factor in cultured renal proximal tubular factor Sp1 gene. J. Biol. Chem., 276: 22126–22132, 2001. cells. FEBS Lett., 488: 154–159, 2001. 59. Borellini, F., and Glazer, R. I. Induction of Sp1–p53 DNA-binding 42. Kuo, C. T., Veselits, M. L., and Leiden, J. M. LKLF: A transcrip- heterocomplexes during granulocyte/macrophage colony-stimulating tional regulator of single-positive T cell quiescence and survival. Sci- factor-dependent proliferation in human erythroleukemia cell line TF-1. ence (Wash. DC), 277: 1986–1990, 1997b. J. Biol. Chem., 268: 7923–7928, 1993. 43. Rieber, M., Strasberg, S., and Rieber, M. Unequal nuclear Sp1/GC 60. Stros, M., Ozaki, T., Bacikova, A., Kageyama, H., and Nakagawara, box DNA binding activity distinguishes proliferating from differentiated A. HMGB1 and HMGB2 cell-specifically down-regulate the p53- and senescent or apoptotic cells. Int. J. Cancer, 83: 359–364, 1999. p73-dependent sequence-specific transactivation from the human Bax 44. Chalaux, E., Lopez-Rovira, T., Rosa, J. L., Pons, G., Boxer, L. M., gene promoter. J. Biol. Chem., 277: 7157–7164, 2002. Bartrons, R., and Ventura, F. A. zinc-finger transcription factor induced by TGF-␤ promotes apoptotic cell death in epithelial Mv1Lu cells. 61. Iliopoulos, O., Levy, A. P., Jiang, C., Kaelin, W. G., Jr., and FEBS Lett., 457: 478–482, 1999. Goldberg, M. A. Negative regulation of hypoxia-inducible genes by the von Hippel Lindau protein. Proc. Natl. Acad. Sci. USA, 93: 10595– 45. Piedrafita, F. J., and Pfahl, M. Retinoid-induced apoptosis and Sp1 10599, 1996. cleavage occur independently of transcription and require caspase activation. Mol. Cell. Biol., 17: 6348–6358, 1997. 62. Chen, L. I., Nishinaka, T., Kwan, K., Kitabayashi, I., Yokoyama, K., Fu, Y. H., Grunwald, S., and Chiu, R. The retinoblastoma gene 46. Finkenzeller, G., Sparacio, A., Technau, A., Marme, D., and Si- product RB stimulates Sp1-mediated transcription by liberating Sp1 emeister, G. Sp1 recognition sites in the proximal promoter of the human vascular endothelial growth factor gene are essential for platelet- from a negative regulator. Mol. Cell. Biol., 14: 4380–4389, 1994. derived growth factor-induced gene expression. Oncogene, 15: 669– 63. Maher, E. R., and Kaelin, W. G., Jr. von Hippel-Lindau disease. 676, 1997. Medicine (Baltimore), 76: 381–391, 1997 47. Luster, T. A., Johnson, L. R., Nowling, T. K., Lamb, K. A., 64. Siemeister, G., Weindel, K., Mohrs, K., Barleon, B., Martinybaron, Philipsen, S., and Rizzino, A. Effects of three Sp1 motifs on the G., and Marme, D. Reversion of deregulated expression of vascular transcription of the FGF-4 gene. Mol. Reprod. Dev., 57: 4–15, 2000. endothelial growth factor in human renal carcinoma cells by von Hippel- 48. Gerwins, P., Skoldenberg, E., and Claesson-Welsh, L. Function of Lindau tumor suppressor protein. Cancer Res. 7: 2299–2301, 1996 fibroblast growth factors and vascular endothelial growth factors and 65. Stratmann, R., Krieg, M., Haas, R., and Plate, K. H. Putative control their receptors in angiogenesis. Crit. Rev. Oncol. Hematol., 34: 185– of angiogenesis in hemangioblastomas by the von Hippel-Lindau tumor 194, 2000. suppressor gene. J. Neuropathol. Exp. Neurol., 56: 1242–1252, 1997. 49. Mukhopadhyay, D., Knebelmann, B., Cohen, H. T., Ananth, S., and 66. Milanini-Mongiat, J., Pouysse´gur, J., and Page`s, G. Identification of Sukhatme, V. P. The von Hippel-Lindau tumor suppressor gene product two Sp1 phosphorylation sites for p42/p44 mitogen-activated protein interacts with Sp1 to repress vascular endothelial growth factor pro- kinases. Their implication in vascular endothelial growth factor gene moter activity. Mol. Cell. Biol., 17: 5629–5639, 1997. transcription. J. Biol. Chem., 277: 20631–20639, 2002. 50. Cieslik, K., Abrams, C. S., and Wu, K. K. Upregulation of endo- 67. Merchant, J. L., Du, M., and Todisco, A. Sp1 phosphorylation by thelial nitric oxide synthase promoter by PI3K{␥}/Jak2/MEK-1-de- Erk 2 stimulates DNA binding. Biochem. Biophys. Res. Commun., 254: pendent pathway. J. Biol. Chem., 276: 1211–1219, 2001. 454–461, 1999. 51. Qin, H., Sun, Y., and Benveniste, E. N. The transcription factors 68. Discher, D. J., Bishopric, N. H., Wu, X., Peterson, C. A., and Sp1, Sp3, and AP-2 are required for constitutive matrix metalloprotein- Webster, K. A. Hypoxia regulates ␤-enolase and pyruvate kinase-M ase-2 gene expression in astroglioma cells. J. Biol. Chem., 274: promoters by modulating Sp1/Sp3 binding to a conserved GC element. 29130–7, 1999. J. Biol. Chem., 273: 26087–26093, 1998. 52. Price, S. J., Greaves, D. R., and Watkins, H. Identification of novel, 69. Xu, Q., Ji, Y. S., and Schmedtje, J. F., Jr. Sp1 increases expression functional genetic variants in the human matrix metalloproteinase-2 of cyclooxygenase-2 in hypoxic vascular endothelium. Implications for gene: role of Sp1 in allele-specific transcriptional regulation. J. Biol. the mechanisms of aortic aneurysm and heart failure. J. Biol. Chem., Chem., 276: 7549–7558, 2001. 275: 24583–24589, 2000. 53. Ibanez-Tallon, I., Ferrai, C., Longobardi, E., Facetti, I., Blasi, F., 70. Sellak, H., Yang, X., Cao, X., Cornwell, T., Soff, G. A., and and Crippa, M. P. Binding of Sp1 to the proximal promoter links Lincoln, T. Sp1 transcription factor as a molecular target for nitric constitutive expression of the human uPA gene and invasive potential of oxide– and cyclic nucleotide–mediated suppression of cGMP-dependent PC3 cells. Blood, 100: 3325–3332, 2002. protein kinase-I␣ expression in vascular smooth muscle cells. Circ. Res., 54. Lee, R. J., Albanese, C., Fu, M., D’Amico, M., Lin, B., Watanabe, 90: 405–412, 2002. G., Haines, G. K. 3rd, Siegel, P. M., Hung, M. C., Yarden, Y., Horowitz, 71. Wang, S., Wang, W., Wesley, R. A., and Danner, R. L. A. Sp1 J. M., Muller, W. J., and Pestell, R. G. Cyclin D1 is required for binding site of the tumor necrosis factor ␣ promoter functions as a nitric transformation by activated Neu and is induced through an E2F-depend- oxide response element. J. Biol. Chem., 274: 33190–33193, 1999. ent signaling pathway. Mol. Cell. Biol., 20: 672–683, 2000. 72. Ammendola, R., Mesuraca, M., Russo, T., and Cimino, F. The 55. Ishibashi, H., Nakagawa, K., Onimaru, M., Castellanous, E. J., DNA-binding efficiency of Sp1 is affected by redox changes. Eur. Kaneda, Y., Nakashima, Y., Shirasuna, K., and Sueishi, K. Sp1 decoy J. Biochem., 225: 483–489, 1994.

Downloaded from clincancerres.aacrjournals.org on September 28, 2021. © 2003 American Association for Cancer Research. Transcription Factor Sp1 Expression Is a Significant Predictor of Survival in Human Gastric Cancer

Liwei Wang, Daoyan Wei, Suyun Huang, et al.

Clin Cancer Res 2003;9:6371-6380.

Updated version Access the most recent version of this article at: http://clincancerres.aacrjournals.org/content/9/17/6371

Cited articles This article cites 72 articles, 35 of which you can access for free at: http://clincancerres.aacrjournals.org/content/9/17/6371.full#ref-list-1

Citing articles This article has been cited by 52 HighWire-hosted articles. Access the articles at: http://clincancerres.aacrjournals.org/content/9/17/6371.full#related-urls

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://clincancerres.aacrjournals.org/content/9/17/6371. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.