点击显示 收起
Institute of Biosciences and Technology, the Texas A&M University System Health Science Center, Houston, Texas (M.A., S.S.)
Department of Veterinary Physiology and Pharmacology, Texas A&M University, College Station, Texas (S.S.)
Abstract
Cyclooxygenase 2 (COX-2) inhibitors, such as celecoxib (Cel), nimesulfide (NM), and NS-398 [NS; N-[2-(cyclohexyloxy)-4-nitrophenyl]methanesulfonamide] and other nonsteroidal anti-inflammatory drugs inhibit colon cancer growth and angiogenesis; however, the mechanism of this response is not well defined. Treatment of SW-480 colon cancer cells with Cel, NS, or NM decreased vascular endothelial growth factor (VEGF) mRNA and immunoreactive protein expression. This was also accompanied by decreased transactivation in cells transfected with constructs containing VEGF gene promoter inserts. Deletion analysis of the VEGF promoter indicated that decreased VEGF expression by COX-2 inhibitors was associated with the proximal -131 to -47 GC-rich region of the VEGF promoter that binds Sp proteins. Treatment of SW-480 cells with Cel, NM, and NS also decreased Sp1 and Sp4 protein expression but not that of Sp2 or Sp3. Similar results were observed in RKO, HT-29, and DLD colon cancer cells demonstrating comparable responses in COX-2-expressing and -nonexpressing colon cancer cell lines. COX-2 inhibitors do not affect Sp1 or Sp4 mRNA levels in SW-480 cells; however, decreased expression of both proteins was accompanied by increased protein ubiquitination and inhibited by the proteasome inhibitor gliotoxin. These results suggest that the antiangiogenic activity of COX-2 inhibitors in colon cancer cells is linked to activation of proteasome-dependent degradation of Sp1 and Sp4 proteins.
Vascular endothelial growth factor (VEGF) is an angiogenic protein that plays an important role in cellular homeostasis and is a key regulator of embryogenesis, the menstrual cycle, and wound repair (Shifren et al., 1996; Torry and Torry, 1997; Carmeliet, 2003; Li et al., 2003). VEGF and VEGF-like proteins stimulate formation of new blood vessels or neovascularization, and this process is also essential in carcinogenesis for tumor growth and metastasis (Folkman, 1995; Bergers and Benjamin, 2003; Carmeliet, 2003). VEGF and related angiogenic factors are overexpressed in multiple tumor types and often serve as negative prognostic indicators for disease-free survival of cancer patients (Maeda et al., 1999; Mineta et al., 2000; Linderholm et al., 2001). Expression of VEGF in a cohort of breast cancer patients correlated with mutant p53 and poor prognosis for recurrence-free survival (Linderholm et al., 2001). Other studies show that high levels of VEGF correlate with poor prognosis for patients with pancreatic cancer, squamous cell carcinoma, gastric carcinoma, and colon cancer (Maeda et al., 1999; Mineta et al., 2000; Linderholm et al., 2001).
Five members of the VEGF family have been identified (VEGF-A, -B, -C, -D, and -F) and these proteins induce angiogenic responses through binding VEGF receptor-1 (VEGFR-1, or Flk-1), VEGFR-2 (KDR/Flk-1), or VEGFR-3 (Flt-4), which are tyrosine kinases that initiate intracellular kinase signaling pathways (Ferrara et al., 2003). Hypoxia is a major pathway for regulation of VEGF in tumor and non-tumor tissue (Shweiki et al., 1992; Rossler et al., 1999; Ziemer et al., 2001), in part because of up-regulation of hypoxia-inducible factor-1 (HIF-1). HIF-1 forms a transcriptionally active HIF-1-HIF-1 complex that interacts with a distal hypoxia-responsive element in the VEGF gene promoter. Enhanced VEGF expression in cells/tissues has also been linked to other factors, including treatment with cytokines, mitogens, activation of kinase signaling pathways, oxidative stress, and hormone stimulation (Finkenzeller et al., 1997; Milanini et al., 1998; Fukuda et al., 2003; Schafer et al., 2003; Stoner et al., 2004). It is interesting that enhanced VEGF stimulation by some of these factors involves direct or indirect activation of members of the specificity protein (Sp) family bound to one or more GC-rich motifs located in the proximal region of the VEGF promoter (Finkenzeller et al., 1997; Milanini et al., 1998; Schafer et al., 2003; Stoner et al., 2004). Platelet-derived growth factor-enhanced expression of VEGF in NIH3T3 cells involves Sp1 and Sp3 bound to the -85 to -50 region of the VEGF promoter (Finkenzeller et al., 1997). Oxidative stress-induced activation of kinases enhance VEGF expression in gastric cancer cells through increased Sp1-dependent activation of the same proximal GC-rich sites (Schafer et al., 2003). Research in this laboratory showed that hormone-induced VEGF expression in ZR-75 breast cancer cells required ER/Sp1 and ER/Sp3 interactions with the proximal GC-rich VEGF promoter elements (Stoner et al., 2004).
Recent studies have demonstrated that Sp1 and other Sp family proteins are required for endogenous expression of VEGF in pancreatic cancer cells (Shi et al., 2001; Abdelrahim et al., 2004), and the cyclooxygenase 2 (COX-2) inhibitor celecoxib also decreased VEGF expression in these cells by suppressing Sp1 activity (Wei et al., 2004). COX-2 inhibitors and other nonsteroidal anti-inflammatory drugs (NSAIDs) inhibit development and growth of colon cancer, and this is due, in part, to their antiangiogenic activity (Williams et al., 1999; Gately and Li, 2004). In this study, we have investigated the effects of three COX-2 inhibitors, celecoxib (Cel), NS-398 (NS), and nimesulide (NM) on VEGF expression in colon cancer cells. The results show that all three COX-2 inhibitors decreased VEGF protein, mRNA, and reporter gene expression, and this was accompanied by down-regulation of both Sp1 and Sp4 but not Sp3 or Sp2 proteins. The effects of the COX-2 inhibitors on VEGF could also be duplicated in colon cancer cells transfected with small inhibitory RNAs (siRNAs) for Sp1 and Sp4, confirming the role of these proteins in mediating VEGF expression. These results and studies with the proteasome inhibitor gliotoxin suggest that the antiangiogenic activity of COX-2 inhibitors in colon cancer may be related to decreased Sp1/Sp4 expression caused by activation of proteasomes and represents a novel mechanism of action for this class of antitumor agents.
Materials and Methods
Cell Lines, Chemicals, Biochemicals, Constructs, and Oligonucleotides. SW-480, DLD, RKO, and HT-29 cells were obtained from the American Type Culture Collection (Manassas, VA). DMEM/Ham's F12 medium with and without phenol red, 100x antibiotic/antimycotic solution was purchased from Sigma-Aldrich (St. Louis, MO). Fetal bovine serum was purchased from Intergen (Purchase, NY). [-32P]ATP (300Ci/mmol) was obtained from PerkinElmer Life Sciences. Poly(dI-dC) and T4 polynucleotide kinase were purchased from Roche Molecular Biochemicals (Indianapolis, IN). Antibodies for Sp1, Sp2, Sp3, Sp4, ubiquitin, COX-2, -tubulin, and VEGF proteins were obtained from Santa Cruz Biotechnology (Santa Cruz, CA). Cel was a kind gift from Dr. Robert Newman (M. D. Anderson Cancer Center, Houston, TX). NM and NS were purchased from Alexis Biochemicals (San Diego, CA). Gliotoxin was kindly provided by Dr. Alan Taylor (National Research Council of Canada, Halifax, NS, Canada). Lysis buffer and luciferase reagent were obtained from Promega (Madison, WI). VEGF promoter constructs have previously been described previously (Abdelrahim et al., 2004; Stoner et al., 2004). siRNA duplexes for Sp1, Sp2, Sp3, Sp4, COX-2, luciferase (GL2), and the scrambled siRNA were prepared by Dharmacon Research (Lafayette, CO). The sequences of these siRNAs are shown in Table 1.
Transfection of Colon Cancer Cells and Preparation of Nuclear Extracts. Cells were cultured in six-well plates in 2 ml of DMEM/Ham's F12 medium supplemented with 5% fetal bovine serum. After 16 to 20 h, when cells were 50 to 60% confluent, iRNA duplexes and/or reporter gene constructs were transfected using Oligofectamine reagent (Invitrogen, Carlsbad, CA). The effects of the selective COX-2 inhibitors (Coxib) and Sp siRNAs on transactivation was investigated in SW-480 cells cotransfected with (500 ng) different VEGF constructs. In brief, iRNA duplex was transfected in each well to give a final concentration of 50 nM. Cells were harvested 48 h later, and luciferase activity of lysates (relative to -galactosidase activity) was determined. For COX-2 inhibitor study, cells were treated with DMSO (control) or with the indicated concentration of celecoxib, nimesulide, and NS-398 for 24 to 30 h. For electrophoretic mobility shift assay, nuclear extracts from SW-480 cells were isolated as described previously, and aliquots were stored at -80°C until used (Abdelrahim et al., 2004; Stoner et al., 2004).
Western Immunoblot. Cells were washed once with PBS and collected by scraping in 200 e of lysis buffer [50 mM HEPES, 0.5 M sodium chloride, 1.5 mM magnesium chloride, 1 mM EGTA, 10% (v/v) glycerol, 1% Triton X-100, and 5 e/ml of protease inhibitor cocktail] (Sigma-Aldrich). The lysates from the cells were incubated on ice for 1 h with intermittent vortexing followed by centrifugation at 40,000 g for 10 min at 4°C. Equal amounts of protein (60 e) from each treatment group were diluted with loading buffer, boiled, and loaded onto 10 and 12.5% SDS-polyacrylamide gel. For VEGF immunoblots, 100 e of protein was used. Samples were electrophoresed, and proteins were detected by incubation with polyclonal primary antibodies Sp1 (PEP2), Sp2 (K-20), Sp3 (D-20), Sp4 (V-20), COX-2 (H-62), VEGF (A-20), and -tubulin (H-235) followed by blotting with appropriate horseradish peroxidase-conjugated secondary antibody as described previously (Abdelrahim et al., 2004). After autoradiography, band intensities were determined by a scanning laser densitometer (Sharp Electronics, Mahwah, NJ) using Zero-D Scanalytics software (Scanalytics, Billerica, MA).
Electrophoretic Mobility Shift Assay. VEGF oligonucleotides were synthesized and annealed, and 5-pmol aliquots were 5'-endlabeled using T4 kinase and [-32P]ATP. A 30-e electrophoretic mobility shift assay reaction mixture contained 100 mM KCl, 3 e of crude nuclear protein, 1 e of poly(dI-dC), with or without unlabeled competitor oligonucleotide, and 10 fmol of radiolabeled probe. After incubation for 20 min on ice, antibodies against Sp1, Sp2, Sp3, or Sp4 proteins were added and incubated another 20 min on ice. Protein-DNA complexes were resolved by 5% polyacrylamide gel electrophoresis as described previously (Abdelrahim et al., 2004; Stoner et al., 2004). Specific DNA-protein and antibody-supershifted complexes were observed as retarded bands in the gel. GC- and GT-box sequence used in gel shift analysis are given below:
Immunocytochemistry. SW-480 cells were seeded in Lab-Tek Chamber slides (Nalge Nunc International, Naperville, IL) at 100,000 cells/well in DMEM/Ham's F12 medium supplemented with 5% fetal bovine serum. Cells were then treated with celecoxib, nimesulide, or NS-398. After 27 h, the media chamber was detached, and the remaining glass slides were washed in Dulbecco's PBS. The immunostaining for Sp2, Sp3, and Sp4 was determined essentially as described previously for Sp1 (Abdelrahim et al., 2004). In brief, the glass slides were fixed with cold (-20°C) methanol for 10 min, and then slides were washed in 0.3% PBS/Tween 20 for 5 min (2x) before blocking with 5% goat serum in antibody dilution buffer (stock solution, 100 ml of PBS/Tween 20, 1 g of bovine serum albumin, and 45 ml of glycerol, pH 8.0) for 1 h at 20°C. After removal of the blocking solution, Sp1, Sp2, Sp3, or Sp4 rabbit polyclonal antibodies were added in antibody dilution buffer (1:200) and incubated for 12 h at 4°C. Blank panels were incubated with antibody dilution buffer only with no primary antibody. Slides were washed for 10 min with 0.3% Tween 20 in 0.02 M PBS (3x) and incubated with fluorescein isothiocyanate-conjugated goat anti-rabbit antibodies (1:1000 dilution) for 2 h at 20°C. Slides were then washed for 10 min in 0.3% PBS-Tween 20 (4x). Slides were mounted in ProLonged antifading medium (Molecular Probes, Inc., Eugene, OR), and coverslips were sealed using Nailslicks fingernail polish (Noxell Corp., Hunt Valley, MD). For VEGF staining, rabbit polyclonal antibody (A-20) was used, and mounting media with DAPI was used for nuclear counterstaining. Fluorescence imaging was performed using Axiophoto 2 (Carl Zeiss, Inc., Thornwood, NY), and Photoshop 5.5 (Adobe Systems, Mountain View, CA) was used to capture the images.
Ubiquitinated Sp Protein Immunoprecipitation. SW-480 cells were seeded into 150-mm tissue culture plates in maintenance medium and allowed to grow to approximately 90% confluence. Cells were treated with DMSO, 30 e Cel, or 60 e NS for 7 h. Whole-cell extracts for each treatment group were obtained using radioimmunoprecipitation assay buffer (50 mM Tris-HCL, 1% Nonidet P-40, 0.25% sodium deoxycholate, 150 mM NaCl, and 1 mM EDTA) with the addition of protease inhibitor cocktail. Duplicate aliquots of 500 e were used for the experiments. Cell extracts were diluted in ice-cold PBS containing protease inhibitor cocktail to a final volume of 1 ml, followed by the addition of 30 e of protein A/G PLUS-agarose beads (Santa Cruz Biotechnology). The reactions were placed on a rocker at 4°C for 3 h, followed by 600 g centrifugation at 4°C for 5 min. A 900-e aliquot of supernatant from each sample was transferred into a new Eppitube on ice. Rabbit polyclonal anti-Sp1 (1 e), Sp4 (1 e), Sp3 (1 e), or normal rabbit IgG (1 e) was added to either treatment set, followed by the addition of 30 e of protein A/G PLUS-agarose beads. The samples were then placed on a rocker at 4°C for 12 h, followed by centrifugation at 600 g at 4°C for 5 min. The supernatant was removed by aspiration and the immunoprecipitates were washed with two cycles of 1 ml of ice-cold radioimmunoprecipitation assay buffer followed by 1 ml of ice-cold PBS using centrifugation at 600g at 4°C for 5 min. The agarose pellet was resuspended in 50 e of loading buffer, boiled, and centrifuged. The supernatant was separated by SDS-10% PAGE, electrophoresed to polyvinylidene difluoride membrane. The polyvinylidene difluoride membrane probed with ubiquitin antibody (P4D1), then stripped and reprobed with Sp1 or Sp3 antibodies. The same membrane was then stripped and reprobed with Sp4 antibody and visualized by enhanced chemiluminescence as described above.
Semiquantitative Reverse Transcription-PCR Analysis. SW-480 cells were treated with DMSO (control) or with the indicated concentration of celecoxib, nimesulide, and NS-398 for 30 h before total RNA collection. Total RNA was obtained with RNAzol B (Tel-Test, Friendswood, TX) according to the manufacturer's protocol. RNA concentration was measured by UV 260:280 nm absorption ratio, and 200 ng/e RNA were used in each reaction for reverse transcription-PCR. RNA was reverse-transcribed at 42°C for 25 min using oligo d(T) primer (Promega) and subsequently PCR-amplified of reverse transcription product using 2 mM MgCl2, 1 e concentrations of each gene-specific primer, 1 mM dNTPs, and 2.5 units of AmpliTaq DNA polymerase (Promega). The gene products were amplified using 22 to 25 cycles (95°C, 30 s; 56°C, 30 s; 72°C, 30 s). The sequence of the oligonucleotide primers used in this study was as shown in Table 2.
After amplification in a PCR express thermal cycler (Hybaid US, Franklin, MA), 20 e of each sample was loaded on a 2% agarose gel containing ethidium bromide. Electrophoresis was performed at 80 V in 1x Tris-acetate/EDTA buffer (0.04 M/0.001 M) for 1 h, and the gel was photographed by UV transillumination using Polaroid film (Waltham, MA). VEGF, GAPDH, Sp1, and Sp4 band intensity values were obtained by scanning the Polaroid print on a Sharp JX-330 scanner. Background signal was subtracted, and densitometric analysis was performed on the inverted image using Zero-D software (Scanalytics). Results were expressed as VEGF, Sp1, or Sp4 band intensity values normalized to GAPDH values and then by averaging three separate determinations for each treatment group.
Statistical Analysis. Statistical significance was determined by analysis of variance and Scheffee's test, and the levels of probability are noted. The results are expressed as means ± S.D. for at least three separate (replicate) experiments for each treatment.
Results
SW-480 colon cancer cells express mutant K-ras and p53, and several COX-2 inhibitors inhibit growth and/or induce apoptosis in SW-480 colon cancer cell lines (Smith et al., 2000; Yamazaki et al., 2002). Although the effects of these compounds are variable, many of the same responses are observed in COX-2-expressing and -nonexpressing cell lines, such as SW-480 cells (Williams et al., 1999; Smith et al., 2000; Yamazaki et al., 2002; Gately and Li, 2004). COX-2 inhibitors also inhibit angiogenesis in colon cancer cells (Gately and Li, 2004), and this study has focused on the effects of the COX-2 inhibitors Cel, NS, and NM on expression of the angiogenic protein VEGF in SW-480 cells. Results illustrated in Fig. 1A show that Cel, NS, and NM significantly decreased VEGF-A mRNA levels in SW-480 cells, and the effects of these compounds on VEGF protein were also investigated by immunostaining in the same cell line. Secreted VEGF exhibits a background cytosolic staining indicative of expression of this protein in SW-480 cells (Fig. 1B). However, after treatment with the COX-2 inhibitors for 27 h, all three compounds decreased immunostaining of VEGF in SW-480 cells. Western blot analysis also confirmed decreased expression of VEGF after treatment with Cel or NS (Fig. 1C). The VEGF promoter contains multiple cis-elements that may be involved in decreased VEGF expression in SW-480 cells treated with the COX-2 inhibitors, and these include activating protein (AP) 1, AP-2, hypoxia-responsive elements, and a cluster of proximal GC-rich sites at -131 to -52 that bind Sp proteins and Egr-1 (Finkenzeller et al., 1997; Abdelrahim et al., 2004; Stoner et al., 2004). Initial studies investigated the effects of Cel (30 e), NS (60 e), and NM (60 e) on transactivation in SW-480 cells transfected with pVEGF-1, pVEGF-2, and pVEGF-3 constructs (Fig. 2A), which contain -2018 to +50, -131 to +54, and -66 to -54 VEGF gene promoter inserts (Fig. 2, B-D). The COX-2 inhibitors decreased transactivation by 40 to 60% in cells transfected with the three constructs. The pattern of inhibition was similar for the three COX-2 inhibitors. A concentration-dependent decrease in luciferase activity in cells treated with the COX-2 inhibitors and transfected with pVEGF-2 indicated that their order of potency for decreasing transactivation was Cel > NS > NM (Fig. 2E).
Results of deletion analysis of the VEGF promoter indicated that luciferase activity was decreased by <50% after deletion of the -2018 to -132 region of the promoter. In contrast, deletion of the cluster of proximal GC-rich sites -131 to -67 (i.e., pVEGF-2 versus pVEGF-3) resulted in a >85% decrease in luciferase activity, and a similar decrease was observed (data not shown) after deletion of the -66 to -50 region, which contains two additional GC-rich sites. Previous studies on VEGF/VEGF promoter expression in pancreatic cancer cells gave similar results, and it was shown that Sp1, Sp3, and Sp4 were the major transcription factors required for VEGF expression (Abdelrahim et al., 2004). Moreover, a recent study indicated that Cel decreased VEGF expression in pancreatic cancer cells and that this was linked to decreased Sp1 protein and phospho-Sp1/Sp1 protein ratios (Wei et al., 2004). Results in Fig. 3A show that Sp1, Sp3, and Sp4 are expressed in SW-480 cells; moreover, we also identified Sp2 protein in SW-480 cells as previously reported in other cancer cell lines. All three COX-2 inhibitors significantly decreased Sp1 and p-Sp1 proteins; however, a significant decrease in p-Sp1/Sp1 ratio was observed only for NS. Sp4 protein expression was also decreased (Fig. 3E) in a manner parallel to that of Sp1, whereas levels of Sp2 and Sp3 were not affected by Cel, NS, or NM (Fig. 3, C and D). Nuclear extracts from SW-480 cells were incubated with
32P-labeled (-66/-47), which contains the two proximal VEGF promoter GC-rich sites (Fig. 4A), and a series of retarded bands were formed (lane 2), whereas in the absence of protein, retarded bands were not observed (lane 1). Incubation with nonspecific IgG (lane 3) or Sp2 antibodies (lane 5) did not affect the retarded band. Antibodies to Sp1 (lane 4), Sp3 (lane 6), and Sp4 (lane 7) supershifted specific retarded bands (as indicated by an arrow) and 100-fold excess unlabeled (-66/-47) decreased retarded band intensities (lane 8). This same 32P-labeled oligonucleotide was incubated alone (lane 1) or with nuclear extracts from untreated SW-480 cells (lanes 2 and 6) and from cells treated with Cel (lanes 3-5), and NS (lanes 7-9). The only consistent decrease that was observed in gel mobility shift assays was the retarded band associated with Sp1 in the high-dose treatment groups. This was due, in part, to the relatively intense (and overlapping) Sp3-DNA bands, which are not affected by the COX-2 inhibitors. Confirmation that the COX-2 inhibitors down-regulated Sp1 and Sp4, but not Sp2 and Sp3 proteins, was confirmed in immunostaining experiments in SW-480 cells (Fig. 5). The blank panels were stained only with the secondary antibody in the presence (right) or absence (left) of the DAPI nuclear stain. All four proteins exhibited nuclear staining, and decreased staining intensities of Sp1 and Sp4 (but not Sp2 and Sp3) were observed after treatment with Cel, NS, and NM.
These results suggest that COX-2 inhibitors specifically decrease Sp1 and Sp4 expression, and previous studies in pancreatic cancer cells show that Sp1, Sp4,, and Sp3 play a critical role in VEGF expression (Abdelrahim et al., 2004).
This was further investigated in SW-480 cells transfected with siRNA for Sp1 (iSp1), Sp2 (iSp2), Sp3 (iSp3), and Sp4 (iSp4) (Fig. 6A) which resulted in >50% knockdown of their targeted protein based on analysis of whole-cell lysates of transfected cells. In cells transfected with siRNA for Sp proteins and pVEGF-2, it was apparent that decreased expression of Sp1, Sp3, and Sp4 decreased luciferase activity. These data confirmed that down-regulation of Sp1 and Sp4 by the COX-2 inhibitors was responsible for decreased VEGF expression in SW-480 cells; moreover, overexpression of Sp1 or Sp4 induced transactivation in cells transfected with pVEGF-2, and this response was inhibited by Cel (Fig. 6, C-E). It was also shown that iSp3, but not iSp2, inhibited transactivation (Fig. 6B). Overexpression of Sp3 also enhanced luciferase activity, and treatment with Cel decreased this response (Fig. 6E). These results confirm that Sp proteins play an important role in regulating VEGF expression in SW-480 cells; therefore, the antiangiogenic responses observed for COX-2 inhibitors is caused by their inhibition of Sp1 and Sp4 protein expression in this cell line.
The effects of Cel and NS on Sp1, Sp3, and Sp4 protein levels were also determined in RKO, HT-29, and DLD cells that were treated for 30 h with 30 and 60 e concentrations of the COX-2 inhibitors, respectively (Fig. 7). COX-2 was expressed in these cells (Fig. 7A). The role of Sp proteins in regulating VEGF in COX-2 overexpressing HT-29 cells was investigated in cells transfected with pVEGF-2, iScr (nonspecific), iGL2 (luciferase), and iSp1, iSp2, iSp3, and iSp4 as well as iCOX-2 (Fig. 7B). The results were similar to those observed in SW-480 cells (Fig. 6A) and demonstrate that Sp1, Sp3, and Sp4 are critical transcription factors required for VEGF expression in HT-29 cells. iCOX-2 significantly decreased COX-2 protein in HT-29 cells but did not affect transactivation in cells transfected with pVEGF2. Moreover, both compounds also decreased Sp1 and Sp4 proteins and luciferase activity in RKO, HT-29, and DLD cells transfected with pVEGF-2 (Fig. 7C), and this was similar to results observed in SW-480 cells (Fig. 2). These results demonstrate that COX-2 inhibitors specifically decrease Sp1 and Sp4 protein expression in several colon cancer cell lines, and this also resulted in decreased VEGF expression.
The effects of these compounds on VEGF expression were not related to decreased p-Sp1/Sp1 ratios, and NS and NM did not decrease Sp1 or Sp4 mRNA levels (Fig. 8, A and B), suggesting a post-transcriptional effect on Sp1 and Sp4 protein expression. Preliminary studies showed that proteasome inhibitors, such as MG-132 and lactacystin, were cytotoxic in SW-480 cells; gliotoxin, however, a fungal metabolite that inhibits chymotrypsin-like activity of the 20S proteasome, significantly inhibited NS-induced degradation of Sp1 and Sp4 (Fig. 8C) and also inhibited the NS-dependent decrease in luciferase activity in SW-480 cells transfected with pvEGF2 (Fig. 8D). Thus, the antiangiogenic activity of COX-2 inhibitors is associated with selective activation of proteasome-dependent degradation of Sp1 and Sp4 proteins.
Proteasome-dependent degradation of proteins is preceded by ubiquitination of the targeted protein or increased conjugation with a ubiquitin-like protein. The effects of Cel and NS on ubiquitination of Sp1, Sp3, and Sp4 were therefore investigated in SW-480 cells (Fig. 9) in which whole-cell lysates from each treatment group were initially immunoprecipitated with IgG or antibodies to Sp1, Sp3, or Sp4. Immunoprecipitated fractions were separated by SDS-polyacrylamide gel electrophoresis and immunoblotted with ubiquitin antibodies (Fig. 9A), and multiple bands associated with ubiquitinated proteins were observed in lysates from NS- or Cel-treated cells and immunoprecipitated with Sp1 or Sp4 (but not Sp3) antibodies. The multiple bands were also not observed in DMSO-treated lysates (D); in addition, an unidentified low molecular weight band was observed in the DMSO, Cel, and NS lysates immunoprecipitated with Sp1, Sp3, or Sp4 antibodies but not with IgG. The membrane was then stripped and probed first with Sp1 antibodies (Fig. 9B); without stripping, the membrane was reprobed with Sp3 antibodies (Fig. 9C), which shows both Sp1 and Sp3 proteins. The membrane was then stripped and reprobed with Sp4 antibodies (Fig. 9D), which appears primarily as a single band on the gel. These latter experiments (Fig. 9, B-D) were used as controls to confirm the initial Sp antibody immunoprecipitation. The results confirm that COX-2 inhibitor-induced degradation of Sp1 and Sp4 is accompanied by enhanced protein ubiquitination, and this is consistent with not only the observed degradation of these proteins but also the inhibition of this response by the proteasome inhibitor gliotoxin (Fig. 8, C and D).
Discussion
COX-2 is one of the rate-limiting steps in the metabolism of arachidonic acid to prostaglandins and thromboxanes, and COX-2 inhibitors, which include a broad spectrum of NSAIDs, are extensively used as anti-inflammatory agents (Williams et al., 1999). COX-2 is also overexpressed in multiple tumor types, including colon cancer, and both NSAIDs and COX-2 inhibitors have been investigated for cancer chemoprevention and chemotherapy (Taketo, 1998a,b; Williams et al., 1999; Gately and Li, 2004). Several studies show that some COX-2 inhibitors, including aspirin, decrease the incidence and/or mortality rate of human colorectal cancer (Giovannucci et al., 1995; Martinez et al., 1995; Peleg et al., 1996). The COX-2 inhibitor sulindac has also been successfully used for treating patients with familial adenomatous polyposis coli, a hereditary colon cancer syndrome in which there is rapid and early development of polyps and tumors (Labayle et al., 1991; Giardiello et al., 1993; Nugent et al., 1993). COX-2 inhibitors are also highly effective chemotherapeutic agents for treating colon cancer in laboratory animals, including carcinogen-induced and transgenic rodent models (Pereira et al., 1994; Boolbol et al., 1996; Chiu et al., 1997).
The mechanisms of COX-2 inhibitor-mediated cancer chemoprevention and chemotherapy are complex, and the inhibition of prostaglandin/thromboxane production (e.g., prostaglandin E2) contributes to these effects (Taketo, 1998a,b; Williams et al., 1999; Gately and Li, 2004). However, it is also clear that COX-2 inhibitors can be effective in cells with minimal COX-2 expression, and many cell growth inhibitory responses induced by these compounds are COX-2-independent (Smith et al., 2000; Liu et al., 2002; Yamazaki et al., 2002). For example, sulindac sulfone induces apoptosis in Caco-2 cells by decreasing polyamine levels through activation of peroxisome proliferator-activated receptor -dependent induction of spermidine/spermine N-acetyltransferase gene expression (Liu et al., 2002).
The cancer chemotherapeutic action of COX-2 inhibitors has also been linked to their antiangiogenic action and inhibition of VEGF expression (Williams et al., 1999; Gately and Li, 2004). COX-2 catalyzes the biosynthesis of PGE2, which induces HIF-1; this transcription factor forms a heterodimer that interacts with the hypoxia-responsive element in the VEGF promoter to activate VEGF gene expression (Masferrer et al., 2000; Fukuda et al., 2003). In this study, we initially used SW-480 cells as a model for investigating the mechanisms of VEGF inhibition by COX-2 inhibitors in COX-2-negative (or under-expressing) cells. Cel, NS, and NM clearly inhibited VEGF mRNA levels (Fig. 1A) and cellular VEGF (Fig. 1B), and this paralleled their inhibitory effects on luciferase activity in SW-480 cells transfected with VEGF constructs (pVEGF-1-3) (Fig. 2). Moreover, using the pVEGF constructs, we also observed comparable decreased transactivation in RKO, DLD, and HT-29 cells (Fig. 7C), and these cell lines exhibited highly variable expression of COX-2 protein (Fig. 7B). The results demonstrate that COX-2 inhibitors down-regulate VEGF, and this response is associated with the proximal -131 to -47 GC-rich region of the VEGF promoter, which binds Sp proteins (Abdelrahim et al., 2004; Stoner et al., 2004).
Reports from a recent study in pancreatic cancer cells indicate that treatment of Panc-1 cells with Cel decreased VEGF expression in vitro and in vivo, and transactivation studies in cells transfected with VEGF constructs indicated that the proximal GC-rich sites were associated with this response (Wei et al., 2004). It was concluded that decreased Sp1 and pSp1/Sp1 ratios were important determinants for this antiangiogenic response in pancreatic cancer. Results of this study in colon cancer cell lines also confirm that COX-2 inhibitors down-regulate Sp1 (Figs. 3-5) but do not consistently affect pSp1/Sp1 ratios (Fig. 3B). Moreover, it was also apparent that COX-2 inhibitors also induced down-regulation of Sp4 protein in colon cancer cells, whereas Sp3 and Sp2 protein levels ere unaffected (Figs. 3-5). These results suggest that the effects of COX-2 inhibitors on VEGF expression in colon cancer cells was related to their down-regulation of Sp1 and Sp4, and previous studies in pancreatic cancer cells have demonstrated that VEGF expression is regulated by Sp1, Sp4, and Sp3 (Sp2 was not investigated) (Abdelrahim et al., 2004). This was also observed in colon cancer cells, because RNA interference assays with iSp1, iSp2, iSp3, and iSp4 showed that transactivation in SW-480 and HT-29 cells transfected with pVEGF-2 was decreased by iSp1, iSp4, and iSp3, but not iSp2 (Figs. 6B and 7B). Moreover, overexpression of Sp1, Sp3, or Sp4 enhanced transactivation in SW-480 cells transfected with pVEGF-2 (± Cel) (Fig. 6C). These results show that VEGF expression in colon cancer cells is primarily dependent on Sp1, Sp4, and Sp3, as reported previously for pancreatic cancer cells (Abdelrahim et al., 2004), and the antiangiogenic activity of COX-2 inhibitors is linked to induced degradation of these transcription factors. It is interesting that treatment of SW-480 cells with NS or NM did not decrease Sp1 or Sp4 mRNA levels (Fig. 8), and gliotoxin, a specific inhibitor of the 20S proteasome complex (Kroll et al., 1999), inhibits NS-induced degradation of both Sp1 and Sp4 and transactivation in SW-480 cells transfected with pVEGF2 (Fig. 8). Moreover, whole-cell lysates from SW-480 cells treated with Cel or NS also exhibit several ubiquitinated bands immunoprecipitated with Sp1 or Sp4 (but not Sp3) antibodies (Fig. 9A). These results now provide a mechanistic basis for the antiangiogenic activity of COX-2 inhibitors in colon cancer cells, where these drugs selectively activate proteasome-dependent degradation of Sp1 and Sp4 proteins. Current studies are focused on potential applications of COX-2 inhibitors alone and in combination with other drugs for treatment of colon and possibly other cancers. In addition, we are also investigating other related compounds that induce degradation of Sp3 protein and thereby enhance their antiangiogenic activities.
References
Abdelrahim M, Smith R 3rd, Burghardt R, and Safe S (2004) Role of Sp proteins in regulation of vascular endothelial growth factor expression and proliferation of pancreatic cancer cells. Cancer Res 64: 6740-6749.
Bergers G and Benjamin LE (2003) Tumorigenesis and the angiogenic switch. Nat Rev Cancer 3: 401-410.
Boolbol SK, Dannenberg AJ, Chadburn A, Martucci C, Guo XJ, Ramonetti JT, Breu-Goris M, Newmark HL, Lipkin ML, DeCosse JJ, et al. (1996) Cyclooxygenase-2 overexpression and tumor formation are blocked by sulindac in a murine model of familial adenomatous polyposis. Cancer Res 56: 2556-2560.
Carmeliet P (2003) Angiogenesis in health and disease. Nat Med 9: 653-660.
Chiu CH, McEntee MF, and Whelan J (1997) Sulindac causes rapid regression of preexisting tumors in Min/+ mice independent of prostaglandin biosynthesis. Cancer Res 57: 4267-4273.
Ferrara N, Gerber HP, and LeCouter J (2003) The biology of VEGF and its receptors. Nat Med 9: 669-676.
Finkenzeller G, Sparacio A, Technau A, Marme D, and Siemeister G (1997) Sp1 recognition sites in the proximal promoter of the human vascular endothelial growth factor gene are essential for platelet-derived growth factor-induced gene expression. Oncogene 15: 669-676.
Folkman J (1995) Angiogenesis in cancer, vascular, rheumatoid and other disease. Nat Med 1: 27-31.
Fukuda R, Kelly B, and Semenza GL (2003) Vascular endothelial growth factor gene expression in colon cancer cells exposed to prostaglandin E2 is mediated by hypoxia-inducible factor 1. Cancer Res 63: 2330-2334.
Gately S and Li WW (2004) Multiple roles of COX-2 in tumor angiogenesis: a target for antiangiogenic therapy. Semin Oncol 31: 2-11.
Giardiello FM, Hamilton SR, Krush AJ, Piantadosi S, Hylind LM, Celano P, Booker SV, Robinson CR, and Offerhaus GJ (1993) Treatment of colonic and rectal adenomas with sulindac in familial adenomatous polyposis. N Engl J Med 328: 1313-1316.
Giovannucci E, Egan KM, Hunter DJ, Stampfer MJ, Colditz GA, Willett WC, and Speizer FE (1995) Aspirin and the risk of colorectal cancer in women. N Engl J Med 333: 609-614.
Kroll M, Renzana-Seisdedos F, Bachelerie F, Thomas D, Friguet B, and Conconi M (1999) The secondary fungal metabolite gliotoxin targets proteolytic activities of the proteasome. Chem Biol 6: 689-698.
Labayle D, Fischer D, Vielh P, Drouhin F, Pariente A, Bories C, Duhamel O, Trousset M, and Attali P (1991) Sulindac causes regression of rectal polyps in familial adenomatous polyposis. Gastroenterology 101: 635-639.
Li J, Zhang YP, and Kirsner RS (2003) Angiogenesis in wound repair: angiogenic growth factors and the extracellular matrix. Microsc Res Tech 60: 107-114.
Linderholm BK, Lindahl T, Holmberg L, Klaar S, Lennerstrand J, Henriksson R, and Bergh J (2001) The expression of vascular endothelial growth factor correlates with mutant p53 and poor prognosis in human breast cancer. Cancer Res 61: 2256-2260.
Liu XH, Kirschenbaum A, Lu M, Yao S, Dosoretz A, Holland JF, and Levine AC (2002) Prostaglandin E2 induces hypoxia-inducible factor-1a stabilization and nuclear localization in a human prostate cancer cell line. J Biol Chem 277: 50081-50086.
Maeda K, Kang SM, Onoda N, Ogawa M, Kato Y, Sawada T, and Chung KH (1999) Vascular endothelial growth factor expression in preoperative biopsy specimens correlates with disease recurrence in patients with early gastric carcinoma. Cancer 86: 566-571.
Martinez ME, McPherson RS, Levin B, and Annegers JF (1995) Aspirin and other nonsteroidal anti-inflammatory drugs and risk of colorectal adenomatous polyps among endoscoped individuals. Cancer Epidemiol Biomarkers Prev 4: 703-707.
Masferrer JL, Leahy KM, Koki AT, Zweifel BS, Settle SL, Woerner BM, Edwards DA, Flickinger AG, Moore RJ, and Seibert K (2000) Antiangiogenic and antitumor activities of cyclooxygenase-2 inhibitors. Cancer Res 60: 1306-1311.
Milanini J, Vinals F, Pouyssegur J, and Pages G (1998) p42/p44 MAP kinase module plays a key role in the transcriptional regulation of the vascular endothelial growth factor gene in fibroblasts. J Biol Chem 273: 18165-18172.
Mineta H, Miura K, Ogino T, Takebayashi S, Misawa K, Ueda Y, Suzuki I, Dictor M, Borg A, and Wennerberg J (2000) Prognostic value of vascular endothelial growth factor (VEGF) in head and neck squamous cell carcinomas. Br J Cancer 83: 775-781.
Nugent KP, Farmer KC, Spigelman AD, Williams CB, and Phillips RK (1993) Randomized controlled trial of the effect of sulindac on duodenal and rectal polyposis and cell proliferation in patients with familial adenomatous polyposis. Br J Surg 80: 1618-1619.
Peleg II, Lubin MF, Cotsonis GA, Clark WS, and Wilcox CM (1996) Long-term use of nonsteroidal antiinflammatory drugs and other chemopreventors and risk of subsequent colorectal neoplasia. Dig Dis Sci 41: 1319-1326.
Pereira MA, Barnes LH, Rassman VL, Kelloff GV, and Steele VE (1994) Use of azoxymethane-induced foci of aberrant crypts in rat colon to identify potential cancer chemopreventive agents. Carcinogenesis 15: 1049-1054.
Rossler J, Breit S, Havers W, and Schweigerer L (1999) Vascular endothelial growth factor expression in human neuroblastoma: up-regulation by hypoxia. Int J Cancer 81: 113-117.
Schafer G, Cramer T, Suske G, Kemmner W, Wiedenmann B, and Hocker M (2003) Oxidative stress regulates vascular endothelial growth factor-A gene transcription through Sp1- and Sp3-dependent activation of two proximal GC-rich promoter elements. J Biol Chem 278: 8190-8198.
Shi Q, Le X, Abbruzzese JL, Peng Z, Qian CN, Tang H, Xiong Q, Wang B, Li XC, and Xie K (2001) Constitutive Sp1 activity is essential for differential constitutive expression of vascular endothelial growth factor in human pancreatic adenocarcinoma. Cancer Res 61: 4143-4154.
Shifren JL, Tseng JF, Zaloudek CJ, Ryan IP, Meng YG, Ferrara N, Jaffe RB, and Taylor RN (1996) Ovarian steroid regulation of vascular endothelial growth factor in the human endometrium: implications for angiogenesis during the menstrual cycle and in the pathogenesis of endometriosis. J Clin Endocrinol Metab 81: 3112-3118.
Shweiki D, Itin A, Soffer D, and Keshet E (1992) Vascular endothelial growth factor induced by hypoxia may mediate hypoxia-initiated angiogenesis. Nature (Lond) 359: 843-845.
Smith ML, Hawcroft G, and Hull MA (2000) The effect of non-steroidal antiinflammatory drugs on human colorectal cancer cells: evidence of different mechanisms of action. Eur J Cancer 36: 664-674.
Stoner M, Wormke M, Saville B, Samudio I, Qin C, Abdelrahim M, and Safe S (2004) Estrogen regulation of vascular endothelial growth factor gene expression in ZR-75 breast cancer cells through interaction of estrogen receptor a and Sp proteins. Oncogene 23: 1052-1063.
Taketo MM (1998a) Cyclooxygenase-2 inhibitors in tumorigenesis (part I). J Natl Cancer Inst 90: 1529-1536.
Taketo MM (1998b) Cyclooxygenase-2 inhibitors in tumorigenesis (part II). J Natl Cancer Inst 90: 1609-1620.
Torry DS and Torry RJ (1997) Angiogenesis and the expression of vascular endothelial growth factor in endometrium and placenta. Am J Reprod Immunol 37: 21-29.
Wei D, Wang L, He Y, Xiong HQ, Abbruzzese JL, and Xie K (2004) Celecoxib inhibits vascular endothelial growth factor expression in and reduces angiogenesis and metastasis of human pancreatic cancer via suppression of Sp1 transcription factor activity. Cancer Res 64: 2030-2038.
Williams CS, Mann M, and DuBois RN (1999) The role of cyclooxygenases in inflammation, cancer and development. Oncogene 18: 7908-7916.
Yamazaki R, Kusunoki N, Matsuzaki T, Hashimoto S, and Kawai S (2002) Selective cyclooxygenase-2 inhibitors show a differential ability to inhibit proliferation and induce apoptosis of colon adenocarcinoma cells. FEBS Lett 531: 278-284.
Ziemer LS, Koch CJ, Maity A, Magarelli DP, Horan AM, and Evans SM (2001) Hypoxia and VEGF mRNA expression in human tumors. Neoplasia 3: 500-508.