Purpose: Our aim was to determine the effects of cyclin D1 inhibition on tumor-associated neovascularization and endothelial cell growth.
Experimental Design: We have generated adenovirus system for antisense to cyclin D1 (AS CyD1) and evaluated in vitro and in vivo effects. Small interfering RNA against cyclin D1 was also used to analyze cyclin D1 inhibition-associated vascular endothelial growth factor (VEGF) regulation.
Results: The xenografts treated with adenoviral AS CyD1 showed less vessel density and displayed smaller tumor size in colon cancer cell lines HCT116 and DLD1. In vitro studies indicated that AS CyD1 decreased VEGF protein expression in DLD1 but not in HCT116. Cyclin D1 small interfering RNA caused a decrease in VEGF expression at protein and RNA levels in DLD1. A modest decrease was noted in the VEGF promoter activity, with inactivation of the STAT3 transcription factor through dephosphorylation. On the hand, the cyclin D1 inhibition plus STAT3 inhibitor markedly decreased VEGF expression in HCT116, although VEGF did not change by the STAT3 inhibitor alone. In cultures of human umbilical vein endothelial cells (HUVEC), VEGF augmented cyclin D1 expression and cell growth. AS CyD1 significantly inhibited HUVEC growth even in the presence of VEGF. AS CyD1 also significantly suppressed in vitro tube formation in VEGF-treated HUVEC and in vivo macroaneurysm formation in VEGF-treated Matrigel plug.
Conclusions: Our results suggest that cyclin D1 may play a role in the maintenance of VEGF expression and that AS CyD1 could be potentially useful for targeting both cancer cells and their microenvironment of tumor vessels.
Cyclin D1, a putative G1 cyclin, preferentially associates with CDK4 and positively regulates the cell cycle transition from G1 to S phase (1). Cyclin D1 is considered an oncogene because forced expression in rodent fibroblasts induces tumorigenicity in nude mice and cyclin D1 transgenic mice develop tumors of breast, esophagus, stomach, and tongue (2–4). In human carcinomas, increased expression of cyclin D1 is one of the most frequent abnormalities because it is detected in ∼60% of breast cancers, 40% of colorectal cancers, 40% of squamous carcinomas of the head and neck, and 20% of prostate cancers (5–8). Furthermore, overexpression of cyclin D1 is associated with poor prognosis of patients with carcinomas of colorectum, esophagus, stomach, pancreas, and liver (9–13). Therefore, cyclin D1 is a crucial target for various types of human malignancies.
To suppress the malignant potential of carcinomas, the strategy of antisense to cyclin D1 (AS CyD1) was first assessed in human esophageal squamous cell carcinoma and colon cancer cells (14, 15). These studies clearly showed that AS CyD1 reversed the transformed phenotype of tumor cells, inhibited cell growth of tumor cells, and resulted in loss of tumorigenicity. Subsequently, AS CyD1 was found to enhance chemosensitivity of 5-fluorouracil, mitoxantrone, and cisplatinum in pancreatic cancer cells and head and neck cancer cells and to induce apoptosis and tumor shrinkage in esophageal squamous carcinoma (16–19). Similar antitumor effects were found in gastric cancer cells and hepatocellular carcinoma cells (20–22). Based on these favorable effects, it was considered that AS CyD1 could be a promising strategy against human pancreatic, colonic, and esophageal cancers (23).
Angiogenesis is essential for tumor growth and expansion because the blood vessels supply malignant cells with sufficient oxygen and nutrients (24, 25). Therefore, interruption of this process is one strategy to prevent invasion and metastasis. Although various biological effects by AS CyD1 have been reported as mentioned above, the antiangiogenic action of AS CyD1 is unknown. To explore this issue, we generated adenoviral AS CyD1 (Ad-AS CyD1) system and examined the effect of AS CyD1 on in vivo tumor-associated neovascularization, with special attention to vascular endothelial growth factor (VEGF), because VEGF is known as a critical growth factor that promotes endothelial cell proliferation and angiogenesis (26). We also examined its direct effect on in vitro and in vivo growth of vascular endothelial cells.
Materials and Methods
Cell lines and animals. HEK293 cells, human colon cancer cells (HCT116, DLD1, and LoVo), and gastric cancer cells (MKN45 and MKN28) were purchased from the American Type Culture Collection (Manassas, VA) or the Japanese Cancer Research Resources Bank (Osaka, Japan). MKN45 was grown in RPMI 1640, whereas the other cell lines were grown in DMEM, both supplemented with 10% fetal bovine serum, 100 units/mL penicillin, and 100 μg/mL streptomycin in 5% CO2 at 37°C. Human umbilical vein endothelial cells (HUVEC) were grown on MCDB131 culture medium (Chlorella, Inc., Tokyo, Japan) supplemented with 10% fetal bovine serum, antibiotics, and 10 ng/mL basic fibroblast growth factor. Female 4-week-old athymic nude mice were purchased from Nihon CREA, Inc. (Tokyo, Japan) and were housed under pathogen-free conditions. The experimental protocol was approved by the Ethics Review Committee for Animal Experimentation of Osaka University School of Medicine.
Reagents and antibodies. Human recombinant VEGF was obtained from IBL Co. (Gunma, Japan). Bromodeoxyuridine (BrdUrd) was purchased from Sigma-Aldrich (St. Louis, MO). STAT3 inhibitor peptide was purchased from Calbiochem (Darmstadt, Germany). This reagent is a cell-permeable analogue of the STAT3-SH2 domain-binding phosphopeptide that contains a COOH-terminal membrane-translocating sequence and acts as a highly selective, potent blocker of STAT3 activation (27). The following antibodies were used at appropriate concentrations as recommended by the manufacturers: (a) anti-human polyclonal antibodies for cyclin D1 (Santa Cruz Biotechnology, Santa Cruz, CA), cyclin A (Upstate Biotechnology, Waltham, MA), VEGF (Santa Cruz Biotechnology), actin (Sigma-Aldrich), phosphorylated STAT3 antibody (Tyr705; Cell Signaling Technology, Beverly, MA), and STAT3 antibody (Cell Signaling Technology); (b) anti-human monoclonal antibodies for cyclin E (BD Biosciences, BD PharMingen, San Diego, CA) and BrdUrd (DAKO, Glostrup, Denmark); and (c) anti-mouse rat monoclonal antibody for CD31 (Santa Cruz Biotechnology).
Western blot analysis. Western blot analysis was done as described previously (28). Briefly, the protein samples (25 μg) were separated by 10% or 12.5% PAGE followed by electroblotting onto a polyvinylidene difluoride membrane. The membrane was incubated with the primary antibodies at the appropriate concentrations (1:200 for cyclin D1 and VEGF and 1:1,000 for total STAT3, phosphorylated STAT3, and actin) for 1 hour. The protein bands were detected using the Amersham enhanced chemiluminescence detection system (Amersham Biosciences Corp., Piscataway, NJ).
Generation of Ad-AS CyD1. Ad-AS CyD1 was constructed using AdEasy Adenoviral Vector System (ref. 29; a generous gift from Dr. Bert Vogelstein, Johns Hopkins University School of Medicine, Baltimore, MD). A 1.1-kb human entire cyclin D1 cDNA that provides 90% homology of mouse cyclin D1 cDNA (2) was cut out from pcDNA3-cyclin D1 plasmid and subcloned into the HindIII site of pShuttle plasmid containing cytomegalovirus promoter in antisense orientation. After confirmation of antisense orientation by appropriate enzymes, recombination with the E1/E3delete adenoviral backbone vector (AdEasy-1) was done in Escherichia coli BJ5183 cells by electroporation. Viral particles were amplified in HEK293 cells and then purified by CsCl banding. Virus titer was measured by Adeno-X rapid titer kit (BD Clontech, Palo Alto, CA). Ad-(cytomegalovirus) Mock and Ad-(cytomegalovirus) green fluorescent protein were prepared as experimental controls.
Infectious efficiency and cytotoxicity of adenovirus in the cell lines. To determine the optimal concentrations that sufficiently realize adenoviral gene transfer, Ad-(cytomegalovirus) green fluorescent protein was infected at various concentrations for 1 hour with gentle shaking and then incubated with the complete medium. The virus titer [multiplicity of infection (MOI)] endowing >90% green fluorescent protein–positive cells at 24 hours after infection was as follows: HUVEC, 20; MKN45, 40; MKN28, 20; HCT116, 10; DLD1, 40; and LoVo, 60. At the respective virus titer, the cell viability of HUVEC, MKN45, MKN28, HCT116, and DLD1 as indicated by the trypan blue exclusion test was >90%, but the cell viability of LoVo was <60%.
Growth assays. Cells were uniformly seeded (1 × 105 per well) into six-well dishes in triplicate. Twenty-four hours later, the culture medium was removed and replaced with 0.5 mL fresh medium containing adenovirus at the optimal concentration for 1 hour. The cells were then grown in the complete medium and counted using a hemocytometer.
Immunohistochemistry. Immunostaining was done as described previously (28). Briefly, after deparaffinization, heat antigen retrieval was done in 10 mmol/L citrate buffer (pH 6.0) at 95°C for 40 minutes. The slides were then processed for immunohistochemistry using the Vectastain Elite avidin-biotin complex kit (Vector Laboratories, Burlingame, CA). Primary antibodies were applied to sections at a dilution of 1:750 for CD31 and incubated overnight at 4°C. For the negative control, nonimmunized immunoglobulin G (Vector Laboratories) was used as a substitute for the primary antibody.
Semiquantitative reverse transcription-PCR. RNA extraction was carried out with TRIzol reagent in a single-step method and cDNA was generated with avian myeloblastosis virus reverse transcriptase (Promega, Madison, WI). Semiquantitative analyses of the expression of VEGF RNA were done using the duplex reverse transcription-PCR technique as described previously (30). β-Actin was used as the internal standard. PCRs were done in a total volume of 25 μL, which consisted of 2 μL cDNA template, 1× Perkin-Elmer PCR buffer (Perkin-Elmer, Foster City, CA), 1.5 mmol/L MgCl2, 0.8 mmol/L deoxynucleotide triphosphates, 1.25 pmol β-actin and 5 pmol VEGF primer, and 1 unit Taq DNA polymerase (AmpliTaq Gold; Roche Molecular Systems, Branchburg, NJ). PCR amplification was done with a GenAmp PCR System 9600 (Perkin-Elmer). The primer sequences were as follows (30, 31): β-actin sense 5′-GAAAATCTGGCACCACACCTT-3′ and antisense 5′-GTTGAAGGTAGTTTCGTGGAT-3′ and VEGF sense 5′-AAGCCATCCTGTGTGCCCCTGATG-3′ and antisense 5′-GCGAATTCCTCCTGCCCGGCTCAC-3′.
BrdUrd labeling index. The cells were incubated with 20 μmol/L BrdUrd (Sigma-Aldrich) at 37°C for 25 minutes and fixed in 70% cold ethanol for 30 minutes. After quenching the endogenous peroxidase activity, the chambers were incubated in 4 N HCl at 37°C for 30 minutes and then neutralized by buffered boric acid (pH 9.0) for 5 minutes. After blocking with 10% rabbit serum, anti-BrdUrd antibody was applied to the chambers at the dilution of 1:20 at room temperature for 2 hours followed by the avidin-biotin complex method.
In vitro angiogenesis assay.In vitro formation of tubular structure in HUVEC was examined using In vitro Angiogenesis Assay kit (Chemicon International, Inc., Temecula, CA). HUVEC infected with Ad-AS CyD1 or Ad-Mock (20 MOI) were seeded on Matrigel-coated well and maintained on complete medium. After attachment of the cells on Matrigel, the medium was changed with fresh medium, either with or without the recombinant VEGF protein (25 ng/mL), and incubated for 10 hours. Cells were then observed under the inverted microscope and the number of capillary connections was counted as reported previously (32).
In vivo Matrigel angiogenesis assay.In vivo angiogenesis was assayed as growth of blood vessels of mouse s.c. tissue in the exogenous Matrigel plug. Matrigel was prepared with 100 ng/mL basic fibroblast growth factor and 64 units/mL heparin with or without 40 ng/mL VEGF. Ad-AS CyD1 or Ad-Mock was included in the Matrigel plug at a concentration of 1 × 109 plaque-forming units/mL. The Matrigel was injected (0.5 mL) into the s.c. tissue of female athymic mice (n = 4 for each group). On day 7, mice were sacrificed and Matrigel plugs were removed and fixed in 10% buffered formalin and embedded in paraffin. Sections were stained with H&E or CD31 antibody and examined under a light microscope. In vivo angiogenesis was scored planimetrically through observation of 10 fields at high-power magnification, and the percentage of vessel area to total Matrigel area was calculated as reported previously (32, 33).
Treatment of established tumor xenografts by intratumoral injection of Ad-AS CyD1. S.c. xenografts of colorectal cancer (HCT116 and DLD1) were established in nude mice (n = 4 for each group) by injection 5 × 106 cells. After 1 week (day 7), when the tumor size reached ∼100 to 150 mm3, Ad-Mock, Ad-AS CyD1 (0.5 × 109 plaque-forming units/injection), and saline were injected into tumors and more than two injections per tumor were done on days 9 and 11. On day 30, the mice were sacrificed.
Transfection. SiTrio cyclin D1 and negative control small interfering RNA (siRNA) were purchased from B-Bridge International, Inc. (Sunnyvale, CA). Each siRNA consisted of three different target sequences and the sequences are as follows: negative control 5′-ATCCGCGCGATAGTACGTA-3′, 5′-TTACGCGTAGCGTAATACG-3′, and 5′-TATTCGCGCGTATAGCGGT-3′ and siRNA CCND1/human 5′-CGCUGGAGCCCGUGAAAAATT-3′, 5′-CCAGAGUGAUCAAGUGUGATT-3′, and 5′-UCCAAUAGGUGUAGGAAAUTT-3′.
Cells were transfected with 100 nmol/L siRNA using LipofectAMINE 2000 (Invitrogen, Carlsbad, CA) in Opti-MEM I Reduced Serum Medium (Invitrogen). After 6 hours, medium was replaced by standard medium. In use of STAT3 inhibitor peptide, it was added at 100 μmol/L 12 hours after transfection.
pcDNA3 (1 μg; Invitrogen) or STAT3 expression vector [pCAG-wtSTAT3-IP (wild-type) and pCAG-dnSTAT3-IP (a dominant-negative mutant of STAT3)] was introduced into cells with LipofectAMINE 2000. The wild-type STAT3 expression vector and the dominant-negative STAT3 plasmid were provided from Prof. T. Yokota (Department of Stem Cell Biology, Kanazawa University, Kanazawa, Japan; ref. 34). These expression vectors are driven by cytomegalovirus enhancer-chicken β-actin hybrid promoter (developed by Prof. Jun-ichi Miyazaki, Department Nutrition and Physiological Chemistry, Osaka University, Osaka, Japan; ref. 35).
Reporter gene assay. The VEGF gene promoter region (provided by Dr. Abraham, SIOS, Inc., CA) was inserted upstream of the luciferase reporter gene in pGVB (Toyo, Inc., Tokyo, Japan) as described previously (36). Cells were transfected with siRNA and 2 μg reporter plasmid pGVB-VEGF and 0.025 μg pRL-SV40 (36) in Opti-MEM I Reduced Serum Medium. After 6 hours, medium were replaced by standard medium. Cells were harvested at 24 hours after transfection for the dual-luciferase assay (Promega) according to the manufacturer's instruction. The firefly luciferase activity of pGVB-VEGF was normalized by the Renilla luciferase activity of pRL-SV40. The level of luciferase in control cells was assigned a value of 1.0, and the relative activities were calculated.
Statistical analysis. Data are expressed as mean ± SD. Differences among three groups were examined by the one-factor ANOVA followed by post-test of Bonferroni/Dunn. Statistical analysis was done using the StatView 5.0 (SAS Institute, Inc., Cary, NC).
Effects of AS CyD1 on cyclin D1 expression and growth of tumor cells. A relatively high expression of the cyclin D1 protein was noted in HCT116, DLD1, and MKN45 cells and VEGF protein was expressed in HCT116, DLD1, LoVo, and MKN45 (Fig. 1A). Compared with Mock control, AS CyD1 decreased cyclin D1 expression in MKN45 cells in a time-dependent manner (Fig. 1B). In HCT116 and DLD1 cells, AS CyD1 showed decreased expression of the cyclin D1 protein after 48 hours (data not shown). In vitro cell proliferation assays showed that AS CyD1 significantly inhibited tumor cell growth compared with Mock control or nontreatment cultures of MKN45, HCT116, and DLD1 (Fig. 1C). AS CyD1 decreased expression of the VEGF protein in DLD1 but not in HCT116 (Fig. 1D).
Effect of AS CyD1 on tumor vascularization. Tumor size on day 30 was significantly smaller in AS CyD1 group compared with Mock control group or saline treatment group in both cell types (P < 0.01; Fig. 2A). Immunohistochemistry of tumor tissues revealed that there was a significant reduction in vessel density in xenografts of both AS CyD1 group and Mock control group (P < 0.01; Fig. 2B). Especially, significant decrease in vessel density was noted in Ad-AS CyD1–injected DLD1 xenografts when compared with saline treatment (P < 0.01; Fig. 2B).
Expression of cyclins in HUVEC. Serial changes in the expression of cyclin D1 were determined in cultures of HUVEC refed with 10% fetal bovine serum after 24 hours serum starvation. Increase in cyclin D1 expression was detected as early as 1 hour after serum addition and a further increase was noted at 12 hours and subsequent time points (Fig. 3A). Cyclin A and cyclin E levels also increased with the highest expression noted at 24 hours (Fig. 3A).
Treatment of HUVEC with recombinant VEGF and Ad-AS CyD1. The addition of recombinant VEGF protein to HUVEC culture further increased the level of cyclin D1 (Fig. 3B). AS CyD1 markedly decreased cyclin D1 expression in VEGF-treated HUVEC compared with Mock control cultures as early as 1 hour (Fig. 3C). When cell growth was assessed by BrdUrd incorporation, VEGF treatment in Mock control cultures caused a significant increase in BrdUrd labeling index than VEGF without Mock control cultures (P < 0.01), which was drawn back to the basal level by AS CyD1 (Fig. 3D). The difference between VEGF treatment and VEGF plus AS CyD1 was significant (P < 0.01; Fig. 3D).
Effects of AS CyD1 on in vitro and in vivo angiogenesis. We then examined effects of AS CyD1 on angiogenesis. In vitro angiogenesis assay showed that HUVEC formed vessel-like structures (tubes) when plated on Matrigel-coated wells (Fig. 4A). The VEGF treatment enhanced HUVEC growth, resulting in thick tubes and increased steady network formation. In contrast, AS CyD1 caused thinner or only faint tube-like structures even in VEGF-treated cultures. There was a significant difference in the number of capillary connections, defined as cross-points consisting of three tubes (32) in each combination (P < 0.01; Fig. 4A).
In vivo angiogenesis assay showed modest vessel formation in Matrigel plug of the Mock control group (Fig. 4B, I). Mean vessel area relative to the Matrigel plug area was calculated (Mock group, 13.2 ± 0.73%) with reference to CD31-stained vascular endothelial cells (Fig. 4B, I, a) and RBC (Fig. 4B, I, b) as guidance for neovascularization in Matrigel plug. With treatment of VEGF, Matrigel was expansively enlarged because of the endothelioma-like structure that grew in the center of Matrigel and contained a giant aneurysm (Fig. 4B, II, mean vessel area, 84.2 ± 13.1%). Although the majority of the aneurysm-like structure dropped off in making paraffin blocks, CD31 staining showed that this structure was lined by endothelium (Fig. 4B, II, a, arrows). There were also some RBC remaining in Matrigel (Fig. 4B, II, b). In contrast, VEGF plus AS CyD1 resulted in production of numerous microaneurysm-like structures in Matrigel (Fig. 4B, III, mean vessel area, 61.0 ± 9.83%). Each small aneurysm consisted of RBC (Fig. 4B, III, a) and few microhemorrhages were noted in this group (Fig. 4B, III, b). Although some small canalized vessels located at the edge of Matrigel were stained with the CD31 antibody, as indicated by arrows (Fig. 4B, III, c), the endothelial cells surrounding small aneurysm did not display a CD31+ finding, suggesting that the vessels in this group might be immature or dying after vascular formation. The mean vessel area of VEGF plus AS CyD1 group was significantly lower than that of VEGF plus Mock group (P < 0.01; Fig. 4C).
Role of cyclin D1 in VEGF regulation. Finally, we did mechanistic studies to elucidate differential effects of AS CyD1 on VEGF expression in DLD1 and HCT116 cells using siRNA. siRNA against cyclin D1 decreased expression of the cyclin D1 protein as early as 24 hours, and decreased level was maintained until 72 hours in both cell lines (Fig. 5A). There was no change in VEGF expression in HCT116 at protein, RNA, and promoter levels (Fig. 5A-C). In DLD1, on the other hand, expression of the VEGF protein apparently decreased with the cyclin D1 inhibition, and VEGF RNA decreased to some extent (Fig. 5A and B). The VEGF promoter activity of cultures treated with cyclin D1 siRNA but not control siRNA was significantly decreased compared with nontreatment cultures (P < 0.01). We then examined expression of the STAT3 transcription factor that is known as an enhancer of the VEGF gene promoter (37, 38). Introduction of wild-type STAT3 cDNA but not dominant-negative construct clearly enhanced VEGF RNA expression in both cell lines (Fig. 6A). The whole STAT3 levels did not change by cyclin D1 inhibition in both cell lines, whereas expression of the phosphorylated STAT3, an activated form of STAT3, changed when the relative ratio to the actin band was calculated by densitometry analyses (Fig. 6B). Thus, in HCT116 cell line, treatment of siRNA against cyclin D1 resulted in a slight decrease in phosphorylated STAT3 as early as 24 hours later, which was largely unchanged at 48 hours, and then slightly enhanced at the late time point of 72 hours, whereas an apparent decrease was consistently seen in DLD1 cell line from 24 to 72 hours (Fig. 6B). When the STAT3 inhibitor was given in the HCT116 control cultures without cyclin D1 inhibition, VEGF expression did not change, whereas the VEGF level decreased at the RNA level at 24 hours (Fig. 6C) and at the protein level at 48 hours (Fig. 6D) in cultures treated with cyclin D1 siRNA and the STAT3 inhibitor.
In this study, we selected DLD1 and HCT116 among tumor cells for investigation of tumor-associated neovascularization because both these cells displayed a high infection efficiency without toxicity by adenovirus and expressed both cyclin D1 and putative angiogenetic factor VEGF. Because AS CyD1 prevents tumorigenicity in nude mice in various tumor cell systems (14, 15), assessment of tumor-associated neovascularization is impossible with tumorigenicity assay. Therefore, in the present study, we employed a “therapeutic model” after establishment of tumor xenograft, which permits detailed survey of vessel formation in the late stage of tumor xenograft. The therapeutic model is also useful with regard to assessment of the value of Ad-AS CyD1 as a clinical tool. It was nice that only three therapeutic administrations of Ad-AS CyD1 were sufficient to inhibit the growth of xenografts of both HCT116 and DLD1. These results represent the direct antitumor effects of AS CyD1 as suggested in tumor growth-inhibitory assay in monolayer cultures. Immunohistochemical survey of the vessel counts using CD31 antibody in xenografts showed that AS CyD1 significantly inhibited tumor vessel formation in both cell lines. It is notable that there was a marked decrease in vessel formation in DLD1-xenografts, which displayed a significant reduction in VEGF expression in in vitro assay.
Our in vitro studies showed that AS CyD1 decreased expression of VEGF protein in DLD1 but not in HCT116. Reverse transcription-PCR assay also showed that AS CyD1 decreased VEGF RNA in DLD1 but not in HCT116 (data not shown). To verify these findings and to explore the underlying mechanism, we used siRNA against cyclin D1. This system allowed us to monitor both early and late events in a time-dependent manner, and it was convenient in combination use with the STAT3 inhibitor because the inhibitor became toxic when used with adenovirus particles. Using siRNA against cyclin D1, we confirmed that VEGF expression was down-regulated at both protein and RNA levels in DLD1 but not in HCT116. The reporter assays also showed reduction of VEGF gene promoter in DLD1. To further investigate the regulation mechanism of VEGF at promoter level, we focused on STAT3 because this transcription factor is considered an enhancer of the VEGF gene promoter in pancreatic carcinomas and other human cancers (37, 38). Transfection assays with STAT3 cDNA indicated that STAT3 enhanced VEGF promoter activity and increased the VEGF RNA level in HCT116 and DLD1 cell lines, suggesting that this pathway is active in both cell lines as an enhancer of the VEGF gene promoter. Of the two cell lines, we found that cyclin D1 inhibition led to inactivation of STAT3 through dephosphorylation solely in DLD1. AS CyD1 adenovirus system also showed similar inactivation of STAT3 in DLD1 (data not shown). Therefore, it seems that cyclin D1 inhibition regulates VEGF expression at the promoter level in DLD1. However, because reduction of the VEGF promoter activity was only modest in DLD1 (∼20%) in contrast to strong inhibition of VEGF expression at the protein level, we cannot rule out the possibility that other mechanisms might be also involved. In this context, the experiments in HCT116 using STAT3 inhibitor provided an important implication. Thus, we used siRNA against cyclin D1 with the STAT3 inhibitor and found that VEGF level decreased in HCT116 as expected. However, of interest was that the VEGF levels did not change in HCT116 control cultures without cyclin D1 inhibition even when the STAT3 inhibitor was applied, suggesting that although STAT3 is an upstream enhancer in the VEGF promoter the inhibition of STAT3 alone may be insufficient to block VEGF expression. Conversely, cyclin D1 inhibition may in part be involved in maintenance of the VEGF expression in HCT116. We are still uncertain why cyclin D1 inhibition led to the inactivation of STAT3 in DLD1 but not in HCT116. However, as this close correlation between cyclin D1 and STAT3 was also reported in head and neck squamous cell carcinoma (39), its occurrence is possibly dependent on the cell context.
One may suppose why AS CyD1 inhibited tumor angiogenesis in HCT116 that maintained VEGF expression. To assess this issue, we examined the direct effects of AS CyD1 on vessel formation because several studies suggest that cyclin D1 may play an essential role in growth of endothelial cells. It is reported that antiangiogenic endostatin-induced G1 arrest occurred through inhibition of cyclin D1 in endothelial cells (40) and that up-regulation of cyclin D1 by cytochrome P450 was associated with proliferation of endothelial cells (41). We found that VEGF up-regulated cyclin D1 as well as cyclin A and cyclin E in HUVEC, being consistent with a recent report that VEGF treatment of HUVEC caused progression of the cell cycle with increases in cyclin D1, cyclin A, and p42/p44 mitogen-activated protein kinase (42). As a result, AS CyD1 was enough to decrease the VEGF-enhanced cyclin D1 expression and significantly inhibited cell proliferation of HUVEC in monolayer culture. It took only a short time to suppress cyclin D1 at protein level in HUVEC with treatment of AS CyD1 compared with that in tumor cells (within 1 hour versus 2-3 days), suggesting that effects of AS CyD1 may be more potent in endothelial cells. To further explore the role of cyclin D1 in neovascularization, we did in vitro and in vivo angiogenesis assays. Even in cells treated with VEGF, AS CyD1 rendered vessel tubes immature and resulted in a decrease in the number of capillary connections in in vitro angiogenesis assay. In nude mice, VEGF-induced Matrigel plug developed macroaneurysms, which represent migration and extended growth of endothelial cells, whereas AS CyD1 produced only microaneurysms even in those treated with VEGF and inhibited macroaneurysm formation. Of interest was that AS CyD1–treated Matrigel showed several microhemorrhages, which could represent the destruction of immature vessels. These findings suggest that cyclin D1 plays a central role in neovascularization not only in vitro but also in vivo. Thus, it is likely that AS CyD1 may contribute to inhibition of tumor angiogenesis via direct inhibitory effects of vascular endothelial cells.
The present results that AS CyD1 could inhibit VEGF-mediated angiogenesis have important clinical implication in therapy of human malignancies. It is known that VEGF is involved in the development of liver metastasis from colorectal cancer and is associated with poor prognosis of patients (43). It is notable that a recent clinical study reported the survival benefits of anti-VEGF reagent when used in combination with conventional chemotherapy in metastatic colorectal cancer (44). As mentioned in Introduction, AS CyD1 have thus far been reported to have various direct effects against tumor cells (e.g., loss of tumorigenicity, reduced growth, induced apoptosis, and enhanced chemosensitivity). The present studies revealed that AS CyD1 inhibits not only tumor cells but also tumor-associated vessel formation. Thus, our data suggest that AS CyD1 is a promising tool against tumor organ that consists of tumor cells and its microenvironment.
Grant support: Grant-in Aid for Cancer Research from the Ministry of Education, Science, Sports, and Culture Technology, Japan (H. Yamamoto) and grant for the Third Term Comprehensive Strategy for Cancer Control from the Ministry of Health Labor and Welfare.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
We thank Dr. Bert Vogelstein for providing the AdEasy system, Prof. T. Yokota for STAT3 expression vectors, Prof. Jun-ichi Miyazaki for cytomegalovirus enhancer-chicken β-actin hybrid promoter, and Dr. H. Nakamura (Division of Hepatobiliary and Pancreatic Medicine, Department of Internal Medicine, Hyogo College of Medicine, Hyogo, Japan) for valuable advice.