Skip Navigation

JNCI Journal of the National Cancer Institute 2003 95(6):437-448; doi:10.1093/jnci/95.6.437
© 2003 by Oxford University Press
This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrow Request Permissions
Google Scholar
Right arrow Articles by von Marschall, Z.
Right arrow Articles by Rosewicz, S.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by von Marschall, Z.
Right arrow Articles by Rosewicz, S.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?

Journal of the National Cancer Institute, Vol. 95, No. 6, 437-448, March 19, 2003
© 2003 Oxford University Press

ARTICLE

Effects of Interferon Alpha on Vascular Endothelial Growth Factor Gene Transcription and Tumor Angiogenesis

Zofia von Marschall, Arne Scholz, Thorsten Cramer, Georgia Schäfer, Michael Schirner, Kjell Öberg, Bertram Wiedenmann, Michael Höcker, Stefan Rosewicz

Affiliations of authors: Z. von Marschall, A. Scholz, T. Cramer, G. Schäfer, B. Wiedenmann, M. Höcker, S. Rosewicz, Department of Hepatology, Gastroenterology, Endocrinology and Metabolism, Charité, Campus Virchow-Klinikum, Humboldt-University, Berlin, Germany; M. Schirner, Corporate Research, Schering AG, Berlin; K. Öberg, Department of Internal Medicine, University Hospital, Uppsala, Sweden.

Correspondence to: Stefan Rosewicz, M.D., Department of Hepatology, Gastroenterology, Endocrinology and Metabolism, Charité, Campus Virchow-Klinikum, Humboldt-University, Augustenburger Platz 1, 13353 Berlin, Germany (e-mail: stefan.rosewicz{at}charite.de).


   ABSTRACT
Top
 Notes
 Abstract
Introduction
 Patients and Methods
 Results
 Discussion
 References
 
Background: Interferon alpha (IFN-{alpha}) has antiangiogenic activity, although the underlying mechanism of action is unclear. Because human neuroendocrine (NE) tumors are highly vascularized and sensitive to IFN-{alpha}, we investigated whether the therapeutic effects of IFN-{alpha} result from an inhibition of angiogenesis mediated by a decrease in vascular endothelial growth factor (VEGF) gene expression. Methods: VEGF gene and protein expression was analyzed in NE tumors by immunohistochemistry and in NE tumor cell lines by quantitative competitive reverse transcription–polymerase chain reaction (RT–PCR) and enzyme-linked immunosorbent assay (ELISA). VEGF promoter–reporter gene constructs containing various deletions or mutations and gel shift assays were used to identify minimal promoter requirements and potential transcription factors. A xenograft nude mouse model (five mice per group) was used to determine the effect of IFN-{alpha} on tumor growth (NE Bon cells and pancreatic Capan-1 cells) and microvessel density. Liver metastases from eight patients with NE tumors were analyzed for microvessel density, VEGF mRNA content, and VEGF plasma levels before and after initiation of IFN-{alpha} therapy. Results: NE tumors and cell lines expressed VEGF mRNA and secreted VEGF protein. In vitro, IFN-{alpha} decreased transcription of VEGF gene expression through an Sp1- and/or Sp3-dependent inhibition of VEGF promoter activity. Compared with vehicle treatment in mice, IFN-{alpha} inhibited tumor growth by 36% and reduced microvessel density from 56 (95% confidence interval [CI] = 49 to 69) to 37 per x400 Field (95% CI = 32 to 41, P = .015). Patients with NE tumors had lower VEGF plasma levels and reduced VEGF mRNA levels and microvessel density in liver metastasis biopsy material after IFN-{alpha} treatment. Conclusion: IFN-{alpha} confers its antitumor activity, at least in part, by its antiangiogenic activity, which results from Sp1- and/or Sp3-mediated inhibition of VEGF gene transcription.


   INTRODUCTION
Top
 Notes
 Abstract
 Introduction
Patients and Methods
 Results
 Discussion
 References
 
Angiogenesis, the formation of new blood vessels, is essential for progression of solid tumors (1–3). Recent evidence suggests that vascular endothelial growth factor (VEGF) is a major regulator of tumor-associated angiogenesis, thereby promoting tumor growth, invasion, and metastasis (4). Accordingly, interfering with VEGF function inhibits angiogenesis and tumor growth in vivo (5).

Interferon alpha (IFN-{alpha}) is a cytokine with pleiotropic cellular functions, including antiviral, antiproliferative, immunomodulatory, and antiangiogenic activities (6,7). Clinical studies demonstrate that IFN-{alpha} treatment can induce impressive responses in angioproliferative diseases, such as Kaposi’s sarcoma and hemangiomas (6). Preclinical data suggest that these antiangiogenic activities may be associated with the regulation of endothelial cell motility (8) and survival (9) and with inhibition of molecules involved in the angiogenic response, such as basic fibroblast growth factor (bFGF) (10), interleukin 8 (11), and matrix metalloproteinase 9 (MMP-9) (12).

In addition to its potential therapeutic effects on angioproliferative diseases, IFN-{alpha} therapy is being clinically evaluated for the treatment of some malignancies, including malignant melanoma, renal cell carcinoma, myeloproliferative disorders [summarized in (13)], and neuroendocrine (NE) tumors (14). NE tumors of the gastroenteropancreatic tract provide a particularly valuable model to study the mechanism of IFN-{alpha} action because most NE tumors are extremely well vascularized (15), and the clinical benefit of IFN-{alpha} treatment in these patients has been well documented (14). Because the VEGF/VEGF receptor (VEGF-R) system is important for the induction of angiogenesis in numerous human solid tumors, we tested the hypothesis that the therapeutic effect of IFN-{alpha} might result from an inhibition of angiogenesis through a decrease in VEGF gene expression. Therefore, we analyzed the effects of IFN-{alpha} on VEGF gene expression in vitro by using human NE tumor cell lines. Biologic and clinical relevance of our findings were further validated in a xenograft mouse model and NE tumor patients treated with IFN-{alpha}.


   PATIENTS AND METHODS
Top
 Notes
 Abstract
 Introduction
 Patients and Methods
Results
 Discussion
 References
 
Patient Data

Eight NE tumor patients with the following characteristics were included in this study: all patients had histologically proven midgut carcinoids with carcinoid syndrome and liver metastasis as documented by computed tomography (CT). The median age was 69.5 years (range = 55–77 years). Six patients were male and two were female. All patients received 5 x 106 IU of IFN-{alpha} subcutaneously three times per week. Distinct liver metastases were detected by ultrasound or CT-guided techniques, and fine-needle biopsy specimens were harvested before and after 6 months of IFN-{alpha} treatment. Tumor staging was performed by CT every 3 months. This study was approved by the ethics committee of the University Hospital of Uppsala, and all patients provided written informed consent before inclusion in the study.

Levels of VEGF and VEGF-R expression and angiogenesis were determined in surgical specimens from a different group of 10 patients with NE tumors. These patients underwent surgery with curative intent, and their diagnoses were proven histologically by a pathologist. Seven patients had NE tumors of the foregut, and three patients had NE tumors of the midgut. This study was approved by the ethics committee of the Charité (Humboldt-University, Berlin, Germany), and all patients gave written informed consent.

Cell Culture

The human pancreatic NE tumor cell line Bon, colon carcinoma cell line HT29, and hepatoma cell line HepG2 (all obtained from the American Type Culture Collection [ATCC], Manassas, VA) were grown in Dulbecco’s modified Eagle’s medium with 10% fetal calf serum. The human NE tumor cell line QGP1 and the human pancreatic carcinoma cell line Capan-1 (obtained from ATCC) were cultured in RPMI-1640 with 10% or 15% fetal calf serum, respectively. The human NE tumor cell line LCC18 was grown as described (16). All cell cultures were grown as subconfluent monolayers in a humidified atmosphere containing 5% CO2 at 37 °C.

Immunohistochemistry and Microvessel Density

Fine-needle biopsy and surgical tissue specimens were fixed in acetone as frozen tissues, and cryosections (5- to 8-µm thick) were stained for immunohistochemistry by the alkaline phosphatase/anti-alkaline phosphastase (APAAP) method with new fuchsin as a developer, as described (17). The following primary antibodies were used: mouse anti-human kinase insert domain-containing region (KDR) (1 : 200; Sigma, Deisenhofen, Germany), mouse anti-human fms-like tyrosine kinase (Flt-1) (1 : 100; Sigma), rabbit anti-human synaptophysin (1 : 200; DAKO Diagnostika GmbH, Hamburg, Germany), and rat anti-mouse pan endothelial cell antigen (MECA-32) (1 : 50; PharMingen, Heidelberg, Germany). If more than 90% of a given cell population stained, then the antigen expression was graded as "strong."

To quantify angiogenesis, microvessel density was determined by staining tissue sections immunohistochemically for the pan endothelial cell antigen as described (18). The entire sections were scanned under low magnification, and vascularization was subjectively graded. Three highly vascularized areas per tumor were then evaluated at high magnification (x400). Any red-staining endothelial cell cluster distinct from adjacent microvessels, tumor cells, or other stromal cells was considered a single countable microvessel. The total number of microvessels was determined for each area, and the average number was recorded for each tumor. Two independent approaches were used to confirm the specificity of the observed immunohistochemical signals: 1) serial dilutions of the primary antibody until the signal disappeared, and 2) substitution of the primary antibody with preimmune rabbit immunoglobulin G. Slides were analyzed by two independent investigators.

In situ Hybridization

Sense and antisense 35S-labeled cRNA probes were prepared from human cDNAs coding for VEGF121, Flt-1, and KDR as described (17). Prehybridization and hybridization conditions, washing procedures, RNase digestion of mismatched sequences, and autoradiography were done on cryosections as described (17).

Quantification of VEGF in Cell Culture Samples and Human Plasma

Bon cells (5 x 104 per well) were cultured in 24-well plates under serum-free conditions in UltraCulture medium (BioWhittaker, Walkersville, MD). The cells were treated with IFN-{alpha}-2a (Hoffmann-La Roche, Grenzach-Wyhlen, Germany) at 1000 IU/mL or vehicle (phosphate-buffered saline [PBS]), and the conditioned medium was collected after 24, 48, 72, or 96 hours. Extracts from IFN-{alpha}- or vehicle-treated cells were prepared as previously described (17). VEGF was analyzed using the Quantikine human VEGF enzyme-linked immunosorbent assay (ELISA) kit (R&D Systems GmbH, Wiesbaden-Nordenstadt, Germany), following the instructions supplied by the manufacturer. VEGF concentration was then normalized to protein content.

Blood samples were collected into tubes containing EDTA from NE tumor patients before and 6 months after initiation of the IFN-{alpha} treatment. Plasma was separated by centrifugation (15 minutes at 1000g at 4 °C) and stored at –80 °C. All samples were measured on the same ELISA plate.

Quantification of VEGF mRNA by Quantitative Competitive Reverse Transcription–Polymerase Chain Reaction

VEGF mRNA concentrations in NE tumor cell lines were analyzed after treatment for 96 hours (unless otherwise indicated) with IFN-{alpha} (1000 IU/mL) or vehicle (PBS). Total RNA from the tissue samples was extracted using the RNeasy Mini Kit (Qiagen, Hilden, Germany) in 30 µL of RNase-free water, and 1 µg of RNA was subjected to reverse transcription. The quantitative competitive reverse transcription–polymerase chain reaction (RT–PCR) for the VEGF isoform VEGF165 (corresponding to the 165-amino-acid splice variant) and {beta}-actin was carried out as described (17). Expression of VEGF mRNA in NE tumor cell lines and in NE tumor tissue was normalized to the expression of {beta}-actin mRNA or to the amount of total cDNA, respectively. cDNA was quantified with Hoechst 33258 DNA binding dye (Sigma), following the manufacturer’s instructions.

Reporter Gene Plasmids

The following plasmids were provided by G. Finkenzeller (GeneScan AG, Freiburg, Germany) and have been previously described (19): human VEGF gene promoter 5'-deletion luciferase constructs -2018-VEGF-Luc, -1286-VEGF-Luc, -789-VEGF-Luc, -414-VEGF-Luc, -267-VEGF-Luc, -85-VEGF-Luc, and -52-VEGF-Luc (previously named pLuc 2068 to pLuc 102); plasmids pTATA and pTATA-85/-50 (sequence -85/-50 in front of a heterologous thymidine kinase promoter luciferase construct); and GCmut and AP-2/Egr-1mut (previously named pLuc 135/Sp1 mut and pLuc 135/Egr-1 mut, respectively).

A reporter gene construct containing the interferon-stimulated response element (ISRE) fused to a luciferase reporter gene (ISRE-Luc) (20) was provided by S. Bhattacharya (Department of Cardiovascular Medicine, University of Oxford, Oxford, U.K.) and was used as a positive control for IFN-{alpha}–induced transactivation in transfection experiments. Vector pT81-based constructs containing various GC box deletions and pGL3-based constructs containing 5' deletions within the –115 to –50 region of the VEGF promoter will be described in detail elsewhere (G. Schäfer and M. Höcker, personal communication). Constructs used for Gal4 reporter assays were Gal4-Luc (five Gal4 binding sites upstream of a minimal promoter linked to luciferase), Sp1-Gal4, and Sp3-Gal4 (activator constructs containing the Gal4 DNA binding domain linked to the Sp1 or Sp3 transactivation domain) and were provided by G. Suske (Institut für Molekularbiologie und Tumorforschung, Marburg, Germany). Two plasmids that do not contain ISREs were used for normalization in cotransfection experiments: the {beta}-galactosidase expression construct {beta}-gal pSV2 (Promega, Mannheim, Germany) and the renilla luciferase expression construct pRL-TK (Promega).

Transient Transfection Experiments

Bon, HT29, and HepG2 cells were transiently transfected with plasmids using the calcium phosphate mammalian cell transfection kit (5 Prime 3 Prime, Boulder, CO), following the manufacturer’s instructions, and Capan-1 cells were transiently transfected using the Effectene Transfection Reagent Kit (Qiagen), according to the instructions supplied by the manufacturer. In brief, cells were plated at a density of 2 x 105 cells/35-mm well and transfected the next day with 2 µg per well of plasmid. To determine the transfection efficiencies of the test plasmids, normalization was performed by cotransfecting 1 µg per well of a {beta}-galactosidase expression construct or 0.01 µg per well of a renilla luciferase expression construct. After the transfection protocols, cells were washed, incubated in serum-free UltraCulture medium, and treated with IFN-{alpha} (1000 IU/mL) or vehicle for 24 hours. Cell extracts were made and subsequently analyzed for luciferase activity (Lumat LB 9501 by EG & G Berthold, Promega, Bad Wildbad, Germany), according to the manufacturer’s recommended protocol. Luciferase activity was expressed as relative light units (RLU).

Electrophoretic Mobility Shift Assays

Approximately 2 x 106 Bon cells per 100-mm dish were incubated in UltraCulture medium with IFN-{alpha} (1000 IU/mL) or vehicle for various times, and nuclear extracts were prepared using a non-ionic detergent method as described (19). Equal amounts (10 µg) of nuclear extracts were incubated with radiolabeled double-stranded (ds)-oligonucleotides –88/–50 (5'-CC CGGGGCGGGCCGGGGGCGGGGTCCCGGCGGGGCGGAG3'), –63/–47 (5'-CCGGCGGGGCGGAGCCA-3'), or –88/–47 (5'-CCGGGGCGGGCCGGGGGCGGGGTCCCGGCGGGGCG GAGCCA-3') of the human VEGF promoter in the presence of 20 µL of binding buffer (20 mM HEPES [pH 8.4], 1 µg of poly(dI-dC), 10 µg of bovine serum albumin, 60 mM KCl, 5 mM dithiothreitol, 1 mM ZnCl2, and 10% glycerol) for 30 minutes at 30 °C. For competition experiments, extracts were preincubated for 30 minutes at 25 °C with a 100x molar excess of wild-type or mutant competitor oligonucleotides (Santa Cruz Biotechnology, Heidelberg, Germany) before the addition of labeled probes. For supershift experiments, extracts were preincubated for 30 minutes at 25 °C with 1 µL of antibodies against human Sp1, Sp3, AP-2, or Egr-1 (Santa Cruz Biotechnology) before the addition of labeled probes. Nuclear extracts derived from HeLa cells (Santa Cruz Biotechnology) served as a positive control for the binding activity of the AP-2 antibody. DNA–nuclear protein complexes were separated and visualized as described (19).

Determination of Nuclear Sp1 and/or Sp3 Levels

Nuclear extracts from Bon cells previously incubated with IFN-{alpha} (1000 IU/mL) in UltraCulture medium for 0, 6, 12, or 24 hours were analyzed by immunoblotting (17) using mouse anti-human Sp1 antibody (1 : 1000; PharMingen) and rabbit anti-human antibodies against Sp3 (1 : 1000; Santa Cruz Biotechnology) and STAT1 (1 : 5000; PharMingen).

Nude Mouse Human Tumor Xenograft Model

Capan-1 or Bon cells (106 cells per animal) were injected subcutaneously into the left flank of female 6- to 8-week-old BALB/c mice (Bomholtgard, Rye, Denmark). Tumor growth was monitored by measuring tumor area, calculated as the product of the largest diameter and its perpendicular diameter. After tumor areas of 25 mm2 were attained (after approximately 10 days), the mice were randomly assigned to one of three groups of five animals, and either a dose of IFN-{alpha} (5 x 104 or 5 x 105 IU) or 0.9% NaCl was injected subcutaneously once a day until the mean tumor area in the control group reached 100 mm2. The mice then were killed, and the tumors were processed for immunohistochemical analysis. Animal care was in accordance with institutional guidelines. The experiments were approved by local animal research authorities.

Statistical Analysis

Nonparametric tests for unpaired and paired observations (Mann–Whitney U test and Wilcoxon matched paired test, respectively) were done using the Statistical Package for Social Sciences 11.0 software package (Statistical Package for Social Sciences, Inc., Chicago, IL). All tests were two-sided, and data were considered to be statistically significant at P<.05. Data represent the mean and 95% confidence intervals (CIs) from at least three independent experiments.


   RESULTS
Top
 Notes
 Abstract
 Introduction
 Patients and Methods
 Results
Discussion
 References
 
Expression of VEGF and VEGF-Rs in Neuroendocrine Tumors

To test the hypothesis that IFN-{alpha} treatment regulates VEGF gene expression, we first analyzed the expression of VEGF and VEGF-Rs in 10 surgically resected NE tumors from humans: seven derived from the foregut and three derived from the midgut. By using immunohistochemistry, we observed strong VEGF protein expression in the cytoplasm of all tumor samples that were positive for the neuroendocrine marker synaptophysin (compare Fig. 1, A, a and bGo). In all tumor specimens, immunohistochemistry for the endothelial cell marker CD31 demonstrated an abundance of endothelial cells surrounding the tumor cells (Fig. 1, A, cGo), which is consistent with the clinical observation of a highly vascularized tumor (15). Nine of the 10 tumor samples also expressed the VEGF-Rs KDR/flk-1 (Fig. 1, A, d and eGo) and Flt-1 (Fig. 1, A, fGo). To confirm the cellular origin of VEGF and the VEGF-Rs, we used in situ hybridization and determined that VEGF mRNA was expressed exclusively in NE tumor cells, whereas KDR and Flt-1 mRNAs were expressed exclusively in tumor endothelia (data not shown).



View larger version (47K):
[in this window]
[in a new window]
  		Fig. 1. Expression of vascular endothelial growth factor (VEGF) and its receptors in neuroendocrine (NE) tumors. A) Immunohistochemical detection of VEGF (a), synaptophysin (b), CD31 (c), mouse anti-human kinase insert domain-containing region (KDR)/flk-1 (d and e), and mouse anti-human fms-like tyrosine kinase (Flt-1) (f) in frozen NE tumor tissue sections. Tissue sections were immunohistochemically stained as described (17) using the alkaline phosphatase/anti-alkaline phosphatase method with new fuchsin as a developer. The antigen–antibody complexes stain red. B) Transcripts of VEGF splice variants analyzed by reverse transcription–polymerase chain reaction (RT–PCR) in the human NE tumor cell lines Bon, LCC18, and QGP1 using primers that recognize all known splice forms (17). PCR was performed without (–) or with (+) prior RT, respectively. RT–PCR for {beta}-actin was used for comparison of input mRNA integrity. C) Human NE tumor cell lines were plated in serum-free medium. After 24 hours, the medium was collected, and cell extracts were prepared as described (17). VEGF protein concentration in supernatants (open bars) and cell extracts (solid bars) was measured by enzyme-linked immunosorbent assay and expressed as nanograms of VEGF per milligram of total protein. Data represent the mean VEGF concentration from three experiments and upper 95% confidence intervals.

 
The abundance of blood vessels in conjunction with substantial expression of VEGF and VEGF-Rs in NE tumors suggests that VEGF may regulate angiogenesis in these tumors. To investigate whether VEGF represents an IFN-{alpha}–regulated target in NE tumors, we sought a suitable in vitro model and first evaluated VEGF expression in the human NE tumor cell lines Bon, LCC18, and QGP1. Using primers designed to amplify all five known splice variants of VEGF, we observed expression of the VEGF121 and VEGF165 isoforms in all three cell lines (Fig. 1, BGo). We also detected VEGF protein in supernatants and cell extracts from these cell lines (Fig. 1, CGo). Because VEGF is constitutively secreted, its concentrations were substantially higher in supernatants than in cell extracts. VEGF synthesis and secretion made these cell lines a suitable in vitro model to study whether IFN-{alpha} regulates VEGF expression.

Effect of IFN-{alpha} on VEGF Gene Expression

To explore whether IFN-{alpha} directly affects VEGF expression in NE tumor cells, we determined VEGF165 mRNA concentrations in Bon, LCC18, and QGP1 cells incubated for 96 hours with IFN-{alpha} at 1000 IU/mL or vehicle by quantitative competitive RT–PCR. This method is highly sensitive and particularly valuable for measuring the relative abundance of a particular mRNA (17). VEGF165 mRNA levels in all three cell lines treated with IFN-{alpha} were lower than those in untreated cells (Fig. 2, AGo). Using Bon cells as a representative cell model, we found that the decrease in mRNA levels was cumulative, occurring in a time-dependent manner and continuing to decrease at 96 hours (Fig. 2, BGo). The IFN-{alpha}-mediated decrease in VEGF mRNA levels was reflected by a concomitant decrease in VEGF protein secretion. Compared with VEGF levels from untreated control cells, those from IFN-{alpha}-treated cells were statistically significantly lower (difference = 38% [95% CI on the difference = 22% to 61%; n = 9; P = .008] after 96 hours) (Fig. 2, CGo).



View larger version (16K):
[in this window]
[in a new window]
  		Fig. 2. Effect of interferon alpha (IFN-{alpha}) on vascular endothelial growth factor (VEGF) gene expression. A) Total RNA was isolated from the human neuroendocrine tumor cell lines Bon, LCC18, and QGP1 that were incubated with vehicle (phosphate-buffered saline) or IFN-{alpha} (1000 IU/mL) for 96 hours. Steady-state VEGF mRNA levels were determined by quantitative competitive reverse transcription–polymerase chain reaction (RT–PCR) as described (17). The level of VEGF165 mRNA was calculated as the ratio of VEGF165 copies to {beta}-actin copies and expressed as a percentage of VEGF165 mRNA from vehicle-treated control cells. Bars represent means of three independent experiments with upper 95% confidence intervals. B) Bon cells were incubated with vehicle or IFN-{alpha} at 1000 IU/mL for the indicated times, and VEGF165 mRNA was quantified by quantitative competitive RT–PCR. Values are expressed as a percentage of the VEGF165 mRNA levels from vehicle-treated cells and are derived from three independent experiments. C) Bon cells were incubated with vehicle or IFN-{alpha} at 1000 IU/mL for the indicated times, and VEGF protein concentration in supernatants was analyzed by enzyme-linked immunosorbent assay and normalized to total protein content. Values are expressed as percentage of VEGF from vehicle-treated cells and are derived from nine experiments. *P = .008, Wilcoxon test. All data represent the mean and upper 95% confidence intervals.

 
Transcriptional Regulation of the VEGF Promoter by IFN-{alpha}

To further elucidate the molecular mechanisms underlying the IFN-{alpha}-mediated decrease in VEGF gene expression, we tested whether IFN-{alpha} inhibits VEGF transcription by using a luciferase reporter gene construct containing the promoter sequences –2018/+50 of the human VEGF gene (–2018-VEGF-Luc) in Bon cells. Compared with the control treatment, IFN-{alpha} treatment resulted in a statistically significant, dose-dependent decrease of VEGF promoter activity, with half-maximal effects of 23% inhibition observed at a therapeutically relevant concentration of 10 IU/mL (P = .012) and maximal effects of 47% (95% CI = 30% to 61%; n = 12; P = .002) observed at 1000 IU/mL (Fig. 3, AGo). This transcriptional inhibition was not the result of a nonspecific inhibition of gene transcription, because IFN-{alpha} increased luciferase activity 3.90-fold (95% CI = 2.78- to 5.02-fold) in Bon cells transfected with the ISRE-Luc reporter construct (Fig. 3, BGo) but had no effect on luciferase activity in Bon cells transfected with the viral thymidine kinase promoter construct that was used to normalize the cotransfection experiments (data not shown).



View larger version (24K):
[in this window]
[in a new window]
  		Fig. 3. Effect of interferon alpha (IFN-{alpha}) on vascular endothelial growth factor (VEGF) promoter activity. A) The human neuroendocrine tumor cell line Bon was transiently transfected with a –2018-VEGF-Luc construct, which contains the VEGF promoter region (–2018 to +50) upstream of the luciferase gene, and incubated for 24 hours with the indicated concentrations of IFN-{alpha}. Luciferase activity was determined and corrected for transfection efficiency. The data are expressed as the percentage of vehicle-treated (phosphate-buffered saline) cells. Bars represent the mean from 12 experiments and the upper 95% confidence intervals. *P = .012 (10, 100, 500 IU/mL), P = .002 (1000 IU/mL), derived from the Wilcoxon test. Comparisons were made with the control group. B) Bon cells were transiently transfected with the indicated 5' deletion reporter gene constructs of the human VEGF promoter and subsequently incubated with vehicle (open bars) or IFN-{alpha} at 1000 IU/mL (filled bars). After 24 hours, luciferase activity was measured in cell extracts. To exclude nonspecific inhibitory effects on transactivation, a construct containing the interferon-stimulated response element (ISRE) fused to a luciferase reporter gene (ISRE-Luc) was used for transfection before IFN-{alpha} or vehicle incubation. Data show the mean relative light units (RLU) of at least nine experiments and the upper 95% confidence intervals. C) Bon cells were transiently transfected with a heterologous enhancerless herpes simplex thymidine kinase promoter reporter construct fused to the region –85/–50 of the human VEGF promoter and incubated with vehicle or IFN-{alpha} at 1000 IU/mL for 24 hours. Luciferase activity was measured in RLU, and data are expressed as the mean and the upper 95% confidence intervals of 12 experiments. *P = .002, Wilcoxon test. D) Pancreatic Capan-1, hepatic HepG2, and colonic HT29 cells were transiently transfected with the –2018-VEGF-Luc construct and treated with vehicle or IFN-{alpha} at 1000 IU/mL for 24 hours. Luciferase activity was measured and expressed as a percentage of vehicle-treated cells. Bars represent the mean of at least five experiments and upper 95% confidence intervals. *P = .031 (Capan-1), P = .002 (HepG2), P = .008 (HT29), derived from the Wilcoxon test.

 
To determine the minimal sequence required for IFN-{alpha} responsiveness, we next tested a series of constructs with different 5' deletions in the VEGF promoter. In Bon cells, neither the basal promoter activity nor the ability to respond to IFN-{alpha} was affected by sequential 5' deletions within the region –2018/–85 (Fig. 3, BGo). By contrast, the basal promoter activity and the ability to respond to IFN-{alpha} were completely abolished by a 5' deletion that removed the region –85/–52. The functional importance of this –85/–52 region was highlighted by the observation that the transfer of this sequence to the heterologous thymidine kinase promoter strongly enhanced basal reporter gene expression and conferred IFN-{alpha} responsiveness in transient transfection experiments (n = 12; P = .002) (Fig. 3, CGo).

Next, we investigated whether IFN-{alpha}-mediated inhibition of VEGF gene transcription is cell type-specific. Using –2018-VEGF-Luc, which contains the promoter sequences –2018/+50 of the human VEGF gene, we observed a statistically significant IFN-{alpha}-mediated decrease in VEGF promoter activity in cancer cell lines from different origins, including those from the pancreas (Capan-1, n = 5, 43% of untreated control; P = .031), liver (HepG2, n = 9, 45% of untreated control; P = .002), and colon (HT29, n = 8, 78% of untreated control; P = .008) (Fig. 3, DGo). These results suggest that the inhibition of VEGF gene transcription by IFN-{alpha} is not restricted to NE tumor cells but can also be observed in tumors from different origins.

Identification of Minimal Promoter Requirements and Involved Transcription Factors

Analysis of the –85/–50 region of the human VEGF gene promoter reveals putative binding sites for the transcription factors AP-2 (21) and Egr-1 (19) and for three GC boxes (19), which are known to bind Sp1 and Sp3 with similar affinities in different promoters (22) (Fig. 4, AGo). To identify transcription factors involved in the IFN-{alpha}-dependent decrease of VEGF promoter activity, we performed reporter gene transfection experiments in which the constructs contained inactivated transcription factor binding sites. The basal promoter activity and the ability to respond to IFN-{alpha} were comparable among Bon cells transfected with constructs containing functionally inactivated Egr-1 and Egr-1/AP-2 binding sites and Bon cells transfected with the intact construct (–85-VEGF-Luc) (Fig. 4, BGo). By contrast, the basal promoter activity and the ability to respond to IFN-{alpha} were reduced in Bon cells transfected with constructs containing functionally inactivated GC boxes (Fig. 4, BGo). Sequential 5' deletion analysis within region –115/–50 (Fig. 4, CGo) and deletions of GC box III, GC boxes II and III, or GC box I (Fig. 4, DGo), showed that the sequence –63/–50 corresponding to the proximal GC box I was sufficient to confer IFN-{alpha} responsiveness to Bon cells.



View larger version (27K):
[in this window]
[in a new window]
  		Fig. 4. Functional analysis of the interferon alpha (IFN-{alpha}) responsive element in the human vascular endothelial growth factor (VEGF) promoter. A) Schematic representation of putative binding sites for transcription factors Egr-1, AP-2, and Sp1 and/or Sp3, which bind to GC boxes within the region -85/-50 of the human VEGF promoter. GC boxes and nucleotide positions are numbered according to their position relative to the transcription start site. B–D) The human neuroendocrine tumor cell line Bon was transiently transfected with the indicated VEGF promoter–luciferase reporter gene constructs encoding inactivated transcription factor binding sites (substituted nucleotides are shown in small letters) in -85-VEGF-Luc-based reporter gene constructs (B), 5' deletions within region -115/ -50 of the human VEGF promoter gene in pGL3-based vectors (C), or single and combined GC box deletions in pT81-based constructs (D). After incubation with IFN-{alpha} at 1000 IU/mL or vehicle (phosphate-buffered saline) for 24 hours, luciferase assays were performed. Luciferase activity was measured in relative light units, and data are expressed as the mean and upper 95% confidence intervals of at least nine experiments.

 
To identify which nuclear proteins bound to the –88/–50 region, we used this 32P-labeled sequence (Fig. 5, AGo) and nuclear extracts derived from Bon cells in gel shift assays (Fig. 5, BGo). Two specific DNA–protein complexes were reproducibly identified. Treatment with IFN-{alpha} for periods from 5 minutes to 12 hours before extracting the nuclear proteins did not change the pattern or the relative amounts of these DNA–protein complexes (lane 3 represents a 20-minute incubation time; other data not shown). These complexes contained Sp-like factors, which disappeared almost completely in competition with a nonradioactively labeled Sp consensus oligonucleotide (lane 4). This competition was specific because a mutant Sp consensus oligonucleotide (lane 5) did not displace the complexes. The slower migrating complex was supershifted by the addition of an antibody against Sp1 (lane 6), whereas the faster migrating complex was supershifted by the addition of an antibody against Sp3 (lane 7). The addition of antibodies to Sp1 and Sp3 supershifted both complexes almost completely (lane 8). The interaction between Sp1 or Sp3 to this promoter region was not affected by IFN-{alpha}, because the binding profiles from nuclear extracts from cells treated with IFN-{alpha} for various time points were similar to those from untreated cells (compare lanes 9–13 with lanes 4–8).



View larger version (42K):
[in this window]
[in a new window]
  		Fig. 5. Transcription factors Sp1 and Sp3 bind vascular endothelial growth factor (VEGF) promoter regions that confer interferon alpha (IFN-{alpha}) responsiveness. A) Schematic representation of radiolabeled sequences used in the gel shift assays below. Putative binding sites for Egr-1, AP-2, and Sp1 and/or Sp3, which bind GC boxes, are marked. Numbers represent nucleotide positions relative to the transcription start site. B) Nuclear extracts derived from Bon cells treated with vehicle (phosphate-buffered saline) or IFN-{alpha} (1000 IU/mL for 20 minutes) were incubated with 32P-labeled oligonucleotides corresponding to regions -88/ -50 (lanes 1–19) and -63/-47 (lanes 20–31) in the DNA–protein complexes separated in gel shift assays. For competition experiments, nuclear extracts were preincubated with 100x molar excess of wild-type (wt) or mutant (m) Sp or Egr-1 oligonucleotides before being incubated with the 32P-labeled oligonucleotides. For supershift experiments to identify specific components, nuclear extracts were incubated with antibodies against Sp1 (aSP-1), Sp3 (aSP-3), or Egr-1 (aEgr-1) before being incubated with the 32P-labeled oligonucleotides. Antibodies against Sp1 supershifted (SS) an Sp1 complex (lanes 6, 8, 11, 13, 24, 26, 29, and 31), and antibodies against Sp3 (aSP-3) supershifted an Sp3-named complex (lanes 7, 8, 12, 13, 25, 26, 30, and 31). By contrast, antibodies against Egr-1 (aEgr-1) did not supershift any complex (lanes 16 and 19). Lanes 1 and 20 represent control lanes in which the oligonucleotide probe was not incubated with any nuclear extracts. Each condition was repeated at least three times; representative gels are shown.

 
We also sought to detect binding of the Egr-1 and AP-2 proteins to the –80/–50 promoter region. In competition or supershift experiments, we were unable to detect binding of Egr-1 or AP-2 proteins to the promoter region under either basal or IFN-{alpha}-stimulated conditions (lanes 14–19 and data not shown). Similar results were obtained from gel shift assays in which we used a slightly extended labeled promoter sequence (–88/–47) to facilitate transcription factor binding (data not shown).

The VEGF promoter–reporter gene experiments had demonstrated that the proximal GC box I (–63/–50) is sufficient to confer IFN-{alpha} responsiveness (Fig. 4, DGo). By using a 32P-labeled –63/–47 promoter region (Fig. 5, AGo) for gel shift assays, we observed two DNA–protein complexes. On the basis of results obtained from competition (lanes 22, 23, 28, and 29) and supershift (lanes 24–26 and 29–31) analyses, Sp1 and Sp3 but not Egr-1 or AP-2 (data not shown) bound to this minimal promoter region required for IFN-{alpha}-mediated promoter inhibition (Fig. 5, BGo). In summary, Sp1 and Sp3 bind to the VEGF promoter sequence located at positions –88/–50 and to the sequence –63/–50, which is sufficient but not absolutely required to confer IFN-{alpha} responsiveness.

IFN-{alpha} treatment did not appear to change Sp1 or Sp3 DNA binding properties (Fig. 5, BGo). To assess whether Sp1- and Sp3-dependent inhibition of VEGF promoter activity resulted from an IFN-{alpha}-induced change of nuclear Sp1 or Sp3 content, we analyzed nuclear levels of both transcription factors by immunoblotting. Compared with Sp1 and Sp3 nuclear protein levels from untreated cells, levels in IFN-{alpha}-treated Bon cells did not change over time (Fig. 6, AGo). By contrast, levels of STAT1, which translocates to the nucleus in response to IFN-{alpha}, did increase over time (Fig. 6, AGo).



View larger version (37K):
[in this window]
[in a new window]
  		Fig. 6. Effects of interferon alpha (IFN-{alpha}) on nuclear Sp1 and/or Sp3 levels and transactivation potential. A) The human neuroendocrine tumor cell line Bon was incubated with IFN-{alpha} at 1000 IU/mL or vehicle (phosphate-buffered saline) for the indicated times. Nuclear extracts were made, and the Sp1 and Sp3 protein levels were subsequently analyzed by immunoblotting with specific mouse anti-human Sp1 and rabbit anti-human Sp3 isoforms p84 and p91 antibodies. Detection of STAT1, which translocates to the nucleus in response to IFN-{alpha}, was used as a positive control. Antibody–antigen complexes were detected by horseradish peroxidase secondary antibodies and enhanced chemiluminescence. A representative immunoblot of four independent experiments is shown. The anti-human Sp3 antibody recognizes two Sp3 isoforms. The anti-STAT1 antibody recognizes the p84 and p91 isoforms. B) Bon cells were transiently cotransfected with a Gal4-Luc plasmid, which contains five Gal4 binding sites in front of a minimal promoter fused to the luciferase gene and the indicated constructs containing the Sp1 or Sp3 transactivation domain fused to a Gal4 DNA-binding domain. The cells were subsequently incubated with IFN-{alpha} at 1000 IU/mL or vehicle for 24 hours before luciferase activity was measured. Luciferase activity was measured in relative light units (RLU) and normalized to renilla RLU of the cotransfected control plasmid pTK-RL. Bars represent the mean and upper 95% confidence intervals from five experiments, each performed in triplicate. *P = .043, Wilcoxon test.

 
By using appropriate Gal4 reporter assay systems, we next analyzed whether IFN-{alpha} influenced the transcriptional activation properties of Sp1 or Sp3. For these assays, we cotransfected constructs containing the transactivation domain of the Sp1 or Sp3 proteins linked to the Gal4 DNA binding domain and a luciferase construct containing five Gal4 binding sites as a reporter. IFN-{alpha} treatment resulted in a statistically significant reduction in the transactivation potential of Sp1-Gal4 (n = 5, P = .043). Although Sp3-mediated transactivation was also decreased by IFN-{alpha}, the decrease did not reach statistical significance (Fig. 6, BGo).

In vivo Effects of INF-{alpha} on VEGF Expression and Angiogenesis

To evaluate the biologic relevance of our in vitro observations, we investigated whether IFN-{alpha} could regulate the angiogenic response elicited by Bon and Capan-1 tumor cells in a nude mouse xenograft model. In response to IFN-{alpha}, constitutive expression of VEGF was transcriptionally inhibited in both cell lines (Figs. 2 and 3GoGo). Compared with tumors in vehicle-treated mice, tumors in INF-{alpha}-treated mice were dose dependently growth inhibited (Fig. 7, AGo, and data not shown). Growth inhibition of Capan-1 xenografts was statistically significant at days 17 and 19 (n = 5; tumor area was reduced by approximately 36% on day 19; P = .008 and P = .016, respectively) for mice who received IFN-{alpha} at 5 x 105 IU/day (Fig. 7, AGo). Compared with microvessel density in tumors from vehicle-treated mice, that in Capan-1 tumors from IFN-{alpha}-treated mice was statistically significantly decreased. The mean microvessel density for the control group was 56 (95% CI = 49 to 69), whereas the mean microvessel density for the 5 x 104 IU/day group and the 5 x 105 IU/day group was 38 (95% CI = 32 to 44; P = .026) and 37 (95% CI = 32 to 41; P = .015), respectively (Fig. 7, B and CGo). The Bon cell xenograft experiments showed similar results regarding IFN-{alpha}-mediated growth inhibition and decreased microvessel density (data not shown). However, Bon cells do not form tumors consistently, which is necessary for appropriate statistical analyses.



View larger version (68K):
[in this window]
[in a new window]
  		Fig. 7. Effect of interferon alpha (IFN-{alpha}) on tumor growth and angiogenesis in vivo. A) Pancreatic Capan-1 cells (106) were injected subcutaneously into the left flank of nude mice. After 10 days, when the tumor areas were approximately 25 mm2, the mice were treated daily with a subcutaneous injection of IFN-{alpha} (at 5 x 104 or 5 x 105 IU/day) or vehicle (0.9% saline). The tumor area was calculated at the indicated times by measuring the largest diameter and its perpendicular diameter and determining the product of the two measurements. Each point represents the mean of measurements obtained from five mice and the upper or lower 95% confidence intervals. * = statistical significance of the difference between IFN-{alpha}-treated mice (5 x 105 IU/day) and vehicle-treated mice (day 17: P = .008; day 19: P = .016; Mann–Whitney U test). B) Tumors harvested from vehicle-treated or IFN-{alpha}-treated mice were sectioned and immunostained with rat anti-mouse pan endothelial cell antigen (MECA)-32 antibody to visualize endothelial cells, and microvessel density was determined by counting three 400x fields per tumor (18). Bars represent the mean microvessel density of five tumors derived from five mice per group and the upper 95% confidence intervals. * indicates statistical significance relative to the vehicle-treated groups (5 x 104 IU/day IFN-{alpha} group: P = .026; 5 x 105 IU/day IFN-{alpha} group: P = .015; Mann–Whitney U test). C) Representative tumor areas from vehicle- or IFN-{alpha}-treated mice after immunostaining for endothelial cells are shown.

 
To determine the biologic significance of our observation for the clinical scenario of IFN-{alpha} treatment as an antineoplastic therapy, we analyzed VEGF plasma levels and VEGF mRNA concentrations in liver metastasis biopsy specimens taken from patients with NE tumors before and during IFN-{alpha} therapy. All patients had midgut tumors with liver metastasis and, in the time between obtaining the two biopsy specimens, had stable disease, as judged by CT and/or abdominal ultrasound. In quantitative competitive RT–PCR analyses, compared with tumor tissues collected before treatment (normalized VEGF165 mRNA = 18.4, 95% CI = –17.6 to 54.4 arbitrary units), VEGF mRNA expression was statistically significantly decreased in tumor tissues collected after treatment with IFN-{alpha} (normalized VEGF165 mRNA = 4.6, 95% CI = –2.5 to 11.7 arbitrary units, n = 5; P = .043) (Fig. 8, A and BGo). We also observed a statistically significant decrease (P = .043) in circulating VEGF plasma levels in these five patients after initiation of the IFN-{alpha} therapy from 38 pg/mL (95% CI = 11.3 to 64.8 pg/mL) to 21 pg/mL (95% CI = 9.4 to 32.7 pg/mL) (Fig. 8, CGo). Circulating levels of bFGF, another angiogenic factor, were not altered (data not shown).



View larger version (53K):
[in this window]
[in a new window]
  		Fig. 8. Effect of interferon alpha (IFN-{alpha}) on vascular endothelial growth factor (VEGF) gene expression and angiogenesis in patients with neuroendocrine (NE) tumors. VEGF165 mRNA expression was examined in NE tumor biopsy specimens collected from the same liver metastasis before treatment and after 6 months of IFN-{alpha} therapy. VEGF165 mRNA levels were quantified by quantitative competitive reverse transcription–polymerase chain reaction (RT–PCR) (16) and expressed as the ratio of VEGF165 to total cDNA concentration. A) Representative image of an ethidium bromide-stained quantitative competitive RT–PCR gel for VEGF165 mRNA levels from a single patient. B) Quantitative analysis of VEGF mRNA levels in biopsy specimens from five patients with NE tumors. Bars represent the mean and the upper 95% confidence intervals. * denotes statistical significance of differences between treatment groups (P = .043, Wilcoxon test). C) Circulating VEGF plasma concentrations before and after IFN-{alpha} therapy in five patients were determined by enzyme-linked immunosorbent assay. Bars represent the mean and the upper 95% confidence intervals. * denotes statistical significance between treatment groups (P = .043, Wilcoxon test). D) Biopsy specimens from identical liver metastases were collected before (a and c) and after (b and d) IFN-{alpha} therapy from patients with NE tumors. The specimens were processed and immunostained for the endothelial cell-specific antigen CD31 as described (17) using the alkaline phosphatase/anti-alkaline phosphatase method with new fuchsin as a developer. Data represent samples from two patients (patient 1: a and b; patient 2: c and d) taken before (a and c) and after (b and d) IFN–{alpha} therapy.

 
To test whether the ability of IFN-{alpha} to inhibit VEGF gene expression in patients with NE tumors translates into an antiangiogenic response, we examined microvessel density in the liver metastasis biopsy specimens before and after IFN-{alpha} treatment. The small amount of biopsy material available and the different areas of viable tumor tissue suitable for analysis prevented an extensive and rigorous microvessel density determination in all samples. However, in comparable tissue samples from several patients, we observed much less CD31 staining in the biopsy specimen collected after initiation of IFN-{alpha} therapy relative to that in the biopsy specimen before treatment from the same patient (Fig. 8, DGo, compare a with b and c with d), which was also paralleled by a decrease of VEGF mRNA expression (Fig. 8, BGo).


   DISCUSSION
Top
 Notes
 Abstract
 Introduction
 Patients and Methods
 Results
 Discussion
References
 
Used clinically for angioproliferative disorders (6) and some neoplastic diseases (13,14), IFN-{alpha} has been perceived as an antiangiogenic agent, although the underlying mechanism of action remains incompletely understood. We examined the functional relationship between IFN-{alpha} and VEGF for several reasons: 1) VEGF contributes to the progression of many solid tumors by promoting the "angiogenic switch" (2,4), 2) tumor cell expression of VEGF is inversely associated with prognosis and survival in several human malignancies (23–25), 3) inhibition of VEGF activity inhibits tumor angiogenesis and reduces tumor growth (26,27), and 4) the VEGF–VEGF-R system is considered to be a promising target for the development of antiangiogenic tumor therapy (4,5,26,27).

We chose human NE tumors as a model to analyze the effects of IFN-{alpha} on VEGF expression and angiogenesis because therapeutic efficacy of IFN-{alpha} in these highly vascularized (15) tumors is well documented (14). VEGF and the VEGF-Rs KDR and flt-1 were strongly overexpressed in all the human tumor samples relative to expression in adjacent nontransformed tissue samples. The tumor samples were also extensively vascularized, suggesting that the VEGF–VEGF-R system plays a central role in the vascularization of NE tumors.

Using several human NE tumor cell lines, we demonstrated that IFN-{alpha} can inhibit VEGF gene transcription. Although many factors have been identified that stimulate VEGF gene expression, including hypoxia, cytokines, growth factors, oncogenes, and tumor suppressor genes (28), IFN-{alpha} is, to the best of our knowledge, the first therapeutically relevant agent to inhibit VEGF gene transcription. Of note, the inhibitory effect of IFN-{alpha} was not restricted to NE tumor cells and could also be demonstrated in other tumor cell types, suggesting that our observations are potentially applicable to a variety of other cancers.

Molecular analysis revealed an unexpected and novel finding regarding the mechanism of IFN-{alpha}-mediated inhibition of VEGF gene transcription. The VEGF promoter does not contain classical IFN-{alpha}-sensitive enhancer elements, such as the gamma-activated sequence (GAS) or ISRE targeted by IFN-stimulated DNA binding factors (7,29). Instead, we found that the proximal GC box I (VEGF promoter region –66/–50) was sufficient to confer IFN-{alpha} responsiveness. Gel shift analysis confirmed that this region specifically bound the transcription factors Sp1 and Sp3. Several activators of the VEGF promoter have been described that act via an Sp1- and/or Sp3-dependent mechanism, such as interleukin 1{beta}, tumor necrosis factor-{alpha}, and platelet-derived growth factor (19); however, Sp1-mediated inhibition of the VEGF promoter has not been previously observed. Furthermore, there is no previous documented evidence describing Sp1 and Sp3 as downstream mediators of IFN-{alpha}-induced inhibition of gene expression. Therefore, the role of the transcription factors Sp1 and Sp3 in IFN-{alpha}-mediated inhibition of VEGF gene expression is novel. Although Sp1 and Sp3 can activate some GC-rich target genes, such as p21waf1/cip1 or the human alpha1 (I) collagen gene (30), Sp3 has been shown to either activate or repress promoter activity of other genes, depending on the cellular and molecular context (22,31,32). We and others [G. Schäfer and M. Höcker, (32)] have recently introduced the VEGF promoter into Sp1- and/or Sp3-deficient Sf9 Schneider insect cells and shown that Sp1 and Sp3 are both able to transactivate the VEGF promoter. Therefore, inhibition of VEGF promoter activity by IFN-{alpha} is most likely mediated by inhibition of Sp1 and/or Sp3 transactivation activity. We tested this hypothesis using Gal4 reporter assays and found that, indeed, IFN-{alpha} inhibits Sp1 and/or Sp3 transactivation activity. The molecular mechanisms underlying Sp1- and/or Sp3-dependent transactivation have not been completely elucidated, although several mechanisms have been proposed: 1) modulation of nuclear Sp1 and/or Sp3 concentrations (22), 2) phosphorylation changes in Sp1 and/or Sp3 proteins (33), 3) glycosylation changes in Sp1 and/or Sp3 proteins (34), and 4) interaction of Sp1 and/or Sp3 with specific co-activators or co-repressors (35). We are currently investigating which of these mechanisms is operative for IFN-{alpha} and, at this point, can exclude alterations in the nuclear concentration of Sp1 and Sp3.

Two of the goals of translational research are the validation of in vitro observations in a preclinical model and then further validation in the clinical setting in patients. To relate our in vitro observations to clinical observations, we analyzed the effects of IFN-{alpha} on VEGF expression and angiogenesis in a xenotransplanted mouse model and on VEGF plasma levels and angiogenesis in liver metastases of patients with NE tumors treated with IFN-{alpha}. In mice, IFN-{alpha} treatment resulted in an inhibition of tumor growth and a marked reduction of angiogenesis demonstrated by reduced microvessel density within the tumor. In humans, patients with NE tumors treated with IFN-{alpha} demonstrated decreased VEGF plasma levels, decreased VEGF mRNA concentrations, and a pronounced reduction of microvessel density in biopsy specimens obtained from liver metastases. Together, these observations strongly suggest that IFN-{alpha} treatment results in inhibition of VEGF gene expression in an experimental in vivo model and in patients, at least in patients with NE tumors. The observations should, however, be considered cautiously because IFN-{alpha} might also regulate other factors involved in the angiogenic response, as suggested by several preclinical models (10–12). However, by contrast with our data, there is currently no evidence for functional relevance of these mechanisms in human therapy. Although we did not examine all angiogenic factors, we did not detect changes in bFGF plasma levels in response to IFN-{alpha} treatment.

Our results have implications for the optimization of some aspects of antiangiogenic therapy. For example, circulating VEGF concentrations paralleled microvessel density during IFN-{alpha} therapy. Therefore, VEGF plasma levels might prove a valuable surrogate marker to optimize response prediction as well as duration and dosage of IFN-{alpha} treatment (36). This monitoring appears essential when balancing the clinical benefit associated with IFN-{alpha} therapy with its costs and side effects. Furthermore, having defined the antiangiogenic mode of action now allows us to develop rationales for combining IFN-{alpha} with other antiangiogenic drugs to achieve synergistic therapeutic benefit.


   NOTES
Top
 Notes
Abstract
 Introduction
 Patients and Methods
 Results
 Discussion
 References
 
Zofia von Marschall and Arne Scholz contributed equally to this article.

S. Rosewicz was supported by grants from the Deutsche Krebshilfe, Wilhelm Sander-Stiftung, Deutsche Forschungsgemeinschaft (DFG), Berliner Krebsgesellschaft, Else Kröner Fresenius Stiftung, and Sonnenfeld Stiftung. M. Höcker was supported by DFG and the Bundesministerium für Bildung und Forschung (BMBF).

We thank G. Finkenzeller (GeneScan AG, Freiburg, Germany), G. Suske (Institut für Molekularbiologie und Tumorforschung, Marburg, Germany), and S. Bhattacharya (Department of Cardiovascular Medicine, University of Oxford, Oxford, U.K.) for the generous gifts of plasmids.


   REFERENCES
Top
 Notes
 Abstract
 Introduction
 Patients and Methods
 Results
 Discussion
 References
 

1 Carmeliet P, Jain RK. Angiogenesis in cancer and other diseases. Nature 2000;407:249–57.[CrossRef][Medline]

2 Hanahan D, Folkman J. Patterns and emerging mechanisms of the angiogenic switch during tumorigenesis. Cell 1996;86:353–64.[CrossRef][ISI][Medline]

3 Risau W. Mechanisms of angiogenesis. Nature 1997;386:671–4.[CrossRef][Medline]

4 Ferrara N. Role of vascular endothelial growth factor in the regulation of angiogenesis. Kidney Int 1999;56:794–814.[CrossRef][ISI][Medline]

5 Ferrara N, Alitalo K. Clinical applications of angiogenic growth factors and their inhibitors. Nat Med 1999;5:1359–64.[CrossRef][ISI][Medline]

6 Gutterman JU. Cytokine therapeutics: lessons from interferon-alpha. Proc Natl Acad Sci U S A 1994;91:1198–205.[Abstract/Free Full Text]

7 Star GA, Kerr IM, Williams BR, Silverman RH, Schreiber RD. How cells respond to interferons. Annu Rev Biochem 1998;67:227–64.[CrossRef][ISI][Medline]

8 Albini A, Marchisone C, Del Grosso F, Benelli R, Masiello L, Tacchetti C, et al. Inhibition of angiogenesis and vascular tumor growth by interferon-producing cells: a gene therapy approach. Am J Pathol 2000;156:1381–93.[Abstract/Free Full Text]

9 Sgonc R, Fuerhapter C, Boeck G, Swerlick R, Fritsch P, Sepp N. Induction of apoptosis in human dermal microvascular endothelial cells and infantile hemangiomas by interferon-alpha. Int Arch Allergy Immunol 1998;117:209–14.[ISI][Medline]

10 Singh RK, Gutman M, Bucana CD, Sanchez R, Llansa N, Fidler IJ. Interferons alpha and beta down-regulate the expression of basic fibroblast growth factor in human carcinomas. Proc Natl Acad Sci U S A 1995;92:4562–6.[Abstract/Free Full Text]

11 Oliveira IC, Sciavolino PJ, Lee TH, Vilcek J. Down-regulation of interleukin-8 gene expression in human fibroblasts: unique mechanism of transcriptional inhibition of interferon. Proc Natl Acad Sci U S A 1992;89:9049–53.[Abstract/Free Full Text]

12 Slaton JW, Perrotte P, Inoue K, Dinney CP, Fidler IJ. Interferon-alpha-mediated down-regulation of angiogenesis-related genes and therapy of bladder cancer are dependent on optimization of biological dose and schedule. Clin Cancer Res 1999;5:2726–34.[Abstract/Free Full Text]

13 Jonasch E, Haluska FG. Interferon in oncological practice: review of interferon biology, clinical applications, and toxicities. Oncologist 2001;6:34–55.[Abstract/Free Full Text]

14 Oberg K. Interferon in the management of neuroendocrine GEP-tumors: a review. Digestion 2000;62:92–7.[CrossRef][Medline]

15 Gaudin PB, Rosai J. Florid vascular proliferation associated with neural and neuroendocrine neoplasms. A diagnostic clue and potential pitfall. Am J Surg Pathol 1995;19:642–52.[ISI][Medline]

16 Zhou Y, Gobl A, Wang S, Jacobsen MB, Janson ET, Haines GK 3rd, et al. Expression of p68 protein kinase and its prognostic significance during IFN-alpha therapy in patients with carcinoid tumors. Eur J Cancer 1998;34:2046–52.[CrossRef][ISI][Medline]

17 von Marschall Z, Cramer T, Höcker M, Finkenzeller G, Wiedenmann B, Rosewicz S. Dual mechanism of vascular endothelial growth factor upregulation by hypoxia in human hepatocellular carcinoma. Gut 2001;48:87–96.[Abstract/Free Full Text]

18 Weidner N, Semple JP, Welch WR, Folkman J. Tumor angiogenesis and metastasis–correlation in invasive breast carcinoma. N Engl J Med 1991;324:1–8.[Abstract]

19 Finkenzeller G, Sparacio A, Technau A, Marme D, Siemeister G. 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 1997;15:669–76.[CrossRef][ISI][Medline]

20 Bhattacharya S, Eckner R, Grossman S, Oldread E, Arany Z, D’Andrea A, et al. Cooperation of Stat2 and p300/CBP in signalling induced by interferon-alpha. Nature 1996;383:344–7.[CrossRef][Medline]

21 Gille J, Swerlick RA, Caughman SW. Transforming growth factor-alpha-induced transcriptional activation of the vascular permeability factor (VPF/VEGF) gene requires AP-2-dependent DNA binding and transactivation. EMBO J 1997;16:750–9.[CrossRef][ISI][Medline]

22 Philipsen S, Suske G. A tale of three fingers: the family of mammalian Sp/XKLF transcription factors. Nucleic Acids Res 1999;27:2991–3000.[Abstract/Free Full Text]

23 Brown LF, Berse B, Jackman RW, Tognazzi K, Manseau EJ, Senger DR, et al. Expression of vascular permeability factor (vascular endothelial growth factor) and its receptors in adenocarcinomas of the gastrointestinal tract. Cancer Res 1993;53:4727–35.[Abstract/Free Full Text]

24 Brown LF, Berse B, Jackman RW, Tognazzi K, Manseau EJ, Dvorak HF, et al. Increased expression of vascular permeability factor (vascular endothelial growth factor) and its receptors in kidney and bladder carcinomas. Am J Pathol 1993;143:1255–62.[Abstract]

25 Plate KH, Breier G, Weich HA, Risau W. Vascular endothelial growth factor is a potential tumor angiogenesis factor in human gliomas in vivo. Nature 1992;359:845–8.[CrossRef][Medline]

26 Kim KJ, Li B, Winer J, Armanini M, Gillett N, Phillips HS, et al. Inhibition of vascular endothelial growth factor-induced angiogenesis suppresses tumor growth in vivo. Nature 1993;362:841–4.[CrossRef][Medline]

27 Millauer B, Shawver LK, Plate KH, Risau W, Ullrich A. Glioblastoma growth inhibited in vivo by a dominant-negative Flk-1 mutant. Nature 1994;367:576–9.[CrossRef][Medline]

28 Neufeld G, Cohen T, Gengrinovitch S, Potorak Z. Vascular endothelial growth factor (VEGF) and its receptors. FASEB J 1999;13:9–22.[Abstract/Free Full Text]

29 Levy DE. Analysis of interferon-regulated protein binding the interferon-alpha-stimulated response element. Methods 1998;15:167–74.[CrossRef][ISI][Medline]

30 Black AR, Black JD, Azizkhan-Clifford J. Sp1 and kruppel-like factor family of transcription factors in cell growth regulation and cancer. J Cell Physiol 2001;188:143–60.[CrossRef][ISI][Medline]

31 Majello B, De Luca P, Lania L. Sp3 is a bifunctional transcription regulator with modular independent activation and repression domains. J Biol Chem 1997;272:4021–6.[Abstract/Free Full Text]

32 Stoner M, Wang F, Wormke M, Nguyen T, Samudio I, Vyhlidal C, et al. Inhibition of vascular endothelial growth factor expression in HEC1A endometrial cancer cells through interactions of estrogen receptor alpha and Sp3 proteins. J Biol Chem 2000;275:22769–79.[Abstract/Free Full Text]

33 Jackson SP, MacDonald JJ, Lees-Miller S, Tjian R. GC box binding induces phosphorylation of Sp1 by a DNA-dependent protein kinase. Cell 1990;63:155–65.[CrossRef][ISI][Medline]

34 Jackson SP, Tjian R. O-Glycosylation of eukaryotic transcription factors: implications for mechanisms of transcriptional regulation. Cell 1988;55:125–33.[Medline]

35 Ryu S, Zhou S, Ladurner AG, Tjian R. The transcriptional cofactor complex CRSP is required for activity of the enhancer-binding protein Sp1. Nature 1999;397:446–50.[CrossRef]