| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
|
FJ
EXPRESS SUMMARY ARTICLE The Full-length version of this article is also available, published online September 19, 2002 as doi:10.1096/fj.02-0109fje. |
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Laboratory of Tumor and Development Biology, and
* Laboratory of Connective Tissues Biology, University of Liège, Sart-Tilman, 4000 Liège, Belgium; and
Department of Oncology, Aventis, Hayward, California, USA
3Correspondence: Laboratory of Tumor and Development Biology, Institute of Pathology B23, University of Liège, Sart-Tilman, 4000 Liège, Belgium. E-mail:agnes.noel{at}ulg.ac.be
SPECIFIC AIMS
Although endostatin and angiostatin are known as tumor-derived angiogenesis inhibitors with potent antitumoral activity, their mechanisms of action are not yet completely understood. To provide further mechanistic insights into endostatin and angiostatin actions, we investigated whether these antiangiogenic factors delivered by adenoviral vectors could regulate vascular endothelial growth factor (VEGF) expression in three different models: the in vitro mouse aortic ring assay, transplantation of malignant keratinocytes in vivo, and the development of highly aggressive and vascularized murine mammary tumors.
PRINCIPAL FINDINGS
1. Endostatin and angiostatin inhibit neovessel formation in the mouse aortic ring assay
To assess the functional activity of adenovirus-expressed endostatin and angiostatin gene products, segments of mouse thoracic aortas exposed or not for 24 h to 4 x 109 pfu recombinant viruses (Ad.END or Ad.ANG) were cultured in collagen gels and microvessel outgrowth was examined. After 7 days of culture, quantification of neovessel formation performed by image analysis demonstrated a significant inhibition of microvessel outgrowth with Ad.ANG and, to a larger extent, with Ad.END (40 and 85% inhibition in the presence of Ad.ANG (P<0.05) and Ad.END (P<0.001), respectively); Ad.LacZ control adenovirus did not interfere with this process. The expression and secretion of angiostatin or endostatin from Ad.ANG and Ad.END-infected aortic rings were confirmed by Western blot analysis.
2. Endostatin and angiostatin inhibit malignant keratinocyte cell invasion and tumor vascularization in vivo
Malignant murine keratinocytes (PDVA cells) transduced in vitro with the different adenoviruses at an MOI of 10 pfu/cell during 24 h were cultured on a collagen gel, then implanted in vivo onto the dorsal muscle fascia of syngeneic mice. When Ad.LacZ-infected PDVA cells were transplanted in vivo, their invasive behavior was similar to that observed previously with parental PDVA cells (Fig. 1
a). As early as 7 days after implantation, host-derived endothelial and stromal cells migrated upward into the collagen gel. Thereafter, malignant cell invasion proceeded downward. After 2 wk, malignant keratinocytes had penetrated deeply into the host mesenchyme and intermingled with host stromal cells. When Ad.END- or Ad.ANG-infected PDVA cells were implanted into mice, they initially proliferated to form a multilayered surface epithelium, but both infected cells failed to invade the host tissue. Tumor cells remained separated from the host tissue by the collagen gel (Fig. 1b, c
). In transplants derived from parental or Ad.LacZ-infected cells, the invading tumor islets were surrounded by a rich capillary network (Fig. 1a
). In sharp contrast, after transplantation of Ad.END- and Ad.ANG-infected cells, host capillaries failed to infiltrate the remodeled matrix and remained below the collagen gel (Fig. 1b, c
).
|
To assess the putative antiangiogenic and anti-invasive actions of in vivo delivery of endostatin and angiostatin, 1 day after in vivo implantation of malignant parental PDVA cells, mice received a single (1x108 pfu) intravenous (i.v.) administration of adenoviruses. Four days after adenovirus injection, endostatin or angiostatin were detected by Western blot in the sera of animals treated with Ad.END or Ad.ANG but not in those treated with control Ad.LacZ. Injection of control viruses did not influence neither the angiogenic nor invasive phenotype of malignant cells (Fig. 1d
). By contrast, tumor invasion and angiogenesis were both inhibited in mice injected with either Ad.END or Ad.ANG (Fig. 1e, f
).
3. Endostatin and angiostatin inhibit in vivo mammary tumor growth
Mouse mammary EF43.fgf-4 cells are highly tumorigenic when injected subcutaneously into syngeneic BALB/c mice and overproduce VEGF. A single i.v. administration of Ad.END or Ad.ANG (1x108 pfu) to mice bearing a 20 mm3 pre-established EF43.fgf-4 tumor resulted in a 50% (P<0.01) and 90% (P<0.001) reduction of tumor volume, respectively. In contrast, a similar injection of control adenovirus (Ad.LacZ) did not affect tumor growth (P>0.05). This antitumoral action of Ad.ANG and Ad.END is associated with an important reduction of tumor vascularization compared with that of control tumors.
4. Endostatin and angiostatin down-regulate VEGF expression
RT-PCR analysis of VEGF expression was carried out on mouse aortic rings infected with adenoviruses. Expression of endostatin in aortic explants was shown to be associated with 3- to 10-fold down-regulation of VEGF188, VEGF164, and VEGF120 mRNA isoforms. No significant modulation of VEGF expression was observed in this model with Ad.ANG.
VEGF expression was then evaluated in PDVA cells infected in vitro with adenoviruses. Endostatin expression was associated with a three- to fivefold down-regulation of VEGF188, VEGF164, and VEGF120 mRNA isoforms (Fig. 2
a). This modulation of VEGF expression by endostatin was confirmed by Western blot analysis (Fig. 2b
). Though expression of angiostatin in PDVA cells had less effect on VEGF mRNA isoforms, Western blot analysis revealed a down-regulation of VEGF at the protein level similar to that observed after treatment with Ad.END (Fig. 2b
). The modulation of VEGF production at protein level was further demonstrated by immunohistochemical analysis of PDVA cells transplanted into mice treated with Ad.END and Ad.ANG.
|
The ability of endostatin and angiostatin to down-regulate VEGF expression was also observed in vivo in EF43.fgf-4 mammary carcinoma by RT-PCR and Western blot analysis. Down-regulation of VEGF mRNA isoforms induced by i.v. administration of Ad.ANG was more pronounced than that mediated by Ad.END.
CONCLUSIONS
VEGF is a strong endothelial cell-specific mitogen that plays an important role in tumor angiogenesis. No evidence of a direct link between endostatin and angiostatin actions and the regulation of VEGF had so far been clearly established. Moreover, controversial data regarding the putative effect of angiostatin on VEGF expression have been reported, depending on the tumor cell type studied. In the present study, we report the antiangiogenic, antitumor and anti-invasive activities of Ad.END and Ad.ANG in three different in vitro and in vivo models and show that the extent of the antiangiogenic activity of endostatin and angiostatin was related to the level of VEGF down-regulation at mRNA and protein levels.
In the mouse aortic ring assay, although angiostatin and endostatin both inhibited endothelial cell migration, only endostatin exerts its antiangiogenic action through a down-regulation of VEGF expression. A down-regulation of VEGF expression at mRNA and protein levels was also observed in the two in vivo experimental cancer models after treatment with each angiogenesis inhibitor. In the EF43.fgf-4 tumor model, the more pronounced antitumoral effect of Ad.ANG vs. Ad.END was associated with a more important reduction of vessel density and a higher down-regulation of VEGF expression. This suggests that reduction of tumor growth is strongly correlated with the degree of VEGF modulation. In the malignant keratinocyte PDVA transplants, local production of endostatin and angiostatin by transplanted, infected PDVA and systemic administration of Ad.END or Ad.ANG resulted in a similar reduction of angiogenesis. In this system, the down-regulation of VEGF expression by endostatin and angiostatin in tumor cells was demonstrated in vitro by RT-PCR and Western blot analysis and in vivo by immunohistochemistry. These results illustrate a direct effect of endostatin and angiostatin on tumor cells themselves in addition to a possible effect on endothelial cells as assessed in the aortic ring model.
Our study provides evidence that angiostatin and endostatin down-regulate the production of VEGF by tumor cells (Fig. 3
). Whatever the mechanisms of endostatin and angiostatin action, our data clearly demonstrate that the activity of angiostatic factors could be extended from an exclusive action on endothelial cells, as initially believed, to action on tumor cells. This effect is related to a transcriptional regulation of VEGF expression and is in accordance with the observation that endostatin treatment of endothelial cells leads to transcriptional control of many genes, including cell cycle-related genes and genes regulating apoptosis. It is known that endostatin can bind cell surface molecules such as 
5- and v-integrins or heparan sulfate proteoglycan (glypican-1 and 4) and that endostatin induces signal transduction in endothelial cells via activation of a tyrosine kinase. Altogether, these observations suggest that endostatin can deliver intracellular signals to antagonize proangiogenic activity and probably to down-regulate VEGF expression.
|
In summary, endostatin and angiostatin may have multiple effects leading to antiangiogenic action. They can directly affect endothelial cell functions by inhibiting their proliferation and migration into the collagen gel. This effect of endostatin and angiostatin on host cells is probably specific to endothelial cells. Indeed, stromal cells other than endothelial cells were frequently observed in the collagen gel (transplantation and aortic ring assays). On the other hand, they can control VEGF production by endothelial cells as assessed in the aortic rings or by tumor cells themselves, as demonstrated in vitro and in vivo. Our findings strongly suggest that these actions of endostatin or angiostatin exerted on tumor cells contribute at least in part to their antiangiogenic and anti-invasive potencies.
FOOTNOTES
1 To read the full text of this article, go to http://www.fasebj.org/cgi/doi/10.1096/fj.02-0109fje; to cite this article, use FASEB J. (September 19, 2002) 10.1096/fj.02-0109fje 
2 The two first authors contributed equally to this work.
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
