From Fibroblast Growth Factors (FGF) and Vascular Endothelial Growth Factors (VEGF) as endothelial cell mitogens, the study of angiogenesis has expanded to include many additional agonists, receptors and inhibitors working in complex and subtle mechanisms (for concise reviews, see references 1-4). The key factors are the five VEGFs, the VEGF receptors, VEGF-R1, -R2 and -R3, and placental growth factors (PlGF). In addition, several newer factors, such as the angiopoietins, ephrins, leptin and chemokines, have been shown to be important in angiogenesis.
Angiogenesis is the formation of new vessels by sprouting of new capillaries from existing vessels, a fundamental phenomenon in diseases such as atherosclerosis, cancer or diabetes and in physiological conditions such as the menstrual cycle and pregnancy. Angiogenesis is closely related to vasculogenesis, the formation of the vascular network from stem cells in the embryo. In each case, the controlling mechanisms are the paracrine regulation of tyrosine kinase receptors, primarily on endothelial cells.1-4
Angiogenic factors are studied in in vivo models (e.g., the appearance of new vessels around an angiogenic factor implanted into an animal cornea) or in roughly analogous in vitro models. In addition, some angiogenic factors are mitogenic for endothelial cells and can be assayed by induction of endothelial cell proliferation. Finally, 'knockout' mice, in which the gene for a single factor has been disrupted, have given crucial information about the roles of angiogenic factors.
The fact that endothelial cell proliferation and organization into tubules are fundamental to either vasculogenesis or angiogenesis has focused attention on endothelial cell-specific ligands and receptors. Two endothelial cell-specific tyrosine kinase receptors were identified, Tie1 and Tie2.5-9 The ligands for Tie2 are Angiopoietin-1 and Angiopoietin-2 (Ang1 and Ang2);10, 11 the ligand(s) for Tie1 has not been identified. The general functions of these receptors have been deduced from studies of 'knockout' mice and over-expressing mice.8, 9, 11-13
The absence of either Tie2 or its agonist Ang1 prevents normal angiogenesis.9, 12 Activation of Tie2 in vitro does not, however, induce mitogenesis in endothelial cells,10 so the mechanism differs from that for VEGF. Ang1 stimulates vessel sprouting in vitro,14 and Ang2 inhibits this effect.11 Similarly, mouse embryos with overexpressed Ang2 have defects similar to those lacking Tie2.114 Thus, it appears that activation of Tie2 by Ang1 is necessary for angiogenesis and that Ang2 is an antagonist.
|Figure 1. A scheme for vessel sprouting (A) and for maturation of the new vessel (B). This figure was developed from models presented in references 1,3,12,13. Vessel structure is maintained by action of Ang1 on Tie2. In A, replacement of Ang1 by Ang2 destabilizes vessel integrity facilitating vessel sprouting in response to VEGF. In B, the new endothelial tubule interacts with surrounding mesenchymal cells in part through Ang1, which acts on endothelial cell Tie2 to promote association of the new tubule with periendothelial cells. The mechanism of this communication must involve other signals, and is postulated to involve growth factors released from endothelial cells in response to activated Tie2.
In an in vivo model154 (corneal implant), neither Ang1 nor Ang2 promoted angiogenesis, but when added with sub-optimal VEGF, Ang1 plus VEGF caused an increased number of new vessels with evidence of a mature vascular system, while Ang2 plus VEGF caused formation of longer, wider and less mature vessels. These data together with those above, suggest that activation of Tie2 by Ang1 is necessary for endothelial-periendothelial cell interactions. Absence of Ang1 or presence of the inhibitor Ang2 destabilizes this cell-cell interaction, facilitating vessel sprouting in response to VEGF (i.e., absence of Ang1 or presence of Ang2 is angiogenic), but absence of Ang1 (or presence of Ang2) also blocks the endothelial-periendothelial cell interactions necessary for formation of a mature vessel (Fig. 1).
The absence of a known agonist for Tie1 complicates analysis of its effects, but from studies of Tie1 knockout mice it appears that Tie1 is required for the development or maintenance of structural integrity of endothelial cells so that they can resist hemodynamic pressure.8,9
Neuropilin is a cell-surface protein that is a receptor for the collapsin/semaphorin family of neuronal guidance proteins.16,17 Over-expression of neuropilin in mice resulted in an excess of capillaries and blood vessels.18 Neuropilin is expressed by endothelial and tumor cells, where it is a receptor specific for VEGF165,19 the longer of two alternatively spliced VEGF family members (the other is VEGF121), and for placental growth factor.20 The binding of VEGF165 to neuropilin on cells that also express VEGF-R2 enhances the binding to VEGF-R2 and the bioactivity of VEGF165, but cells expressing neuropilin alone showed no bioactivity.19 These results suggest surprising roles for neuropilin in angiogenesis.
Equally surprising is the observation that from the earliest stages of angiogenesis, endothelial cells destined to become arteries express ephrin-B2, while the cognate receptor, eph B4, is expressed on endothelial cells destined to become veins.21 The Ephrin/Eph family of cell-surface proteins is important in the cell-cell recognition and signaling of nervous system patterning.22 Their specific location on venous vs. arterial endothelial cells suggests that the formation of a vascular system may be appreciably more complicated than predicted.
More surprising yet was the report that endothelial cells express OB-R beta, the leptin receptor.23 In addition, leptin, the anti-obesity hormone, is an endothelial cell mitogen and chemoattractant, and it induces angiogenesis in a cornea implant model.23
Finally, in addition to these factors that act early in angiogenesis, Tachibana et al.24 demonstrated that a chemokine receptor, CXCR-4, and its ligand, PBSF/SDF-1, are required for later stages of development of the vascularization of the gastrointestinal tract.
- Folkman J. and P.A. D'Amore (1996) Cell 87:1153.
- Risau, W. (1997) Nature 386:671.
- Hanahan, D. (1997) Science 277:48.
- Peters, K.G. (1998) Circ. Res. 83:342.
- Dumont, D.J. et al. (1992) Oncogene 7:1471.
- Maisonpierre, P.C. (1993) Oncogene 8:1631.
- Mustonen, T. and K. Alitalo (1995) J. Cell Biol. 129:895.
- Puri, M.C. et al. (1995) EMBO J. 14:5884.
- Sato, T.N. et al. (1995) Nature 376:70.
- Davis, S. et al. (1996) Cell 87:1161.
- Maisonpierre, P.C. (1997) Science 277:55.
- Suri, C. et al. (1996) Cell 87:1171.
- Vikkula, M. et al. (1996) Cell 87:1181.
- Koblizek, T.I. et al. (1998) Current Biol. 8:529.
- Asahara, T. et al. (1998) Circ. Res. 83:233.
- He, Z. et al. (1997) Cell 90:739.
- Kolodkin, A.L. et al. (1997) Cell 90:753.
- Kitsukawa, T. et al. (1995) Development 121:4309.
- Soker, S. et al. (1998) Cell 92:735.
- Migdal, M. et al. (1998) J. Biol. Chem. 35:22272.
- Wang, H.U. et al. (1998) Cell 93:741.
- Drescher, U. (1997) Curr. Biol. 7:799.
- Sierra-Honigmann, M.R. et al. (1998) Science 281:1683.
- Tachibana, K. et al. (1998) 393:591.