insight progress Vascular-specific growth factors and blood vessel formation George D. Yancopoulos, Samuel Davis, Nicholas W. Gale, John S. Rudge, Stanley J. Wiegand &amp; Jocelyn Holash Regeneron Pharmaceuticals, Inc., 777 Old Saw Mill River Road, Tarrytown, New York 10591, USA A recent explosion in newly discovered vascular growth factors has coincided with exploitation of powerful new genetic approaches for studying vascular development. An emerging rule is that all of these factors must be used in perfect harmony to form functional vessels. These new findings also demand re-evaluation of therapeutic efforts aimed at regulating blood vessel growth in ischaemia, cancer and other pathological settings. ntil recently, <a href="/keyword/vascular-endothelial/" >vascular endothelial</a> growth factor (VEGF) was the only growth factor proven to be specific and critical for blood vessel formation1 3. Other long-known factors, such as the fibroblast growth factors (FGFs), had profound effects in various endothelial cell assays4. But such factors were also known to be nonspecific in that they could act on many other cell types, and it was questionable whether the assays used to evaluate them were physiologically relevant. For example, the most widely used assays involved adding putative angiogenic agents to cornea pocket models, or to chick chorioallantoic membranes5,6. In such assays, FGFs could robustly induce new vessel growth, but there was limited ability to evaluate the induced vessels functionally, or to determine the relevance of these inductions for normal vascular development. A recent explosion of newly discovered growth factors acting on the vascular endothelium has coincided with application of powerful new genetic approaches to the problem of vascular development7,8. The vascular endothelium-specific growth factors now include five members of the VEGF family, four members of the angiopoietin family, and at least one member of the large ephrin family (Fig. 1). For almost all of these and their receptors, mouse models involving genetic disruption and/or transgenic misexpression have contributed to an understanding of their normal physiological roles, as well as of their pathological capabilities. A rule that is emerging is that all of these factors must be used in perfect harmony, in a complementary and coordinated manner, to form functional vessels7. In addition, many other growth factors that are not vascular endothelium-specific are also required for blood vessel formation, such as members of the platelet-derived growth factor or transforming growth factor- families, although these factors also have critical roles for many other systems as well8 10. Furthermore, there are myriad other gene products -- ranging from transcription factors to members of the Notch family -- that have been shown crucial for vessel formation8. In an attempt to do justice to the topic, this review will focus only on the vascular endothelium-specific growth factors, and how they are involved in vessel formation. The recent explosion in identifying and characterizing physiological regulators of blood vessel growth demands reevaluation of therapeutic efforts aimed at regulating blood vessel growth -- whether it be promoting vascular ingrowth to replenish ischaemic tissue, blocking vessel growth in order to blunt tumours, or repairing damaged and leaky 242 U vessels during inflammation or other pathological settings. The privilege of hindsight makes some of the bold, early therapeutic efforts directed towards ischaemic disease, based on random delivery of a single growth factor to grow an entirely new functional network of vessels, now appear somewhat naive and even misguided. On the other hand, recent insights continue to support the notion that blockade of even a single growth factor might limit diseaseinduced vascular growth, with the most compelling evidence supporting approaches based on blockade of VEGF. Furthermore, recent advances indicate previously unanticipated clinical applications for vascular growth factors, such as the use of angiopoietin-1 (Ang1) for the repair of damaged and leaky vessels. Vasculogenesis and angiogenic remodelling Vessel formation can occur by a number of different processes4. Early in development, vessel formation occurs by a process referred to as vasculogenesis (Fig. 2, stage A), in which endothelial cells differentiate and proliferate in situ within a previously avascular tissue, and then coalesce to form a primitive tubular network. This primary network includes some of the major vessels in the embryo, such as the aorta and major veins, as well as a honeycomb-like plexus connecting these major vessels. Angiogenic remodelling refers to the process by which this initial network is modified -- through both pruning and vessel enlargement -- to form the interconnecting branching patterns characteristic of the mature vasculature (Fig. 2, stage B). During this time, vessel walls also mature, as endothelial cells integrate tightly with supporting cells (such as smooth muscle cells and pericytes) and surrounding matrix (Fig. 2, stage C). A different process, referred to as angiogenic sprouting, involves the sprouting from existing vessels into a previously avascular tissue. In some cases, it seems as if mature vessels must first be destabilized to allow for subsequent sprouting (Fig. 2, stages D, F); once again, vessels formed by sprouting are initially immature and must further develop. Angiogenic sprouting is responsible for vascularizing certain structures during normal development, such as the neural tube or the retina, and for most new vessel formation in the adult. Destabilization of vessels can also apparently lead to vascular regression (Fig. 2, stage E), as described below. Emerging model of vascular formation Recent insights have led to a model of vascular formation that attempts to incorporate the known vascular-specific growth NATURE | VOL 407 | 14 SEPTEMBER 2000 | www.nature.com 2000 Macmillan Magazines Ltd insight progress Figure 1 Schematic representation of three families of vascular growth factors and their receptor interactions. a, VEGFs; b, angiopoietins; c, ephrins. The four factors that are discussed in detail in this review are highlighted in red. In b, `+' or ` ' indicates whether the particular angiopoietin activates or blocks the Tie2 receptor, whereas `?' indicates that a potential interaction has not yet been confirmed experimentally. In c, only those members of the large ephrin ligand family (and only their counterpart Eph receptors) that have been implicated in vascular growth are shown. a PlGF VEGF-A VEGF-B VEGF-C VEGF-D b Ang1 Ang2 Ang3 Ang4 c Ephrin-B1 Ephrin-B2 Ephrin-A1 ? ? VEGFR-1 VEGFR-2 VEGFR-3 (Flt-1) (KDR/Flk-1) (Flt-4) Tie1 Tie2 EphB2 EphB3 EphB4 EphA2 factors7,11 14, and the details of this model will be a major subject of this review. According to this model, the first characterized vascularspecific growth factor, VEGF, maintains its position as the most critical driver of vascular formation, as it is required to initiate the formation of immature vessels by vasculogenesis or angiogenic sprouting (Fig. 2, stages A, F), during development as well as in the adult. Ang1 and ephrin-B2 are subsequently required for further remodelling and maturation of this initially immature vasculature (Fig. 2, stages B, C), with ephrin-B2 being particularly important in distinguishing developing arterial and venous vessels, as will be discussed in more detail below. Following vessel maturation, Ang1 seems to continue to be important in maintaining the quiescence and stability of the mature vasculature (Fig. 2, stage C). Disruption of this stabilizing signal coincides with reinitiation of vascular remodelling in the adult -- as occurs in the adult female reproductive system or in tumours (Fig. 2, stage D, and see below). Such de-stabilization seems to involve the autocrine induction -- by the endothelium to be remodelled -- of a natural antagonist of Ang1, termed Ang2 (Fig. 2, stage D). VEGFs, angiopoietins and ephrinB2 apparently recapitulate their developmental roles during vascular remodelling in the adult, and administration of individual factors to the adult allows them to reprise these roles but not to trigger the entire process (see below). Thus VEGF administration can initiate vessel formation in adult animals, but by itself promotes formation of only leaky, immature and unstable vessels. In contrast, Ang1 administration seemingly further stabilizes and protects the adult vasculature, making it resistant to the damage and leak induced by VEGF or inflammatory challenges. Altogether, it is becoming clear that precise understanding of the normal developmental roles of the VEGFs, the angiopoietins and the ephrins will greatly aid in understanding how to manipulate these growth factor systems for therapeutic benefit. VEGF, its relatives, and their receptors VEGF was initially defined, characterized and purified for its ability to induce vascular leak and permeability, as well as for its ability to promote <a href="/keyword/vascular-endothelial/" >vascular endothelial</a> cell proliferation1,2. Thus, it was originally termed vascular permeability factor as well as VEGF. Although most research efforts have focused on its growth-promoting ability, recent findings are once again highlighting its potent permeability-inducing effects, and in particular their role in disease. Other members of the VEGF family were identified based on their homology to VEGF3. The various members of the VEGF family have NATURE | VOL 407 | 14 SEPTEMBER 2000 | www.nature.com overlapping abilities to interact with a set of cell-surface receptors3 that trigger responses to these factors (Fig. 1a). The main receptors that seem to be involved in initiating signal transduction cascades in response to the VEGFs comprise a family of closely related receptor tyrosine kinases consisting of three members now termed VEGFR-1 (previously known as Flt-1), VEGFR-2 (previously known as KDR or Flk-1) and VEGFR-3 (previously known as Flt-3). In addition, there are a number of accessory receptors such as the neuropilins15 which seem to be involved primarily in modulating binding to the main receptors, although roles in signalling have not been ruled out. VEGFR-2 seems to mediate the major growth and permeability actions of VEGF, whereas VEGFR-1 may have a negative role, either by acting as a decoy receptor or by suppressing signalling through VEGFR2. Thus, mice engineered to lack VEGFR-2 fail to develop a vasculature and have very few endothelial cells16, whereas mice lacking VEGFR-1 seem to have excess formation of endothelial cells which abnormally coalesce into disorganized tubules17. Mice engineered to express only a truncated form of VEGFR-1, lacking its kinase domain, appear rather normal, consistent with the notion that the primary role of VEGFR-1 may be that of a decoy receptor18. VEGFR-3 may be important during blood vessel development, but is most unique based on its expression on lymphatic vessels, for which it seems to be critical19. The first VEGF relative identified is known as placental growth factor (PlGF), and until recently little was known about its normal function, in part because mice engineered to lack PlGF were overtly normal8,20. Recent findings indicate that adult mice lacking PlGF exhibit deficiencies in certain models of adult vascular remodelling, raising the interesting possibility that the activity of PlGF may be limited to these settings8. VEGF-C, based on its ability to bind the lymphatic-specific VEGFR-3, seems to be important for lymphatic development, and transgenic overexpression of VEGF-C leads to lymphatic hyperplasia21. Mice lacking VEGF-B are overtly normal and fertile, but their hearts are reduced in size, suggesting that VEGF-B may have a role in coronary vascularization and growth22. Little is known about the normal physiological role of VEGF-D3. VEGF must be well regulated Compared to its more recently discovered relatives, much more is known about VEGF. It is now quite clear that VEGF is such a potent and critical vascular regulator that its dosage must be exquisitely regulated in spatial, temporal and quantitative manner to avoid vascular disaster. Disruption of both VEGF alleles in mice mimicks 243 2000 Macmillan Magazines Ltd insight progress Figure 2 Schematic representation of the roles of VEGF, Ang1, Ang2 and ephrin-B2 during vessel formation. The processes include vasculogenesis (stage A), angiogenic remodelling (B), stabilization and maturation (C), destabilization (D), regression (E) and sprouting (F), as described in detail in the text. An attempt is made to assign the indicated vascular growth factors to the various processes, and to indicate their expression patterns. Although not noted in the figure, expression of ephrin-B2 marks arterial vessels from the earliest developmental times. Primary vasculature Remodelled vasculature (A) V asculogenesis VEGF (B) Angiogenic remodelling VEGF Ang1 &amp; EphrinB2 Artery Vein (C) Maturation, Stabilization &amp; Quiescence Ang1 Artery Vein Ang1 EphrinB2 Immature unstable vessels VEGFR1/2 Tie1/2 EphrinB2 Stable mature vessels (E) Regression N VEGo F (D) Destabilization Ang2 Unstable vessel Stable vessel E +V GF (F) Angiogenic sprouting knockout of VEGFR-2, resulting in almost complete absence of a vasculature23,24. Disruption of even a single VEGF allele in mice leads to embryonic lethality due to severe vascular abnormalities, providing perhaps the only example of embryonic lethality due to a simple halfdosage effect23,24. Even more subtle alterations in VEGF expression during embryonic development result in profound abnormalities, leading to embryonic or early post-natal death25,26. VEGF continues to be critical during early post-natal growth and development, as evidenced by post-natal VEGF inactivation using Cre-loxP-mediated VEGF gene deletion, or by administration of a soluble VEGF receptor that effectively blocks VEGF action27. Although VEGF inactivation is lethal during the first few post-natal weeks, VEGF inactivation in older animals is much less traumatic, seemingly affecting only those structures that continue to undergo vascular remodelling, such as bone growth plates or ovarian corpus lutei27 29. Thus, VEGF does not seem to have a continuous maintenance function for much of the adult vasculature. The most elegant demonstration of the need for exquisite VEGF regulation involves retinal vascularization, which occurs post-natally in rodents. Angiogenic sprouting into the initially avascular and hypoxic rodent retina depends upon its VEGF expression30 32. Any perturbation of normal VEGF expression patterns destroys retinal vascularization patterns, with dire results for retinal function; subsequent restoration of VEGF expression does not correct the problem, but rather exacerbates it. A simple way to perturb VEGF expression involves exposing post-natal rodents to a brief period of hyperoxia31,33,34, which transiently suppresses retinal VEGF, resulting in cessation of vessel growth and even causing vascular regression31,33,34. When the rodents are returned to normoxia, the now undervascularized retina becomes hypoxic, causing an abnormal burst of VEGF, which promotes robust new angiogenesis, but of haemorrhagic and 244 leaky vessels growing in totally abnormal patterns that wreak havoc upon the retina. This model reflects the ability of oxygen therapy in premature infants to cause retinopathy of prematurity, and shows the need for precise regulation of VEGF. Similarly, diabetic retinopathy initiates with damage and loss of healthy vessels, followed by retinal hypoxia and resulting VEGF induction, once again leading to an abnormal angiogenic response with leaky and haemorrhagic vessels35,36. These findings show that inappropriate induction of VEGF, in the absence of the entire angiogenic programme, leads to formation of immature and leaky vessels that cause disease. These findings also show that tissue hypoxia cannot necessarily induce a useful angiogenic response. Consistent with the above findings concerning the devastating consequences of unregulated VEGF expression, several studies have delivered excess VEGF to adult tissues -- to adult muscle using retrovirally engineered myoblasts37, to skin using transgenic or adenoviral delivery38 41, or to whole animals using acute adenoviral delivery42 -- and found that leaky and haemorrhagic vessels were formed, often associated with an inflammatory response, resulting in pronounced tissue swelling and oedema. The angiopoietins and their Tie receptors Despite its requisite role in vascular formation, VEGF must work in concert with other factors. The angiopoietins (Fig. 1b) seem to be some of VEGF's most important partners (Fig. 2). The angiopoietins were discovered as ligands for the Ties, a family of receptor tyrosine kinases that are as selectively expressed within the vascular endothelium (despite expression in some other cells, such as in the haemopoietic lineage) as are the VEGF receptors43 47. There are now four definitive members of the angiopoietin family, although Ang3 and Ang4 may represent widely diverged counterparts of the same gene locus in mouse NATURE | VOL 407 | 14 SEPTEMBER 2000 | www.nature.com 2000 Macmillan Magazines Ltd insight progress and man12,48,49. All of the known angiopoietins bind primarily to Tie2, and it is unclear whether there are independent ligands for the second Tie receptor, Tie1, or -- as currently seems more likely -- whether the known angiopoietins can in some way or under some circumstances also engage Tie1, perhaps as a second component in a heteromerized complex. The rest of this review will deal only with Ang1 and Ang2, since little more can be said at this time about Ang3 and Ang4. vivo, at least under some circumstances12. This possibility became even more intriguing when Ang2 expression profiles were examined. In adult animals, Ang2 was induced in the endothelium of vessels undergoing active remodelling, such as sprouting or regressing vessels in the ovary12,53, or in tumours13,14,54,55 (as will be discussed in detail below). These findings, together with the possibility that Ang2 could act as a Tie2 antagonist, led to the hypothesis that Ang2 might provide a key de-stabilizing signal involved in initiating angiogenic remodelling12 14,55. That is, based on previous evidence that Ang1 engagement of the Tie2 receptor was constitutive in the adult vasculature and indeed necessary to maintain its quiescence (Fig. 2, stage C), it was proposed that autocrine induction of Ang2 in endothelium blocked this constitutive stabilizing influence of paracrine Ang1, allowing the endothelial cells to revert to a more plastic and destabilized state reminiscent of developing vessels (Fig. 2, stage D). Such destabilized vessels could then be prone to two fates. On the one hand, these destabilized vessels would be prone to regression in the absence of associated growth factors, as also occurs with primitive vessels during development (Fig. 2, stage E). On the other hand, they would be more sensitive to angiogenic changes induced by simultaneously available angiogenic factors such as VEGF, essentially recapitulating an early embryonic situation in which VEGF acts prior to the involvement of Ang1 (Fig. 2, stage F). This model of Ang2 as a destabilizing signal that reverts vessels to a more plastic and tenuous state, initially developed based on observations in the remodelling ovary12, is consistent with more recent data in tumours (see below) as well as emerging data from knockout mice lacking Ang2. One of the best characterized settings of post-natal vascular regression and remodelling in mice involves the eye, in which regression of the hyaloid vasculature encasing the lens is coupled to angiogenic sprouting that leads to vascularization of the initially avascular retina, as described above. Neither regression of the hyaloid vasculature nor vascularization of the retina occur in mice lacking Ang2 (S.J.W., R. Tzekova, Q. Wong, N.W.G., C. Suri &amp; G.D.Y., unpublished results). These data show that Ang2 is required for some post-natal vascular remodelling events, and support the notion that Ang2 provides a key role in destabilizing the vasculature in a manner that is necessary for its subsequent remodelling. However, other defects in the Ang2-knockout mice suggest that it may in some cases also have an agonistic role. That is, it is highly expressed in the developing aortic wall, which does not develop properly in mice lacking Ang2. Similarly, lymphatic development is perturbed in these mice. Ang1 stabilizes vessel walls The most important insights into the normal roles of Ang1 and its Tie2 receptor came from the analysis of mice engineered to lack these gene products11,50,51. Unlike mouse embryos lacking VEGF or VEGFR-2, embryos lacking Ang1 or Tie2 develop a rather normal primary vasculature. However, this vasculature fails to undergo normal further remodelling. The most prominent defects are in the heart, with problems in the associations between the endocardium and underlying myocardium as well as in trabeculae formation, and also in the remodelling of many vascular beds into large and small vessels. In these vascular beds, as in the heart, ultrastructural analysis indicates that endothelial cells fail to associate appropriately with underlying support cells, which are the cells that provide the Ang1 protein that acts on endothelial Tie2 receptors11. This finding led to the suggestion that Ang1 does not supply an instructive signal that actually directs specific vascular remodelling events, but rather has more of a permissive role by optimizing the manner in which endothelial cells integrate with supporting cells, thus allowing them to receive other critical signals from their environment11. Transgenic overexpression of Ang1 in skin results in pronounced hypervacularization40,52. Although there are modest increases in vessel number, the most marked increase is in vessel size. In contrast, VEGF in similar models primarily increases vessel number38 40. These findings indicate that Ang1 might promote circumferential as opposed to sproutive growth. Combining transgenic Ang1 and VEGF leads to unprecedented hypervascularity resulting from increases in both vessel size and number40. The vascular patterns induced by the combination are still obviously abnormal morphologically, suggesting that much must be learned about exploiting even this growth factor combination in therapeutic settings so as to grow normal vessels. In addition to their effects on vascular morphology, transgenic overexpression of Ang1 and VEGF had distinct effects on vascular function and integrity. As had been expected, VEGF led to immature, leaky and haemorrhagic vessels38 40. On the other hand, Ang1 led to vessels that were actually resistant to leak, whether the leak was induced by VEGF or inflammatory agents40. This resistance seems related to the ability of Ang1 to maximize interactions between endothelial cells and their surrounding support cells and matrix, as the Ang1 vessels were resistant to treatments that normally created holes in the endothelial cell barrier40. These findings indicated that Ang1 might counter the effect of VEGF on permeability, raising multiple therapeutic possibilities40. There are numerous disease processes -- ranging from diabetic retinopathy to inflammation to brain oedema following ischaemic stroke -- in which vessels become damaged and leaky, and an agent that could repair the damage and prevent the leak could have enormous therapeutic benefit. Supporting the clinical potential of Ang1, acute adenoviral administration of Ang1 to adult animals showed that Ang1 can indeed protect the adult vasculature from vascular leak, without inducing immediate changes in vascular morphology42. The ephrins The Eph receptor tyrosine kinases comprise the largest known family of growth factor receptors (Fig. 1c), and use the similarly numerous ephrins as their ligands7,56. The ephrins are unlike ligands for other receptor tyrosine kinases in that they must be tethered to the membrane to activate their Eph receptors7,57. Although initially characterized in the nervous system7,56, recent knockout studies have suggested key roles for ephrin-B2 and its EphB4 receptor during vascular development58 60. Mouse embryos lacking ephrin-B2 and EphB4 suffer fatal defects in early angiogenic remodelling that are somewhat reminiscent of those seen in mice lacking Ang1 or Tie258 60. Moreover, ephrin-B2 and EphB4 display remarkably reciprocal distribution patterns during vascular development, with ephrin-B2 marking the endothelium of primordial arterial vessels while EphB4 marks the endothelium of primordial venous vessels58 60. These distributions suggested that ephrin-B2 and EphB4 are involved in establishing arterial versus venous identity, perhaps in fusing arterial and venous vessels at their junctions, and that defects in these processes might account for the early lethality observed in mouse embryos lacking these proteins58 60 (Fig. 2, stage A). Ephrin-B2 continues to selectively mark arteries during later embryonic development as well as in the adult, although this expression extends progressively from the arterial endothelium to the 245 Ang2: agonist and antagonist? Ang2 was cloned based on its homology to Ang1, and displayed similarly high affinity for Tie2, but -- depending on the cell examined -- Ang2 could either activate or antagonize Tie2 (ref. 12). Transgenic overexpression of Ang2 in the embryonic endothelium resulted in embryonic death due to defects resembling those of Ang1 or Tie2 knockouts, demonstrating that Ang2 could act as a Tie2 antagonist in NATURE | VOL 407 | 14 SEPTEMBER 2000 | www.nature.com 2000 Macmillan Magazines Ltd insight progress a Artery Vein Artery Vein Artery Vein Avascular tumour growth Tumour Hypoxic tumour: VEGF Angiogenic sprouting into tumour (VEGF&amp;Ang2) Ang2 VEGF Ang2 EphrinB2 b VEGF Vessel regression due to Ang2 leads to tumour regression &amp; VEGF Ang2 Ang2 EphrinB2 Angiogenic sprouting (VEGF &amp; Ang2) VEGF Figure 3 Models of tumour angiogenesis. a, Model of avascular tumour initiation contrasted with b, tumour initiation involving host vessel co-option. An attempt is made to assign the indicated vascular growth factors to roles in the various indicated steps in tumour development, and to indicate their expression patterns. surrounding arterial smooth muscle and to pericytes (N.W.G. and G.D.Y., unpublished results; D. Shin and D. J. Anderson, unpublished results). Thus, ephrin-B2 is apparently not only required during the earliest stages of arterial/venous determination, but may continue to be important during the development of arteries, perhaps by regulating interactions between endothelial and smooth muscle cells involved in the formation of arterial muscular walls (Fig. 2, stage B). In adult settings of angiogenesis, as in tumours or in the female reproductive system, the endothelium of new vessels strongly re-expresses ephrin-B2 (N.W.G. and G.D.Y., unpublished results; D. Shin and D. J. Anderson, unpublished results) (Fig. 3a,b). The finding that angiogenic sprouting in the adult and in tumours involves re-expression of the ephrin-B2 arterial marker challenges existing dogma that such sprouting primarily involves venous or uncommitted vessels, and also suggests that ephrin-B2 may be important in these angiogenic settings. VEGF and Ang2 in tumour angiogenesis Much has been made of the notion that tumours and metastases initiate as small avascular masses, which only subsequently induce the angiogenic ingrowth that is required to allow further growth of the early tumour61 63 (Fig. 3a). It is clear that many natural tumours initially arise in this manner, particularly primary epithelial tumours that are initially separated from underlying vessels by a basement membrane that must be broken before tumour cells can access the vasculature. In addition, many artificial model systems forcibly create initially avascular tumours by placing tumour cells in a space that is normally devoid of vessels -- such as the subcutaneous space, the cornea pocket or the vitreous or the tumour window -- thus requiring angiogenesis to get vessels to the tumour. Despite all the attention directed towards avascular tumour growth, recent findings14,55 have refocused attention on previous observations64 66 that many tumours, and metastases in particular, do not initiate in an avascular manner (Fig. 3b). Rather, tumour cells can initially home in on and grow by co-opting existing host vessels, 246 and thus start off as well-vascularized small tumours13,14 (Fig. 3b, left). In response to co-option, the host vessels mount a defence -- sensing inappropriate co-option, they regress, choking off the tumour and resulting in a secondarily avascular and hypoxic tumour (Fig. 3b, middle). However, successful tumours seem to overcome host vessel regression by inducing robust new angiogenesis (Fig. 3b, right). Ang2 and VEGF inductions correlate remarkably well with the above processes13,14,55. That is, soon after tumour co-option, host vessels start expressing high autocrine levels of Ang2; thus Ang2 is one of the earliest tumour markers described, and one of the most general because it marks co-opted vessels and not the tumour cells themselves (Fig. 3b, left). Consistent with the possibility that autocrine Ang2 expression can destabilize vessels (Fig. 2, stage D), the co-opted vessels begin to die by an apoptotic process shortly after expressing Ang2 (Fig. 3b, middle). As vessels die, the tumour becomes secondarily avascular and hypoxic, resulting in marked induction of tumour-derived VEGF (Fig. 3b, middle). These high levels of VEGF correlate with cessation of regression of the destabilized co-opted vessels, and onset of robust new angiogenesis sprouting from these vessels, allowing for tumour survival and further growth (Fig. 3b, right). Thus, in such settings, endothelial Ang2 expression seems to correlate with vessel destabilization, apparently leading to vessel regression in the absence of tumour-derived VEGF, or robust new angiogenesis following induction of tumour-derived VEGF (stage D in Fig. 2, and Fig. 3b). The possibility that tumour vessel Tie2 receptors are blocked continuously by Ang2 and thus have an imbalance towards VEGF may well explain long-standing observations that tumour vessels fail to mature, exhibit poor associations between endothelial cells and their supporting cells, and are characterized by their leaky and haemorrhagic state. One practical prediction, which applies whether tumour growth initiates avascularly or through co-option, is that anti-VEGF therapy should ultimately blunt tumour growth. Early studies using an anti-VEGF antibody provided the first support for this notion67. This NATURE | VOL 407 | 14 SEPTEMBER 2000 | www.nature.com 2000 Macmillan Magazines Ltd insight progress has subsequently been confirmed in many laboratories using numerous approaches ranging from antibodies that bind and block VEGF, to those that bind and block VEGFR-2, to small molecules that block the activity of the VEGF-2 kinase domain, to genetic ablation of VEGF in tumour cells68. Thus, blockade of VEGF represents the best validated and most compelling anti-angiogenesis approach described so far. model for in vivo research on angiogenesis. Int. J. Dev. Biol. 40, 1189 1197 (1996). 7. Gale, N. W. &amp; Yancopoulos, G. D. Growth factors acting via endothelial cell-specific receptor tyrosine kinases: VEGFs, angiopoietins, and ephrins in vascular development. Genes Dev. 13, 1055 1066 (1999). 8. Carmeliet, P. Mechanisms of angiogenesis and arteriogenesis. Nature Med. 6, 389 395 (2000). 9. Hellstrom, M., Kaln, M., Lindahl, P., Abramsson, A. &amp; Betsholtz, C. Role of PDGF-B and PDGFRbeta in recruitment of vascular smooth muscle cells and pericytes during embryonic blood vessel formation in the mouse. Development 126, 3047 3055 (1999). 10. Hirschi, K. K., Rohovsky, S. A., Beck, L. H., Smith, S. R. &amp; D'Amore, P. A. Endothelial cells modulate the proliferation of mural cell precursors via platelet-derived growth factor-BB and heterotypic cell contact. Circ. Res. 84, 298 305 (1999). 11. Suri, C. et al. Requisite role of Angiopoietin-1, a ligand for the Tie2 receptor, during embryonic angiogenesis. Cell 87, 1171 1180 (1996). 12. Maisonpierre, P. C. et al. Angiopoietin-2, a natural antagonist for Tie2 that disrupts in vivo angiogenesis. Science 277, 55 60 (1997). 13. Holash, J., Wiegand, S. J. &amp; Yancopoulos, G. D. New model of tumor angiogenesis: dynamic balance between vessel regression and growth mediated by angiopoietins and VEGF. Oncogene 18, 5356 5362 (1999). 14. Holash, J. et al. Vessel cooption, regression, and growth in tumors mediated by angiopoietins and VEGF. Science 284, 1994 1998 (1999). 15. Soker, S., Takashima, S., Miao, H., Neufeld, G. &amp; Klagsbrun, M. Neuropilin-1 is expressed by endothelial and tumor cells as an isoform-specific receptor for <a href="/keyword/vascular-endothelial/" >vascular endothelial</a> growth factor. Cell 92, 735 745 (1998). 16. Shalaby, F. et al. Failure of blood-island formation and vasculogenesis in Flk-1-deficient mice. Nature 376, 62 66 (1995). 17. Fong, G. H., Rossant, J., Gertenstein, M. &amp; Breitman, M. L. Role of the Flt-1 receptor tyrosine kinase in regulating assembly of vascular endothelium. Nature 376, 66 70 (1995). 18. Hiratsuka, S., Minowa, O., Kuno, J., Noda, T. &amp; Shibuya, M. Flt-1 lacking the tyrosine kinase domain is sufficient for normal development and angiogenesis in mice. Proc. Natl Acad. Sci. USA 95, 9349 9354 (1998). 19. Taipale, J. et al. <a href="/keyword/vascular-endothelial/" >vascular endothelial</a> growth factor receptor-3. Curr. Top. Microbiol. Immunol. 237, 85 96 (1999). 20. Persico, M. G., Vincenti, V. &amp; DiPalma, T. Structure, expression and receptor-binding properties of placenta growth factor (PlGF). Curr. Top. Microbiol. Immunol. 237, 31 40 (1999). 21. Olofsson, B., Jeltsch, M., Eriksson, U. &amp; Alitalo, K. Current biology of VEGF-B and VEGF-C. Curr. Opin. Biotechnol. 10, 528 535 (1999). 22. Bellomo, D. et al. Mice lacking the <a href="/keyword/vascular-endothelial/" >vascular endothelial</a> growth factor-B gene (Vegfb) have smaller hearts, dysfunctional coronary vasculature, and impaired recovery from cardiac ischemia. Circ. Res. 86, E29 E35 (2000). 23. Carmeliet, P. et al. Abnormal blood vessel development and lethality in embryos lacking a single VEGF allele. Nature 380, 435 439 (1996). 24. Ferrara, N. et al. Heterozygous embryonic lethality induced by targeted inactivation of the VEGF gene. Nature 380, 439 442 (1996). 25. Carmeliet, P. et al. Impaired myocardial angiogenesis and ischemic cardiomyopathy in mice lacking the <a href="/keyword/vascular-endothelial/" >vascular endothelial</a> growth factor isoforms VEGF164 and VEGF188. Nature Med. 5, 495 502 (1999). 26. Miquerol, L., Langille, B. L. &amp; Nagy, A. Embryonic development is disrupted by modest increases in <a href="/keyword/vascular-endothelial/" >vascular endothelial</a> growth factor gene expression. Development 127, 3941 3946 (2000). 27. Gerber, H. P. et al. VEGF is required for growth and survival in neonatal mice. Development 126, 1149 1159 (1999). 28. Ferrara, N. et al. <a href="/keyword/vascular-endothelial/" >vascular endothelial</a> growth factor is essential for corpus luteum angiogenesis. Nature Med. 4, 336 340 (1998). 29. Gerber, H. P. et al. VEGF couples hypertrophic cartilage remodeling, ossification and angiogenesis during endochondral bone formation. Nature Med. 5, 623 628 (1999). 30. Stone, J. et al. Development of retinal vasculature is mediated by hypoxia-induced <a href="/keyword/vascular-endothelial/" >vascular endothelial</a> growth factor (VEGF) expression by neuroglia. J. Neurosci. 15, 4738 4747 (1995). 31. Alon, T. et al. <a href="/keyword/vascular-endothelial/" >vascular endothelial</a> growth factor acts as a survival factor for newly formed retinal vessels and has implications for retinopathy of prematurity. Nature Med. 1, 1024 1028 (1995). 32. Benjamin, L. E., Hemo, I. &amp; Keshet, E. A plasticity window for blood vessel remodelling is defined by pericyte coverage of the preformed endothelial network and is regulated by PDGF-B and VEGF. Development 125, 1591 1598 (1998). 33. Stone, J. et al. Roles of <a href="/keyword/vascular-endothelial/" >vascular endothelial</a> growth factor and astrocyte degeneration in the genesis of retinopathy of prematurity. Invest. Ophthalmol. Vis. Sci. 37, 290 299 (1996). 34. Pierce, E. A., Foley, E. D. &amp; Smith, L. E. Regulation of <a href="/keyword/vascular-endothelial/" >vascular endothelial</a> growth factor by oxygen in a model of retinopathy of prematurity. Arch. Ophthalmol. 114, 1219 1228 (1996). [Published erratum appears in Arch. Ophthalmol. 115, 427 (1997).] 35. Aiello, L. P. et al. <a href="/keyword/vascular-endothelial/" >vascular endothelial</a> growth factor in ocular fluid of patients with diabetic retinopathy and other retinal disorders. N. Engl. J. Med. 331, 1480 1487 (1994). 36. Adamis, A. P. et al. Increased <a href="/keyword/vascular-endothelial/" >vascular endothelial</a> growth factor levels in the vitreous of eyes with proliferative diabetic retinopathy. Am. J. Ophthalmol. 118, 445 450 (1994). 37. Springer, M. L., Chen, A. S., Kraft, P. E., Bednarski, M. &amp; Blau, H. M. VEGF gene delivery to muscle: potential role for vasculogenesis in adults. Mol. Cell 2, 549 558 (1998). 38. Detmar, M. et al. Increased microvascular density and enhanced leukocyte rolling and adhesion in the skin of VEGF transgenic mice. J. Invest. Dermatol. 111, 1 6 (1998). 39. Larcher, F., Murillas, R., Bolontrade, M., Conti, C. J. &amp; Jorcano, J. L. VEGF/VPF overexpression in skin of transgenic mice induces angiogenesis, vascular hyperpermeability and accelerated tumor development. Oncogene 17, 303 311 (1998). 40. Thurston, G. et al. Leakage-resistant blood vessels in mice transgenically overexpressing angiopoietin-1. Science 286, 2511 2514 (1999). 41. Pettersson, A. et al. Heterogeneity of the angiogenic response induced in different normal adult tissues by vascular permeability factor/<a href="/keyword/vascular-endothelial/" >vascular endothelial</a> growth factor. Lab. Invest. 80, 99 115 (2000). 42. Thurston, G. et al. Angiopoietin-1 protects the adult vasculature against plasma leakage. Nature Med. 6, 460 463 (2000). 43. Korhonen, J. et al. Enhanced expression of the tie receptor tyrosine kinase in endothelial cells during Perspectives and therapeutic possibilities There are many critical growth factors in...