Once the physical barriers are dissolved, proliferating ECs migrate to distant sites. This is a complex, tightly regulated process, requiring the involvement of numerous stimulators and inhibitors. For reasons of brevity, we will only review some key signals. The most important signalling of all involves VEGF, which via binding its receptor VEGFR-2, regulates embryonic, neonatal and pathological angiogenesis in a strict dose-dependent manner. The latter phenomenon is exemplified by genetic studies. Indeed, loss of a single VEGF allele results in lethality due to early embryonic vascular defects (Carmeliet et al. 1996; Ferrara et al. 1996), while reduction of VEGF levels by only 25% impairs spinal cord perfusion and causes motor neuron degeneration, reminiscent of amyotrophic lateral sclerosis (Oosthuyse et al. 2001). Several additional gene-manipulating studies in mice, zebrafish and Xenopus have documented the principal role of VEGF in vascular development and illustrated its potential to stimulate new vessel growth (Cleaver and Krieg 1998; Nasevicius et al. 2000; Liang et al. 2001; Stalmans et al. 2003). Conditional inactivation of VEGF after birth or expression of particular VEGF isoforms in knock-in mice revealed that VEGF is requisite for vascular expansion during post-natal growth in various organs (e.g. kidney, bone, heart and retina) and, when insufficiently available, causes tissue ischaemia, impaired growth and organ failure (Carmeliet et al. 1999b; Haigh et al. 2000; Maes et al. 2002; Mattot et al. 2002; Stalmans et al. 2002; Eremina et al. 2003). On the other hand, over-expression of VEGF, for instance in the skin of transgenic mice, stimulates abundant cutaneous capillary growth and an inflammatory skin condition resembling psoriasis (Xia et al. 2003). Numerous studies also established VEGF as a key angiogenic player in cancer (Ferrara 2002). Because of its predominant role, VEGF is currently being evaluated for both pro- and anti-angiogenic therapy (Ferrara 2000a, b).
Gene targeting studies in mice have elucidated the functional role of PlGF, a homologue of VEGF, which binds to VEGFR-1. Loss of PlGF-while not causing any vascular defects during embryonic development, reproduction or normal adult life-impaired angiogenesis and plasma extravasation in pathological conditions, including ischaemia, inflammation and cancer (Carmeliet et al. 2001). The important role of PlGF in pathological angiogenesis is further evidenced by findings that PlGF stimulates angiogenesis and collateral growth (see below)
in the ischaemic heart and limb of wild-type mice (Luttun et al. 2002) and, in combination with VEGF, in the ischaemic heart of a mouse model resistant to the VEGF treatment alone (Autiero et al. 2003). PlGF contributes to the angio-genic switch in pathological conditions by affecting, directly and indirectly, multiple cell types (Fig. 4).
First, PlGF has direct effects on ECs by inducing its own signalling via VEGFR-1 and by amplifying VEGF-driven angiogenesis (Carmeliet et al. 2001; Autiero et al. 2003). Second, PlGF-by directly stimulating SMCs and fibroblasts, which express VEGFR-1-recruits SMC around nascent vessels and thus promotes vessel maturation and stabilisation (Green et al. 2001; Ishida et al. 2001; Luttun et al. 2002; see Sect. 5). Third, PlGF stimulates the mobilisation of VEGFR-1-positive HSC/HPCs from the bone marrow (Carmeliet et al. 2001; Lyden et al. 2001) and, indirectly via up-regulation of VEGF expression, recruitment of VEGFR-2-positive EPCs to the site of neovascularisation (Hattori et al. 2001, Luttun et al. 2002). At such sites they promote new vessel growth by directly incorporating into the vessel wall and/or by creating a pro-angiogenic microenvironment through the release of angiogenic molecules (Rehman et al. 2003). Furthermore, PlGF can also recruit HPCs to distant sites to form pre-metastatic niches. Fourth, PlGF is chemo-attractive for monocytes and macrophages, which express VEGFR-1 (Sawano et al. 2001; Luttun
Fig. 4 PlGF is a master switch of pathological angiogenesis and stimulates tumour vascu-larisation and growth by affecting multiple cell types. Within the tumour environment, PlGF stimulates either vascular cells (endothelial and smooth muscle cells) or non-vascular cells (monocytes/macrophages, stromal cells and dendritic cells). PlGF may also affect VEGFR-1-expressing tumour cells directly. In addition, PlGF stimulates the mobilisation and recruitment of VEGFR-1-positive HSC/HPCs from the bone marrow to the primary tumour and pre-metastatic niches. It remains to be determined whether PlGF also directly affects EPC recruitment by interacting with its receptor VEGFR-1 on EPCs
Fig. 4 PlGF is a master switch of pathological angiogenesis and stimulates tumour vascu-larisation and growth by affecting multiple cell types. Within the tumour environment, PlGF stimulates either vascular cells (endothelial and smooth muscle cells) or non-vascular cells (monocytes/macrophages, stromal cells and dendritic cells). PlGF may also affect VEGFR-1-expressing tumour cells directly. In addition, PlGF stimulates the mobilisation and recruitment of VEGFR-1-positive HSC/HPCs from the bone marrow to the primary tumour and pre-metastatic niches. It remains to be determined whether PlGF also directly affects EPC recruitment by interacting with its receptor VEGFR-1 on EPCs prevents anti-tumour immune attack
Pre-metastatic niche prevents anti-tumour immune attack
Pre-metastatic niche et al. 2002)-activated macrophages are a rich source of a variety of angiogenic molecules (Autiero et al. 2003) and also produce PlGF, thereby providing a positive feedback.
The role of PlGF and VEGFR-1 in both endothelial and haematopoietic lineages explains why blocking VEGFR-1 more efficiently suppresses inflammatory angiogenic disorders (atherosclerosis, arthritis) than blocking VEGFR-2 (Luttun et al. 2002). Similar effects would thus be expected when VEGFR-1 activation is prevented by PlGF inhibitors or antibodies. VEGF-B is another homologue of VEGF, but its angiogenic activities remain to be determined.
Another angiogenic signalling system involved in vessel growth and stabilisation comprises the Tie-2 receptor, which binds the angiopoietins (Ang-1 and Ang-2). Ang-1, via phosphorylation of Tie-2, is chemotactic for ECs, induces vascular sprouting, stimulates EC survival, mobilises EPCs and HSC/HPCs, and stabilises networks initiated by VEGF, presumably by stimulating the interaction between endothelial and peri-endothelial cells (Suri et al. 1996,1998; Gale and Yancopoulos 1999; Hattori et al. 2001). All these activities may explain why Ang-1 stimulates vessel growth in skin, ischaemic limbs, gastric ulcers and in some tumours (Suri et al. 1998; Shim et al. 2002; Plank et al. 2004). However, Ang-1 also suppresses angiogenesis in other tumours and the heart (Ahmad et al. 2001; Visconti et al. 2002). In fact, Ang-1 may restrain vessel sprouting by tightening vessels via effects on junctional molecules (Thurston et al. 2000), by recruiting pericytes and by promoting endothelial-mural cell interactions as an adhesive protein (Carlson et al. 2001). Ang-2 in concert with VEGF is also angiogenic and has been proposed to stimulate the growth of immature (SMC-poor) tumour vessels by loosening the endothelial-peri-endothelial cell interactions and degrading the ECM via up-regulation of proteinases, thereby counteracting the activity of Ang-1 (Ahmad et al. 2001; Etoh et al. 2001; Gale et al. 2002). However, the angiogenic activity of Ang-2 seems to be contextual as well since, in the absence of VEGF, Ang-2 causes EC death and induces vessel regression (Maisonpierre et al. 1997).
Several additional factors regulate the proliferation of ECs. Integrins are heterodimeric cell surface receptors of specific ECM molecules which, by bi-directionally transmitting signals between the outside and inside of vascular cells, assist vascular cells to build new vessels in co-ordination with their surroundings (Hood and Cheresh 2002; Hynes 2002). The av^3 and av^5 in-tegrins have long been considered to regulate the angiogenic switch positively (Lee and Juliano 2004), because their pharmacological antagonists which are currently being evaluated in clinical trials suppress pathological (i.e. tumour) angiogenesis (McNeel et al. 2005). Furthermore, a combination of antibodies against a101 and a201 integrins reduces tumour vascularisation (Senger et al. 2002). However, genetic deletion studies suggest that vascular integrins inhibit, rather than stimulate, tumour angiogenesis (Reynolds et al. 2002). This inhibitory activity may be attributable to the ability of these integrins to suppress VEGFR-2-mediated EC survival, trans-dominantly block other inte-
grins, or mediate the anti-angiogenic activity of angiogenesis inhibitors such as tumstatin, endostatin, and canstatin (Carmeliet 2002; Reynolds et al. 2002; Hamano et al. 2003; Sudhakar et al. 2003,2005; Lee and Juliano 2004, Magnon et al. 2005). Thus, while these genetic insights do not invalidate the promising (pre)clinical results obtained with integrin antagonists for cancer treatment, a better understanding of whether and in which conditions integrins play positive or negative roles in tumour angiogenesis is desirable.
FGFs stimulate EC growth directly and, by recruiting pro-angiogenic mes-enchymal and inflammatory cells, also indirectly (Carmeliet 2000a). Though PDGF-BB has been documented to stimulate microvascular sprouting of ECs, its main activity is to recruit pericytes and SMCs around nascent vessel sprouts, thereby stimulating vessel maturation and stabilisation, and increasing vessel perfusion (Lindahl et al 1998, 1999; see Sect. 5). Molecules such as TGF-01, activin-A and TNF-a stimulate or inhibit EC growth, depending on the context (Pepper 1997; Gohongi et al 1999; Guo et al. 2000).
Chemokines are another interesting class of molecules, capable of stimulating or inhibiting EC growth, depending on the type of receptor they activate. Chemokines binding CXCR2 and CXCR4 are angiogenic (e.g. GRO-a, GRO-y, ENA-78, GCP-2, IL-8, SDF-1a, 9E3, eotaxin, I-309, MCP-1, fractalkine), while chemokines binding CXCR3 (e.g. PF-4, MIG, IP-10, ITAC, BCA-1, SLC/6Ckine) have angiostatic activity (Bernardini et al. 2003). At least two of those have received increasing recognition. IL-8 is expressed in several tumours and inflammatory conditions, and is even up-regulated in tumours after anti-VEGF therapy, while anti-IL-8 antibodies block tumour growth (Mizukami et al. 2005). Furthermore, emerging evidence indicates that SDF-1a stimulates angiogenesis via direct effects on ECs, as well as via recruitment of bone marrow-derived EPCs and HPCs both in ischaemic and malignant tissues (Ceradini et al. 2004; Butler et al. 2005); antagonists of SDF-1a block tumour growth (Guleng et al. 2005; see above).
EGF is a mitogen for epithelial cells and is over-expressed in various tumours. While it does not regulate vascular development, it has been implicated in tumour angiogenesis. Indeed, EGF induces the expression of its own receptors in ECs and is mitogenic for EGFR-positive ECs. In addition, EGF indirectly stimulates tumour angiogenesis by inducing the release of VEGF and the expression of VEGF receptors in tumour vessels (van Cruijsen et al. 2005). Another growth factor, hepatocyte growth factor (HGF), stimulates angiogenesis when exogenously administered (Jiang et al. 2005). Other molecules are capable of stimulating EC growth in vitro or angiogenesis in experimental models, but their endogenous role in angiogenesis during development or disease often remains incompletely determined. Some examples include ery-thropoietin, IGF-1, neuropeptide-Y, leptin, Thy-1, tissue factor, interleukins and others (Carmeliet 2003a).
Angiogenic sprouting is a complex process, requiring a finely tuned balance between activators and inhibitors. Some of the endogenous angiogenesis inhibitors that are currently being evaluated for clinical use include angio-statin, endostatin, anti-thrombin III, interferon-^, leukaemia inhibitory factor and platelet factor 4, tumstatin, C-terminal hemopexin-like domain of MMP-2 (PEX) andvasostatin (O'Reilly et al. 1994,1997; Carmeliet 2000b, 2003a; Nyberg et al. 2005).
Was this article helpful?