Vasculogenesis

The angioblasts that give rise to endothelial cells in the first stages of developmental vascularisation originate in the mesoderm under the influence of fibroblast growth factor2 (FGF-2).² These differentiate into endothelial cells, which proliferate and aggregate into cords, and become lumenized to form primitive vascular plexuses.¹ Signalling through vascular endothelial growth factor receptor 2 (VEGF-R2) is crucial for angioblast survival and establishment of these first vessels.³

Endothelial cell formation from precursor cells, and in situ differentiation into vessels, was thought to be confined to developmental vascularisation. However, it has now been shown that circulating endothelial pro genitor cells (EPCs) exist in the adult and that they can contribute to vessel formation.⁴ These cells originate in the bone marrow and express CD34 and VEGF R2 as well as the orphan receptor AC133.⁵ EPCs can incorporate into neovessels formed in healing wounds, ischaemic tissue and tumours.⁶ The mechanisms controlling the incorporation of EPCs into neovessels are at present poorly defined although a number of growth factors, including VEGF, FGF2, granulocytemacrophage colony stimulating factor (GMCSF), angiopoietins and cytokines have all been shown to increase the mobilization of EPCs from bone marrow.⁷ Crude cell fractions that include EPCs can be expanded ex vivo and transplanted into ischaemic tissue in animal models where they have been reported to incorporate into neovessels.⁴ Importantly, in some situations neovessels comprising only EPC-derived cells were observed. Several studies have highlighted that the increase in the number of circulating progenitors, induced by cell transfusion or enhanced mobilization, can also enhance restoration and integrity of the endothelial lining, suppress neointimal formation, and increase blood flow to ischaemic sites.⁸ Whilst it is possible that EPCs, or EPC-derived cells, can form vessels in situ, by a process similar to vasculogenesis, current thinking suggests that the principal role of vascular stem and progenitor cells is paracrine, that is, these cells promote proliferation and migration of existing endothelial cells, as well as produce additional cytokines and chemokines to continue stem and progenitor mobilization, trafficking and adhesion.⁹

Angiogenesis

Establishment of the vascular network in development, as well as new vessel formation in the adult, requires angiogenesis. The two processes of sprouting and non-sprouting angiogenesis are responsible for remodelling of the primitive vascular plexus into a complex functional network. In sprouting angiogenesis, endothelial cells are activated by growth factors to undergo migration, proliferation and morphogenesis into new vessels, and VEGF is the major physiological activator.

Ischaemia is the primary initiator of sprouting angiogenic growth. Low oxygen tension activates expression of a wide range of angiogenic factors including VEGF, VEGF receptors, angiopoietin-2 and platelet-derived growth factor (PDGF). Genes for these factors contain hypoxia responsive elements in their promoters and some, like VEGF, have been shown to be direct targets of the transcriptional regulator hypoxia inducible factor (HIF). HIF-1 is a heterodimer composed of HIFα and HIFβ subunits.¹⁰ Under normoxic conditions HIF-1α is held at low intracellular concentrations by proteosomal degradation. With decreased oxygen tension HIF-1α becomes hydroxylated preventing its association with the von Hippel-Lindau ubiquitin ligase complex that is responsible for directing HIF-1α for degradation.¹¹ Thus, HIF-1α accumulates in the cell allowing it to activate transcription of hypoxia-inducible genes via the HIFα:β dimer.

VEGF expressed in response to HIF is secreted by ischaemic cells and acts on endothelial cells in adjacent microvessels. In these previously quiescent microvessels, endothelial activation, proliferation and migration are normally suppressed by signals from abluminal perivascular support cells; pericytes. This suppression needs to be relieved before angiogenesis can proceed.¹² The close interaction between endothelial cells and pericytes is promoted by the ligand angiopoietin-1, which is produced by the pericyte and acts on the endothelial receptor Tie2.¹³ Disruption of this signalling interaction is likely to involve angiopoietin-2, another hypoxiaresponsive molecule.¹⁴ Angiopoietin-2 can act to inhibit angiopoietin-1-induced activation of Tie2.¹⁵ Once released from the prostabilizing effects of pericytes, the endothelial cells are free to invade the perivascular space, aided by proteases that degrade extracellular matrix. Many metalloproteinases have been implicated in angiogenic sprouting, including matrix metalloproteinases 2, 3, 7 and 9 as well as other proteolytic enzymes such as urokinase-type plasminogen activator.¹⁶ Migration of the activated endothelial cells is aided by plasma proteins that extravasate from the activated microvessels in response to the vasodilatory and pro-permeability effects of VEGF. This growth factor is a potent chemoattractant and mitogen for endothelial cells, and directs their migration and proliferation. Interestingly, in vivo endothelial cells in the developing vascular sprouts respond differentially to VEGF, with the cells at the tip migrating and those behind the tip proliferating.¹⁷ Migration and proliferation give rise to endothelial cords that become lumenized, a process which is poorly understood, though is known to be enhanced by angio poietin-1.¹⁸

Neovessel maturation

Whether by vasculogenesis or angiogenesis, newly formed blood vessels are highly unstable, and are prone to haemorrhage, thrombosis and spontaneous regression in the absence of elevated growth factors.¹⁹ Such vessels are characteristically found in pathological vascularisation and their phenotype can directly contribute to the disease process. Newly formed primitive vascular channels are maintained by local high concentrations of VEGF, withdrawal of which leads to endothelial apoptosis and neovessel regression.²⁰ Transformation to a functional vessel requires interaction of endothelial cells in the nascent vessel with pericytes, which originate as mesenchymal cells that are recruited to the developing vessel and differentiate into pericytes on contact with the endothelium.²¹ Proliferation and migration of partially or fully differentiated pericytes along established microvessels also contributes to mural cell acquisition by sprouting neovessels, and sprouts can themselves recruit mesenchymal cells.²² Migration and proliferation of pericytes is regulated mainly by platelet-derived growth factor-β (PDGFβ) secreted by endothelial cells. Interaction between mesenchymal cells and endothelial cells in a developing vessel produces phenotypic changes in both cell types. The mesenchymal cell is directed toward a pericyte or smooth muscle phenotype and the endothelial cell adopts the phenotype required for formation of a stabilized microvessel.¹ Pericytes supply anti apoptotic ligands, including angiopoietin-1,²³ to underlying endothelial cells allowing neovessels to survive the decrease in VEGF concentrations that occur as the ischaemia is relieved by increased perfusion. In addition, pericytes suppress endothelial proliferation and migration, and increase deposition of the perivascular basement membrane, all of which contribute to switching the fragile nascent vessel into a quiescent functional microvessel.²⁴ Transforming growth factor-β, produced by proteolytic cleavage from a precursor form as a result of endothelial:pericyte interaction,²⁵ has a central role in these effects.

Microvascular network maturation

Maturation of vascular channels into functional vessels is accompanied by maturation of the neovessel network. This involves optimization of the new vessel configuration, density, branching pattern and vessel hierarchy. Spatial distributions of angiogenic initiators, like VEGF, have a major influence on the direction of the initial branches. There are six members of the VEGF family and VEGF-A is expressed as a number of alternatively spliced variants; in humans these are mainly forms with 121, 165, 189 and 206 amino acids.²⁶ VEGF165, 189 and 206 possess heparin-binding domains that allow these forms to interact with the extracellular matrix. The ability of VEGF isoforms to be retained by the matrix is important in regulating the spatial organization of vessel branching.²⁷

Patterning also occurs through selective loss of certain vessels by regression. Another major determinant of network maturation is the branching and ‘splitting’ of vessels by non-sprouting angiogenesis. This occurs by the process of intussusception in which vessel lumens are internally divided by insertion, and subsequent growth and stabilization, of transcapillary tissue pillars.²⁸ The mechanism of intussusception and factors that regulate it are poorly understood, though the Tie receptors are known to have a role.²⁹

Arteriogenesis

Further expansion and muscularization of vessels occurs by the process of arteriogenesis; the development of large calibre collateral arteries from a pre existing network, in response to occlusive disease. This involves recruitment of additional mural cells and their proliferation, as well as expansion of the abluminal extracellular matrix. Whilst angiogenesis is driven by ischaemia, the initiator of arteriogenesis is increased fluid shear stress, which acts directly upon gene expression involved in the endothelial cell cycle.³⁰ PDGFβ with its effects on smooth muscle cell recruitment and proliferation, and TGFβ, a known regulator of vascular extracellular matrix synthesis, are likely to be key regulators.

Arteriogenesis of collateral vessels has been demonstrated in a number of animal models following the occlusion of major vessels.³¹ In humans, well-developed collateral vessels that bypass occluded arteries have been found frequently and those patients with the best developed collaterals often have minimal symptoms.³² Arteriogenesis of collaterals in response to the occlusion of primary supply vessels occurs in two phases. An initial increase in the lumen size occurs within a few days and this is followed by a slower remodelling of smooth muscle cell and extracellular matrix cover.³¹ The early phase of this adaptive arteriogenesis is associated with inflammation. There is monocyte attachment to endothelium secondary to release of mono cyte chemoattractant protein-1 (MCP-1), GMCSF and stromal-cell-derived factor-1 (SDF-1), which recruit CD14+ monocytes to the activated endothelial cell surface as well as extravasation and accumulation in the adventitia and perivascular space, with mast cells and T lymphocytes.³³ Monocytes have a critical role in adaptive arteriogenesis as experimental suppression of monocyte numbers decreases arteriogenesis.³⁴ These cells provide growth factors to stimulate vessel enlargement and proteases that can act on the extracellular matrix to accommodate increased vessel size. Changes to the wall of the vessel involve remodelling of the media, with increased turnover of medial smooth muscle cells and a shift towards a synthetic phenotype, assuming a contractile phenotype after enlargement of arterial diameter.³³ A new internal elastic lamina is established and vessel wall thickening results from increased extracellular matrix deposition.

Delivery of molecular activators of vascular growth

Induction of angiogenesis in the appropriate ischaemic areas and in future local arteriogenesis at suitable sites for collateral development, requires activating agents to be delivered in a manner that ensures controllable local activity and minimizes systemic side effects. Growth factors are readily degraded and if administered systemically would have to be used at high concentrations in order to ensure enough active growth factor reached the appropriate site. There are risks associated with non-local delivery, for example, systemically delivered VEGF produces severe hypotension in animal models.³⁵ In addition to localization, activators must be present for sufficient time to induce optimal vascular growth.

The two principal methods used to deliver angiogenic activators in pre-clinical studies as well as clinical trials have been as recombinant proteins or as the genes that encode these proteins. Activators can be delivered locally via intramuscular catheters, direct injection into the muscle or use of coated stents. An important consideration in the use of peptide growth factors is ensuring sufficient longevity of the molecules at the desired site. Where recombinant proteins are injected this would necessitate multiple injections during the course of treatment. An alternative strategy is the use of local reservoirs of recombinant protein, such as bio degradable microspheres.³⁶ A major limitation to the use of recombinant protein, however, is the expense and difficulty in obtaining large enough quantities of appropriate purity, especially when cocktails of growth factors are required.

Delivery of angiogenic factors by gene transfer has significant advantages over the administration of recombinant protein. It is relatively easy to produce high purity DNA in large quantities and the transfected genes remain active over a period of several days to several weeks. In contrast to gene therapy aimed at correcting genetic diseases, gene transfer as a means of providing short term local expression of therapeutic proteins has been successful. Surprisingly, small amounts of DNA plasmid vectors can be taken up by muscle cells in vivo and are reported to result in significant gene expression in humans.³⁷ Improvements in transfer efficiency have been sought by the use of liposomal carriers and viral vectors. Adenovirus is the most common viral vector used for the delivery of angiogenic genes, and since genes transferred in this way do not integrate into chromosomes of transduced cells, transient expression is provided.

There have been reports of an inflammatory reaction to adenoviral vectors in human trials, but no long term safety problems at doses appropriate for angiogenic therapy. Second generation adenoviral vectors with deletions of E1 and E4 regions have better transfection efficiency and elicit a decreased inflammatory response³⁸ and further improved adenoviral vectors can be expected.³⁹ The adeno-associated viruses (AAV) offer an alternative viral means for gene delivery. AAV efficiently transduce skeletal muscle and vasculature.⁴⁰ However, along with retroviruses and lentiviruses, AAV integrate into the recipient genome requiring the development of regulatory systems if they are to provide controllable expression of vascular growth genes. As with recombinant protein, local delivery of genes can be accomplished by direct intramuscular injection, implantation of coated stents or catheters. It may also be possible to utilize tissue-specific endothelial surface molecules for targeting vectors to particular vascular beds. Implantation of cells transfected ex vivo offers an additional route of local delivery.

Angiogenic activators

Perhaps the simplest approach to activating angiogenesis is the administration of a soluble angiogenic activator. VEGF is relatively specific for endothelial cells and physiologically relevant. Although VEGF-A, -B, -C, -D and –E, as well as the VEGF-R1 ligand placental growth factor, have all been shown to activate angiogenesis when administered in animal models, most studies have concentrated on VEGF-A. Administration of VEGF alone results in a high percentage of malformed capillaries in animal models.⁴¹ This growth factor also induces vessel permeability resulting in local hypotension and oedema.³⁵ Neovessels induced by VEGF are transient, with regression occurring on growth factor withdrawal.⁴² These data indicate formation of a sustained microvessel network may require relatively long term exposure to the angiogenic initiator.

The fibroblast growth factors have also been examined as potential therapeutic agents to induce angiogenesis in clinical trials. There are 23 members of the FGF family and FGF-2 and FGF-4 have been used in clinical trials. FGF-1, -2, - 4 and -9 are highly mitogenic for endothelial cells, although these growth factors are also active on non endothelial cells.⁴³ Another growth factor that induces angiogenesis and is in early trials is hepatocyte growth factor; again its effects are not confined to endothelial cells.⁴⁴

With the recognition that physiological angiogenesis requires a spatially and temporally co-ordinated repertoire of signals, attempts have been made to improve capillary formation by providing cocktails of growth factors. Indeed combination of VEGF with the pro-stabilizing Tie2 agonist angiopoietin-1 in a mouse model does produce microvessels with increased lumen size, less thrombosis and increased perfusion compared with VEGF alone.¹⁸ These vessels are also less permeable than those formed in response to VEGF alone.⁴⁵ Another approach aimed at providing a more physiological range of angiogenic factors is the targeting of HIF. Expression of a form of HIF-1α that is resistant to oxygen induced degradation in mouse skin led to up-regulation of HIF-sensitive angiogenic genes and the stimulation of microvessel formation.⁴⁶ Again the vessels produced were not associated with oedema. The cell permeable peptide, PR39, inhibits proteosomal degradation and can stabilize HIF-1α.⁴⁷ PR39 stimulates angiogenesis in mouse heart, although further studies are required to determine its specificity.⁴⁷

Pre-clinical and early clinical studies have shown that angiogenesis can be induced in vivo by a variety of approaches. The challenge is to devise a means to stimulate the conversion of these neovessels into optimally organized, persistent and functional microvascular networks.

Arteriogenic activators

Little is known about the molecular mechanisms of arteriogenesis. In contrast to the ischaemic tissue microenvironment in which angiogenesis occurs, collateral arteriogenesis in the limb takes place in normoxic conditions.⁴⁸ In adaptive arteriogenesis studied in animal models, the biochemical effects of the increased flow that the collaterals experience as a result of occlusion of conductance vessels plays a major role. These effects include increased wall shear stress as well as tangential and axial stresses. Increased shear can up-regulate vascular cell adhesion molecule-1 and intracellular adhesion molecule-1, as well as MCP-1 which contributes to monocyte recruitment.⁴⁹ Growth factors are undoubtedly involved in adaptive arteriogenesis, but again more work is required to identify the key factors and their roles. In animal models FGF-1 and FGF-2 were found to be unchanged during adaptive arteriogenesis, although there was a transient increase in expression of FGFreceptor1.⁵⁰

TGFβ is increased during collateral development and can enhance arteriogenesis in animal models. Several factors have been found to enhance arteriogenesis when administered to animals, including FGF-2, VEGF, placental growth factor, angiopoietin-1 and MCP-1, although the exact mechanisms of action remain undefined.³¹ Many appear to have indirect actions, for example, VEGF and placental growth factor infusions are likely to enhance arteriogenesis via their monocyte chemoattractive activity. Given the great potential of therapeutic induction of collateral arteriogenesis for the treatment of peripheral vascular disease, it is important that a better understanding of the molecular mechanisms be gained. Identification of key regulators that can induce or enhance arteriogenesis of collaterals is a priority.

Clinical trials for angiogenic therapy of peripheral vascular disease

There have been a number of phase 2 and 3 clinical trials aimed at relieving peripheral vascular disease by angiogenic therapy. The Therapeutic Angiogenesis with Recombinant Fibroblast Growth Factor-2 for Intermittent Claudication (TRAFFIC) study used single or repeated doses of recombinant FGF-2 delivered by arterial puncture and crossover catheter in patients with intermittent claudication. Patients receiving a single FGF-2 dose showed a significant improvement in peak walking time at 90 days.⁵¹ In the Regional Angiogenesis with Vascular Endothelial Growth Factor (RAVE) study, VEGF121 gene transfer by adenovirus was utilized but failed to produce a significant improvement in peak walking time at 12 weeks.⁵² In contrast, in a different study adenoviral delivered VEGF165 given during angioplasty did produce an increased angiographically-assessed vascularity at 3 months.⁵³ The Vascular Endothelial Growth Factor in Ischemia for Vascular Angiogenesis (VIVA) trial failed to show a difference between the treatment and placebo groups for the primary endpoint of walking time.⁵⁴ The TALISMAN collaborators evaluated a plasmid-based angiogenic gene delivery system for the local expression of FGF-1 in patients with non-healing ulcers who were not suitable for invasive re-vascularisation. Whilst there was no difference observed in the primary endpoint of complete ulcer healing, there was a reduced risk of major amputation.⁵⁵ Initial reports from the ongoing Bone Marrow Outcome Trial in Critical Limb Ischemia (BONMOT-CLI), show promise for the intramuscular injection of autologous bone marrow stem cells into ischaemic limbs, with significant improvements in Rutherford categories and a trend against major limb amputation.⁵⁶

A meta-analysis of 5 randomised controlled trials investigating the role of gene therapy as an option for the treatment of peripheral vascular disease concluded that the current literature does not demonstrate a clinical benefit for patients with peripheral vascular disease.⁵⁷ Although these trials have met with limited success, together with phase 1 studies and earlier small trials, they demonstrate the feasibility and safety of molecular approaches to therapeutic modulation of vascular growth. The trials have also been valuable in aiding development of techniques for delivering therapeutic agents, as well as helping clinicians refine aspects of trial design for future clinical work on arteriogenesis and angiogenesis.

The realization that therapeutic induction of collateralization by arteriogenesis would be most appropriate for occlusive disease, whilst angiogenic therapy would benefit patients with microvascular defects, should help improve selection of the most appropriate populations for use in future trials. Clear clinical end-points are required in such work. Where angiogenesis is the aim, establishment of optimal treatment modalities will depend on further preclinical work focused on determining ways to establish mature, correctly patterned vascular networks. This may involve defined cocktails of stimulators, or activation of transcriptional factors triggering coordinated expression of stimulators. In both cases distinct spatial and temporal expression patterns are likely to be required.