Article Text

Download PDFPDF

Therapeutic angiogenesis for cardiovascular disease: biological context, challenges, prospects
  1. Ian Zachary,
  2. Robert D Morgan
  1. Centre for Cardiovascular Biology and Medicine, Department of Cardiovascular Science, Division of Medicine, University College London, London, UK
  1. Correspondence to Professor Ian Zachary, University College London, Rayne Institute, Department of Medicine, 5 University Street, London WC1E 6JJ, UK; i.zachary{at}


Since the discovery of vascular endothelial growth factor (VEGF), therapeutic angiogenesis has attracted interest as an alternative treatment for ischaemic heart and peripheral disease. In parallel, the view has also gained ground that angiogenesis has an important role in the pathogenesis of atherosclerotic disease and its clinical sequelae. These conflicting perspectives have both been based on a large quantity of preclinical data obtained from mainly small rodent models of disease. However, in recent years further research and the results of clinical trials of pro-angiogenic and anti-angiogenic treatments have provided new insights into the impact of VEGF and other angiogenesis-based approaches on human health and disease. This review discusses therapeutic angiogenesis in the light of recent scientific advances and clinical evidence, and considers the future challenges and prospects for therapeutic angiogenesis.

  • Coronary artery disease (CAD)
  • public health
  • gene therapy
  • growth factors
View Full Text

Statistics from


The molecular understanding of blood vessel formation has been a source of recent innovation in the treatment of common human diseases. In the case of many human pathologies, cancer and ocular disease being the best understood examples, angiogenesis is the target for new treatments and the goal is to inhibit blood vessel formation. In ischaemic heart and peripheral disease, where the problem is essentially one of underperfusion and vascular insufficiency, the challenge has been the opposite: to stimulate angiogenesis in order to bypass diseased or occluded vessels, and revascularise ischaemic tissues in the heart and periphery, an approach called therapeutic angiogenesis. The discovery of vascular endothelial growth factor (VEGF or VEGF-A), an essential angiogenic factor in development and in many disease states, made it possible to limit or promote blood vessel formation, and this has led directly to the development of several anti-angiogenic treatments targeted at VEGF or its receptors, which are now in routine clinical use for various cancers and for the major eye disease, wet age-related macular degeneration. In contrast, there is currently no angiogenic treatment for cardiovascular disease, and translation of the attractively simple idea that stimulating angiogenesis can give benefit to patients with heart disease has proved highly problematic, and has been dogged by controversy surrounding the role of angiogenesis in atherosclerosis. This review will try to explain why the early promise of VEGF and other angiogenic cytokines as novel treatments for cardiovascular disease has not so far been fulfilled, and discuss the future prospects for this approach. However, before considering therapeutic angiogenesis itself, it is important to review recent important developments to understand how VEGF impacts on the adult cardiovascular system.

Role of VEGF in the adult cardiovascular system and cardiovascular disease

The concept that VEGF-stimulated angiogenesis would be therapeutically beneficial for ischaemic heart disease is supported by a large amount of data from animal models.1 Furthermore, as well as being essential for angiogenesis in development and in disease,2 it is also now recognised that VEGF has an important role in maintaining healthy adult vascular function and homoeostasis (figure 1). This is a new paradigm for VEGF activity, mechanistically distinct from its angiogenic role, and one with obvious relevance for therapeutic manipulation of VEGF in cardiovascular disease. We originally hypothesised that VEGF was arterioprotective based on studies showing that VEGF stimulated endothelial production of nitric oxide (NO) and prostacyclin (PGI2), two vasodilatory agents also known to inhibit platelet aggregation and vascular smooth muscle cell (VSMC) proliferation, and that VEGF gene transfer inhibited neointima formation in vivo.1 3–6 Furthermore, the most salient short-term effects of VEGF in vivo in animal models and human studies are rapid vasodilatation and an acute hypotensive response, both probably mediated through NO, possibly in combination with PGI2.1 5

Figure 1

Dual role of vascular endothelial growth factor (VEGF). VEGF is essential for angiogenesis in embryonic development and in many human diseases (right panel), but is also increasingly recognised to have an important role in homoeostasis and the maintenance of healthy endothelial function in the adult vasculature through the actions of circulating VEGF and intracrine actions of endothelial-derived VEGF (left panel). See text for further details. EC, endothelial cell; VSMC, vascular smooth muscle cell.

Compelling evidence has emerged from studies in murine models that VEGF has a more fundamental role in the maintenance of adult vascular integrity and health, particularly in the kidney,7 8 and has tissue-specific roles in the maintenance of capillary beds—for example, in pancreatic islets and the choroid plexus.9 The need for VEGF in the maintenance of adult vascular integrity is strikingly demonstrated by the phenotype of mice with an endothelial-specific loss of VEGF. Though endothelial-derived VEGF is thought to constitute a small proportion of total body VEGF, endothelial-specific deletion of VEGF caused a progressive endothelial degeneration due to systemic endothelial apoptosis. This, in turn, resulted in a range of vascular pathologies in multiple organs, including haemorrhaging, intestinal perforations and multiple infarcts, culminating in sudden death of 55% of mice surviving to adulthood by 25 weeks of postnatal life.10 Interestingly, these defects could not be rescued by exogenous VEGF, indicating a surprising requirement for endothelial-specific autocrine or intracrine VEGF signalling which cannot be compensated by paracrine actions of VEGF derived from other tissue sources (see figure 1).

VEGF signalling is also implicated in adaptive cardiac growth. Treatment with a soluble decoy form of Flk-1 (murine equivalent of VEGFR2/KDR), in mouse models of adaptive cardiac hypertrophy and pressure overload, reduced cardiac capillary density and induced phenotypic characteristics of heart failure, including contractile dysfunction, left ventricular dilatation, and inhibition of cardiac growth.11 12 These studies suggest that an active angiogenic response is essential for adaptive cardiac growth, and that disruption of VEGF signalling impairs heart function under stress.

While these studies predict that VEGF would be beneficial for cardiovascular health either by promoting therapeutic angiogenesis or by enhancing protective vascular functions, this straightforward view is tempered by much evidence suggesting, paradoxically, that microvessels are an important feature of human atherosclerotic plaques, and that the extent of vessel formation is associated with disease progression and specifically with plaques vulnerable to rupture and haemorrhage.13 14 In addition, some, though not all, studies of VEGF or other angiogenic factors in mouse models of atherosclerosis seem to favour a pro-atherogenic role of angiogenesis.14 15 The role of VEGF in atherosclerosis is a highly contentious issue, and much of the data obtained from larger animal species in models of balloon injury-induced neointima formation, or of peripheral and myocardial ischaemia and infarction have reported beneficial effects of VEGF.1 14 15 The current state of our knowledge can tentatively be summarised as follows. Despite strong empirical support for an association between neovascularisation and severity of human atherosclerotic disease, the causal links between angiogenesis, atherosclerosis, plaque rupture, arteriothrombosis and cardiovascular disease in humans remain elusive. Experimental evidence from a variety of animal studies modelling different aspects of human cardiovascular disease has produced mixed, and in some cases apparently contradictory, results, and the relevance of neovascularisation for plaque formation in animals is both unproved and difficult to assess empirically. Much of the relevant work supporting the contending arguments has been critically and comprehensively reviewed and the reader is referred to those articles and the primary literature cited therein.1 13–15

The theory that VEGF is a homoeostatic, vasculoprotective cytokine is seemingly at odds with the argument that VEGF-dependent angiogenesis is pro-atherogenic in the plaque. However, these two paradigms may be reconciled, if it is argued that different VEGF concentration ranges and/or distinct tissue or cellular pools, or regional patterns of bioavailability, are important for determining the specific response profile and biological outcome. There appear to be at least three distinct modes of action for VEGF in vivo: (i) angiogenic, probably requiring higher VEGF concentrations, though such effects may usefully be further subdivided dependent on tissue, and physiological or disease setting; (ii) acute effects of exogenous or circulating VEGF on the vasculature—for example, vasodilatation and hypotension, and perhaps also maintenance of some vascular beds, likely to be dependent on lower ‘constitutive’ levels of VEGF, as suggested by the hypertensive effects of VEGF inhibitors (discussed below); (iii) the autocrine, or intracrine, role of endothelial VEGF in vascular homoeostasis, protection and endothelial survival. At present our understanding of the mechanisms and parameters determining these modes of action of VEGF is very limited, and is an important question to be addressed in future research.

The protective and pro-atherogenic angiogenesis paradigms clearly have very different implications for cardiovascular treatments targeted at either inhibiting or stimulating angiogenesis. A pro-atherogenic role predicts that inhibition of VEGF or its receptors will limit or reduce atherosclerosis and its clinical consequences, while treatment with VEGF may accelerate the disease. Conversely, the vascular protective paradigm suggests that inhibiting VEGF will cause a loss of endothelial integrity and thus tend to enhance disease progression, whereas treatment with VEGF might promote healthy vascular function, but possibly through a non-angiogenic mechanism, and probably also only at a level insufficient to stimulate a robust angiogenic response.

Some light has been thrown onto these issues by the clinical use of anti-VEGF antibodies and VEGF receptor inhibitors to treat cancer. The humanised monoclonal anti-VEGF antibody, bevacizumab (Avastin), in combination with standard chemotherapy, is associated with a higher incidence of hypertension of any grade, including severe hypertension, compared with standard care alone.16–18 Indeed, hypertension is now such a well-recognised side effect of bevacizumab in patients with cancer that it has been suggested as a potential prognostic indicator of clinical outcome.19 Bevacizumab is also associated with a significant increase in the incidence of arterial thromboembolic events in several trials, though not venous thromboembolism.16 20 The small molecule receptor tyrosine kinase inhibitors of VEGFRs, sunitinib (Sutent) and sorafenib (Nexavar), are also associated with cardiovascular side effects such as hypertension, myocardial infarction and markers of cardiotoxicity, including cardiac ischaemia, abnormal echocardiography associated with decline in left ventricular ejection fraction and changes in ECG.21 22

The aetiology of hypertension and arteriothrombosis associated with anti-VEGF treatments remains unknown and is likely to be complex and varied, but one mechanism that may be involved is decreased levels of the VEGF-stimulated vasodilators, NO and PGI2. Reduced endothelial production of these factors would both promote hypertension and vasoconstriction, and would also remove or dampen an important endogenous anti-thrombotic mechanism that inhibits platelet aggregation. It is also possible that hypertensive effects of inhibiting VEGF may be secondary to impaired kidney function, as indicated by an increased incidence in proteinuria, resulting, in turn, from endothelial dysfunction and/or capillary rarefaction.23

Trials of bevacizumab and VEGFR inhibitors suggest that systemic inhibition of VEGF causes an increase in blood pressure, and some increased risk of arteriothromboembolic events, outcomes that would be predicted if VEGF were essential for maintenance of normal vascular function. These and other findings considered so far predict that delivery of VEGF would stimulate an angiogenesis in the ischaemic heart and periphery and could also, as a collateral benefit, promote healthy cardiovascular function, both with therapeutic potential. We turn now to consider what clinical trials have disclosed about the effects of angiogenic factors on cardiovascular disease.

Clinical trials of therapeutic angiogenesis

Although conventional revascularisation by percutaneous coronary intervention and coronary artery bypass graft, is an effective treatment for coronary artery disease, some patients are unsuitable for these procedures, and a substantial proportion experience incomplete revascularisation. Thus, there is an important unmet clinical need for additional treatments for ischaemic heart and peripheral disease. The aim of therapeutic angiogenesis is to use angiogenic factors to induce the formation of a collateral blood supply, effectively bypassing an occluded diseased blood vessel in patients with coronary or peripheral artery disease, thereby revascularising ischaemic or vulnerable myocardial or other tissue (figure 2).24

Figure 2

Therapeutic angiogenesis. The stimulation of angiogenesis or more precisely arteriogenesis is predicted to promote formation of collateral blood vessels (right panel), thus bypassing a thrombotic occlusion commonly in the coronary artery (left panel) and revascularising ischaemic cardiac tissue. Various routes of administration of therapeutic angiogenic cytokines can potentially be used as shown on the right.

Angiogenesis is a complex highly coordinated process, involving cooperation between an array of diverse factors acting in a spatiotemporally orchestrated manner. However, though many candidates for therapeutic stimulation of angiogenesis have been proposed,25 in practice clinical translation of therapeutic angiogenesis has focused largely on VEGF-A and, additionally, on other angiogenic growth factors. The therapeutic angiogenesis approach is supported by an impressive body of preclinical evidence showing improvements in blood flow, revascularisation and myocardial function after angiogenic cytokine delivery in different animal models of myocardial and limb ischaemic disease.1 To date, there have been more than 25 phase II or phase III clinical trials involving approximately 2500 patients, assessing the efficacy of angiogenic cytokine protein and gene therapy involving VEGFs, fibroblast growth factors (FGFs) and other factors in patients with ischaemic heart and peripheral disease (see tables 1–3).

Table 1

Clinical phase I/II randomised placebo-controlled therapeutic angiogenesis trials in coronary artery disease: protein therapy

Table 2

Clinical phase II/III randomised controlled therapeutic angiogenesis trials in coronary and peripheral artery disease: plasmid DNA gene therapy

Table 3

Clinical phase II/III randomized controlled therapeutic angiogenesis trials in coronary and peripheral artery disease: adenoviral gene therapy

Protein therapy

Phase I trials in patients with stable angina not suitable for conventional revascularisation established that intracoronary and intravenous recombinant VEGF-A165 (rhVEGF) infusion was safe and well tolerated in humans and suggested that treatment improved myocardial perfusion imaging and collateralisation.30 31 VEGF-induced hypotension was the major criterion for non-tolerability of dosing with higher concentrations,31 consistent with vasodilatory effects observed in preclinical studies.5 Pharmacokinetic analysis of rhVEGF protein delivery following infusion into the coronary circulation indicated rapid and extensive tissue uptake of rhVEGF consistent with binding to specific endothelial receptors and other binding sites such as heparan sulphate proteoglycans, followed by a more prolonged elimination phase, with a half-life after a 4 h intravenous perfusion at 50 ng/kg/min of approximately 34 min.30 32

The subsequent VIVA (VEGF in Ischaemia for Vascular Angiogenesis) phase II trial examined the effects of rhVEGF in 178 patients, randomised to receive a 20-min intracoronary infusion of placebo, low-dose rhVEGF (17 ng/kg/min) or high-dose rhVEGF (50 ng/kg/min) followed by 4 h intravenous infusions on days 3, 6 and 9, with assessment after 60 and 120 days. No significant differences in exercise treadmill time, angina score or in myocardial perfusion were demonstrated between treatment groups after 60 days.26 After 120 days, the marked placebo effect observed at 60 day follow-up had diminished and high-dose rhVEGF caused a significant improvement in angina class and non-significant trends to improvement in treadmill times and angina frequency compared with placebo. Despite these modest signs of efficacy, and the general safety and tolerability of the treatment, the VIVA trial was a setback for therapeutic angiogenesis, and, to date, no further large-scale trials of rhVEGF have been performed or appear to be planned.

Similarly disappointing results were obtained with other angiogenic protein treatments. While phase I/II trials of intracoronary FGF-2 infusion showed signs of therapeutic efficacy,27 a phase II trial in 337 patients revealed no significant improvement in either treadmill times at 90-day follow-up or in ischaemic areas with FGF-2 treatment versus placebo, though a significant improvement in angina class and frequency was seen, most prominently in sicker patients.28 Granulocyte/macrophage colony stimulating factor (GM-CSF), which promotes angiogenesis by mobilising monocytic cells, was investigated in 21 patients by intracoronary injection, followed after 2 weeks by subcutaneous GM-CSF administration. GM-CSF significantly improved collateral blood flow compared with placebo and caused a non-significant trend to improvement in ECG signs of myocardial ischaemia, but as yet no larger clinical studies have been conducted.29

Overall, the results of clinical trials of angiogenic cytokine protein treatment were disappointing (table 1). The large placebo effect, suboptimal dosing concentrations and schedules, ineffective route of administration, choice of inadequate or subjective end points, and use of an insufficiently differentiated target patient population are all possible explanations for the lack of efficacy of VEGF protein.33 However, perhaps the key limitations of protein therapy are the relatively short half-life of VEGF and other cytokines in blood, coupled with dose-limiting hypotension, which together may prevent exposure of the target tissue to VEGF at concentrations and for times sufficient to promote a viable and sustained improvement in collateral blood flow.

Gene therapy

Gene therapy is an alternative way of delivering therapeutic cytokines to achieve a more sustained therapeutic effect after a single application, obviating the need for repeat administrations of short-lived proteins, and avoiding the risks of short-term systemic exposure to relatively high concentrations of cytokines.25 34 Preclinical studies of VEGF and FGF gene delivery using plasmid and adenoviral vectors showed impressive improvements in heart function and perfusion in different animal models of myocardial ischaemia, such as the porcine ameroid model of stress-induced ischaemia.35 These encouraging data provided the impetus for several clinical gene therapy trials of angiogenic cytokines encoded either in plasmid or adenoviral vectors.36

Plasmid vectors

The Isner group first established the feasibility of VEGF gene therapy in phase I investigations of cDNA plasmids encoding VEGF-A165 (phVEGF165), delivered by direct intramyocardial injection via minimally invasive thoracotomy.37–39 These early studies, though demonstrating safety and with promising signs of clinical benefit, were small, lacked controls and also raised ethical concerns about the route of administration when used to treat with an unproven agent. Furthermore, use of a thoracotomy to gain access to the heart was found incompatible with a double-blind study design. A catheter-guided injection system was then developed which circumvents the need for more invasive methods and made a blinded design feasible. Placebo-controlled phase I/II clinical trials have provided proof of concept for this approach, and again showed significant improvements in myocardial perfusion and ischaemia compared with placebo control treatment.40 41

Two randomised, double-blind, placebo-controlled trials of phVEGF165, EUROINJECT-ONE42 and NORTHERN,43 involving, respectively, 80 and 93 patients with severe stable ischaemic heart disease, assessed the effects of plasmid delivered intramyocardially using a NOGA guidance catheter. Neither trial found a significant difference between the VEGF-treated and the placebo groups in the primary end point of change in myocardial perfusion, assessed by single photon emission computed tomography, this despite the use of a high plasmid dose (2 mg) in the NORTHERN study. Improvements in exercise treadmill time and anginal symptoms were observed, but with no significant differences between the VEGF and placebo groups. The larger Genesis trial of VEGF-C plasmid was terminated after enrolment of 295 out of a planned 404 patients, owing to complications associated with the catheter delivery system and the expectation that the results would be negative.36

Angiogenic factors have also been tested in phase II trials of plasmid DNA delivery for peripheral artery disease or critical limb ischaemia (CLI). Intramuscular injection of phVEGF-A165 produced no significant change in the primary objective of significant amputation reduction compared with the placebo control group in a study of 54 patients with CLI.44 The TALISMAN phase II trial of FGF-1 plasmid in 125 patients showed a significantly reduced risk (by twofold) of major amputation in subjects receiving FGF-1 compared with the placebo control group and a non-significant trend towards reduced mortality, though no differences were seen in the primary end point of ulcer healing between the treatment groups.45 This study is the basis for the phase III TAMARIS trial (Therapeutic Angiogenesis for the Management of Arteriopathy in a Randomised International Study) evaluating efficacy of FGF-1 plasmid versus placebo in 490 patients with CLI using the combined primary end point of amputation or death.45 A phase I/II trial (HGF-STAT) of intramuscular injection of hepatocyte growth factor in 104 patients with CLI demonstrated safety and tolerability but showed no differences in measures of clinical efficacy in the hepatocyte growth factor (HGF)-treated group compared with placebo.46 An indirect approach to stimulating angiogenesis therapeutically has examined the effect of plasmid encoding the matrix protein Del-1 (developmentally regulated endothelial locus 1), which promotes angiogenesis by upregulating the integrins αυβ5 and αυβ3. A phase II investigation in 105 patients with peripheral arterial disease of Del-1 plasmid, showed no significant differences in primary or secondary end points between Del-1 and control groups.47

Most phase II/III trials of therapeutic angiogenesis in coronary artery or peripheral arterial disease using plasmid vectors have yielded negative results (see table 2). Though easy to manufacture and generally considered safe, a major disadvantage of plasmid vectors is their inherently low transfection efficiency. Translation of the impressive therapeutic effects found in healthy animal models into the clinical situation has to overcome the formidable challenge of achieving similar effects in blood vessels and hearts that both have much larger tissue volumes, and are diseased, a factor that may further reduce DNA transfection efficiency. Even though therapeutic angiogenesis has employed mainly secreted diffusible cytokines, which theoretically achieve a territorially widespread ‘by-stander’ effect even if relatively few cells are transfected, it has been argued that the only way to attain levels of expression sufficient for a therapeutic effect in larger organisms is to use viral vectors able to infect a much larger proportion of cells in the target tissue.25 36


Though several types of viral vector are available, adenoviruses have so far been the preferred vector for studies of cardiovascular disease because they display a high transduction efficiency in endothelial cells and VSMC in culture and in vivo, they can be produced at high titres and can transduce quiescent or non-proliferating cells. Although concern has been raised about the safety of adenoviruses, the fact that they are expressed episomally without integration into the host genome, and therefore relatively transient, mitigates to some extent the fear that they may give rise to long-term and potentially harmful side effects. Furthermore, in the case of many serotypes, these viruses are only mildly pathological.

Intramyocardial AdVEGF121 delivery by intramyocardial injection via thoracotomy was shown to be safe and well tolerated in a phase I study of 21 patients, either given alone or as an adjunct to coronary artery bypass grafting.48 Another pilot study demonstrated that percutaneous delivery of AdVEGF121 using a NOGA-guided injection catheter was also safe and feasible.49 In the REVASC phase II trial 67 patients with untreatable angina were randomised to receive either AdVEGF121 by intramyocardial injection via minithoracotomy, or best medical treatment, with no placebo.50 A statistically significant increase in the primary end point, the exercise time to 1 mm ST-segment depression, was found in the AdVEGF121 group at 26 weeks, compared with controls. Other secondary end points, such as total exercise duration and time to moderate angina after 12 and 26 weeks, as well as angina symptoms, were also improved in the treatment group compared with controls. The REVASC trial subsequently led to the phase II NORTHERN trial of AdVEGF-A121 plasmid delivered by percutaneous injection catheter, which as mentioned above, did not provide evidence of clinical benefit.43 A phase II trial (RAVE) of Ad.VEGF-A121 in 105 patients with peripheral artery disease failed to demonstrate any improvement in peak walking time or in other measures of quality of life.51

The 103 patient randomised placebo-controlled Kuopio angiogenesis trial investigated the effect of AdVEGF165 on restenosis and myocardial perfusion compared with either plasmid VEGF or placebo, the vectors delivered by an infusion-perfusion catheter during angioplasty and stent placement. After 6 months, no significant difference was seen in the rates of restenosis between the treatment groups, but myocardial perfusion improved in the AdVEGF165 group compared with plasmid VEGF.52 In an 8-year follow-up study of patients in the Kuopio angiogenesis trial, the incidence of major adverse cardiovascular events, cancer or diabetes did not differ between the treatment groups.53 A similar comparison of adenoviral with plasmid delivery of VEGF-A165 in patients with CLI, demonstrated an improvement in the primary end point of vascularity, determined by digital subtraction angiography, with both Ad.VEGF-A165 and phVEGF-A165 compared with control, but no differences in secondary end points.54

The FGF family member, FGF-4, has also attracted considerable interest as a therapeutic angiogenic agent, based on very promising preclinical data. FGF-4 shares ∼40% homology with the more intensively studied angiogenic factors, FGF-1 and FGF-2, and unlike them contains a signal sequence that mediates its secretion. It has an essential role in embryonic development and stimulates endothelial cell function and angiogenesis in vivo.55 Studies in the pig ameroid constrictor model of stress-induced myocardial ischaemia, showed sustained improvement in coronary blood flow after a single Ad.FGF-4 injection.55 The Angiogenic Gene Therapy (AGENT-1/2) phase I/II trials of Ad.FGF-4 produced improvement in exercise treadmill times and a significant reduction in myocardial ischaemic areas, but a larger phase III trial in 116 patients (AGENT-4) did not replicate the earlier promising findings, and showed no significant increase in exercise tolerance tests.55 56 This and the AGENT-3 trial were halted when it became apparent that the primary end point—change in exercise treadmill time at 12 weeks—was not reached. The AGENT-3/4 trials of FGF-4 involved the largest group of patients (355) so far to receive a single angiogenic cytokine therapy (compared with 177 receiving placebo). The results of an analysis of data pooled from the AGENT-3 and -4 trials showed that the treatment was safe, but the only statistically significant beneficial effect was an improvement in Canadian Cardiovascular Society angina class, and this was mainly due to a response in the female participants, who constituted 14% of the total patient group.56 This gender-specific difference in response to treatment may be attributable to a more pronounced placebo effect in the male patients than in the female cohort.56 The AWARE (Angiogenesis in Women with Angina pectoris who are not candidates for REvascularisation) trial in 300 women with stable angina is examining the possibility that women may preferentially derive benefit from Ad.FGF-4 treatment, and is ongoing.55


After treatment of >2500 patients, what do trials of therapeutic angiogenesis tell us about the impact of angiogenic cytokines on human cardiovascular disease? First, no indications have been found of any long-term effect of VEGF protein or gene delivery, in some cases up to several years after the initial treatment, on cardiovascular disease, cancer or diabetes. These trials therefore suggest strongly that VEGF and other pro-angiogenic treatments do not promote or accelerate atherosclerosis and its clinical consequences. While this will certainly not be the last word on arguments about the role of angiogenesis in atherosclerosis, it suggests that VEGF-dependent angiogenesis is unlikely to be the crucial tipping-point in plaque rupture leading to myocardial infarction. Angiogenesis may well be important for plaque pathogenesis during the earlier stages of lesion growth, though it is also plausible that angiogenic stimuli may have a stabilising role in the mature plaque. The second major conclusion to emerge from therapeutic angiogenesis trials is that no large clinical trial has yet shown a substantial benefit for cardiovascular disease, and judged from this standpoint, the findings from these trials have been disappointing, and the early promise of therapeutic angiogenesis shown in preclinical studies is yet to be translated to the clinic.

Is the lack of clinical benefit due to the failure of current approaches to produce the desired biological response—namely, angiogenesis, or to stimulate it sufficiently; or is it that the angiogenic modality—that is, the type of vessels, is not therapeutically optimal; or, more pessimistically, does the treatment do what the hypothesis predicts, but the theory is simply wrong? Given the impressive body of preclinical evidence supporting the concept, the most judicious view would seem to be that the lack of striking effects in trials of angiogenic cytokine therapy is due to shortcomings in existing approaches, and that progress in translation to the clinic will follow advances in delivery, expression and biological efficacy of angiogenic stimuli.

Perspectives on the future of therapeutic angiogenesis

The main obstacle to the successful application of a therapeutic angiogenesis approach to ischaemic disease remains an inability to deliver an effective angiogenic stimulus to the ischaemic human heart. The safety of angiogenic cytokines, particularly in the long term, is still a cause for concern, but the good safety record of VEGF and FGF gene and protein delivery in patients with heart disease indicates that this is not the major obstacle to clinical application of angiogenic cytokines it was originally feared to be. Furthermore, recent evidence supports the view that, at least for cardiovascular health, a lack, rather than abundance, of VEGF, is the more likely to cause adverse effects.

The best way forward for therapeutic angiogenesis may be a gene therapy approach, involving a gene or genes encoding a secreted, biologically potent, angiogenic factor, delivered in a viral vector affording high transduction efficiency. However, a major limitation of all current approaches is the relatively low transduction efficiency of available vectors in human target tissues. It is estimated that using the most effective vectors available, approximately 5–10% of the target cell population in humans can be transduced, considerably less than levels attainable in the smaller mammalian species generally used to generate preclinical data.25 Another more difficult problem, lies in achieving the optimal balance between, on the one hand, the longevity of expression necessary to sustain a therapeutic biological response, and on the other, the desirability of avoiding indefinite expression of cytokines with unpredictable long-term consequences. Though there is considerable activity in the area of vector platform technology, at present there do not appear to be easily translatable solutions to these problems, at least in clinical trials for cardiovascular diseases.

While VEGF-A remains probably the best hope for achieving a therapeutically effective angiogenic response in the heart, the disappointing results of clinical trials with VEGF and other cytokines as monotherapies suggest that single cytokine treatments may be insufficient to promote therapeutically effective collateral growth or angiogenesis. Treatments more specifically tailored to stimulate collateral arteriogenesis as opposed to capillary formation might involve combinations of cytokines, performing complementary functions in arteriogenesis, such as endothelial tubulogenesis, combined with pericyte recruitment and VSMC proliferation and envelopment. At the moment, combination therapy for cardiovascular disease has been little explored in the preclinical setting. The transfer of AdPDGF-B alone and in combination with AdVEGF-A induced prominent proliferation of alpha-smooth muscle actin-, CD31-, RAM11-, HAM56- and VEGF- positive cells.57 Combination gene transfer induced a longer-lasting increase in perfusion in both intact and ischaemic muscles than AdVEGF-A gene transfer alone, indicating that the combined treatment caused a more prolonged angiogenic response, though these effects may have resulted from paracrine mechanisms rather than an increase in vascular pericyte or VSMC coverage. An attempt has also been made to combine an angiogenic cytokine with a factor which mobilises stem cells.58 Ripa et al investigated the effect of intramyocardial injection of phVEGF-A165 plasmid followed by treatment with G-CSF in 16 patients, with two control groups of 16 patients each receiving VEGF-A165 plasmid alone, or placebo. After 3 months, the combined treatment appeared safe, but no improvement was seen either in myocardial perfusion using single photon emission computed tomography, or in clinical symptoms.59

An alternative approach to stimulating multiple angiogenic pathways is to manipulate the mechanisms regulating angiogenic cytokine expression. The most important of these is hypoxia inducible factor-1α (HIF-1α), the key mediator of hypoxia-induced VEGF-A expression. A phase I dose-escalation study of an adenoviral vector expressing constitutively active HIF-1α (Ad2.HIF-1α/VP16) in 34 no-option patients with CLI demonstrated safety and tolerability,60 and has been the basis for an ongoing phase II randomised, double-blind, placebo-controlled study in 300 patients with severe intermittent claudication (the WALK trial).61 Recently, the transcriptional coactivator, peroxisome proliferator-activated receptor-γ-coactivator 1α (PGC-1α), was identified as a potentially important mediator of angiogenesis in ischaemic muscle. Like HIF-1α, PGC-1α is a sensor of oxygen and nutrient deprivation and can upregulate VEGF expression, but also promotes PDGF-BB, the key promoter of VSMC/pericyte recruitment to growing vessels, and angiopoietin-2, which facilitates capillary sprouting.62 A limitation of delivering genes encoding transcription factors is that the intracellular mode of action demands a high level of transduction efficiency. However, identification of PGC-1α as an important coordinator of the angiogenic response to ischaemia offers a potentially powerful new approach to therapeutic angiogenesis.

Despite the failure so far to fulfil the earlier, perhaps overly simplistic, hopes for therapeutic angiogenesis, the cardiovascular research and clinical communities should resist the ebbing away of enthusiasm for new molecularly targeted therapeutic strategies for heart disease. It is essential that efforts continue to bring about the technical innovations necessary to translate these approaches to the clinic and develop others; in this regard, it is encouraging that there continues to be considerable interest in using angiogenic stimuli therapeutically, with trials still ongoing and others planned or recently initiated.


View Abstract


  • Funding British Heart Foundation, Ark Therapeutics Ltd.

  • Competing interests IZ is a consultant to Ark Therapeutics Ltd, a company with an interest in developing gene therapy approaches for cardiovascular disease.

  • Provenance and peer review Commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.