Angiogenesis: Bench to Bedside,Have We Learned Anything?

Abstract

End-stage ischemic cardiomyopathy patients are an ever-increasing group of coronary artery disease patients, often with no options in our current treatment armamentarium. Angiogenesis therapy pre-clinical and phase I clinical trials showed great promise, however, the benefits of single growth factor treatments have not been borne out in the larger phase II randomized trials. The complexity of angiogenesis process and the challenges in creating animal models to replicate and study this process in ischemic adult human myocardium have been major limitations to progress in this field. In addition failure to control for the powerful placebo effect in the clinical trials and inadequate methods of outcomes measures assessment have created difficult to overcome road blocks in establishing the efficacy of angiogenic strategies. Herein we review the challenges of angiogenesis research and development of treatment strategies. We also propose a structured model for further investigations of angiogenic therapies. The adherence to such a regimented approach as proposed here is, in our opinion, the only way to achieve success in angiogenesis approach development to treatment of patients with end-stage cardiac ischemia refractory to other established therapies.

The magnitude of the problem

Despite advances in preventive health care, medical management, interventional cardiology, and cardiovascular surgery, atherosclerotic disease remains the leading cause of morbidity and mortality in the Western Hemisphere. Cardiovascular disease accounted for 38.5% of all deaths or 1 of every 2.6 deaths in the United States in 2001. Mortality due to cardiovascular causes was about 60% of “total mortality,” i.e., out of 2,400,000 deaths from all causes, and was listed as a primary or contributing cause of death of about 1,408,000 death certificates. Since 1900 cardiovascular disease has been the number one killer in the United States every year except 1918 (AHA, 2004). Treatment of coronary artery disease includes risk factor modification, use of anti-platelet agents, medical therapy (decreasing myocardial oxygen demand and coronary vasodilatation), and restoring myocardial perfusion using percutaneous coronary interventions (PCI) or coronary artery bypass grafting (CABG).

Although significant advances have reduced the mortality of cardiovascular disease, the number of cardiac interventions continues to grow: A total of 1.3 million inpatient cardiac catheterizations, 561,000 PTCA procedures, and 519,000 coronary artery bypass procedures were performed in 2000 in the US alone. This is due to the progressive nature of atherosclerotic disease and the fact that effects of many cardiovascular procedures are not permanent. Finally, the cost of cardiovascular disease and stroke treatment in the United States in 2004 is estimated at $368.4 billion. This figure includes health expenditures and lost productivity resulting from morbidity and mortality.

In addition, ischemic heart disease remains the leading cause of congestive heart failure (CHF), which has reached epidemic proportion in the United States. Based on the 44-year follow-up of the NHLBI’s Framingham Heart Study CHF incidence approaches 10 per 1,000 population after age 65 with 22% of male and 46% of female patients with myocardial infarction becoming disabled with heart failure. Hospital discharges for CHF rose from 377,000 in 1979 to 995,000 in 2001 (AHA, 2004).

A significant number of patients (5–21%) with ischemic heart disease are not optimal candidates for revascularization (PCI/CABG) or receive incomplete revascularizations with these procedures (McNeer et al., 1974; Jones et al., 1983; de Feyter, 1992; Mukherjee et al., 1999). Many of them have residual angina and congestive heart failure symptoms despite maximal medical therapy. Thus an alternative treatment strategy is needed, and therapeutic angiogenesis may play that role by providing new avenues for increasing blood supply to ischemic myocardium (Isner and Feldman, 1994; Isner, 1996, 1997; Isner et al., 1996; Isner, 1997; Laham et al., 1999a, 1999b; Laham and Simons, 1999; Laham et al., 2001b, 2002d; 2000d; Laham and Simons, 2000; Laham et al., 2001a, 2001b, 2002; Laham and Oettgen, 2003). Similarly, a significant number of patients with peripheral vascular disease (5%) have residual symptoms despite medical and surgical therapy and may benefit from treatment with angiogenesis factors (Bauters et al., 1994; Isner and Feldman, 1994; Bauters et al., 1995; Isner, 1996; Isner et al., 1996; Isner, 1997) Furthermore, congestive heart failure is a progressive disease resulting from irreversible myocyte loss, and may potentially benefit from strategies involving myocyte regeneration, i.e., myogenesis. The lack of sufficient organs for transplantation and the slow development of mechanical assist devices make myogenesis the major therapeutic option for these patients, particularly that implantable cardioverters-defibrillators while improving survival had no positive effect on their quality of life.

However, it is important first to define the target patient population for such therapies, patients more commonly known as “no-option” patients. For angiogenesis application, one study of 500 patients yielded 59 patients (12%) who were considered ineligible for PCI/CABG (Mukherjee et al., 1999; Hennebry and Saucedo, 2004; Rosinberg et al., 2004) However, wide regional and institutional variability in treatment patterns of coronary disease including more or less aggressive revascularization practices contributes to different estimates of the magnitude of the problem. No-option patients have been estimated to constitute anywhere between 5 to 21% of all patients with coronary artery disease. The most common reasons for residual unrevascularized yet ischemic myocardial territories are the following: recurrent restenosis (less frequent with the introduction of drug eluting stents), prohibitive expected failure, chronic total occlusion, poor targets for CABG/PCI, saphenous graft total occlusion with patent left internal mammary artery graft, degenerated saphenous vein grafts (less frequent with introduction of distal protection), no conduits or calcified aorta, and co-morbidities such as renal failure, cancer, and cerebrovascular disease.

Current management strategies for these patients are limited. Patients are often maintained on a cocktail of medications to help control symptoms including antiplatelet agents, nitrates, beta-blockers, angiotensin converting enzyme inhibitors, angiotensin II receptor blockers, and calcium channel blockers. Many of these patients continue to have symptoms despite maximal medical therapy. The treatment options are also a moving target since advances in interventional and surgical techniques have helped improve the quality of life. Most notably, the development of drug (sirolimus and paclitaxel) eluting stents has all but solved the problem of recurrent restenosis (Grube et al., 2002; Doggrell, 2004; Grube and Buellesfeld, 2004; Hoye et al., 2004; Waugh and Wagstaff, 2004). The target population for myogenesis application is easier to define since it includes patients with Class III and IV New York Heart Association (NYHA) symptoms refractory to currently available treatment options including angiotensin converting enzyme inhibitors, beta-blockers, diuretics, venodilators and inotropes.

The need to develop treatment strategies led to extensive preclinical testing of various agents, vectors, and cells to achieve therapeutic angiogenesis and myogenesis, rapidly followed by clinical investigations. This paper focuses on angiogenesis. It seems that the same mistakes committed during the development of angiogenesis strategies are being repeated in the field of myogenesis. Driven by the hype and the need for rapid development of these strategies, the scientific basis and careful controls are often lacking leading to misleading results and consequently resulting in negative phase II trials.

Therapeutic Angiogenesis

Angiogenesis is a complex process that involves multiple steps of endothelial cell proliferation and migration, extracellular matrix breakdown, attraction of pericytes and macrophages, followed by smooth muscle cell proliferation and migration, formation and “sealing” of new vascular structures, and finally deposition of new matrix (Folkman, 1998a, 1998b; Laham et al., 1999b, 2001b). It is likely that a coordinated action of several mitogens, cascades, and inhibitors is needed to achieve this process. Gradual occlusion of coronary arteries is frequently associated with development of collateral circulation in patients with atherosclerosis (Figure 1). Although the existence of the collateral circulation in such patients is associated with improved clinical outcomes, the large number of revascularization procedures performed attests to the inadequacy of native collateralization in most patients. Myocardial ischemia is a potent angiogenic stimulus and a number of growth factors and chemo-attractants have been shown to increase during ischemia suggesting that these molecules may play a role in ischemia-induced angiogenesis. Among these growth factors, fibroblast growth factors [acidic (aFGF) and basic (bFGF)] and vascular endothelial growth factor (VEGF) are the most widely known and studied (Laham et al., 1999c, 1999d, 2003, 2005).

First it is important to define angiogenesis. Although often referred to as angiogenesis, the process of neovascularization can occur via 3 different mechanisms: vasculogenesis, angiogenesis and arteriogenesis. Vasculogenesis is the formation of new vascular structures from stem cells during embryogenesis and may contribute to adult neovascularization (Asahara and Isner, 2002). Angiogenesis refers to the formation of thin-walled endothelium lined structures lacking a smooth muscle layer from preexisting vessels (sprouting from post-capillary venules). For example, angiogenesis is the manner by which capillaries proliferate in healing wounds and along the border of myocardial infarctions. Arteriogenesis is the formation of vessels with a complete smooth muscle wall as seen in the development of angiographically visible collaterals in patients with advanced obstructive arterial disease. Arteriogenesis is believed to result from remodeling of existing collateral vessels as well as formation of new vessels (not well established). In addition, significant variability has been observed in intrinsic angiogenesis in response to ischemia between patients, with some having robust collaterals while others having none.

A number of growth factors, gene therapy strategies and cell-based therapies have been evaluated for their angiogenic potential. These included fibroblast growth factors, vascular endothelial growth factors, hepatocyte growth/scatter factor (HGF/SF), and chemokines such as IL-8 and MCP-1. In addition growth factors involved in maturation of vascular tree such as angiopoietins and platelet derived growth factor (PDGF), as well as transcription factors that stimulate expression of angiogenic cytokines and their receptors such as hypoxia-induced factor (HIF)-1α, and cell-based approaches have been studied. The first two to be assessed were FGF-2 and VEGF₁₆₅.

VEGF is a heparin-binding glycoprotein, which acts as a mitogen for vascular endothelial cells as well as stimulates endothelial progenitor cell mobilization. VEGF is expressed in cardiac myocytes and vascular smooth muscle cells, with increased expression in the setting of vascular injury, acute and chronic ischemia, and hypoxia (Tofukuji et al., 1998). VEGF administration resulted in functionally significant angiogenesis in animal models of ischemia (Harada et al., 1996; Lopez et al., 1998) but caused hypotension (nitric oxide dependent vasodilatation), limiting the maximally tolerated dose. Similarly FGF-2, which is present in the normal myocardium, and becomes up-regulated in response to stimuli of hypoxia and hemodynamic stress. FGF-2 is a pluripotent mitogen able to induce migration and proliferation of a variety of cell types including endothelial cells, smooth muscle cells macrophages and fibroblasts. It stimulates endothelial cells to produce a variety of proteases, including plasminogen activator and matrix metalloproteinases, and induces significant vasodilatation through nitric oxide release as well as promotes chemotaxis. FGF-2 was shown to induce functionally significant angiogenesis in a variety of animal models using various delivery strategies (Battler et al., 1993; Laham et al., 2000a; Sato et al., 2000; Ruel et al., 2003). Several other cell transplantation, growth factors delivered as proteins and gene transfer strategies were all studied in animal models with encouraging results. However, one suspects possible publication bias with positive studies being preferentially published.

Clinical Investigation

The promising preclinical results and the hype surrounding the prospect of growing new blood vessels to the heart heralding the era of biologic therapy propelled the investigation of several agents for angiogenesis, with some trials based on a small amount of data and a large leap of faith.

Several of the studies were phase I open label studies with small numbers of patients. Almost all open-label studies showed both good safety (below the MTD) and evidence of improvement in most patients which was considered, by the scientific and lay community, as sufficient proof of efficacy. The three major phase I studies which were adequately powered randomized placebo controlled studies will be discussed here.

Several phase I trials of safety and tolerability of VEGF delivered as a protein or gene therapy were carried out. VEGF was infused using the intracoronary and intravenous routes in a dose escalation study (Hendel et al., 2000; Henry et al., 2001). It resulted in a dose dependent hypotension limiting the maximal dose. Most patients improved not only symptomatically, but they also had evidence of improved perfusion (Hendel et al., 2000; Henry et al., 2001). However, when this strategy was studied in a randomized, double-blind, placebo controlled study of 178 patients, the VIVA study (Henry et al., 2003), the improvements in symptoms, quality of life, and exercise time between treated (high and low dose) and control patients was equivalent.

For FGF-2, the first study was a randomized placebo-controlled study of 24 patients undergoing CABG with a viable but ungraftable myocardial area. Heparin alginate beads (for sustained delivery over 3 weeks) were loaded with FGF-2 (10 and 100 μg) or vehicle and implanted in ischemic area (Sellke et al., 1998; Laham et al., 1999d). Follow-up averaged 16 months with clinical assessment and nuclear perfusion imaging. At 90 days, there was a significant improvement in symptoms and reduction in nuclear defect size using nuclear perfusion scans and magnetic resonance imaging in the 100 μg group compared to controls. However, with longer follow-up, patients in the control group caught up to the treated group underscoring the potency of the intrinsic collateralization process (Ruel et al., 2002). Polymer based delivery though promising was difficult to implement and the CABG -plus design limited the ability to investigate stand alone therapy. Therefore, FGF-2 was studied using intracoronary and intravenous administration in no-option patients as stand-alone therapy. Fifty-two patients with coronary artery disease and inducible ischemia who were deemed suboptimal candidates for either PCI or CABG received intracoronary FGF-2 at doses ranging from 0.33 μg/kg to 48 μg/kg (Laham et al., 2000c). At 6-month follow-up, there was a significant improvement in quality of life measures, exercise time, target wall thickening and motion on magnetic resonance imaging, and defect size on nuclear imaging (Laham et al., 2000c; Udelson et al., 2000).

The lack of a control group and the open label design of the study precluded conclusions as to the efficacy of the treatment. However, marked improvement in physiologic measures such as exercise capacity, regional left ventricular function, and ischemic areas was interpreted as equating efficacy. As with VEGF, The “FIRST” was a multicenter, randomized, double-blind, placebo-controlled phase II study designed to examine the safety, pharmacokinetics, and efficacy of FGF-2. A total of 337 patients who were poor candidates for percutaneous or surgical revascularization were randomized to treatment with 0, 0.3, 3, or 30 μg/kg doses of FGF-2 by intracoronary route. There were no significant differences in exercise tolerance testing (ETT, primary endpoint), quality of life, or nuclear perfusion imaging defect size between the treatment and control groups at 180 days, with all groups showing improvements (albeit less than that seen in phase I studies; Simons et al., 2002; Laham, 2005).

Therapeutic angiogenesis using gene therapy approaches did not fare much better. Small open-label studies increased the hype with 5–20 patient studies presented as evidence of efficacy. The Angiogenic Gene Therapy trial (AGENT) was a “blinded,” phase I/II trial using intracoronary infusion of increasing doses of adenovirus encoding for FGF-4. Seventy-nine patients were randomized to receive either placebo or 1 of 5 doses of Ad5-FGF-4, resulting in very small numbers in each group. (Laham et al., 2001a; Grines et al., 2002) At 12-week follow-up, exercise tolerance was not significantly increased in the test groups over placebo. However, subgroup analysis of patients with initial ETTs of 10 minutes or less did show an improvement (Grines et al., 2002). However, this pivotal study stopped enrollment when an interim analysis of half of the 450 projected patients showed no possibility of detecting efficacy. Thus, to date, most agents studied were shown to induce functionally significant angiogenesis in animal models of myocardial ischemia and have been safe in pilot clinical studies with observed benefits. These benefits were no longer sustained in larger controlled studies (Rosinberg et al., 2004; Laham, 2005).

Why the Failures and Disapointments

The field of angiogenesis (proangiogenesis for heart disease and antiangiogenesis for cancer) has followed the usual path of novel technology: incredible results→unrealistic expectations →sobering disappointments →cautious optimism. This is despite the scientific community following the “righteous path” for any development strategy (Figure 2): understanding the biology →developing therapeutic agents, vectors, and animal models →site specific delivery →adequate outcome measures (Laham, 2005). Why then the failures and disappointments despite promising initial results? First, angiogenesis is a complex process that requires the concerted, sequential, and sustained action of multiple growth factors, angiogenesis inhibitors, and modulators. It is naïve and simplistic to think or expect that a single growth factor delivered for a short duration, is capable of resulting in a sustained angiogenic response. In addition, these growth factors are already present in the ischemic myocardium, and several lines of evidence suggest that advancing age, diabetes, elevated cholesterol, and their resultant endothelial dysfunction (contributors to coronary artery disease), all result in impaired angiogenic response (Xu et al., 2001; Nisanci et al., 2002; Ruel et al., 2003). In one of the studies a porcine model of chronic myocardial ischemia, animals were fed either a high cholesterol or normal diet. Four weeks after placement of an ameroid constrictor on the left coronary circumflex artery, FGF-2 loaded in heparin alginate beads for slow release was implanted in the circumflex territory. The hypercholesterolemic group showed significant endothelial dysfunction and impaired angiogenesis manifest as decreased circumflex perfusion compared to the control, normal diet group (Ruel et al., 2003).

This brings up the choice of the appropriate animal models. Most preclinical studies to date have been performed in juvenile animals. The predominant animal models for angiogenesis are the porcine, rabbit, and murine models. The most commonly used model for cardiac angiogenesis research is the porcine ameroid constrictor model. Juvenile Yorkshire pigs have been traditionally used. These animals are pre-pubertal, normocholesterolemic, and have normal endothelial function unlike our patients who are older, diabetic, hypercholesterolemic, and have endothelial dysfunction. This does explain why agents that have been shown to be effective in preclinical studies were not effective when studied in clinical trials. In addition, effective site specific delivery remains a major limiting factor, given the limited and short lived myocardial distribution and retention with most studied strategies. Intravenous and intracoronary delivery have limited myocardial tissue distribution and retention. The majority of the growth factor delivered intravenously is taken up by the liver (Laham et al., 1999c). Intrapericardial delivery resulted in improved myocardial distribution and retention, but endocardial penetration was poor (Laham et al., 2003b). Intramyocardial delivery resulted in the best myocardial deposition and retention (still less than 20%), with injections localized to administration site (Laham et al., 2005). What is clear is that prior to studying a specific agent, cell, or vector, it is crucial to optimize the delivery strategy and determine the best route of administration for that specific agent.

Another major caveat is that the outcome measures that have been used to assess angiogenesis strategies in patients have been adopted from cardiology and cardiac surgery studies, and may not be sensitive enough to detect the small changes seen with therapeutic neovascularization. In animal studies, the improvement in blood flow in the ischemic territory has been ~20–40%, far less than that seen with angioplasty or bypass surgery. Thus the outcome measures designed for these revascularization strategies cannot be used without being modified. For example, nuclear perfusion scans having a spatial resolution of ~8–10 mm cannot be expected to detect a small tissue-level increase in perfusion seen with angiogenic strategies, even though these small changes may be what is needed to improve the quality of life of these “no-option” patients. The development of novel outcome measures capable of assessing angiogenesis is as essential as developing angiogenic agents. Typically, outcome measures consist of hard endpoints (such as death, myocardial infarction, stroke, and recurrent ischemia—MACE) and soft endpoints (such as angina class and quality of life measures). Hard endpoints are preferable for any investigation, however, the rarity of these events even in high-risk “no-option” patients would necessitate large scale studies. Independent assessment and use of validated questionnaires make these softer endpoints slightly more objective. Surrogate endpoints play an important part in reducing numbers of patients needed for preliminary efficacy and in providing a mechanistic insight into the treatment. These include exercise assessment, nuclear perfusion scans (SPECT and PET), magnetic resonance functional and perfusion imaging, multidetector computer tomography, and echocardiography. Magnetic resonance imaging, in particular, is very promising (Pearlman et al., 1997; Laham et al., 1998; Pearlman et al., 2002).

Finally, appropriate study design is essential. Small open-label studies can be used for safety and tolerability, but no efficacy data should be reported or claimed. The placebo effect is very powerful in patients with end stage heart disease and is not only associated with improvement in symptoms, but also improvement in physiologic measures such as exercise time and perfusion scans (Rana et al., 2005). This placebo effect is only sustained at up to two years of follow-up (Rana et al., 2005). Thus, adequately powered (to avoid chance findings seen in small studies), randomized, double-blinded, placebo-controlled studies are essential before any claims of efficacy can be made.

Can Angiogenesis be Achieved and How

Therapeutic angiogenesis can and will be achieved. However, this can only be done with a robust translational model and several basic principles:

1. Single growth factors and short-term delivery are unlikely to result in functionally significant angiogenesis. To that purpose, sustained delivery of multiple growth factors, Master switch molecules (HIF-1α, RTEF-1; Deindl et al., 2001; Shie et al., 2004), cell-based therapy (Laham and Oettgen, 2003), and microtissue transplantations are all candidate agents.

2. Optimized organ specific and agent specific delivery strategy. Prior to preclinical and clinical studies, the optimal delivery modality should be studied to avoid ineffective delivery and agent inactivation (e.g., adenoviral vector inactivation with catheter delivery; Marshall et al., 2000). Only after a systematic evaluation of best available delivery modality is performed, should further development proceed.

3. Cell-based strategies and tissue transplantation should be performed only after ensuring adequate survival and confirmed treatment specific trans-differentiation.

4. Animal models should reflect disease state. Preliminary results in juvenile normal animals should be confirmed in adult, aging, hypercholesterolemic animals. In addition, the preclinical studies should be adequately powered to account for the marked variability in most animal models. In addition, multiple outcome measures should be assessed in preclinical studies to avoid a chance finding.

5. Well-designed studies and outcome measures:

Only by following a strict and well thought-out translational development scheme can angiogenesis as a treatment for cardiovascular disease become a reality. Our proposed scheme include the following set of investigations for any pro-angiogenic candidate: (Figure 3)

1. Understanding biology of proposed agent with mechanism of action and role in physiologic and pathologic angiogenesis

2. In vitro testing with endothelial cell proliferation (cell lines followed by primary human endothelial cells), migration (wounding assay), and tube formation (matrigel plates) (Figure 3A, B)

3. In vivo testing in matrigel mouse plug model (Figure 3C, D)

4. In vivo testing using murine myocardial infarction or hind limb ischemia model assessing infarct or limb salvage, perfusion, and endothelial cell density (CD31 staining). (Figure 3E, F)

5. Delivery optimization in large animal model (Figure 3G, H)

6. Preclinical study in juvenile normal animals for proof of concept

7. Preclinical study in adult or aging diseased animals

8. Pilot Phase I clinical study for safety and tolerability assessment without efficacy assessment

9. Phase II, randomized, double-blind, placebo-controlled study with adequate outcome measures and mechanistic endpoints.