by Ellis Unger and Stephen E. Epstein (Senior Research Investigator and Chief,respectively, Cardiology Branch, NHLBI)
Richard Lower provided the first description of coronary collaterals in humans in 1669, when he noted that fluid injected into one coronary artery emerged from another. We now know that coronary collaterals exist at only a rudimentary stage of development in the normal heart; however, with the hemodynamic alterations and/or metabolic derangements that accompany occlusive coronary artery disease, their development is stimulated, and they provide an alternative source of myocardial perfusion to the territory of the impeded artery.
The presence of functional collaterals enables the heart to maintain essentially normal perfusion under resting conditions, despite near-total occlusion of a major coronary artery. During the stress of exercise or the infusion of a vasodilator, however, the limitations of these vessels become apparent: vasodilator reserve is attenuated, maximal perfusion is curtailed, and myocardial ischemia results if the metabolic demands of the heart are unmet. The importance of coronary collaterals is underscored by the demonstration that the degree of myocardial dysfunction that develops in patients following acute occlusion of a coronary artery (during balloon angioplasty, for example) is inversely related to the extent of collateral vessels present (1). Moreover, myocardial viability in patients with recent myocardial infarction (MI) is correlated with the extent of collateral blood flow within the territory of the infarct-related artery (2).
Given the clinical importance of collaterals, two important questions emerge: what factors lead to their development, and what is the possibility that their growth can be enhanced pharmacologically? These questions have been a focus of the Physiology and Pharmacology Section of the Cardiology Branch, and based on our investigations in an animal model, we have now concluded that coronary collateral growth can be stimulated by using angiogenic peptides, at least in experimental animals (3,4).
"Angiogenesis" refers to the growth and/or development of blood vessels. In 1971, in the course of exploring the mechanisms responsible for vascular growth in malignant tumors, Judah Folkman and his colleagues discovered a diffusible substance that stimulated the growth of blood vessels (5). Subsequently, Schaper and co-workers provided support for the concept of a diffusible mediator of collateral growth in the heart (6). In 1985, Vallee et al. were the first to purify a human angiogenesis factor (angiogenin), and shortly thereafter, several other angiogenic peptides were isolated (see below).
The stimuli leading to the synthesis, release, and activation of angiogenic factors are probably multiple. Evidence suggests that ischemia itself can provide adequate stimulus for vascular growth; however, a competing (but not mutually exclusive) theory is that mechanical or hemodynamic factors initiate collateral development. In the normal heart, there may be many immature, undeveloped anastomoses between coronary arteries that carry little, if any, perfusion in the absence of a pressure gradient driving flow across them. With the development of proximal obstruction in one of the coronary arteries, there is the simultaneous development of a pressure gradient between the nonobstructed and obstructed arteries, inducing flow across the collateral vessel (Fig. 1). With the increase in flow, augmented collateral shear stress and tangential wall stress (stretch) may be important triggers in the initiation of angiogenesis (7). We think it is likely that both metabolic and hemodynamic factors are responsible for initiating and maintaining collateral development, and the cellular, biochemical, and molecular events that transduce these influences into vascular growth are currently under intense investigation.
In early 1985, we embarked on a series of experiments in which we hoped to facilitate the development of coronary collaterals by using angiogenic substances. We thought it likely that growth factors could serve as endogenous biological "distress signals," initiating the angiogenic response. Such mediators might be synthesized or released by the vascular wall and/or the myocardium in response to the development of significant coronary artery occlusions. Given the imperfect nature of collaterals -- that is, their inability to provide adequate perfusion under conditions of stress -- we then hypothesized that exogenous administration of such mediators, adding to endogenous stores, might enhance angiogenesis and improve myocardial perfusion. Initially, these hypotheses were difficult to test because 1) the angiogenic polypeptide growth factors had not been well characterized, 2) there was no practical method for administering growth factors to the heart, and 3) the assessment of biological endpoints relevant to angiogenesis was difficult.
Several polypeptide growth factors were eventually characterized and purified, including acidic fibroblast growth factor (FGF), basic FGF, vascular endothelial growth factor (VEGF), insulin-like growth factor I, scatter factor (hepatocyte growth factor, a glycoprotein), and others. Acidic FGF, basic FGF, and VEGF became available in large quantities through recombinant-DNA technology, making possible studies of their effects on coronary angiogenesis.
Our recent studies have focused on two peptides: basic FGF and VEGF. Basic FGF is the most extensively characterized member of the FGF family, a group of angiogenic heparin-binding polypeptides. This growth factor is produced by diverse cell types, including endothelial cells and cardiac myocytes. Basic FGF stimulates the proliferation of cells of mesodermal and neuroectodermal origin, targeting vascular endothelial cells, fibroblasts, smooth muscle cells, neuroblasts, osteoblasts, and melanocytes (8,9). VEGF is an angiogenic dimeric peptide with sequence homology to the A and B chains of platelet- derived growth factor (10). It was independently isolated as (and is identical to) vascular permeability factor (VPF). Its potency as an inducer of permeability exceeds that of histamine by orders of magnitude in some systems. It is functionally similar to basic FGF in its ability to stimulate the proliferation of vascular endothelial cells and induce angiogenesis; however, it differs from basic FGF in that its trophic effects are specific for endothelial cells.
Initial published reports suggested that the half-life of basic FGF was approximately 2 min. after intravascular administration. On the basis of this observation, we believed it necessary to administer basic FGF and other growth factors on a continuous basis by direct infusion (11) or via a sustained release mechanism (12) in order to achieve and maintain adequate tissue concentrations. Subsequently, studies by Vlodavsky et al. (13) demonstrated that basic FGF is sequestered by glycosaminoglycans in the extracellular matrix, the latter serving as a repository for the peptide. More recently, we found that the elimination half-life of pharmacologic doses of basic FGF was on the order of 60 min., not 2 min. as previously reported. These observations suggested to us that intermittent injection of large boluses of basic FGF could yield therapeutic concentrations in tissue, knowledge that vastly simplified our experimental models.
Our primary interest has been to foster collateral growth in patients with chronic myocardial ischemia. Accordingly, our animal models have involved the gradual occlusion of one or more coronary arteries, mimicking the pathophysiology of occlusive coronary artery disease in humans. Thus, we have implanted ameroid constrictors on the proximal left circumflex coronary artery (LCX) of dogs (Fig. 1). In the course of 10 - 20 days, these devices cause progressive arterial compression and ultimately, thrombotic occlusion. Dogs have a natural tendency to develop collaterals under these circumstances (infarcts are small and tend to be the exception rather than the rule), and the goal of our studies has been to promote the development of these collaterals.
In recent studies in which we assessed the effects of basic FGF and VEGF, the polypeptides (or placebo) were injected as a daily bolus directly into the LCX at a point just distal to the obstruction (3,4). Collateral blood flow to the LCX territory was quantified on a weekly basis during pharmacologically induced maximal coronary vasodilation. Both basic FGF (110 ug/d) and VEGF (45 ug/d) increased collateral flow by 40% after 4 wk. of treatment, and both peptides increased the number of blood vessels in the collateral-dependent myocardium. We also found a significant increase in cell proliferation in the collateral-dependent zone of basic FGF-treated dogs, supporting the suggestion that the increases in vessel number and myocardial perfusion were the result of an angiogenic mechanism.
Having determined that intracoronary administration of basic FGF and VEGF enhanced collateral development, we evaluated the effects of systemic basic FGF administration, and found that left atrial injection of basic FGF (1.74 mg/d for 4 wk) accelerated collateral development without major adverse effects. In a more ambitious study in which basic FGF was given systemically at the same dose for 5 or 9 wk (14), we found that treatment during the period of most pronounced ischemia (10 - 17 days after implantation of the constrictor) was important in enhancing collateral development, whereas treatment beyond this interval was not of additional benefit. We also found that the effects of basic FGF were sustained, persisting after withdrawal of treatment. More recently, therefore, we have limited the interval of basic FGF treatment to 7 days, and still obtained substantial increases in collateral blood flow (15).
We had largely discounted a potential role for growth factors in acute MI, because reperfusion of acutely ischemic myocardium must occur within 4 - 6 hours of coronary occlusion in order to avert infarction, whereas the angiogenic process governed by growth factors requires far longer to reestablish perfusion. Despite these theoretical concerns, intriguing studies by Yanagisawa-Miwa et al. (16) have demonstrated salutary effects of basic FGF on aspects of left ventricular function and infarct size after acute MI in dogs, and several groups have observed transcription of growth factor mRNA or growth factor bioactivity at various times after acute coronary occlusion in animals. These studies suggest a physiologic role for basic FGF in acute MI, both as a cardioprotective agent and as a mediator of infarct healing and remodeling.
To date, we have studied the effects of basic FGF in four independent studies in 87 dogs, and the results consistently demonstrate that the peptide stimulates coronary collateral development in this species. A major question, however, is whether these results can be extrapolated to humans with obstructive coronary artery disease. Tissue specificity is an important concern when basic FGF therapy for humans is being considered. As a nonspecific stimulator of proliferation in mesenchyme-derived cells, basic FGF has the potential to cause renal mesangial cell proliferation and myelophthisis with prolonged exposure to high doses, and this has been borne out in toxicology studies in animals. We are hopeful that such adverse effects can be avoided by limiting the dose and duration of treatment, a focus of current studies. The potential of basic FGF to accelerate tumor formation also needs to be considered. Obviously, the use of basic FGF in patients with known tumors would be contraindicated. Basic FGF does not have the ability to transform cells; however, it could potentially facilitate the growth of tumors in which an inadequate blood supply is the rate-limiting step. Another conceivable pitfall of angiogenic therapy relates to the potential of basic FGF to induce vascular smooth muscle cell proliferation, because neointimal smooth muscle cell hyperplasia is a fundamental component of atherosclerosis. Thus, basic FGF treatment could be a two-edged sword, enhancing collateral growth while accelerating atherosclerosis. We have preliminary data to suggest that this is not the case (15) and are planning additional studies to evaluate this issue further.
Clearly, the investigative area in which most work is needed is in elucidating the physiologic role of each growth factor and the intricacies of growth factor - growth factor and growth factor - receptor interactions. These factors may function through a complex cascade, in much the same way as the clotting factors do. Our current ignorance with respect to the growth factors is analogous to knowing the effect of placing thrombin on a bleeding wound but understanding nothing of the coagulation pathways. Nevertheless, we are cautiously optimistic about the potential clinical ramifications of these data and are currently planning phase I clinical trials to determine the pharmacokinetics and safety of basic FGF in humans.
Figure 1. Left: Schema of experimental model used to engender coronary collateral formation. A constrictor placed on the left circumflex coronary artery (LCX) causes progressive obstruction of the proximal vessel. With the decline in LCX flow, there is the potential development of ischemia in the LCX territory, as well as a pressure gradient between the LCX territory and other normally perfused regions of the heart. In the shaded area, a collateral vessel is depicted, having developed between the branches of the LCX and the left anterior descending coronary artery, the other principal coronary artery of the dog. Stimuli for angiogenesis may originate from within the vascular wall (i.e., shear stress and tangential wall stress), the ischemic myocardium, or both. Right: Stages of angiogenesis are represented from the shaded area (enlarged) on the left: A) quiescent vessel, B) degradation of basement membrane, C) endothelial cell migration, D) further endothelial cell migration and proliferation, and E) tube formation. Later, the collateral will become invested with layers of vascular smooth muscle cells.
1 K. P. Rentrop, J. C. Thornton, F. Feit, and M. Van Buskirk. "Determinants and protective potential of coronary arterial collaterals as assessed by an angioplasty model." Am. J. Cardiol. 61, 677 - 84 (1988).
2 P. J. Sabia, E. R. Powers, M. Ragosta, I. J. Sarembock, L. R. Burwell, and S. Kaul. "An association between collateral blood flow and myocardial viability in patients with recent myocardial infarction." N. Engl. J. Med. 327, 1825 - 31 (1992).
3 E. F. Unger, S. Banai, M. Shou, D. F. Lazarous, M. T. Jaklitsch, M. Scheinowitz, et al. "Basic fibroblast growth factor enhances myocardial collateral flow in a canine model." Am. J. Physiol. 266, H1588 - 95 (1994).
4 S. Banai, M. T. Jaklitsch, M. Shou, D. F. Lazarous, M. Scheinowitz, S. Biro, et al. "Angiogenic-induced enhancement of collateral blood flow to ischemic myocardium by vascular endothelial growth factor in dogs." Circulation 89, 2183 - 89 (1994).
5 J. Folkman, E. Merler, C. Abernathy, and G. Williams. "Isolation of a tumor factor responsible or angiogenesis." J. Exp. Med. 133, 275 - 88 (1971).
6 W. Schaper and B. Wüsten. "Collateral circulation." In: The Pathophysiology of Myocardial Perfusion, W. Schaper, ed. Amsterdam: Elsevier North-Holland Biomedical Press, pp. 415 - 70 (1979).
7 K. W. Scheel, E. M. Fitzgerald, R. O. Martin, and R. A. Larsen. "The possible role of mechanical stresses on coronary collateral development during gradual coronary occlusion." In: The Pathophysiology of Myocardial Perfusion, W. Schaper, ed. Amsterdam: Elsevier North-Holland Biomedical Press, pp 489 - 518 (1979).
8 W. Casscells, E. Speir, J. Sasse, M. Klagsburn, P. Allen, M. Lee, et al. "Isolation, characterization, and localization of heparin-binding growth factors in the heart." J. Clin. Invest. 85, 433 - 41 (1990).
9 M. Klagsburn and E. R. Edelman. "Biological and biochemical properties of fibroblast growth factors; implications for the pathogenesis of atherosclerosis." Arteriosclerosis 9, 269 - 78 (1989).
10 P. J. Keck, S. D. Hauser, G. Krivi, K. Sanzo, T. Warren, J. Feder, and D. T. Connolly. "Vascular permeability factor, an endothelial cell mitogen related to PDGF." Science 246, 1309 - 12 (1989).
11 E. F. Unger, S. Banai, M. Shou, M. T. Jaklitsch, E. Hodge, R. Correa, et al. "A model to assess interventions to improve collateral blood flow: continuous administration of agents into the left coronary artery in dogs." Cardiovasc. Res. 27, 785 - 91 (1993).
12 S. Banai, M. T. Jaklitsch, W. Casscells, M. Shou, S. Shrivastav, R. Correa, et al. "Effects of acidic fibroblast growth factor on normal and ischemic myocardium." Circ. Res. 69, 76 - 85 (1991).
13 I. Vlodavsky, Z. Fuks, R. Ishai- Michaeli, P. Bashkin, E. Levi, G. Korner, et al. "Extracellular matrix- resident basic fibroblast growth factor: implication for the control of angiogenesis." J. Cell. Biochem. 45, 167 - 76 (1991).
14 D. F. Lazarous, M. Shou, M. Scheinowitz, E. Hodge, S. E. Epstein, and E. F. Unger. "Effects of "long term" (9 week) administration of basic fibroblast growth factor on collateral development in ischemic canine myocardium." Circulation 88 (suppl I), I - 375 (1993).
15 D. F. Lazarous, M. Shou, M. Scheinowitz, S. E. Epstein, and E. F. Unger. "Comparative effects of basic fibroblast growth factor and vascular endothelial growth factor on coronary collateral development and the arterial response to injury." J. Am. Coll. Cardiol. 23 (suppl A), 177A (1994).
16 A. Yanagisawa-Miwa, Y. Uchida, F. Nakamura, T. Tomaru, H. Kido, T. Kamijo, et al. "Salvage of infarcted myocardium by angiogenic action of basic fibroblast growth factor." Science 257, 1401 - 3 (1992).