Author + information
- Received September 3, 1996
- Revision received December 11, 1996
- Accepted December 20, 1996
- Published online April 1, 1997.
- Matie Shou, MDA,
- Venugopal Thirumurti, MDA,
- M.A.Sharmini Rajanayagam, PhDA,
- Daisy F Lazarous, MDA,
- Everett HodgeA,
- Jonathan A Stiber, BSA,
- Mary Pettiford, BSA,
- Elizabeth Elliott, BSA,
- Sanjay M Shah, BAA and
- Ellis F Unger, MDA,* ()
- ↵*Dr. Ellis F. Unger, Physiology and Pharmacology Section, Cardiology Branch, National Heart, Lung, and Blood Institute, Building 10, Room 7B15, 10 Center Drive MSC 1650, Bethesda, Maryland 20892-1650.
Objectives. We sought to evaluate the potential of basic fibroblast growth factor (bFGF) to enhance coronary collateral perfusion in dogs with chronic single-vessel coronary occlusion. A secondary goal was to examine whether the salutary effects of bFGF treatment, previously proved effective in the short term, would be maintained in the long term (6 months).
Background. bFGF, an angiogenic growth factor, is currently the subject of a Phase I trial in patients with ischemic heart disease. It has been shown to promote collateral development in dogs with progressive coronary occlusion when given during the period of natural collateralization. The effect of bFGF on quiescent collateral vessels, a subject of significant clinical importance, is uncertain.
Methods. Dogs were subjected to ameroid-induced occlusion of the left circumflex coronary artery and randomized to bFGF (1.74 mg/day for 7 days), a regimen previously proved effective, or to saline solution. Maximal collateral perfusion was assessed 6 months later, and the dogs were reassigned to a course of bFGF or saline solution. Collateral perfusion was reevaluated after the second treatment course.
Results. At 6 months, collateral function was identical in the groups treated initially with bFGF and saline solution. The subsequent course of bFGF did not induce further collateralization.
Conclusions. Although we previously demonstrated the salutary effects of this bFGF regimen in the short term (5 weeks), collateral flow in control dogs reached parity with that of bFGF-treated dogs after 6 months. bFGF did not induce further collateralization in dogs with mature collateral vessels, underscoring the priming role of ischemia for bFGF-induced collateral development.
(J Am Coll Cardiol 1997;29:1102–6)
© 1997 by the American College of Cardiology
Basic fibroblast growth factor (bFGF, FGF-2) is a 154–amino acid heparin-binding polypeptide growth factor that is angiogenic in vitro and in vivo (). We ([2–5]) have previously demonstrated the potential of bFGF to induce myocardial angiogenesis in the ischemic canine heart, and the growth factor has entered into a Phase I trial in patients with ischemic heart disease. In preclinical studies ([4, 5]), bFGF enhanced collateral development when administered during the 10- to 16-day interval after ameroid placement, a period of heightened myocardial ischemia during which collateral formation is known to occur rapidly and spontaneously in the dog-ameroid model. Thus, ischemia (or probably the hemodynamic alterations associated with ischemia, i.e., a transcollateral pressure gradient) may constitute the substrate on which exogenously administered growth factors act; however, this is by no means proved. More important, however, the effect of these peptides on quiescent collateral vessels, a matter of significant clinical import, is uncertain. Thus, we evaluated the effect of a 7-day course of bFGF in dogs with mature collateral vessels—dogs with normal, collateral-derived perfusion under basal conditions but compromised perfusion under stress. An additional study goal was to determine whether a 7-day regimen of bFGF, administered during the rapid phase of collateral development and previously proved effective in the short term (5 weeks) (), would maintain its effectiveness over 6 months.
Forty male or female adult mongrel dogs (weight 15 to 25 kg) were used for the study. These animals were used for this protocol only, and their data have not been reported elsewhere as a subset of a larger group. The study protocol was approved by the National Heart, Lung, and Blood Institute Animal Care and Use Committee, and procedures were conducted in accordance with the National Institutes of Health “Guide for the Care and Use of Laboratory Animals” (Department of Health and Human Services Publication No. [NIH] 86-23, revised 1992) and NIH issuance 3040-2, “Animal Care and Use in the Intramural Program.” Expert veterinary care was provided by the Surgery and Radiology Unit of the Veterinary Resources Program of the National Institutes of Health.
1.1 Aseptic surgical techniques.
The methods have been described previously in detail ([2–4]) and will be summarized herein briefly. Dogs were anesthetized with acepromazine (0.2 mg/kg body weight intramuscularly), thiopental sodium (15 mg/kg intravenously) and 0.5% to 2% inhaled methoxyflurane. A thoracotomy was performed at the left fifth interspace by sterile technique. The origin of the left circumflex coronary artery (LCx) was isolated and a 2- to 2.5-mm ameroid constrictor (Research Instruments Manufacturing) was fitted on the artery proximal to the first marginal branch. A Silastic catheter was placed in the left atrial appendage for drug delivery and subsequent injections of radiolabeled microspheres. The terminal infusion port of the catheter was positioned subcutaneously for percutaneous access. The chest incision was approximated and the animals were allowed to recover. Postoperatively, buprenorphine (6 μg/kg intramuscularly) was given as needed for pain, and intravenous lidocaine was administered prophylactically for 24 h. A broad spectrum antibiotic and procainamide (500 mg) were given twice daily by mouth for 7 days. The left atrial catheter was flushed weekly with saline solution to maintain patency.
1.2 Study design/randomization scheme.
The selected bFGF dose duplicated the doses used in our previous systemic arterial studies ([3, 4]). Clinical grade human recombinant bFGF was obtained from Scios Nova, Inc. and prepared as previously described (). All dogs were randomized to receive daily boluses of bFGF (1.74 mg/day) or saline solution into the systemic arterial circulation by way of the left atrial catheter. Treatment duration was 7 days, 10 to 16 days after implantation of the ameroid constrictor (Fig. 1). The first nine study dogs received a single course of treatment, and collateral perfusion was assessed by using radiolabeled microspheres 6 months later (see later). The remaining dogs were randomized to a second course of treatment after collateral perfusion measurements were completed at 6 months, such that there were four treatment groups: saline solution followed by saline solution (n = 7); saline solution followed by bFGF (n = 6); bFGF followed by saline solution (n = 8); bFGF followed by bFGF (n = 5). The second treatment regimen was identical to the first.
1.3 Collateral perfusion studies.
Maximal collateral perfusion was assessed with microspheres 6 months (183 ± 1 days) after ameroid placement, ∼5.5 months after the first treatment period. To induce maximal coronary vasodilation, chromonar (8 mg/kg) (Intensain, Hoechst-Roussel Pharmaceuticals) () was administered as a 30-min left atrial infusion as we ([2–4]) have done previously. Two weeks after completion of the second course of treatment (204 ± 1 days after ameroid placement), perfusion measurements were repeated. Again, maximal collateral perfusion was quantified after chromonar infusion. A period of 1 to 2 days was allowed to elapse for elimination of the chromonar, and the dogs were returned to the laboratory for additional microsphere studies under basal conditions (in the absence of a vasodilator). The dogs were conscious during all microsphere studies, mildly sedated with diazepam (2 to 3 mg/kg). A 5F Cordis sheath was inserted percutaneously into a femoral artery under local anesthesia. The central lumen was used to obtain arterial reference samples for myocardial blood flow determinations (); the side port was connected to a transducer for blood pressure monitoring. Three to six million tracer microspheres (15 μM in diameter) were injected into the left atrium as previously described (). Isotopes were randomly selected from one of four labels: 141Ce, 85Sr, 95Nb and 46Sc (New England Nuclear).
1.4 Data analysis.
The heart was arrested with KCl after administration of a bolus of adenosine (250 mg) and heparin (5,000 U). The coronary arteries were perfusion-fixed with 10% formalin at 100 mm Hg pressure for 1 h. After fixation was completed by immersion, the left ventricle was divided into 7-mm slices in the short-axis plane. The two central slices were cut into eight wedges. Each of the 16 transmural wedges were then subdivided into endocardial and epicardial segments. Regional myocardial blood flow was calculated, and the ischemic zone (IZ) and normal zone (NZ) were defined as previously described. The four transmural wedges with the lowest perfusion during maximal coronary vasodilation on day 183 were defined as the IZ; the four transmural wedges with the highest perfusion were defined as the NZ (), such that eight tissue samples (four transmural wedges times 2 segments per wedge) comprised both the IZ and NZ. Mean perfusion (in ml/min per 100 g) for each set of eight samples was calculated as: Vascular conductance was computed as the quotient of perfusion and mean arterial pressure (units = ml/min per 100 g per mm Hg).
In 10 dogs, the slice basal to the two central slices was analyzed for infarct size. This slice was divided into eight transmural wedges, paraffin-embedded and stained by using Masson’s trichrome method to delineate viable myocardium (red) versus scar (blue). Each section was examined by using a microscope fitted with a 1× objective lens. Rectangular fields (area 50.0 mm2) were digitized in true color (512 × 480 pixel resolution), and areas of scarred myocardium were quantified and summed by using a computer-based image analysis system (Advanced Imaging Concepts, Inc.). The total tissue wedge area was assessed by using a 35-mm macro lens (field area 745 mm2) interfaced with a video camera and the computer-based system. Infarct area was computed as the quotient of scarred area/total tissue area × 100%.
Complete blood count and blood chemistry values were analyzed weekly in all dogs. Urine samples were collected for analysis of urine volume, creatinine and protein before, during and after completion of 7 days of treatment.
The investigators responsible for the performance of microsphere blood flow determinations and analyses of study end points were naive to treatment group. Data were analyzed using a two-way analysis of variance for repeated measures. All data are expressed as mean value ± SEM.
Four (10%) of 40 dogs died within 1 week of ameroid placement, probably because of coronary spasm or ischemia-related arrhythmias, or both. One additional dog died 1 month after ameroid placement (bFGF-treated), and an extensive myocardial infarction was noted at necropsy. Thirty-five (88%) of the 40 dogs completed the studies and constitute the basis of this report.
2.1 Systemic hemodynamic measurements.
Heart rate and blood pressure were recorded ∼5.5 months after the initial treatment period, and again 11 days after the second treatment period. There were no significant differences between groups (data not shown).
2.2 Infarct size.
In our experience ([2–4]), the dog LCx ameroid model produces infarcts of minimal size (1% to 2%). Ten dogs underwent an analysis of infarct size in the present study. Infarct size was 0.88 ± 0.29% in dogs that received an initial course of bFGF and 1.49 ± 0.62% in control dogs (p = NS).
2.3 Results of initial treatment.
Vascular conductance during maximal vasodilation, quantified 6 months after ameroid placement and ∼5.5 months after the initial treatment period, was identical in dogs that received initial treatment with bFGF and dogs that received the vehicle (Fig. 2, left). Maximal IZ conductance was 2.39 ± 0.21 ml/min per 100 g per mm Hg in bFGF-treated dogs and 2.24 ± 0.21 ml/min per 100 g per mm Hg in control dogs, with IZ/NZ perfusion ratios of 0.51 ± 0.02 and 0.49 ± 0.01, respectively. In bFGF-treated dogs, the endocardial/epicardial perfusion ratio (during maximal vasodilation) was 0.74 ± 0.03 in the IZ and 0.98 ± 0.02 in the NZ. In control dogs, the corresponding ratios were quite similar (0.76 ± 0.04 and 1.01 ± 0.02, respectively, p = NS for both the IZ and the NZ).
2.4 Results of second treatment.
The dogs were randomized to a second course of treatment at 6 months (bFGF vs. saline solution). For analysis of this intervention, the dogs were segregated into two groups on the basis of the second treatment group assignment (irrespective of whether their first course of treatment was bFGF or vehicle), and there was no evidence of a bFGF effect (Fig. 2, right). Under basal conditions, IZ collateral perfusion was normal (i.e., equivalent to NZ perfusion) and similar in bFGF-and saline solution–treated dogs (0.72 ± 0.10 and 0.65 ± 0.05 ml/min per 100 g per mm Hg, respectively). Basal IZ/NZ ratios were 1.00 ± 0.05 in bFGF-treated dogs and 1.06 ± 0.06 in control dogs. In the IZ, the endocardial/epicardial perfusion ratios were similar in bFGF-treated and control dogs (1.20 ± 0.06 vs. 1.16 ± 0.07, respectively, p = NS). During chromonar-induced maximal vasodilation, there was a 3.5-fold increase in IZ perfusion, with a 7-fold increase in NZ perfusion. Maximal IZ perfusion was similar in bFGF-treated and control dogs (2.70 ± 0.21 and 2.44 ± 0.07 ml/min per 100 g per mm Hg, respectively). Expressed as an IZ/NZ ratio, maximal collateral perfusion in bFGF-treated dogs was 0.51 ± 0.01 (it had been 0.50 ± 0.01 before the second course of treatment). In saline solution–treated control dogs, the IZ/NZ ratio during maximal vasodilation was 0.53 ± 0.03 (it had been 0.50 ± 0.02 before the second course of treatment). The endocardial/epicardial perfusion ratios were also similar to those obtained before the second course of treatment, and there was no difference between groups.
2.5 Hematologic and biochemical data.
Total leukocyte count, hematocrit, calcium, phosphate, creatinine, urea nitrogen, albumin, alkaline phosphatase, serum glutamic oxaloacetic transaminase, serum glutamic pyruvic transaminase, bilirubin, glucose and creatine kinase were similar in all groups throughout time (data not shown). In bFGF-treated dogs, the platelet count decreased from a pretreatment value of 314 ± 28 × 103/mm3to 151 ± 18 × 103/mm31 week after treatment (p < 0.0005) but returned to normal within 1 week (305 ± 22 × 103/mm3). The second course of bFGF did not depress the platelet count significantly. No clinical problems were apparent as a result of thrombocytopenia. Urinary protein was measured weekly, beginning 1 week before treatment with bFGF or saline solution and continuing for 3 weeks. Expressed as a urinary protein/creatinine ratio, there were no significant increases in urinary protein excretion in bFGF-treated dogs.
We ([2–5]) have shown previously that bFGF improves collateral perfusion in dogs with ameroid-induced single-vessel coronary occlusion, and a Phase I clinical trial is ongoing at this institution on the basis of these investigations. It is well established that coronary collateral vessels develop in dogs as an adaptive response to experimentally induced coronary artery occlusion, a process that involves neovascularization as well as remodeling of preexisting vessels (). In dogs subjected to progressive ameroid-induced coronary occlusion, we ([4, 5]) have shown that administration of bFGF, synchronized with the interval of maximal adaptive collateral development, enhances myocardial angiogenesis and collateral perfusion. We () have also demonstrated that dogs so treated do not maintain their responsiveness to bFGF: Additional treatment 5 to 9 weeks after ameroid placement does not elicit further collateral development. Related studies have shown the benefit of bFGF () and of recombinant adenovirus expressing human fibroblast growth factor 5 () in ischemic porcine myocardium, also when the growth factors were available during the phase of programmed collateral development brought about in response to coronary occlusion. These peptides, therefore, appear to amplify the adaptive response of collateral vessels to coronary occlusion. On the basis of these and other studies, ischemia (or the hemodynamic conditions associated therewith) appears to be requisite for growth factor–enhanced collateral expansion; however, this supposition is unproved. For example, angiogenic growth factors induce focal angiogenesis in the apparent absence of ischemia or a pressure gradient in the chorioallantoic membrane (CAM) assay system (), a standard assay for angiogenesis.
The relative importance of ischemia (or a transcollateral pressure gradient, or both) as a primer for pharmacologically driven collateralization has significant ramifications for the treatment of patients with ischemic heart disease and peripheral vascular disease. If ischemia were not fundamentally important for pharmacologically enhanced collateral development, systemic administration of angiogenic growth factors might be expected to cause uncontrolled vascular proliferation throughout the body, with potentially deleterious consequences. Conversely, if ischemia (or a transcollateral gradient) were essential for pharmacologically enhanced collateral development, systemic administration of one or more angiogenic growth factors could lead to site-specific collateralization in areas of ischemia. The latter concept would also predict that therapeutic benefit would be realized only if ischemia were present or induced coincident with the administration of the agent. If this were true, ischemia would play an important adjunctive role in angiogenic treatment, an implicit assumption made in two clinical studies in which the potential salutary effects of heparin () and low molecular weight heparin () were evaluated in patients with ischemic heart disease. In both studies, it was assumed that the induction of myocardial ischemia (through treadmill exercise) was essential to facilitate coronary collateral development.
3.1 bFGF and mature collateral vessels.
In the present study, bFGF was administered to dogs with mature collateral vessels in which ischemia was unlikely to occur with any regularity. Before treatment, these dogs exhibited normal rest IZ perfusion (IZ/NZ ratio 1.00 ± 0.05), although coronary reserve was severely compromised (IZ/NZ during chromonar 0.51 ± 0.01). In this setting, bFGF failed to elicit further collateral development. The implication of this finding is that bFGF would not be expected to induce collateral formation in the absence of ischemia (or a transcollateral pressure gradient), an observation that has major ramifications for its potential clinical use, as noted earlier. We obtained similar results in a related dog-ameroid model in which acidic fibroblast growth factor (aFGF) was infused continuously into the left main coronary artery (). In that study, aFGF was not administered until 5 weeks after ameroid placement, when collateral development was well advanced. Although aFGF was administered for 4 weeks, there was no effect on collateral perfusion. The present results are also concordant with those of Yang et al. (), who found that heparin exerted an angiogenic effect in a rat femoral artery ligation model in the presence but not in the absence of exercise.
Whereas bFGF did not promote further collateral development in dogs with well developed collateral vessels, it would be interesting to assess the potential effect of adjunctive exercise on bFGF-induced collateralization in dogs with mature collateral vessels. A similar consideration might be relevant in human angiogenesis trials, as noted earlier.
3.2 Long-term effect of bFGF on collateral development.
bFGF had been shown previously to improve coronary collateral perfusion in dogs in the short term (4 weeks after ameroid placement) ([2, 4, 5]) and in the intermediate term (up to 9 weeks after ameroid placement) (). In this investigation, we were interested in the state of collateral vessels 6 months after the termination of bFGF treatment to ascertain whether the increase in collateral function would be maintained. We found that maximal collateral perfusion was identical in dogs that had received bFGF or saline solution, roughly 5.5 months earlier. Thus, substantial collateral vessels develop in dogs after 6 months irrespective of dogs’ exposure to exogenous bFGF during the period of rapid collateral development. This “catch-up” phenomenon is probably a manifestation of the dog’s inherent propensity for collateral growth.
3.3 Study limitations.
It would not necessarily be valid to extrapolate the lack of effect of bFGF on mature collateral vessels to patients with atherosclerotic cardiovascular disease and chronic ischemia and to presume that bFGF would accelerate collateral development and improve perfusion in the short term without affecting long-term results (as in our animal experiments). Symptoms of patients with ischemic cardiovascular disease generally do not lessen spontaneously, suggesting that, in these patients, a mechanism for adequate collateral development is lacking. The effect of bFGF may be of considerable importance in such patients. Because no animal model can accurately predict the response in humans, the testing of this important hypothesis must await clinical trials.
We were unable to quantify collateral perfusion 5 weeks after ameroid placement; therefore, our assessment of the effectiveness of the first course of bFGF therapy is based not on the present data but on our previous study (). Thus, the conclusion that collateral development in control animals reached parity with that of bFGF-treated animals after 6 months is based partially on historical data, because it was not demonstrated that bFGF enhanced collateral development in this experiment in the short term. The extreme length of the study posed practical limitations to isotope use early in the study protocol because of radioactive decay. Such data might have been obtained by using the colored microsphere technique; however, this modality was not available to us at the time the study was performed.
3.4 Clinical implications.
Extrapolation of these data to patients with ischemic heart disease suggests that bFGF would not enhance collateral development in patients with ischemic heart disease unless ischemia (or a transcollateral pressure gradient) is elicited intentionally through exercise or stress or occurs spontaneously, or both. Thus, clinical studies should be designed in such a way as to ensure that ischemia, if not present spontaneously, is induced routinely during the course of treatment. Conversely, bFGF would not be expected to induce vascular development in tissues in the absence of ischemia (or a transcollateral pressure gradient), suggesting that systemic bFGF administration would be reasonably safe in patients.
We gratefully acknowledge the assistance of Victoria Hampshire, DVM, John Bacher, DVM and their staffs for providing expert veterinary care.
This work was presented in part at the 45th Annual Scientific Session of the American College of Cardiology, Orlando, Florida, March 1996.
- acidic fibroblast growth factor
- basic fibroblast growth factor
- ischemic zone
- left circumflex coronary artery
- normal zone
- Received September 3, 1996.
- Revision received December 11, 1996.
- Accepted December 20, 1996.
- The American College of Cardiology
- Folkman J,
- Klagsburn M
- Unger EF,
- Banai S,
- Shou M,
- et al.
- Lazarous DF,
- Scheinowitz M,
- Shou M,
- et al.
- Lazarous DF,
- Shou M,
- Scheinowitz M,
- et al.
- Rajanayagam S,
- Shou M,
- Thirumurti V,
- et al.
- Baer RW,
- Payne BD,
- Verrier ED,
- et al.
- Schaper W,
- Ito WD
- Fujita M,
- Sasayama S,
- Asanoi H,
- Nakajima H,
- Sakai O,
- Ohno A
- Quyyumi AA,
- Diodati JG,
- Lakatos E,
- Bonow RO,
- Epstein SE
- Unger EF,
- Banai S,
- Shou M,
- et al.
- Yang HT,
- Ogilvie RW,
- Terjung RL