Effects of Chronic Systemic Administration of Basic Fibroblast Growth Factor on Collateral Development in the Canine Heart
Background Recently we reported that intracoronary administration of basic fibroblast growth factor (bFGF), a potent angiogenic peptide, increases collateral blood flow in dogs subjected to progressive left circumflex coronary artery (LCx) occlusion. The aim of the present study was to examine the effect of systemically administered bFGF on collateral blood flow and to assess its pharmacokinetics and potential side effects.
Methods and Results Forty-seven dogs were subjected to progressive ameroid-induced occlusion of the LCx, an intervention known to induce the development of collateral vessels. In phase I of the investigation, dogs were randomized to receive bFGF 1.74 mg/d (n=10) or saline (n=9) as a left atrial injection for 4 weeks. Relative collateral blood flow was assessed serially with radiolabeled microspheres in the conscious state during maximal coronary vasodilatation. Initiation of bFGF treatment was temporally associated with a marked acceleration of collateral development; however, collateral flow in control dogs improved toward the end of the study, approaching that of bFGF-treated dogs at the 38-day end point. Phase II of the investigation was a three-armed study of extended duration to determine whether bFGF caused a sustained increase in collateral function. Dogs were randomized to receive bFGF 1.74 mg/d for 9 weeks (n=7), bFGF 1.74 mg/d for 5 weeks followed by placebo for 4 weeks (n=11), or placebo for 9 weeks (n=10). Relative and absolute collateral blood flow were assessed serially with microspheres during maximal coronary vasodilatation. Between the 10th and 17th days after ameroid placement, bFGF-treated dogs exhibited marked improvement in collateral flow such that maximal collateral conductance exceeded that of controls by 24% at the 5-week crossover point. Final collateral conductance was similar in dogs receiving bFGF for 5 and 9 weeks despite withdrawal of treatment in the former group. bFGF administration was associated with a 21% increase in final collateral conductance as well as a 49% increase in collateral zone vascular density. Prolonged bFGF administration was also associated with a decrease in arterial pressure, moderate thrombocytopenia, and moderate, reversible anemia.
Conclusions Systemic administration of bFGF enhanced collateral conductance in dogs with progressive single-vessel coronary occlusion. The beneficial effect of bFGF occurred primarily between the 7th and 14th days of therapy, and regression of collateral development was not noted after withdrawal of treatment. The present investigation provides impetus to the concept that collateral development can be enhanced pharmacologically–specifically by bFGF–raising the possibility that such an intervention might eventually be applied clinically.
Coronary collaterals have the potential to ameliorate myocardial ischemia and sustain contractile function when flow is impaired as a result of coronary occlusive disease.1 2 Moreover, myocardial viability after acute infarction has been correlated with the extent of collateral blood flow within the territory of the infarct-related artery.3 Collateral development occurs as an adaptive response to coronary occlusive disease; however, individuals vary widely in their propensity for collateral growth, and the possibility exists that collateral development could be enhanced pharmacologically. A number of angiogenic peptide growth factors have been identified that may be involved in myocardial vascular homeostasis as well as the circulatory adaptation to pathological states.4 Basic fibroblast growth factor (bFGF) is one such peptide. A member of a family of heparin-binding polypeptides, bFGF is widely distributed and has been identified in many tissues of neuroectodermal and mesodermal origin. The peptide stimulates proliferation of the three principal vascular cell types (endothelial cells, smooth muscle cells, and fibroblasts) and is chemotactic for endothelial cells in vitro, inducing capillary tube formation. bFGF has been isolated from the heart5 6 and has been localized to adult cardiac myocytes in vitro,7 suggesting a role in coronary vascular homeostasis.
In a previous study,8 we administered bFGF via the intracoronary route to dogs with progressive ameroid-induced occlusion of the left circumflex coronary artery (LCx) in an attempt to promote coronary collateral development. bFGF was injected directly into the LCx distal to the point of occlusion (110 μg/d for 4 weeks), an intervention associated with a 40% increase in collateral blood flow and concomitant increases in collateral zone vascular density and endothelial cell proliferation. The aims of the present investigation were (1) to determine whether systemic bFGF increases collateral blood flow, (2) to determine whether any beneficial effect is maintained after cessation of therapy, (3) to determine the pharmacokinetics of systemic bFGF, and (4) to evaluate potential adverse effects of systemic bFGF administration.
Fifty-eight adult mongrel dogs of either sex (weight, 15 to 25 kg) were used for the study. The investigation was conducted in two phases. Phase I (principal investigator, M. Scheinowitz) was designed to assess bFGF pharmacokinetics and the feasibility of chronic systemic bFGF administration. Phase II (principal investigator, D.F. Lazarous) was a more comprehensive study designed to evaluate the long-term effects of bFGF on absolute (nonindexed) coronary collateral blood flow, myocardial vascular density, and systemic hemodynamics as well as to assess potential toxicity. The experimental protocols were approved by the Animal Care and Use Committee of the National Heart, Lung, and Blood Institute, and animal procedures were conducted in accordance with the National Institutes of Health (NIH) “Guide for the Care and Use of Laboratory Animals” (Department of Health and Human Services Publication No. [NIH] 86-23, revised 1992) and NIH manual 3040-2, “Animal Care and Use in the Intramural Program.” Veterinary care was provided by the Veterinary Resources Branch of the National Institutes of Health.
The methods have been described previously and will be summarized briefly.8 9 Dogs were anesthetized with acepromazine (0.2 mg/kg IM), thiopental sodium (15 mg/kg IV), and 0.5% to 2% inhaled methoxyflurane for the aseptic surgical procedure. A thoracotomy was performed using a left-sided approach, and a 2.5- to 3.5-mm ameroid constrictor was fitted on the LCx proximal to the first marginal branch. A hydraulic balloon occluder was placed on the LCx immediately distal to the ameroid, and a Silastic catheter was placed in the left atrial appendage for drug delivery and subsequent radiolabeled microsphere injections. The termini of the balloon occluder and left atrial catheter were secured subcutaneously at the dorsum of the neck. The chest was closed, and the animals were allowed to recover. The analgesic buprenorphine was administered as needed. A broad-spectrum antibiotic and procainamide (500 mg) were given twice daily by mouth for 7 days. The left atrial catheter was flushed daily with normal saline to maintain patency. No dog received heparin on any occasion because the anticoagulant has been shown to have angiogenic effects in dogs.10
Phase I Studies
Estimation of the systemic bFGF dose was based on the dose used in our previous intracoronary study8 and the fact that approximately 7% of cardiac output is distributed to the myocardium. In our previous study, daily intracoronary injection of 0.11 mg bFGF enhanced collateral flow. Thus, to achieve a comparable myocardial concentration in the present study, we estimated the daily systemic bFGF dose to be 0.11 mg÷0.07 or approximately 1.6 mg. Dogs were randomized to receive human recombinant bFGF 1.74 mg/day (n=10) or vehicle (n=9) beginning 10 days after placement of the ameroid constrictor and continuing for 4 weeks. Injections were made on a daily basis, Monday through Friday. Clinical grade human recombinant bFGF was obtained from Scios Nova, Inc. The peptide was provided in citrate buffer (20 mmol/L sodium citrate, 1 mmol/L EDTA, 9% sucrose, pH 5) at a protein concentration of 10.0 mg/mL. Endotoxin content was negligible (<0.45 endotoxin U/mg bFGF, or approximately 0.04 endotoxin U/kg/d). bFGF aliquots were prepared by diluting this solution 1:3 in PBS to a volume of 0.5 mL. Control dogs received PBS aliquots of equivalent volume. All aliquots were stored at −20°C until use. The investigators responsible for analyses of study end points were blinded to treatment group. Complete blood counts and blood chemistries (total leukocyte count, hemoglobin, hematocrit, platelet count, sodium, potassium, chloride, calcium, phosphate, creatinine, urea nitrogen, albumin, total protein, lactate dehydrogenase, alkaline phosphatase, serum glutamic oxaloacetic transaminase [SGOT], serum glutamic pyruvic transaminase [SGPT], bilirubin, glucose, and CPK) were performed weekly.
Relative collateral blood flow was assessed in the conscious state during pharmacologically induced maximal vasodilatation, as we have done previously.8 9 Both relative and absolute blood flows were quantified at the final time point. Baseline measurements were made 3 and 10 days after ameroid placement. Subsequent measurements were made weekly during the treatment period (days 17, 24, 31, and 38). Ameroid constrictors require a range of approximately 10 to 20 days to effect total coronary occlusion. Thus, the LCx hydraulic balloon occluder was used to temporarily arrest all antegrade LCx flow during the microsphere injections on days 3, 10, and 17, such that collateral flow was measured exclusively. The balloon occluder was permanently inflated on day 24 to ensure complete and timely occlusion of the LCx. In all cases collateral blood flow was evaluated before the daily dose of bFGF (or placebo) was administered. The dogs were mildly sedated with diazepam 1 to 2 mg/kg, and maximal coronary vasodilatation was induced with chromonar (Hoechst-Roussel Pharmaceuticals) 8 mg/kg through the left atrial catheter.11 Approximately 3 to 6×106 radiolabeled microspheres, 15 μm in diameter, were injected into the left atrium for each blood flow measurement. On day 38, an arterial catheter inserted into a femoral artery under local anesthesia provided a reference sample for quantification of absolute myocardial blood flow.12 After the final blood flow measurement, dogs were anesthetized with sodium pentobarbital, the heart was arrested with KCl, and the coronary arteries were perfusion-fixed at 100 mm Hg pressure for 2 hours.
The LCx was examined to verify occlusion at the site of the ameroid. Myocardial blood flow was calculated using standard methods.12 Bilateral renal cortical samples were analyzed to ensure adequate mixing of microspheres. Blood flow was assessed in 32 myocardial samples obtained from two short-axis slices, each divided into eight wedges and subdivided into epicardial and endocardial segments. Ischemic zone (IZ) and normal zone (NZ) samples were selected, and mean IZ and NZ blood flows were calculated as previously described.8 9 Infarct size was quantified in an adjacent left ventricular slice basal to the two slices used for blood flow analyses, as we have done previously.8 9 Tissue samples were removed from the skin, small intestine, lung, liver, kidney, and spleen, stained with hematoxylin and eosin, and reviewed by a pathologist blinded to treatment group.
Pharmacokinetics of Systemic bFGF
In one dog, 1.74 mg bFGF was injected into the left atrium, and venous blood samples were obtained at multiple time points thereafter. Samples were centrifuged at 4°C for 10 minutes, and serum was assayed for human bFGF using a solid-phase ELISA kit according to the manufacturer’s instructions (catalog No. DBF00, R&D Systems).
Phase II Studies
Surgical preparation of the animals was as described in phase I. Baseline microsphere blood flow measurements were made on day 3, after which the dogs were randomized to receive bFGF 1.74 mg/d as a single dose 5 days per week for 9 weeks (n=7), bFGF 1.74 mg/d 5 days per week for 5 weeks followed by placebo for 4 weeks (n=11), or placebo for 9 weeks (n=10). Additional blood flow measurements were obtained on days 10, 17, 38, 52, and 66. As in phase I, inflation of the LCx hydraulic balloon occluder during microsphere injections on days 3, 10, and 17 provided a means to appraise IZ collateral flow exclusive of residual antegrade LCx flow, and the balloon occluder was inflated permanently on day 24. Blood flow was quantified in the conscious state before the daily dose of bFGF (or placebo) was administered and during pharmacologically induced maximal coronary vasodilatation. On days 10, 38, and 66, a femoral artery catheter was inserted using local lidocaine anesthesia for calculation of absolute myocardial blood flow using the reference sample technique.12 At alternate time points (days 3, 17, and 52), an arterial line was not inserted and microspheres were injected without a reference sample such that the IZ/NZ blood flow ratio alone was calculated. Hemodynamic measurements were obtained in the conscious state on days 10, 38, and 66. Baseline heart rate and blood pressure were recorded to assess potential long-term hemodynamic effects of bFGF administration. After these measurements, chromonar was administered and collateral blood flow was quantified with microspheres (as described above). Approximately 60 minutes was allowed for the effect of chromonar to subside before bFGF (or placebo) was administered with continuous monitoring of heart rate and mean arterial pressure. Arterial pressure was monitored with a strain gauge, and heart rate was obtained from the ECG. On day 66, a solid-state transducer (SPC-370, Millar Instruments, Inc) was passed retrograde from the femoral artery into the left ventricle for measurement of left ventricular pressure and dP/dt. Complete blood counts and serum chemistries were performed weekly as in phase I. After the final blood flow measurement, the heart was potassium-arrested and fixed as before.
The IZ and NZ myocardial blood flows were calculated as in phase I. The transmural wedge from the central IZ was subjected to vessel counting. Masson’s trichrome–stained sections were systematically analyzed in their entirety using computer-based image analysis, as we have done previously.8 9 The sections were examined at two magnifications to optimize match between vessel size and system magnification: vessels in the 10- to 20-μm and 20- to 60-μm range were counted at ×10; vessels 60 to 120 μm, 120 to 180 μm, and >180 μm were analyzed at ×4. Total tissue area was quantified at low magnification using a macro lens. Percent infarct area was calculated as before.8 9 Tissue samples were removed from the skin, small intestine, lung, liver, kidney, spleen, and bone marrow, stained with hematoxylin and eosin, and reviewed by a veterinary pathologist blinded to treatment group. The eyes were removed immediately after death, and elastase digest preparations of the whole retinal vasculature were examined for capillary dilation, tortuosity, loss or increase in endothelial cells or pericytes, microaneurysms, signs of vascular budding, and changes in length and number.13 This analysis was carried out independently by two observers unaware of treatment group (W.G.R. and N.M.L, see “Acknowledgments”).
Data are expressed as mean±SEM. Two-tailed Student’s t tests for unpaired data were performed to assess differences between two groups. Collateral blood flow data, hematological data, and biochemical data were analyzed using an ANOVA for repeated measures model. Potential interactions between groups were assessed by treatment, time, and time by treatment interaction. Collateral blood flow measurements obtained before randomization to treatment in phase I (days 3 and 10) were considered to constitute baseline measurements for each dog and were subtracted from subsequent values on days 17, 24, 31, and 38. Thus, collateral flow was compared at four time points and subjected to the correction method of Bonferroni.14 In phase II, collateral blood flow measurements on day 3 were considered baseline and subtracted from the following five measurements. For the analyses of blood chemistries and hematological studies, tests were considered statistically significant at P<.005 to account for the fact that multiple tests were performed at each time point.
Microsphere blood flow data are summarized on 19 of 24 dogs that completed the phase I studies successfully: 10 bFGF-treated dogs and 9 control dogs. Two dogs died and one developed an extensive myocardial infarction before randomization, presumably related to technical factors or ameroid-induced LCx spasm. No deaths occurred during treatment with bFGF or placebo. One dog was excluded because the LCx was determined to be partially patent at the site of the ameroid and another was excluded because of unsatisfactory blood flow results presumed to be due to leakage of the left atrial catheter. No data were excluded after the phase I randomization code was broken.
On day 38, basal heart rates were 157±13 and 132±7 beats per minute in bFGF-treated and control dogs, respectively (P=NS). Mean arterial pressure was significantly lower in bFGF-treated dogs than in control dogs (113±4 versus 128±3 mm Hg, respectively, P<.05), although differences in systolic and diastolic blood pressures were not individually significant.
IZ/NZ Blood Flow Ratio, Final Collateral Conductance
Fig 1⇓ shows the changes in IZ/NZ blood flow ratio as a function of time. Before the institution of therapy (days 3 and 10), the IZ/NZ ratio was similar in both groups. Marked improvement in the IZ/NZ ratio was observed in bFGF-treated dogs coinciding with the first week of treatment, when the IZ/NZ ratio improved from 0.06±0.02 (day 10) to 0.34±0.06 (day 17). Smaller increments in the IZ/NZ ratio occurred during the final 3 weeks of the study, such that the final IZ/NZ ratio was 0.48±0.04 in bFGF-treated dogs. In control dogs, only minimal increases in the IZ/NZ ratio were observed through day 24, with IZ/NZ ratios of 0.13±0.04, 0.16±0.04, and 0.19±0.04 on days 10, 17, and 24, respectively. Analysis of repeated measures demonstrated a significant difference in transmural blood flow between the two groups (P=.0007), with significant differences detected in both the endocardial and epicardial layers (P<.0001 for both). The analysis of repeated measures was suggestive of a time by group interaction (P=.18), consistent with a projected intersection of the two lines. Such an intersection implies equivalence of the IZ/NZ ratios of the two groups at some point after day 38. Indeed, there was a substantial increase in the IZ/NZ ratio in control dogs between days 24 and 31 with a further increase recorded on day 38, such that there was no significant difference in IZ/NZ blood flow ratio between the two groups at the final time point. Total myocardial conductance was calculated from microsphere data obtained on day 38 as the quotient of myocardial blood flow and mean arterial pressure. Within the IZ, collateral conductance was 3.23±0.38 mL/min per 100 g/mm Hg in bFGF-treated dogs versus 2.49±0.26 mL/min per 100 g/mm Hg in control dogs (P=.11). In the NZ, the coronary conductances were 6.76±0.50 and 6.26±0.88 mL/min per 100 g/mm Hg in bFGF-treated and control dogs, respectively (P=NS).
Pharmacokinetics of Systemic bFGF
After systemic injection of 1.74 mg bFGF, there was a rapid distribution phase followed by an elimination phase with first order kinetics (Fig 2⇓). The distribution half-life was approximately 3 minutes, whereas the elimination half-life was 50 minutes (elimination rate constant, 0.013 min−1), with a volume of distribution of 290 L.
Infarct Size, Myocardial Histology
Infarct size was negligible, averaging 0.34±0.18% of a representative left ventricular slice in bFGF-treated dogs and 0.56±0.25% in control dogs (P=NS). An unexpected finding was the presence of a neointima in occasional intramyocardial coronary arteries in two dogs. Neointimal hyperplasia was moderate in one dog and marked in the other. Although both dogs had been treated with bFGF, this finding was not significantly related to treatment group (χ2 test, P=.50).
Hematological and Biochemical Data, Histology of Noncardiac Organs
Baseline laboratory values were similar in treated and control dogs. Hematocrit in control dogs remained stable throughout the study; however, hematocrit decreased from 46±1% at baseline to 38±1% during the fourth week of treatment in bFGF-treated dogs (P<.001). Baseline platelet counts were also similar in the two groups (3.2±0.5×105/mm3 in bFGF-treated dogs versus 3.0±0.2×105/mm3 in control dogs); however, the platelet count decreased significantly during the third and fourth weeks of treatment in bFGF-treated dogs to 1.7±0.2×105/mm3 and 1.5±0.1×105/mm3, respectively (P<.001). Week-by-week comparison of biochemical data did not demonstrate any significant abnormalities or differences between groups. No significant pathological findings were observed in skin, intestine, lung, liver, kidney, or spleen in either group.
Data were obtained successfully on 28 of 34 dogs. Three sudden deaths occurred within 2 weeks of surgery, presumably related to abrupt LCx spasm or occlusion. A fourth dog sustained a myocardial infarction within 3 days of surgery and was excluded. Two dogs were excluded before breaking the code because of unsatisfactory blood flow data (>10% discrepancy between left and right renal cortical blood flows in one dog, failure of the left atrial line in the other).
Table 1⇓ provides data reflecting both the acute and chronic effects of bFGF administration. Values obtained before (pre) treatment are the measurements obtained just before bolus injection of bFGF, 24 hours after the last dose administered; hence, they reflect the chronic effects of drug administration. Comparison of values obtained before and after (pre and post) drug injection were used to assess the acute effects of drug administration.
Chronic Effects of bFGF Administration
Heart rate was similar in bFGF-treated and control dogs throughout the study. Mean arterial pressure was similar in all treatment groups on day 10; however, mean arterial pressure was decreased by approximately 8% in bFGF-treated dogs on day 38. In the group of dogs in which bFGF treatment was discontinued on day 38, arterial pressure recovered, whereas arterial pressure was persistently depressed in dogs that continued to receive bFGF until day 66. On day 66, left ventricular (LV) dP/dt and LV end-diastolic pressure (LVEDP) were similar in all groups.
Acute Effects of bFGF Administration
Acute administration of bFGF decreased mean arterial pressure as determined on days 38 and 66; the occurrence of a hypotensive response on day 66 indicates lack of tolerance despite daily bFGF administration (Table 1⇑). In general, the onset of the hypotensive effect occurred approximately 1 minute after bFGF administration, with a peak effect at 5 minutes. The arterial pressure recovered to baseline after 20 to 30 minutes. In the 7 dogs that continued to receive bFGF until day 66, LV dP/dt (as determined on day 66) was unaffected by acute bFGF administration despite a decrease in arterial pressure. With acute bFGF administration, the change in LVEDP (−2.9±1.2 mm Hg) was not significant.
Regional Myocardial Blood Flow
Collateral blood flow was assessed as maximal collateral conductance at three time points (days 10, 38, and 66) and as the IZ/NZ ratio at all six time points. These data are summarized in Table 2⇓ and Fig 3⇓. In bFGF-treated dogs, marked improvement in collateral flow occurred in the interval 10 to 17 days after ameroid placement (as was the case in phase I). After 5 weeks of treatment (day 38), maximal collateral conductance in bFGF-treated dogs significantly exceeded that of control dogs (P<.05). At the 9-week study conclusion, collateral conductance was similar in the two bFGF-treated groups: one group treated with bFGF for the entire 9 weeks, the other treated for only 5 weeks with no bFGF administered during the last 4 weeks of the study. However, collateral conductance in control dogs was significantly lower than in the bFGF-treated dogs (P<.05). Similarly, the IZ/NZ ratio at the conclusion of the study in dogs treated with bFGF for 9 weeks was virtually identical to that of dogs treated for 5 weeks (0.51±0.03 versus 0.50±0.03, respectively). In control dogs, the final IZ/NZ ratio was 0.40±0.03 (P<.005, bFGF versus control). Significant differences between bFGF-treated and control dogs were present in both the endocardium and epicardium.
The numerical density of vessels >10 μm in diameter was 56.9±5.0 vessels/mm2 in the IZ of bFGF-treated dogs and 38.1±6.3 vessels/mm2 in control dogs (P<.05). The increase in vascular density was not confined to any size range of vessels analyzed (10 to 20 μm, 20 to 60 μm, 60 to 120 μm, 120 to 180 μm, or >180 μm).
Infarct Size, Myocardial Histology
Infarct area was quite limited, averaging 0.43±0.12%, 0.40±0.23%, and 1.04±0.50% of the area of the mid LV slices in bFGF (5 weeks), bFGF (9 weeks), and control dogs, respectively (P=NS). For all dogs, mean infarct area was 0.64±0.20% (range, 0% to 4.95%). Neointimal hyperplasia was observed in occasional epicardial and intramyocardial arteries, randomly distributed in the IZ and NZ of both groups. Intimal proliferation was noted in 2 of 10 control dogs and in 10 of 18 treated dogs (χ2 test, P=.15).
Hematological and Biochemical Data
Prolonged bFGF administration was associated with significant decreases in hematocrit (P=.0001) and platelet count (P=.0007). In control dogs, hematocrit was 49±2% before treatment and remained essentially constant throughout the study (Fig 4⇓). In dogs treated with bFGF for 9 weeks, hematocrit decreased from 51±1% (before treatment) to 34±1% during the first 6 weeks of treatment (P<.001) and remained at this level during the final 3 weeks. In dogs treated with bFGF for 5 weeks, hematocrit dropped from a pretreatment value of 49±2% to a nadir of 31±2% 1 week after treatment was suspended, gradually improving to 40±2% at the conclusion of the study. Platelet count (Fig 4⇓) was virtually constant across time in control dogs. In bFGF-treated dogs, the platelet count decreased from 2.5±0.1×105/mm3 before treatment to 1.7±0.1×105/mm3 on the fourth week of treatment. At the conclusion of the study, platelet count was virtually the same in both bFGF-treated groups irrespective of whether treatment was maintained or discontinued (2.0±0.4×105/mm3 versus 2.0±0.6×105/mm3, respectively). The circulating platelet count decreased to <105/mm3 in 9 bFGF-treated dogs. In 6 of these dogs, the greatest effect was observed between the second and fourth weeks of treatment, and the platelet count generally improved spontaneously as the study progressed (despite continued treatment with bFGF in 3 of 6 of these dogs). In 2 additional dogs treated with bFGF for 9 weeks and in 1 dog in which bFGF was discontinued after 5 weeks, the platelet count decreased to <105/mm3 during the final week of the study. Total leucocyte counts and serum chemistries were similar in treated and control animals.
Histology of Noncardiac Organs
Mild to moderate chronic interstitial nephritis and mild thickening of Bowman’s capsules were observed in both bFGF-treated and control animals. This was interpreted as a nonspecific reaction to the presence of microspheres within the renal parenchyma, and there was no difference between groups. Retinal digests were analyzed in 27 of 28 dogs. In all cases, the retinal vasculature was considered to be in the normal range with respect to vascular density; however, 5 of 7 dogs in the 9-week treatment group, 3 of 10 dogs in the 5-week treatment group, and 3 of 10 control dogs were identified as demonstrating an apparent slight increase in capillary tortuosity, albeit within the “normal range” (χ2 test, P=.16). The liver, lung, kidney, skin, spleen, small intestine, and bone marrow were free of significant abnormalities in all dogs.
The impetus for these investigations was our previous observation that bFGF promoted myocardial collateral development in dogs when injected directly into a coronary artery beyond the point of obstruction.8 Because of the clinical impracticality of administering bFGF distal to a stenosis in the setting of total or nearly total coronary obstruction, the primary aim of this investigation was to determine whether systemic bFGF administration also increased coronary collateral development and if so, whether this effect was sustained even after termination of bFGF delivery. We also sought to examine the pharmacokinetics, pharmacodynamics, and potential toxic effects of systemic bFGF administration. These goals were met in two phases. In phase I, it was found that systemic bFGF markedly increased collateral blood flow during the first week of administration (day 10 to 17, Fig 1⇑); however, in contradistinction to our previous investigation using intracoronary bFGF,8 the benefit attributable to bFGF was not maintained throughout the duration of the study. Thus, our initial study and phase I of the present study were concordant in that both demonstrated that bFGF accelerated collateral development; however, the advantage of bFGF treatment was only transient in phase I of the present study. Phase II of the present investigation was a more rigorous and extended study designed to better evaluate whether bFGF enhanced only the rate of collateral development or whether the peptide also induced a sustained increase in collateral perfusion. The phase II study design used two bFGF treatment arms (5 weeks and 9 weeks) to ascertain whether extension of the treatment duration could lead to further improvement in collateral flow and, conversely, to determine if withdrawal of treatment resulted in collateral regression. In addition, bFGF treatment was initiated 1 week earlier than in phase I (day 3 in lieu of day 10) to determine what effect, if any, earlier exposure to bFGF might have on collateral development.
Timing of bFGF Administration
Data of phase II suggest that the major effect of bFGF on collateral flow occurred during the period of maximal ischemia, during the 7-day interval 10 to 17 days after ameroid placement. An additional effect of bFGF was not apparent after this period in that the relation between collateral flow and time remained essentially parallel in treated and control dogs after day 17 (Fig 3⇑). Furthermore, comparison of the two bFGF-treated subgroups (5-week treatment plus 4-week placebo versus 9-week treatment) demonstrates that continued treatment with bFGF after 5 weeks imparted no additional benefit. These data suggest that maximal benefit can be expected when there is active ischemia, with little or no effect in its absence. Administration of bFGF during the interval 3 to 10 days after ameroid placement did not immediately improve collateral flow. It should be noted, however, that early bFGF treatment before the development of ischemia (days 3 to 10) could have had a delayed effect, resulting in tissue storage with salutary effects on collateral growth as ischemia progressed. Additional experiments in which treatment was limited to this 3- to 10-day interval would be required to test this hypothesis.
Enhancement of Collateral Flow: Short-term Versus Long-term Benefit
Relative blood flow data from phase I of this investigation demonstrate that bFGF accelerates collateral development with the advantage of treatment disappearing within 38 days. Such a “catch-up” phenomenon was observed neither in our previous study8 nor in phase II of the present study. These were three independent studies using different experimental designs; however, the experimental preparation was a virtual constant. If the IZ/NZ ratios from the 68 dogs of all three studies are combined and analyzed together, the difference between bFGF-treated dogs (administered by the intracoronary or systemic route) and control dogs is highly significant on days 17, 24, 31, 38, and 66 (Table 3⇓). Although the difference between groups was not significant on day 52, it must be noted that data for days 52 and 66 were derived solely from phase II of the present study, whereas data for the earlier time points were derived from all three studies. Taken together, these data strongly suggest that bFGF has a sustained beneficial effect on collateral perfusion.
Regression of Collaterals
Of note, collateral flow did not regress in the 5-week treatment group after treatment was discontinued. It is likely that once established, collaterals are maintained by mechanical forces such as increased flow velocity, pressure, or shear stress15 despite withdrawal of the growth factor. The fact that virtually all of the treatment benefit of bFGF was manifested during the 10- to 17-day interval after ameroid placement, coupled with the fact that bFGF-stimulated collaterals did not regress after withdrawal of treatment, suggest that treatment limited to 1 week might provide a maximal inducement for collateral development in this system. We have preliminary data suggesting that this is, in fact, the case.16
bFGF Is a Vasodilator
Systemic administration of bFGF has been reported to decrease arterial pressure in rats and rabbits,17 and we observed an acute hypotensive effect in dogs. Moreover, daily administration of bFGF had a persistent hypotensive effect (both in phase I and phase II), such that a significant decrease in blood pressure was apparent fully 24 hours after the preceding dose of bFGF was administered. This was particularly striking in dogs receiving bFGF for 9 weeks in phase II (Table 1⇑). The recovery in arterial pressure that occurred in the subgroup of dogs after bFGF was discontinued provides further evidence of a chronic hypotensive effect of the peptide. It had been reported previously that the elimination half-life of small doses of intravenous bFGF is in the range of 30 to 90 seconds,18 19 in which case a chronic hypotensive effect of bFGF is quite puzzling. Using a much larger dose of bFGF in this study and a solid-phase ELISA assay, we found the elimination half-life of the peptide to be 50 minutes, with a large volume of distribution consistent with extensive tissue binding. The latter could account for the chronic hemodynamic effects of the peptide in the present study.
On the surface, it might appear that the increase in collateral conductance (and IZ/NZ ratio) in bFGF-treated dogs merely reflected the vasodilator effect of bFGF; however, this is not plausible. First, blood flow measurements were made under the influence of chromonar, a potent coronary vasodilator that would override potential vasodilator effects of bFGF. Second, a temporary vasodilator effect of the peptide could not account for the increase in vascular density detected in bFGF-treated dogs. Third, collateral blood flow at the final time point was similarly improved in dogs treated with bFGF for 5 and 9 weeks despite the fact that bFGF treatment had been discontinued in the former group 4 weeks before final measurements of collateral flow were obtained. Thus, these data confirm that bFGF induced a permanent increase in maximal collateral flow independent of any temporary vasodilator effect.
Although the vasodilatory effect of bFGF could not have confounded our measurements of collateral flow (or certainly those obtained on day 66 in dogs in which bFGF was discontinued on day 38), it is possible that bFGF-induced vasodilatation may have played a causal role in promoting collateral formation. Dipyridamole (an adenosine-dependent vasodilator) has been demonstrated to augment myocardial capillary density in rats20 and enhance coronary collateral flow in swine.21 Thus, bFGF may have promoted collateral formation by its direct angiogenic effect as well as through a secondary vasodilator effect. Whether one or both of these mechanisms played a role in the development of collaterals and the relative contributions of these two potential mechanisms cannot be discriminated on the basis of our data.
Effect of Hematocrit on Collateral Flow
Severe chronic anemia (hematocrit of ≈11%×4 weeks) has been shown to decrease both minimum coronary vascular resistance and acute collateral resistance in dogs.22 These effects may be mediated by tissue hypoxia22 or decreased blood viscosity with increased shear stress.15 In comparison, the decrease in hematocrit in the present study was only moderate, and we have observed a salutary effect of bFGF on collateral blood flow in the absence of changes in hematocrit (using a lower dose of bFGF8 or more abbreviated treatment period16 ). Moreover, in both phases of the present study, collateral flow was significantly increased in the treated dogs on day 17, a time at which no significant change was observed in hematocrit. Therefore, it is unlikely that anemia played a major role in enhancing collateral growth in the present study.
Effect of bFGF on Infarct Size
Yanagisawa-Miwa et al23 have reported that acute bFGF administration reduces infarct size and preserves left ventricular function after acute myocardial infarction in dogs; however, bFGF did not affect infarct size significantly in our study. It must be pointed out that the study of Yanagisawa-Miwa et al used a model of acute coronary occlusion/ myocardial infarction. In the present study, the method and dynamics of coronary occlusion were entirely different, such that myocardial infarction was not intentionally created. Indeed, the percent of infarcted left ventricle was minimal, and the lack of an effect of bFGF administration on infarct size in our model is not surprising.
Toxicity of bFGF
Mazué et al24 25 have identified the kidney, long bones, and red blood cells as the main targets of toxicity of bFGF. The effects on these systems were dose dependent and reversed completely after treatment was discontinued. In the present investigation, the bFGF dose used was similar to the highest dose tested by Mazué et al, although we maintained treatment for as long as 9 weeks, in contrast to 4 weeks of administration used by Mazué et al. We observed anemia in bFGF-treated dogs with decreases in hematocrit of 8 and 12 vol% in phases I and II, respectively. In phase II, these changes began to resolve 2 weeks after treatment was discontinued. In addition, bFGF was associated with a significant decrease in circulating platelets (mean decrease of approximately 105/mm3). Although it was not statistically significant, there was a suggestion of a subtle increase in retinal vascular tortuosity observed in dogs treated with bFGF for 9 weeks. Thus, prolonged bFGF treatment was associated with untoward effects that might be avoided if the dose or duration of treatment were reduced. Certainly, the potential long-range toxicity of this peptide remains to be defined.
bFGF stimulates proliferation of the three primary vascular cell types (endothelial cells, smooth muscle cells, and fibroblasts) and has been found to induce vascular cell proliferation after arterial injury26 27 28 as well as endothelial cell regrowth after endothelial denudation injury.29 If atherosclerosis is construed as a form of low-grade vascular injury, then the possibility exists that bFGF could incite intimal smooth muscle cell proliferation in this setting. Moreover, intimal hyperplasia has been observed in coronary collateral vessels that develop in response to ameroid-induced coronary occlusion,30 raising the possibility that bFGF could foster collateral formation but ultimately impair collateral function by inducing intimal proliferation within these vessels. We observed intimal proliferation in isolated epicardial and intramyocardial coronary arteries in both treatment groups. Statistical analyses dictate that a difference between groups could not be ascribed to bFGF; however, we believe that further studies are needed to elucidate the long-term effects of bFGF on intimal smooth muscle cell proliferation, a pertinent issue in atherosclerosis and restenosis after angioplasty.
We made no attempt to ascertain the timing of ameroid-induced LCx occlusion in these animals, although we attempted to exert some measure of control on this process by permanently inflating the hydraulic occluder on day 24. On the basis of the hypothetical concerns raised above, it is possible that bFGF induced intimal hyperplasia at the site of the ameroid, thereby accelerating the rate of LCx occlusion. If this were the only effect of bFGF, the peptide would be expected to shift the IZ/NZ versus time relation to the left, which is essentially what was observed in the phase I study. Such a mechanism could not explain the sustained disparity between groups observed in phase II of this study, however. If bFGF had no effect other than to accelerate LCx closure, then collateral flow in control dogs should have reached parity with that of treated dogs at some point after LCx occlusion occurred.
Systemic administration of bFGF enhanced coronary collateral conductance in dogs with progressive single-vessel coronary occlusion. Prolonged treatment with bFGF was associated with decreases in hematocrit and circulating platelets. The possibility that bFGF could induce intimal hyperplasia in atherosclerotic vessels, arteries subjected to recent angioplasty, or collaterals themselves remains an unsettled issue. However, the data demonstrate that treatment beyond the period of active myocardial ischemia is not of further benefit and suggest in fact that bFGF treatment might be effective if reduced to as few as 7 days. Thus, the present investigation provides further impetus to the concept that collateral development can be enhanced pharmacologically, specifically by bFGF, raising the possibility that such an intervention might eventually be applied clinically. Before then, however, additional studies are warranted to determine whether therapy limited to shorter time periods would enhance collateral flow while limiting or avoiding potential adverse effects.
The authors wish to express their gratitude to Victoria Hampshire, DVM, John Bacher, DVM, and their staffs for providing expert veterinary care, to Nora M. Laver, MD, National Eye Institute, NIH, Bethesda, Md, for analyzing retinal digests, and to Arlene Lobo for technical assistance. We would also like to acknowledge William O. Ballew for clinical laboratory support and Phyllis Sholinsky, MSc, Epidemiology Branch, National Heart, Lung, and Blood Institute, for providing statistical analyses of ancillary (non–blood flow) data of phase I. We also would like to acknowledge Scios Nova and Hoechst-Roussel for supplying bFGF and chromonar, respectively.
1 Principal Investigators.
Presented in part at the 66th Annual Scientific Sessions of the American Heart Association, Atlanta, Ga, November 8-11, 1993.
- Received April 27, 1994.
- Accepted August 8, 1994.
- Copyright © 1995 by American Heart Association
Levin DC. Pathways and functional significance of coronary collateral circulation. Circulation. 1974;50:831-837.
Kardami E, Fandrich RR. Basic fibroblast growth factor in atria and ventricles of the vertebrate heart. J Cell Biol. 1989;109:1865-1875.
Speir E, Tanner V, Gonzalez AM, Farris J, Baird A, Casscells W. Acidic and basic fibroblast growth factors in adult rat heart myocytes: localization, regulation in culture, and effects on DNA synthesis. Circ Res. 1992;71:251-259.
Banai S, Jaklitsch MT, Shou M, Lazarous DF, Scheinowitz M, Biro S, Ferrara N, Epstein SE, Unger EF. Angiogenic induced enhancement of collateral blood flow to ischemic myocardium by vascular endothelial growth factor in dogs. Circulation. 1994;89:2183-2189.
Laver NM, Robison WG Jr, Pfeffer BA. Novel procedures for isolating intact retinal vascular beds from diabetic humans and animal models. Invest Ophthalmol Vis Sci. 1993;34:2097-2104.
Miller RG Jr. Simultaneous Statistical Inference. New York, NY: Springer Verlag; 1981.
Scheel KW, Fritzgerald EM, Martin RO, Larsen RA. The possible role of mechanical stresses on coronary collateral development during gradual coronary occlusion. In: Schaper W, ed. The Pathophysiology of Myocardial Perfusion. Amsterdam: Elsevier North-Holland Biomedical Press; 1979:489-518.
Lazarous DF, Shou M, Scheinowitz M, Epstein SE, Unger EF. 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. 1994;23(suppl A):177A. Abstract.
Cuevas P, Carceller F, Ortega S, Zazo M, Nieto I, Giménez-Gallego G. Hypotensive activity of fibroblast growth factor. Science. 1991;254:1208-1210.
Whalen GF, Shing Y, Folkman J. The fate of intravenously administered bFGF and the effect of heparin. Growth Factors. 1989; 1:157-164.
Symons JD, Firoozmand E, Longhurst JC. Repeated dipyridamole administration enhances collateral-dependent flow and regional function during exercise: a role for adenosine. Circ Res. 1993;73:503-513.
Yanagisawa-Miwa A, Uchida Y, Nakamura F, Tomaru T, Kido H, Kamijo T, Sugimoto T, Kaji K, Utsuyama M, Kurashima C, et al. Salvage of infarcted myocardium by angiogenic action of basic fibroblast growth factor. Science. 1992;257:1401-1403.
Mazué G, Bertolero F, Garofano L, Brughera M, Carminati P. Experience with the preclinical assessment of basic fibroblast growth factor (bFGF). Toxicol Lett. 1992;64/65:329-338.
Lindner V, Majack RA, Reidy MA. Basic FGF induces the proliferation of vascular cells in injured arteries. FASEB J. 1990;4:A625. Abstract.
Lindner V, Reidy MA. Proliferation of smooth muscle cells after vascular injury is inhibited by an antibody against basic fibroblast growth factor. Proc Natl Acad Sci U S A. 1991;88:3739-3743.
Cohen MV, ed. Coronary Collaterals: Clinical and Experimental Observations. Mount Kisco, NY: Futura; 1985:289-294.