Introduction

Top Abstract Introduction Methods Results Discussion References

Several members of the fibroblast growth factor family (Klagsbrun, 1991, Folkman and Shing, 1992) including bFGF, acidic fibroblast growth factor (aFGF) and fibroblast growth factor 5 (FGF-5) have been suggested as potential therapeutic agents for amelioration of chronic myocardial ischemia (Unger et al., 1994, Harada et al., 1994, Engelmann et al., 1993, Giordano et al., 1996). bFGF has been studied most extensively in this regard, and we have previously demonstrated that local administration of 5 µg of bFGF in a sustained release preparation results in better coronary blood flow in the ameroid constrictor-compromised territory during pacing stress that was associated with preservation of regional left ventricular regional function in the treated, compared to untreated (control) animals (Harada et al., 1994). It is important to note, however, that the observed beneficial effects of bFGF in this study were manifested as a lack of decline in the measured parameter (coronary flow, regional wall motion) during pacing stress. It is conceivable, therefore, that higher dosages of the growth factor might not only result in the preservation of function during stress but in improvement in resting parameters.

However, little is known about the dose-response properties of this growth factor. This consideration is particularly important given broad spectrum of biological activity associated with bFGF and its well known vasoactive properties (Cuevas et al., 1991a; Sellke et al., 1994, 1996). Therefore the desire to obtain maximal functional benefits in treatment of myocardial ischemia by administration of large doses of bFGF must be tempered by considerations of its potential toxicity (Mazue et al., 1991). These effects may include hypotension, anemia related to bone marrow suppression and renal damage due to glomerular toxicity (Mazue et al., 1991). With these considerations in mind, we undertook our study to determine whether a dose-response relationship exists in regard to bFGF effects on myocardial flow and function.

	Methods

Top Abstract Introduction Methods Results Discussion References

Initial surgery. Twenty two Yorkshire pigs (25-35 lbs, Pine Acres Farm, Norwood, MA) were intubated and mechanically ventilated under general Halothane anesthesia after administration of ketamine (10 mg/kg i.m.) and pentobarbital (30 mg/kg i.v.). After left lateral thoracotomy the left circumflex artery was isolated after its takeoff from the left main artery proximally to any major arterial branches. As previously described (Harada et al., 1994), a 2.75- to 3.0-mm ameroid occluder (Research Instruments, Corvallis OR), matched to the LCX diameter, was placed around the vessel. All animals were randomly assigned to one of three treatments: perivascular administration of bFGF at 10 µg (n = 7) or 100 µg (n = 8) via heparin-alginate microspheres, or control administration of inert EVAc polymer alone (n = 7), without active growth factor. Postoperatively all animals were treated with antibiotics for 48 hr, and narcotic analgesics were used as needed. All animals were cared for according to National Institute of Health guidelines for the care and use of laboratory animals and the protocol was approved by IACUC.

Growth factor and delivery system preparation. Heparin-alginate microspheres were prepared as previously described (Harada et al., 1994, Edelman et al., 1993). Briefly, heparin-Sepharose beads (Pharmacia LKB, Piscataway, NJ) sterilized under ultraviolet light were mixed with filter-sterilized sodium alginate (1.2%, w/v; Sigma Chemical Co., St. Louis, MO). The slurry was then dropped through a needle into a beaker containing a hardened solution of CaCl₂ (1.5% w/v) leading to instantaneous bead formation. The beads were washed three times with sterile water and stored in 0.9% NaCl/l mM CaCl₂ at 4°C. Each capsule in its hydrated state contained 0.05 mg heparin-Sepharose, 0.18 mg of alginate and 11 mg of water. For heparin-alginate capsule loading with bFGF, on the evening before surgery, 13 or 130 µg of sterile bFGF (Scios Nova Inc., Mountain View, CA) were added to a gelatin coated cryotube containing five to six sterile microspheres, incubated overnight at 4°C with gentle agitation and washed before use. Previous work has demonstrated 80% incorporation of bFGF into microspheres via this method, thus resulting in approximately 10 or 100 µg of bFGF/set of five microspheres (Edelman et al., 1992).

Follow-up evaluation. Coronary angiography was performed on all animals at about 8 wk after initial surgery. After administering pentobarbitol and halothane inhalation anesthesia, animals were mechanically ventilated under constant hemodynamic monitoring. A 7F JR4 diagnostic angiography catheter (Cordis Corp, Miami, FL) was introduced over a 0.035-inch J wire via a femoral artery cutdown and selective coronary angiography was performed on the right and left coronary artery in multiple LAO and RAO projections, with ionic contrast (Renograffin, Squibb Diagnostics, Princeton, NJ).

Angiographic analysis of collateral density. Evaluation of angiographic collateral density was performed through cine film review by two experienced angiographers, blinded to treatment group. Angiographic analysis consisted of 1) documentation of complete vessel occlusion in the proximal LCX artery at the site of the ameroid, 2) assessment of collateral vessel development in the left circumflex region by the "collateral index" and 3) determination of TIMI grade flow within the left circumflex vessel. The collateral index is a well-established scale for assessing collateral vessel density, where collateral vessels within a specified region are described on a 0 to 3 scale (0, no visible collateral vessels; 1, faint filling of side branches of the main epicardial vessel, without filling the main vessel; 2, partial filling of the main epicardial vessel and 3, complete filling of the main vessel) (Rentrop et al., 1985, Fujita et al., 1988). TIMI grade flow within the proximally occluded vessel was used to assess the adequacy of flow within the given myocardial region (TIMI Study Group, 1988) TIMI flow was assessed using the following gradations: TIMI 0, no flow within the native vessel; 1, faint, slow filling of the native vessel, without opacification of the distal vessel; 2, slow filling of the entire vessel length and 3, brisk, normal flow within the entire vessel. On several occasions, differences existed in Cine film interpretation, and in these instances, collateral index and TIMI grade flow results were adjudicated between readers.

Echocardiographic analysis of regional and global myocardial function. Echocardiographic parameters were assessed from standard M-mode and two-dimensional echocardiographic images (Hewlett Packard, Andover, MA) obtained in the open-chest state. Images were compared using apical four-chamber and midventricular short-axis planes, before and after 2 min pacing. Infarct size was determined as the % akinetic/contractile endocardial circumference in the short axis plane. Global ejection fraction was determined from the four-chamber view using a modified Simpson's algorithm (Stamm et al., 1982). Regional wall thickening was determined from short axis M-mode recordings images in the mid left ventricular region oriented to include the lateral wall.

Evaluation of regional coronary flow. Colored microspheres (15 ± 0.1 µm diameter, Triton Technology Inc. San Diego, CA) were used to determine coronary blood flow (Kowallik et al., 1991) at the time of ameroid placement as well as during the final study as previously described (Harada et al., 1996). For data presentation, coronary flow is given as a weighted average of flows per gram of tissue, in subendocardial, midmyocardial and subepicardial regions.

Statistics. All data are expressed as mean ± S.D. P  .05 was considered significant. Comparison of angiographic collateral density via the collateral index and TIMI grade flow was assessed between groups by using two-sided Kruskal-Wallis (multiple group comparison) and Wilcoxon rank-sum (two group comparison) tests for nonparametric, ordinal data. continuous variables including echocardiographic comparison of regional and global left ventricular function, regional coronary flow and coronary resistance were compared by using one way analyses of variance and a Student's t test with Bonferroni correction. All t tests were two-tailed. Statistics were calculated with the commercially available statistical software packages SigmaStat 2.0 (Jandel Scientific, San Francisco, CA) and Origin 4.1 (Microcal Software, Northampton, MA).

	Results

Top Abstract Introduction Methods Results Discussion References

Study groups. A total of 22 animals survived the initial surgery. Two animals (bFGF-100 µg group) died suddenly 1 to 3 wk after surgery and were excluded from the analysis; two more animals (10 and 100 µg bFGF groups) died during anesthesia induction during the final study. Thus, seven animals in the control group 6 in the 10 µg bFGF group, and five in the 100 µg bFGF group, are included in this analysis. The goals of this study were to assess the dose response of bFGF administration in this ameroid constrictor model. Therefore, the data are presented as comparisons between bFGF groups and controls.

Coronary angiography. Coronary angiography was used to document occlusion of the ameroid-instrumented artery, to evaluate the extent of collateral formation around the occlusion (collateral index) and to assess the flow in the distal portion of occluded LCX (TIMI grade). Angiography documented LCX occlusion in all 18 pigs included in this analysis. Analysis of the collateral index (fig. 1) demonstrated a difference between the three groups (Kruskal-Wallis test, P = .04), with combined bFGF animals and the 10 µg bFGF group differing from the control group on multiple comparison testing (Dunn's method and Wilcoxon rank sum test, P < .05). However, there was no difference between the two bFGF groups by this analysis. Analysis of TIMI flow in the distal segment of the occluded LCX artery (fig. 2), also revealed that TIMI flow grade was different between groups (Kruskal-Wallis test, P < .01), with the combined bFGF animals and the 100 µg bFGF group differing from the control animals (Dunn's method and Wilcoxon rank sum test, P < .05). Again, there was no statistical difference between the two bFGF groups.

View larger version (K): [in this window] [in a new window]   Fig. 1.   Angiographic assessment of collateral index. Shown is a graph of angiographic collateral vessel evaluation using the collateral index scale. Each treatment group is graphed as the % of observations at each level from grade 0 (no visible collateral vessels) to grade 3 (complete filling of the main vessel). Black bars, controls; stippled bars, 10 µg bFGF-treated animals; gray bars, 100 µg bFGF-treated animals.

View larger version (K): [in this window] [in a new window]   Fig. 2.   Angiographic assessment of coronary flow in the occluded artery. Graph of angiographic collateral vessel evaluation using TIMI flow scale. As in figure 3, each treatment group is graphed as the % of observations at each level from TIMI grade 0 (no visible flow) to grade 3 (brisk flow). Black bars, controls; stippled bars, 10 µg bFGF-treated animals; gray bars, 100 µg bFGF-treated animals.

Assessment of regional myocardial flow. At the time of final study, there was no significant difference in left ventricular systolic pressure (control: 111 ± 6, bFGF-10: 109 ± 9, bFGF-100: 110 ± 6 mm Hg) or heart rate (control: 107 ± 9, bFGF-10: 110 ± 10, bFGF-100: 106 ± 7 beats/min) between the groups. Coronary blood flow in the LAD (nonischemic) territory at rest was similar in all three groups [coronary blood flow (ml/min · g): control: 0.80 ± 0.09; bFGF-10: 0.92 ± 0.11, bFGF-100: 1.14 ± 0.14, P = NS]. However, coronary flow in the LCX territory at rest was significantly higher in bFGF-treated compared to control animals (analysis of variance, P < .001) with LCX flow in both bFGF groups (fig. 3) significantly higher than controls (Bonferroni t test, P < .05). However, there was no significant differences between the two bFGF groups.

View larger version (K): [in this window] [in a new window]   Fig. 3.   Coronary blood flow. Bar graph demonstrating coronary blood flow (ml min1 g1) in the left anterior descending (LAD, gray bars) and the left circumflex (LCX; dark bars) territory for each treatment group. All comparisons are shown against the control group (*P < .05).

View larger version (K): [in this window] [in a new window]   Fig. 4.   Coronary resistance in the LCX territory. Scatter plot of coronary resistance in the LCX territory (mm Hg/ml1 min1 g1) in controls (0 mg bFGF) and in animals treated with 5 µg bFGF (data from Harada et al., 1994) and 10 and 100 µg bFGF (our study).

Echocardiographic evaluation of regional and global myocardial function. Two-dimensional and M-mode echocardiography were used to measure left ventricular global and regional function (figs. 5 and 6)) in open-chest animals. Treatment with both bFGF doses significantly increased left ventricular ejection fraction (fig. 5) compared to controls, both at rest (analysis of variance, P = .004), as well as during rapid pacing (analysis of variance, P = .006). Although treatment with both bFGF doses resulted in significantly higher ejection fractions than controls (Bonferroni t test, P < .05 for both bFGF vs. control), there was no significant difference between the two bFGF groups either at rest or during pacing.

View larger version (K): [in this window] [in a new window]   Fig. 5.   Left ventricular ejection fraction. Global left ventricular ejection fraction, obtained using open-chest echocardiography, for each treatment group at rest (dark bars) and during pacing stress (gray bars) evaluation. All comparisons are shown against the control group (*P < .05).

View larger version (K): [in this window] [in a new window]   Fig. 6.   Regional left ventricular wall function. Left ventricular wall thickening in the posterior wall (LCX) region, obtained using open-chest echocardiography, for each treatment group at rest (dark bars) and during pacing stress (gray bars) evaluation.

	Discussion

Top Abstract Introduction Methods Results Discussion References

A number of recent investigations has described therapeutic applications of heparin binding growth factors including aFGF, bFGF, FGF-5 and VEGF. Basic FGF have received the most attention with a number of investigators demonstrating improvement in coronary flow (Unger et al., 1994, Harada et al., 1994), left ventricular function (Harada et al., 1994) and myocardial infarct sizes (Battler et al., 1993; Yanagisawa-Miwa et al., 1992). Although the precise mechanism of this beneficial effect of bFGF therapy has not been defined, investigations have centered on bFGF's ability to induce growth of new vessels although other biological activities attributable to this growth factor including a cardioprotective effect in ischemia and/or hypoxia (Padua et al., 1995) or its ability to induce vasodilation (Cuevas et al., 1991a-b; Sellke et al., 1994) may well contribute to the observed therapeutic effects.

However, little is known about the dose-response properties of bFGF and to date there have been no angiographic studies assessing its ability to stimulate myocardial angiogenesis. These considerations are particularly important given the pluripotent biological properties of bFGF and known toxicity associated with high-dose bFGF administration, which could preclude its clinical use. Thus, determination of incremental benefits with larger doses of locally delivered bFGF is important in any effort to maximize the advantages of local delivery while minimizing systemic toxicity. Therefore the major goal of our study was to examine whether a dose-response effect exists in regard to local heparin-alginate delivery of bFGF.

The benefits of local heparin-alginate growth factor administration include relative ease of polymer preparation and surgical manipulation, as well as zero order kinetics of release that provides sustained delivery over a period of several weeks (Edelman et al., 1993; Lopez et al., 1996). Growth factor release studies in this model demonstrated detectable bFGF serum levels that increased within 15 min of growth factor application and remained elevated for up to 4 wk (Lopez et al., 1996). However, these detectable serum levels were not associated with untoward hemodynamic effects or any evidence of systemic toxicity (Lopez et al., 1996).

Animals treated with both low (10 µg) and high (100 µg) doses of bFGF demonstrated improvement in resting coronary flow, collateral resistance, global and regional left ventricular function as well as angiographically determined collateral index and TIMI flow grade. In all of these parameters, both bFGF dosages produced fairly consistent and similar results, although improvement in regional wall motion during pacing was significantly better in the 100 µg bFGF-treated compared to the 10 µg bFGF-treated animals. Comparison of these two bFGF groups with our previous study that used a 5-µg bFGF dose (Harada et al., 1994) demonstrates a significant dose-related improvement in LCX coronary resistance with each increase in dosage. Furthermore, animals treated with the 5-µg dose demonstrated improved LCX territory perfusion compared with control animals only during pacing stress (Harada et al., 1994), although both 10 and 100 µg doses in our study resulted in detectable improvement in coronary flow at rest. Overall these results suggest that a dose-response effect is present with an increasing dose of locally delivered bFGF regarding physiological functional measurements and collateral vessel flow. However, analysis of coronary resistance data, although limited by small numbers of dose data points, suggests a plateau effect on resting coronary resistance between 20 to 30 µg of bFGF.

This improvement in resting LCX perfusion correlated with angiographic observation of increased collateral density and increased TIMI flow grade in both bFGF groups. Of interest, angiographic collateral formation was observed distal to the ameroid occluder with most of the collaterals originating from mid-LAD with an occasional collateral arising from the distal RCA, although peri-ameroid collaterals were also observed. A combination of factors probably accounts for this pattern. Heparin-alginate delivery is known to result in the distal transport of bFGF along the vessel wall from the delivery site (Edelman et al., 1993). In addition, more severe ischemia in the distal myocardial segments may have also influenced collateral formation.

Several issues need to be considered in evaluating the results of our study. A control group used in the study received inert EVAc and not heparin-alginate pellets although study results in this group do not differ significantly from previous control groups that have used non-FGF-treated heparin-alginate microspheres or other inactive polymers (Harada et al., 1994, 1996). In addition, although all animals had completely occluded LCX artery, we cannot exclude the possibility that bFGF may have influenced the rate of ameroid-induced vessel closure compared to control animals. This possibility is potentially relevant given known vasoactive properties of this growth factor and its already discussed ability to induce vasodilation in coronary beds as well as its potential cardioprotective activity (Padua et al., 1995). Thus, bFGF-mediated delay in the time of vessel occlusion may have influenced results in the treatment groups. The absence of significant differences in the left ventricular infarct size between bFGF and control groups makes this possibility relatively unlikely.

In summary, we have demonstrated that local perivascular delivery of bFGF results in a dose-dependent improvement in regional coronary flow and myocardial function. Whether this incremental improvement in myocardial function and perfusion warrants a potential increase in risk associated with higher-dose bFGF therapy will require further investigation.