INTRODUCTION

Top Abstract Introduction Methods Results Discussion References

IT IS WELL DOCUMENTED that enhanced physical activity is an important means of managing patients with peripheral arterial insufficiency (18, 19). Exercise training can induce adaptations that enhance exercise tolerance in patients with intermittent claudication (12, 14, 19, 23, 40). Two aspects of vascular adaptations are probably important. First, increased physical activity enhances capillary density within active muscles (20). This could involve the action of angiogenic growth factors, as Breen and co-workers (9) observed an increase in the expression of vascular endothelial growth factor (VEGF) after acute exercise. The enhanced capillarity should establish a shorter average diffusion path length, a greater capillary surface area for nutrient exchange, and/or a prolonged capillary red blood cell transit time. These changes could contribute to the increase in blood/tissue oxygen exchange observed previously (35, 36) and account for the greater oxygen extraction observed across the active limbs of claudicant patients after exercise training, even in the absence of an increased in limb blood flow (30, 40).

Second, collateral blood flow is increased by exercise training in experimental peripheral insufficiency (37, 38) and in many, but not all, clinical studies evaluating patients with claudication (28). This latter response is most significant, since it implicates vascular remodeling of the upstream conduit vessels (34, 37) that become manifest upon vascular occlusion of a primary supply artery (2, 5-7, 11, 32). Factors responsible for inducing vascular remodeling of these upstream collateral vessels within the thigh are not likely due to frank ischemia, since the surrounding tissue has a relatively high blood flow, well in excess of the more ischemic distal calf muscle (21, 22). Rather, it has been argued that altered flow dynamics in the existing collateral vessels impart a stimulus to enlarge the vessels (21). This could come about by an increase in pressure and/or flow (destined for the lower leg) and result in a greater wall tension and/or shear. A similar process could also be important in the basic fibroblast growth factor (bFGF)-induced increase in collateral blood flow observed in claudicant rats (34). Angiography of the upper thigh vasculature has identified expanded conduit vessels that likely served as the collateral channels that increased flow delivery downstream. If enhanced flow through the collateral vessels of the thigh is a critical determinant, then exercise training could amplify the bFGF-induced collateral vessel expansion. Thus the purpose of this study was to evaluate whether the increased blood flow obligated by daily exercise could enhance collateral vessel development in response to bFGF after bilateral ligation of the femoral arteries.

	METHODS

Top Abstract Introduction Methods Results Discussion References

Animal care. Adult male Sprague-Dawley rats (~325 g; from Taconic Farms, Germantown, NY) were housed in a temperature-controlled room (20 ± 10°C) with a 12:12-h light-dark cycle and had free access to Purina Lab Chow and tap water ad libitum. One day after arrival, all animals were run 5-10 min/day for 5 days to become familiarized with the treadmill protocol.

Experimental design. The effect of intra-arterial infusion of bFGF on collateral-dependent blood flow was assessed in rats that were kept sedentary or exercised by treadmill walking after bilateral ligation of the femoral arteries. Femoral artery ligation markedly reduces blood flow reserve in the distal hindlimb tissue by ~80% without affecting resting blood flow needs (37-39).

Ligated rats were divided into two treatment groups: carrier controls (n = 13) and bFGF (1 µg/day)-infused group (n = 15); animals from each group were either kept sedentary, by being limited to cage activity, or trained by treadmill running for ~4 wk. bFGF was infused into the left iliac artery near the source of the collateral vessels at an effective dose for 2 wk, since this infusion period was shown to be optimal for collateral development (34). At the end of exercise training, rats were randomly taken from each group to evaluate calf muscle performance and collateral-dependent blood flow using an isolated perfused hindquarter preparation.

Femoral artery ligation and osmotic pump installation. The surgical procedure for ligation of the femoral artery has been described in detail previously (34). In brief, under ketamine-acepromazine anesthesia (100 and 0.5 mg/kg body wt), the right femoral artery was ligated just distal to the inguinal ligament; topical antibiotic powder (Neo-Predef; Upjohn) was placed on the wound before closure. The left artery was similarly ligated, except that a catheter (PE-60) was inserted upstream through the ligature; the tip of the catheter was near the opening of the hypogastric artery in the iliac artery. The catheter was connected to an osmotic pump containing either carrier or bFGF. A subcutaneous tunnel was made under the abdominal skin for placement of the pump before catheterization. The incision was closed with skin clips after application of topical antibiotic powder.

Miniosmotic pumps delivering 0.50 ± 0.02 µl/h with 14-day capacity (Alzet model 2002; Alza, Palo Alto, CA) were used after preparation according to the manufacturer's instructions. The absence of any residual pump volume (<5%) at the end of the experiment verified delivery of the contents. The bFGF pump contained human recombinant bFGF at a dose of 1.0 µg/day for the bFGF-infused group (~325 g rat) or the same volume of phosphate-buffered saline for carrier controls. Each pump also contained heparin (100 IU · day¹ · rat¹) to complex with the bFGF, 5% glycerol as a protein stabilizer, and 0.02% sodium azide as a bacteriostatic agent, as done previously (34).

Exercise training. Exercised rats from carrier control (n = 6) and bFGF-treated group (n = 8) were run on a rodent treadmill (Quinton model 42-15) daily for 4 wk. On the second day after femoral ligation and insertion of the infusion pump, rats were exercised two times a day at 20 m/min up a 15% incline until exhibiting signs of fatigue; fatigue is preceded by a characteristic gait change to a gallop with exaggerated hops. The time to fatigue was recorded in minutes.

Hindquarter preparation. At the designated time, rats were randomly taken from each group for hindquarter perfusion preparation as described in detail previously (17). In brief, after anesthesia with pentobarbital sodium (60 mg/kg), a PE-50 catheter was placed into the left carotid artery of the rat for later fluid infusion. An incision on the abdominal midline allowed access to and removal of the testes, bladder, seminal vesicles, prostate gland, and small and large intestines to expose the descending aorta and vena cava. Approximately 1.5-cm lengths of the descending aorta and vena cava, just distal to the renal artery branch, were dissected free and prepared for cannulation. Next, the skin was removed from both hindlimbs. The left hindlimb was prepared for muscle stimulation (17); the left sciatic nerve was exposed in its course through the thigh via an incision made laterally. The trunk of the sciatic nerve near the gluteus muscle was ligated, cut, and ready for placement on the electrodes. When perfusion began, the free end of the nerve was placed over a platinum bipolar electrode connected to a Grass model S48 stimulator. Isometric tension of the gastrocnemius-soleus-plantaris (GPS) muscle group was measured by attaching the Achilles tendon to a Cambridge level system (Series 300 B; Cambridge Technology, Watertown, MA). The left distal hindlimb was immobilized by a bone pin drilled though the distal end of the femur.

Perfusion procedure. After surgical preparation, the animal was transferred to a temperature-regulated (37°C) Plexiglas chamber. After an intra-arterial (carotid) injection of heparin (2,000 IU), the abdominal aorta or vena cava was cannulated with 16- and 14-gauge Teflon catheters, respectively (IV catheter; Becton-Dickinson, Rutherford, NJ). Once the perfusion catheters had been inserted and secured in place, perfusion was begun immediately; the animals were then killed with 0.5 ml pentobarbital sodium injected via the catheter in the carotid artery. The period of ischemia before initiating perfusate flow was generally <15 s. The perfusion medium (~300 ml) was recirculated after an initial volume of ~30 ml of venous effluent had been discarded. Flow rate and perfusion pressure were gradually increased until the net aortic pressure, measured by an on-line transducer, reached 100 mmHg. When perfusion pressure and perfusion rate were stable, muscle contractions were started. At the end of the muscle contraction sequence, blood flow was determined with radiolabeled microspheres. Ligatures were put at the base of the rat tail and the right and left hind feet to block perfusion flow to these areas. Tissues of both hindlimbs and trunk, between the kidney and the base of the tail, were dissected, as previously described (34-38). All tissues were counted at ~1% error with a Gamma 8000 counter to determine radioactivity in each individual muscle sample.

Perfusion medium. The perfusion medium consisted of Krebs-Henseleit bicarbonate buffer (17) containing washed bovine red blood cells, 4 g/100 ml bovine serum albumin, 100 µU/ml bovine insulin, amino acids at concentrations typical for rat blood, and 5 mmol/l glucose (maintained constant with periodic additions of glucose during perfusion). Hematocrit of the perfusion medium was at ~39-40%.

Muscle stimulation. Passive tension (~150 g) was adjusted to establish maximal active tension. Maximal tetanic contractions were elicited with supramaximal square-wave pulses (8 V at 0.1-ms duration) delivered in 100-ms trains at 100 Hz (Grass model S48 stimulator; Grass Instruments, Quincy, MA). Ten-minute periods of contractions progressing over a broad range of energy demands was elicited using 4, 8, 15, 30, and 45 tetani/min stimulation conditions. Tension output of the GPS muscle group was constantly recorded with a Gould five-channel polygraph (Gould, Cleveland, OH). Average tension for each contraction period was calculated from the tension measured at the 5th and 10th min of contractions.

Blood flow distribution. The distribution of blood flow in the hindlimb was determined with radiolabeled ⁸⁵Sr microspheres at the end of calf muscle contraction as described previously (17). The fraction of radioactivity found in a tissue, relative to the total activity infused, is a measure of its fractional blood flow. Absolute blood flow was calculated as follows where CPM_tissue is counts per minute in tissue and CPM_{total injected} is total counts per minute injected. Muscle blood flow is expressed per 100 g of tissue weight.

Muscle sampling and histochemical analyses. The superficial white (predominately type IIb) and deep red (predominately type IIa) sections of the gastrocnemius muscle and the soleus muscle were dissected from the noncontracting hindlimb (right side) immediately before hindquarter perfusion. Five-millimeter-thick cross sections were cut with a sharp blade, fixed to pieces of cork, and frozen in isopentane cooled by liquid nitrogen. These muscle samples were stored at 80°C for histochemical analysis.

Muscle capillarity was determined by first reacting for alkaline phosphatase, followed by the periodic acid-Schiff reaction for glycogen and glycoprotein after removal of glycogen with diastase and subsequently by counterstaining with metanil yellow, as done previously (34-36). The capillaries are rendered blue-black, the fibers yellow, and the external lamina just outside the sarcolemma magenta. Muscle sections from carrier controls and bFGF-treated animals were paired on the same slide. Myocyte and capillary number and number of capillaries surrounding fibers (capillary-to-fiber contacts) were determined from 20 nonoverlapping fields (0.06 mm² each). The capillaries around fibers were tabulated for at least 100 fibers for each section.

Biochemical analysis. The right plantaris muscle (mixed-muscle fiber-type region) was resected before hindquarter perfusion and stored at 80°C until analyzed for citrate synthase activity, a marker enzyme for aerobic capacity, as done previously (36).

Materials. Human recombinant bFGF was obtained from Scios (Mountainview, CA). Radioactive ⁸⁵Sr-labeled microspheres (14.2 ± 0.92 µm) with a specific activity of 9.6 mCi/g were obtained from NEN (Boston, MA) in a suspension of 10% dextran containing 0.05% Tween 80 surfactant. Beef lung heparin sodium (10,000 U/vial) was obtained from Upjohn (Kalamazoo, MI). Bovine serum albumin (fraction V) and the reagents used for biochemical analyses were obtained from Sigma Chemical (St. Louis, MO). Fresh bovine red blood cells were prepared by extensive washing (>20 vol) of acid-citrate-dextrose titrated blood collected at a local meat packer.

Analysis of the data. The results are presented as means ± SE. Two-by-two analyses of variance were applied, with individual mean comparisons made by Tukey's procedure (31). Significant differences for main treatment effects and individual comparisons were recognized at P < 0.05.

	RESULTS

Top Abstract Introduction Methods Results Discussion References

Training response. Running performance increased markedly over the first week of running to ~30 min/session (cf. Fig. 1); thereafter, exercise performance remained unchanged over time. Total run time per day was ~1 h with exercise sessions in the morning and afternoon. Exercise tolerance was not different between the carrier and bFGF groups (F = 0.49); therefore, the response has been combined in Fig. 1 for clarity of presentation. Because we have repeatedly shown that the exercise tolerance of untrained animals with peripheral arterial insufficiency remains relatively unchanged over time (36, 38, 39), we did not evaluate exercise tolerance in the sedentary groups in this study. An increase (P < 0.001) in citrate synthase activity in an active muscle (e.g., plantaris), typical of endurance-type exercise training (3), was observed in the carrier (41 ± 2.3 µmol · min¹ · g¹) and bFGF (33 ± 2.1 µmol · min¹ · g¹)-infused trained groups. The citrate synthase activity was not different between the sedentary carrier (26 ± 1.8 µmol · min¹ · g¹) and bFGF-infused groups (22 ± 1.8 µmol · min¹ · g¹).

View larger version (19K): [in this window] [in a new window]   Fig. 1.   Treadmill run time for each exercise session (rats were run 2 times/day). Response of the carrier and basic fibroblast growth factor (bFGF)-infused groups were not different and have been combined for clarity.

Body weights of the carrier and bFGF-trained rats were less (P < 0.001) than the corresponding sedentary groups limited to cage activity. However, hindlimb tissue weights were not different between the groups (Table 1); this is consistent with a reduced fat accrual that occurs in adult male rats that are exercise trained (35, 37).

Perfusion conditions. The perfusion pressures were essentially maintained at ~100 mmHg for all rats (Table 1). However, a greater perfusion inflow rate was necessary to establish this same 100 mmHg aortic pressure in the bFGF-treated groups (sedentary and trained) compared with the carrier groups (P < 0.001, Table 1). The similar aortic pressures should establish the same pressure head at the source of the collateral vessels that feed the distal hindlimb tissue in all of the animals.

Hindlimb blood flow. There were significant main treatment effects for both bFGF administration and exercise training in blood flows to the hindlimb tissues; however, the nature and/or magnitude of the differences were dependent on specific tissues. As illustrated in Table 2, bFGF significantly increased (P < 0.05) blood flow to the total hindlimb; the addition of training further increased (P < 0.01) the bFGF effect. On the other hand, significant increases in blood flows to the proximal hindlimb tissues of the bFGF-treated groups were dependent on training (i.e., a significant treatment interaction). Thus the increase in blood flow in this region was found only with a combination of training and bFGF administration. Two tissue sections of the proximal hindlimb are notable (quadriceps and adductor group; cf. Table 3) as exhibiting a training-dependent bFGF increase in blood flow.

Blood flows to the distal tissues were increased by bFGF administration (P < 0.001) and by exercise training in an independent manner (P < 0.001; cf. Table 2). For example, collateral-dependent blood flow to the contracting calf muscle group increased ~140% in carrier-trained (P < 0.001), ~180% in bFGF sedentary (P < 0.001), and ~240% in the bFGF-trained (P < 0.001) compared with the carrier sedentary group (Fig. 2). As shown in Table 3, the increases in blood flows are well distributed among the individual muscles of the distal hindlimb. However, for all of the tissues of the distal hindlimb, the training enhancement of the bFGF increase in blood flow was marginal in the statistical analysis (0.10 > P > 0.05).

View larger version (18K): [in this window] [in a new window]   Fig. 2.   Collateral-dependent blood flow to calf muscles. * Significantly less than all other groups (P < 0.001);  significantly greater than carrier-trained group (P < 0.05).

Muscle performance. Initial force development of the GPS muscle group was similar across treatment groups (~11-12 N/g; cf. Table 1). However, the ability to maintain tension was much better in bFGF-treated animals (both sedentary and trained) and carrier-trained rats than in the carrier sedentary group (P < 0.025) throughout the entire muscle contraction sequence (cf. Fig. 3A). The best performance was exhibited by the bFGF-trained group (e.g., even significantly greater than the carrier-trained rats at 15 tetani/min; cf. Fig. 3), whereas the carrier-trained and bFGF sedentary groups exhibit fairly similar performance. The greatest differences among the groups were apparent at the lower energy demand conditions established by the relatively low contraction frequency (e.g., 8-15 tetani/min).

View larger version (26K): [in this window] [in a new window]   Fig. 3.   Tension development of the calf muscles (A). Estimated tension development for low-oxidative (white gastrocnemius; B) and high-oxidative (red gastrocnemius; C) regions of the calf muscles. See text for details. * Significantly different from carrier sedentary group;  significantly different from bFGF sedentary group;  significantly different from carrier-trained group (P < 0.025).

Capillary contact per fiber. The distribution of capillary contacts per fiber was not different among the high-oxidative red gastrocnemius muscle sections or among the soleus muscles of all the groups (Fig. 4). Similarly, bFGF infusion in the absence of training did not alter the capillary contact per fiber distribution in the low-oxidative white gastrocnemius muscle section (Fig. 4); however, the distribution of capillary contacts per fiber shifted to the right (P < 0.05) in both trained groups (carrier and bFGF). The same results are obtained when evaluated on a simple capillary per fiber ratio basis (data not presented).

View larger version (21K): [in this window] [in a new window]   Fig. 4.   Distributions of capillary contacts per fiber for white gastrocnemius (A), red gastrocnemius (B), and soleus (C).

	DISCUSSION

Top Abstract Introduction Methods Results Discussion References

The present study provides novel findings on cytokine-induced collateral blood development in experimental peripheral arterial insufficiency; exogenous bFGF administration and physical activity interact to impart an enhanced improvement in collateral blood flow development beyond that observed with either treatment. Our findings again confirm that enhanced physical activity can increase collateral-dependent blood flow in animals with experimental (37, 38) and in humans with acquired (15, 29) peripheral arterial insufficiency. In addition, our findings illustrate the potent influence of angiogenic growth factors in expanding collateral vessels (2, 5-7, 11, 32) and improving downstream blood flow (33, 34) reported previously. Blood flow to the entire hindlimb of the trained animals that received bFGF infusion was significantly greater (30-50%) than that for either the bFGF only (P < 0.05) or training only (P < 0.01) groups (cf. Table 2). Thus there was an "additive" effect of exercise to increase the bFGF improvement in blood flow. Because perfusion pressures in the descending aorta were maintained in these experiments at 100 mmHg, the greater flow was achieved by a reduction in peripheral resistance. This is most meaningful in the distal hindlimb, where tissue blood flow is completely collateral dependent, since the occluded femoral artery is the primary arterial source. Recall that contractions of the distal muscles were employed to ensure that distal resistance was minimal. Thus any increase in blood flow to the lower limb should reflect a decrease in upstream collateral vessel resistance. Both exercise and bFGF are effective treatments, as collateral blood flow to the distal tissue was increased by ~115-140%; the combination of physical activity and bFGF administration proved to be the most effective, increasing collateral blood flow ~190% over that observed in the saline-infused sedentary group.

The large increase in collateral flow observed with the combination of training and bFGF administration may actually be an underestimate of the full extent of vascular remodeling. The blood flow increase realized in the upper hindlimb tissues in this group may have taken flow destined for the lower limb tissues ("steal" phenomenon). For example, the greater flow measured in the quadriceps and adductor muscles of the thigh occurred in the area where existing vessels enlarged, due to bFGF (34) and/or training (37), and likely served as collateral vessels. Thus these vessels are expected to support local tissue needs as well as serve as conduits for collateral flow to the distal tissue. Because we have not established a dose response for bFGF, it is unclear whether exercise exerts an influence independent of bFGF, as might be expected by synergism if exercise prompts VEGF involvement (27). Furthermore, it is not presently known whether the exercise program used in the present study provided the maximal stimulus; previous evidence indicates that exercise intensity is an important variable in establishing the magnitude of the collateral blood flow increase (37). Although it is clear that the present combination of exercise and bFGF imparted the greatest response, further work is needed to better define the interactions between the physiological events established by exercise and those prompted by exogenous growth factor administration.

Factors responsible for the observed collateral vessel expansion are presently not well understood. Hypoxia has been implicated in the angiogenic response of the microvasculature (4). However, the vascular remodeling responsible for the enhanced collateral-dependent blood flow after femoral artery occlusion involves enlargement of conduit vessels in the upper thigh (34) that is unlikely related to tissue hypoxia (21). Blood flow to this region is least affected by femoral artery occlusion, compared with the blood flow impairment to the distal hindlimb tissue (21, 22, 38, 39). Even during treadmill exercise, the muscles of the upper leg receive approximately three to four times greater blood flow than the tissues of the lower leg (39). Thus, consistent with the compelling evidence reported by Ito and co-workers (21, 22), factors other than hypoxia are implicated. The results of the present study may provide some insight. Vascular occlusion of the femoral artery is expected to reduce distal pressure, alter pressure gradients in the adjacent vasculature, and redirect flow through existing vessels in the upper thigh (20) that may be unaccustomed to such sustained hemodynamic alterations. This should increase flow velocity and shear stress, events that are known to induce bFGF gene expression in vascular endothelium (26). An upregulation of angiogenic growth factor could contribute to the endothelial cell proliferation and remodeling of vessels that enlarge to become conduits for collateral flow, in the absence of a local blood flow inadequacy (21, 22). Routine daily exercise enhances this vascular remodeling to improve blood flow down stream. The sustained increase in flow velocity and/or shear stress of these collateral channels during exercise could be a determining factor for this greater neovascular development. It is interesting to speculate further that the greater collateral blood flow in the exercise plus bFGF group was made possible by the added stimulus of the sustained flow increase in these channels obligated by exercise. If this proves to be the case, then the expansion of the collateral channels should become self-limiting as the increased diameter accommodates to lessen flow shear. As a practical matter, enhanced physical activity appears to be of great value in optimizing collateral development stimulated by exogenous angiogenic growth factor administration. This could be an important consideration should management of patients with peripheral arterial insufficiency with exogenous angiogenic growth factor administration become realized.

The enhanced collateral blood flow, observed in the treated groups in this study, is functionally significant and leads to direct improvements in muscle performance. This is apparent by the ability of the calf muscles with higher blood flows (cf. Fig. 2) to sustain higher tension developments over the domain of muscle contraction conditions (cf. Fig. 3A). However, the relationship between blood flow and muscle performance for a whole muscle of mixed-fiber composition is complex and influenced by the heterogeneity in aerobic capacity and blood flow distribution among muscle fiber regions within the whole muscle and by muscle-specific training adaptations. For example, low-oxidative fast-twitch white fibers, which comprise most of the gastrocnemius muscle (1), receive a relatively meager blood flow (24, 25, 34, 37, 39), are easily overwhelmed to metabolic failure (13), and exhibit poor fatigue resistance (10). In contrast, the high-oxidative fast-twitch red fibers of the gastrocnemius receive a relatively high blood flow (24, 25, 34, 37, 39), manifest excellent energy balance during high rates of energy expenditure (13), and are fatigue resistant (10). Thus, during our contraction conditions, the fast-twitch white fibers should exhibit tension loss first, during the less intense contraction conditions, whereas the fast-twitch red fibers should function well until the extremely high energy demands of the higher-frequency contraction conditions are reached (16, 35). This heterogeneity must be taken into account when critically evaluating overall muscle performance. Furthermore, exercise training adaptations increase the aerobic capacity of the active fibers (13), enhance the blood/tissue oxygen exchange properties (8, 35, 36), and alter blood flow distribution among fiber sections (35-38). These adaptations are most pronounced in the fast-twitch white regions of the gastrocnemius in claudicant animals that are exercised (35-38) and are evident in this study by the enhanced capillarity observed only in the white gastrocnemius region (cf. Fig. 4).

As done previously (16, 35), it is possible to estimate the performance of the fast-twitch white and fast-twitch red fiber composition of the gastrocnemius muscle from that measured for the whole muscle (cf. Fig. 3). As shown in Fig. 3B, the white gastrocnemius fibers are expected to exhibit extensive fatigue during the 8-15 tetani/min phase of contractions.¹ In contrast, the fast-twitch red fibers (Fig. 3C) are expected to provide peak tension until the energy demands become excessive at the higher contraction frequencies. More appropriate to the present study, a comparison of the respective measured blood flows and corresponding muscle performances expected for the muscle fiber sections is shown in Fig. 5. There is a fairly good correspondence between muscle fiber section blood flow and muscle performance expected for each of the muscle fiber types. It is noteworthy to recognize that the intensity of the contraction condition, which established extensive fatigue for the high-oxidative red section, was three times greater than that for the low-oxidative white section (cf. Fig. 5). Thus improvements in collateral blood flow, induced by exercise and/or exogenous bFGF administration, led to corresponding improvements in muscle function. This would directly improve mobility in activities that involve utilization of the relatively ischemic distal leg muscles.

View larger version (18K): [in this window] [in a new window]   Fig. 5.   Relationship between measured blood flows and estimated tension development for low-oxidative (white gastrocnemius at 15 tetani/min; y = 6.0 + 1.57x; r = 0.983; ) and high-oxidative fibers (red gastrocnemius at 45 tetani/min; y = 28.4 + 0.45x; r = 0.987; ) of the calf muscle.

In the present study, experimental peripheral arterial insufficiency was created with bilateral ligation of the femoral arteries. This procedure results in a large reduction in flow reserve (~80-90%) to the tissues of the distal hindlimb and renders the tissue collateral dependent. However, flow reserve established by existing collateral channels is sufficient for resting tissue flow needs. For example, blood flows to the muscle sections of the lower leg during treadmill walking increase to approximately two to four times that measured in quiescent anesthetized rats (25) instead of the typical 20- to 30-fold elevation observed in the absence of femoral occlusion (39). Thus the animals used in this study are examples of intermittent claudication and are not subject to the complications caused by ischemia at rest, muscle inflammation, and/or tissue necrosis produced when vascular occlusion is introduced at a more proximal site (11). It should be recognized, however, that the acute occlusion established in this study does not represent the broad spectrum of peripheral arterial insufficiency seen clinically; rather, it represents simply those conditions which present with an acute discrete proximal occlusion of a major supply artery.

In summary, the present study demonstrates that exogenous bFGF administration in combination with an exercise program of moderate intensity greatly increases collateral-dependent blood flow and improves muscle performance. That physical activity enriched the bFGF response is consistent with, but does not prove that, hemodynamic factors are important contributors to collateral vessel enlargement. Our findings support the potential use of bFGF as a therapeutic agent in treating peripheral arterial insufficiency, especially patients with intermittent claudication caused by large-vessel proximal obstructions. The potential enhancement of the collateral response by adding physical activity should be evaluated further. However, even when the patient is incapable of exercise or it is contraindicated, exogenous administration of bFGF may be a valuable treatment option in improving limb circulation that could potentially combat disease progression in claudicants.