Vasoresponsiveness of collateral vessels in the rat hindlimb: influence of training

Abstract

Exercise training is known to be an effective means of improving functional capacity and quality of life in patients with peripheral arterial insufficiency (PAI). However, the specific training-induced physiological adaptations occurring within collateral vessels remain to be clearly defined. The purpose of this study was to determine the effect of exercise training on vasomotor properties of isolated peripheral collateral arteries. We hypothesized that daily treadmill exercise would improve the poor vasodilatory capacity of collateral arteries isolated from rats exposed to surgical occlusion of the femoral artery. Following femoral artery ligation, animals were either kept sedentary or exercise trained daily for a period of 3 weeks. Hindlimb collateral arteries were then isolated, cannulated and pressurized via hydrostatic reservoirs to an intravascular pressure of either 45 or 120 cmH₂O. Non-occluded contralateral vessels of the sedentary animals served as normal Control. Vasodilatory responses to acetylcholine (ACh; 1 × 10⁻⁹–1 × 10⁻⁵m) and sodium nitroprusside (SNP; 1 × 10⁻⁹–1 × 10⁻⁴m), constrictor responses to phenylephrine (PE; 1 × 10⁻⁹–1 × 10⁻⁴m), and flow-induced vasodilatation were determined. Endothelium-mediated vasodilatation responses were significantly greater to either ACh (P < 0.02) or intravascular flow (P < 0.001) in collateral arteries of trained rats. Neither blockade of cyclooxygenase with indomethacin (Indo; 5 μm) nor blockade of endothelial nitric oxide synthase with N^G-nitro-l-arginine methyl ester (l-NAME; 300 μm) eliminated this ACh- or flow-induced vasodilatation. The depressed vasodilatory response to SNP caused by vascular occlusion was reversed with training. These data indicate that exercise training improves endothelium-mediated vasodilatory capacity of hindlimb collateral arteries, apparently by enhanced production of the putative endothelium-derived hyperpolarizing factor(s). If these findings were applicable to patients with PAI, they could contribute to an improved collateral vessel function and enhance exercise tolerance during routine physical activity.

Introduction

Peripheral arterial insufficiency (PAI), resulting in a reduced blood flow capacity to the lower limbs and its resultant symptom of intermittent claudication, is a common vascular malady that is associated with increased morbidity and mortality (Sieminski & Gardner, 1997; Breek et al. 2001; Stewart et al. 2002; Aquarius et al. 2005). Enhanced physical activity has been shown to be an effective means of managing patients with PAI. For example, exercise therapy for patients with intermittent claudication increases pain-free walking time, maximal walking time and overall physical capacity (Gardner & Poehlman, 1995; Leng et al. 2000; Stewart et al. 2002). Although the mechanisms underlying the exercise-induced reduction of claudication symptoms remain to be completely defined, there is evidence that an increased blood flow to ischaemic skeletal muscle can contribute to this improved ambulatory function (Brendle et al. 2001; Gardner et al. 2001). Thus, enhancing daily physical activity has the potential of establishing a significant collateral vascular circuit, as extensively studied in preclinical models of intermittent claudication (Lash et al. 1995; Yang et al. 1995a; Lloyd et al. 2001, 2003; Prior et al. 2004).

While the structural enlargment of peripheral collateral vessels has been fairly well characterized in animal models (Ito et al. 1997; Scholz et al. 2001; Herzog et al. 2002; Prior et al. 2003, 2004), there is relatively little appreciation of the functional behaviour of these collateral vessels. Arterial occlusion results in a diminished transmural pressure among downstream vessels (Rosenthal & Guyton, 1968; Unthank et al. 1994, 1995, 1996a) that is likely to account for the reduced endothelium-mediated vasodilatation evident, when even normal vessels are evaluated at the low collateral-dependent pressures (Taylor et al. 2008). Little is known about the long-term effects of the low pressure on collateral vessels. Unfortunately, insights from studies using vessels from collateral-dependent regions of the coronary circuit are limited, as enhanced (Sellke et al. 1990, 1992; Dulas et al. 1996), unchanged (Angus et al. 1991; Flynn et al. 1991), or diminished (Sellke et al. 1990, 1992; Rapps et al. 1998) capacity for endothelial regulation have been reported. On the other hand, the low luminal pressure, caused by an upstream obstruction, modifies the flow distribution to enhance blood flow through the collateral circuit (Rosenthal & Guyton, 1968; Conrad et al. 1971; Paskins-Hurlburt & Hollenberg, 1992; Unthank et al. 1994, 1995, 1996) in a nitric oxide (NO)-dependent manner (Unthank et al. 1994, 1996a). Thus, there is expected to be a significant increase in shear stress, which is likely to contribute to the enlargement of the collateral vessels (Tuttle et al. 2001; Prior et al. 2003) and would be exaggerated by exercise training (Prior et al. 2004). In addition, it is likely that the enhanced blood flow during physical activity would impart adaptations in vasomotor control of collateral vessels (Sun et al. 1994; Delp, 1995; Koller et al. 1995; Lash & Bohlen, 1997; McAllister & Laughlin, 1997; Sun et al. 1998, 2002; Laughlin et al. 2004; McAllister et al. 2005) that could countermand the consequences of vascular occlusion. This establishes the hypothesis that the reduced endothelium-mediated vasodilatation, evident upon exposure to low luminal pressure typical of that observe in vivo, would be reversed with time post-occlusion and enhanced by exercise training (Griffin et al. 1999; Gielen et al. 2001; Griffin et al. 2001; Fogarty et al. 2004; Thengchaisri et al. 2007).

Therefore the purpose of this study was to evaluate the vasomotor characteristics of collateral vessels following occlusion of the femoral artery and the adaptations induced by exercise training. The vasoresponsiveness to endothelium-dependent and endothelium-independent pharmacological agents, and incremental increases in intraluminal flow were determined. Selective inhibition of nitric oxide synthase (NOS) and cyclooxygenase (COX) were used to determine the contributions of these pathways to endothelial regulation of collateral vascular function. We observed a reduced endothelium-mediated vasodilatation, upon exposure to the low pressure typically experienced by collateral vessels, but adaptations occurred within 3 weeks to recover vasodilatory responsiveness which included NOS- and COX-independent pathways. Exercise training enhanced these responses, especially vessel vasodilatation at low shear rates.

Methods

Ethical approval

The care and treatment of all animals and experimental procedures were conducted in accordance with the Guide for the Care and Use of Laboratory Animals published by the US National Institutes of Health (NIH Publication no. 85-23, revised 1996) and approved by the Animal Care and Use Committee of the University of Missouri. All animals were killed under deep anaesthesia by a pneumothorax.

Experimental design

Vasomotor responses of collateral vessels isolated from sedentary and exercise trained rats were examined 3 weeks following unilateral femoral artery occlusion. The perforating artery was used, as it is a pre-existing peripheral vessel that becomes part of the collateral circuit of the distal hindlimb (Prior et al. 2004). The perforating artery from the non-occluded contralateral limb of the Sedentary group served as Control for normal vessels. Preliminary experiments indicated that the response of the vessels from the contralateral non-occluded limb was not different from the response of vessels obtained from naive animals that were not subjected to any vascular occlusion. Each set of experiments was performed to accommodate differences in intraluminal pressure existing in vessels with and without occlusion of the femoral artery in vivo. Thus, every experiment was conducted at intraluminal pressures of both 120 cmH₂O and 45 cmH₂O, corresponding to normal and post-occlusion pressures measured within the hindlimb collateral circulation as previously reported (Yang et al. 2002; Taylor et al. 2008).

Two series of experiments were performed. In series I, vasomotor responses to acetylcholine (ACh), sodium nitroprusside (SNP), and phenylephrine (PE) were characterized at both 45 and 120 cmH₂O. In series II, the vasodilatory responses to intraluminal flow and ACh were determined in the absence and presence of NOS (l-NAME), COX (indomethacin), and both inhibitors in combination. Each artery in series II was used in flow-response experiments at either 45 cmH₂O or 120 cmH₂O, but not both.

Animals

Sprague–Dawley rats weighing ∼300 g were obtained (Taconic Farms; Germantown, NY, USA) and housed in a temperature controlled (21°C) room with a 12 h–12 h light–dark cycle. Commercial rat chow and water were provided ad libitum. Under ketamine–acepromazine anaesthesia (100 mg kg⁻¹–0.5 mg kg⁻¹) rats had experimentally induced hindlimb ischaemia produced by unilateral occlusion of the left femoral artery. This was achieved by surgical ligation with 3-0 silk sutures approximately 5–6 mm distal to the inguinal ligament, following access by a 1 cm skin incision above the inguinal canal. Topical antibiotic powder (Neo-Predef, Upjohn) was placed on the wound prior to closure with skin clips. The surgical procedure was brief, could be achieved with a 100% success rate, and the animals recovered rapidly. The rats were kept warm post-surgery and observed until they had recovered from anaesthesia, as routinely done (Yang et al. 1995a, 1996).

Exercise training

Following unilateral femoral artery ligation, animals were randomly assigned to the Sedentary or Trained group. Exercised was performed on a motor driven treadmill (Quinton) twice daily, 6 days per week at 20 m min⁻¹, 15% grade until fatigue. The exercise sessions began the day following femoral artery occlusion and were performed in the morning and afternoon, at least 4 h apart. The modestly intense exercise effort could be easily achieved by each animal, with only the duration of each exercise bout increasing to progressively establish the trained condition. Total duration of training was 3 weeks. Animals in the Sedentary group were limited to cage activity.

Isolation of collateral arteries

Animals were anaesthetized by an excess dose of ketamine–acepromazine (1.5 times the normal dose of 100 mg kg⁻¹–0.5 mg kg⁻¹) and following the absence of response to external stimuli killed by a pneumothorax. The hindlimbs were then removed and placed in 4°C Krebs buffer solution for isolation of the collateral artery. With the use of a dissecting microscope, we identified the distal portion of the collateral circuit to be the perforating artery described by Green (1935). Originating from the femoral/popliteal artery in the distal hindlimb, the perforating artery is the distal portion of a pre-existing arterial anastomosis connecting the popliteal to the internal iliac artery in the proximal hindlimb via the hypogastric trunk (Green, 1935). This pre-existing anastomosis contributes to the collateral circuit that circumvents the site of occlusion following ligation of the femoral artery (Herzog et al. 2002; Prior et al. 2004). Perforating arteries, easily identified from their origin at the femoral/popliteal artery extending to the insertion in the hamstrings muscle (∼8–10 mm), were carefully dissected free of adjacent muscle and transferred to a Lucite vessel chamber containing filtered physiological saline solution (PSS) containing (mm): 145.0 NaCl, 4.7 KCl, 1.2 NaH₂PO₄, 1.17 MgSO₄, 2.0 CaCl₂, 5.0 glucose, 2.0 pyruvate, 0.02 EDTA, 3.0 Mops, and 1 g (100 ml)⁻¹ BSA, pH 7.4. Each end of the artery was cannulated with resistance matched glass micropipettes containing PSS with albumin and secured with 11-0 nylon ophthamalic suture. The vessel chamber was then transferred to the stage of an inverted microscope (Olympus) equipped with a video camera (Panasonic), videomicrometer (Microcirculation Research Institute, Texas A&M University), videotape recorder (Panasonic), and data acquisition system (Apple Macintosh with ADInstruments MacLab) for on-line recording of intraluminal diameter. Hydrostatic pressure reservoirs were attached to each micropipette to maintain constant intraluminal pressure, and the vessel was checked for leaks. If no leaks were detected, vessels were treated with 80 mm KCl to verify viability, washed with PSS, and allowed to equilibrate for 1 h at 37°C to develop spontaneous tone; the bathing medium was changed every 15 min during the equilibration period. Intraluminal diameter was monitored continuously throughout the experiment with the use of the videomicrometer.

In vitro studies

In the first series of experiments (see Fig. 1), cannulated arteries were pressurized at 45 cmH₂O, similar to the pressure measured in vivo (Yang et al. 2002; Taylor et al. 2008), and allowed to reach equilibrium over a 1 h period; the physiological saline solution (PSS-bathing medium) was replaced every 15 min during the equilibration period. To assess endothelium-dependent vasodilatation, a concentration–response relationship to the cumulative addition of ACh (1 × 10⁻⁹–1 × 10⁻⁵m) was determined. Upon completion of this experimental protocol, intraluminal pressure was increased to 120 cmH₂O. The vessel was allowed to equilibrate, and the protocol was repeated at the higher pressure.

Figure 1.

Schemata for Series I and Series II protocols

Vasoconstrictor responses to cumulative addition of phenylephrine (PE, 1 × 10⁻⁹–1 × 10⁻⁴m), an α₁-adrenergic agonist, and vasodilatory responses to SNP (1 × 10⁻⁹–1 × 10⁻⁴m) were established at 120 cmH₂O to evaluate vascular smooth muscle function. The vessels were washed with bathing medium and allowed to equilibrate for 20 min between each dose–response sequence. At the conclusion of each experiment, maximal passive diameter was determined at 120 cmH₂O after a 60 min incubation period in calcium-free PSS. The calcium-free PSS was similar to the PSS except that it contained 2 mm EDTA and CaCl₂ was replaced with 2.0 mm NaCl.

In the second series of experiments (see Fig. 1), vasodilatory responses of the perforating arteries to intraluminal flow were determined at intraluminal pressures of either 45 or 120 cmH₂O. Vessels were mounted for flow to occur in the direction that occurs in vivo (i.e. retrograde, relative to normal, in vessels following femoral artery occlusion). Intraluminal flow was induced by creating a pressure gradient across the vessel by changing the height of the pressure reservoirs in equal but opposite directions while maintaining a constant intraluminal pressure (Jasperse & Laughlin, 1997). Intraluminal flow rates were increased by 2 cmH₂O increments from 0 to a final pressure difference of 20 cmH₂O, producing flow rates of 3, 9, 22, 27, 34, 38, 42, 51, 55, 61, 66 and 4, 10, 25, 34, 38, 42, 47, 51, 54, 58, 62 μl min⁻¹ at intraluminal pressures of 45 and 120 cmH₂O, respectively. Flow at each pressure gradient was maintained for 3 min to insure a steady state response was achieved. The same procedure was then repeated in each artery after a 20 min incubation period in the presence of the cyclooxygenase inhibitor indomethacin (Indo; 5 μm), endothelial nitric oxide synthase inhibitor N^G-nitro-l-arginine methyl ester (l-NAME; 300 μm), or Indo +l-NAME. The vessels were washed and allowed to equilibrate between each successive flow sequence measurement. Vasodilatation in response to acetylcholine (1 × 10⁻⁹–1 × 10⁻⁴m) in the presence of Indo +l-NAME was also evaluated. Vessels were then incubated in calcium-free PSS at the conclusion of each experiment to determine maximal passive diameters.

Data and statistical analyses

Vascular responses were recorded as actual diameter (μm) and presented as actual diameter, as diameter relative to maximal diameter, and as the percentage of possible vasodilatation according to the formula:

where D_s is the steady-state luminal diameter measured following each addition of each vasodilatory agonist and D_b is the initial baseline luminal diameter measured prior to the onset of the dose–response curve. D_m is defined as the maximal diameter observed for the vessel at 120 cmH₂O, which is frequently observed at the onset of the experiment rather than following incubation in calcium-free PSS at the conclusion. Expressing the vasodilatory response relative to the maximal diameter observed at 120 cmH₂O permits a comparison of the relative responses between treatment groups as well as differences caused by intraluminal pressure.

The concentration–response or flow–response curves were analysed using a two-way repeated-measures ANOVA. Post hoc analyses were performed using the Student–Neuman–Keuls test where appropriate. Student's t test was used to determine the significance of differences in developed tone and maximal vessel diameter. EC₅₀ values were calculated using a sigmoidal curve fitting model (Prism, GraphPad Software Inc., La Jolla, CA, USA). All values are presented as means ±s.e.m. The acceptable level of significance was defined as P < 0.05.

Results

Exercise training response

Daily run time increased progressively, in a modestly curvilinear fashion, over the course of the 3 week training protocol from approximately 35 min day⁻¹ initially to approximately 150 min day⁻¹ during the two run bouts per day.

Vessel characteristics

Maximal vessel diameters are given in Table 1. Femoral artery occlusion resulted in a greater maximal luminal diameter of the collateral vessel (P < 0.001). As expected, maximal luminal diameters of all vessels were significantly less at the lower distending pressure (i.e. 45 vs. 120 cmH₂O). Baseline tone present prior to each concentration–response relationship performed in series I is shown in Table 2. The amount of baseline tone present prior to ACh and PE exposure was essentially similar across groups. The amount of baseline tone present following PE and prior to the SNP concentration–response curve was greatest for all three groups (P < 0.001). Baseline tone prior to and following each pharmacological intervention in the second series of flow–response experiments is shown in Table 3. In general, vascular tone was greater at 120 cmH₂O, as compared to 45 cmH₂O (significant main treatment effect, P < 0.025). There were no significant differences in tone among the groups at all conditions at 45 cmH₂O. However, tone for the Trained group at 120 cmH₂O was significantly lower than some values for the Sedentary and Control vessels (significant main treatment effect, P < 0.025; Table 3).

Table 1.

Maximal vessel diameters

	Control	Sedentary	Trained
45 cmH2O	227 ± 6.7	312 ± 15.7†	366 ± 14†
	(18)	(16)	(17)
120 cmH2O	310 ± 9.2*	373 ± 9.8*†	385 ± 9†
	(17)	(16)	(18)

All data expressed as means ±s.e.m.

^*Significantly greater than corresponding values at 45 cmH₂O (P < 0.05).

^†Significantly greater than corresponding Control values (P < 0.05). Number of arteries indicated in parentheses.

Table 2.

Series I. Baseline tone (% of maximal diameter) prior to dose–response curves

	Control	Sedentary	Trained
Prior to acetylcholine
45 cmH2O	36 ± 1.8	35 ± 2.4	34 ± 2.0
	(10)	(10)	(11)
120 cmH2O	31 ± 2.4	33 ± 3.0	32 ± 2.5
	(10)	(10)	(11)
Prior to phenylephrine
120 cmH2O	23 ± 2.5	24 ± 2.4	29 ± 3.5
	(10)	(10)	(11)
Prior to sodium nitroprusside*
120 cmH2O	36 ± 3.4*	38 ± 3.2*	39 ± 2.8*
	(10)	(10)	(11)

All data expressed as means ±s.e.m. Basal tone =[1 − (vessel size/maximum)]× 100.

^*Significant main treatment effect for a greater tone compared to its prior measurement (P < 0.001); number of arteries studied are given in the parentheses.

Table 3.

Series II. Baseline tone (% of maximal diameter) and tone prior to and following each antagonist, but prior to flow determination

	Control		Sedentary		Trained
	45 cmH2O	120 cmH2O	45 cmH2O	120 cmH2O	45 cmH2O	120 cmH2O
Baseline tone (initial)	18 ± 8.1	46 ± 5.6*	22 ± 5.8	54 ± 11*	31 ± 8.9	44 ± 4.7
	(8)	(7)	(6)	(6)	(8)	(7)
Pre-indomethacin	31 ± 10	53 ± 9.2*	36 ± 11	61 ± 9.5*‡	30 ± 10.1	28 ± 7.0
	(8)	(6)	(6)	(6)	(8)	(6)
Indomethacin	34 ± 9.8	55 ± 8.9*‡	34 ± 11	61 ± 9.2*‡	29 ± 9.7	34 ± 6.6
	(8)	(7)	(6)	(6)	(8)	(7)
Pre- l-NAME	30 ± 11	52 ± 11*‡	33 ± 7.7	61 ± 8.7*‡	26 ± 9.1	29 ± 7.5
	(8)	(6)	(6)	(6)	(8)	(6)
l-NAME	36 ± 9.6†	61 ± 7.7*‡	36 ± 7.7	65 ± 7.2*‡	28 ± 8.9	39 ± 9.2
	(8)	(6)	(6)	(6)	(8)	(7)
Pre-indomethacin +l-NAME	41 ± 8.2†	58 ± 7.6	54 ± 10†	68 ± 7.7‡	39 ± 9.6	32 ± 8.8
	(8)	(7)	(6)	(6)	(8)	(7)
Indomethacin +l-NAME	48 ± 9.0†	53 ± 5.7	55 ± 11†	74 ± 5.5*†‡	41 ± 9.3	42 ± 11
	(8)	(7)	(6)	(6)	(8)	(7)
Indomethacin +l-NAME (prior to ACh)	47 ± 8.7†	61 ± 5.3	58 ± 11†	67 ± 6.2	46 ± 8.1	48 ± 11
	(8)	(7)	(6)	(6)	(8)	(7)

Data expressed as means ±s.e.m. Basal tone =[1 − (vessel size/maximum)]× 100.

^*Significant main treatment effect (P < 0.025) of luminal pressure, with 120 cmH₂O > 45 cmH₂O (mean difference needed = 17.7).

^†Significantly greater than baseline tone within same group (P < 0.05).

^‡Significantly greater than tone at 120 cmH₂O for the same treatment (P < 0.05).

Vasodilatory responses to ACh

Cumulative increases in ACh produced increases in vessel diameters in a dose-dependent manner as illustrated in Fig. 2. At the low pressure, the percentage possible vasodilatory response to ACh was greater in the Sedentary and Trained groups (P < 0.05), as compared to the Control vessels. Interestingly, vessels from the Trained group exhibited a shift to the left in the dose–response during the low-pressure condition (P < 0.02; EC₅₀ 1.5 × 10⁻⁸mvs. 5.9 × 10⁻⁷m). The meagre vasodilatory response to ACh observed in the Control group at 45 cmH₂O reverted to a robust response at 120 cmH₂O, illustrating the importance of radial wall tension, which was not different from that observed for the Sedentary and Trained groups.

Figure 2. Concentration–response to acetylcholine of isolated collateral vessels.

Concentration–response to acetylcholine of isolated collateral vessels from animals with unilateral occlusion of the femoral artery and kept Sedentary or Trained for 3 weeks, as compared to the vessel of the Control non-occluded contralateral limb of the Sedentary group, evaluated at 45 cmH₂O (A and B) and at 120 cmH₂O (C and D). Values are means ±s.e.m.*Significantly greater than Control non-occluded group (P < 0.05); **EC₅₀ of Trained group is significantly shifted to the left (P < 0.02), as compared to the Sedentary group.

In a second series of experiments, as illustrated in Fig. 3, dual inhibition of NOS (l-NAME; 300 mm) and COX (Indo; 5 μm) effectively eliminated the vasodilatory response to ACh at 120 cmH₂O in vessels from the Control animals. On the other hand, the vasodilatory response was appreciably reduced, but not eliminated, in vessels from the occluded animals that were kept Sedentary or Trained. This response was more marked at 120 cmH₂O, again illustrating the importance of radial wall tension. Thus, endothelium-dependent vasodilatory responses in the collateral vessels from the occluded animals appear to include increased contribution of NOS- and COX-independent processes.

Figure 3. Effects of combined NOS inhibition with l-NAME and of COX inhibition with indomethacin.

Effects of combined NOS inhibition with l-NAME and of COX inhibition with indomethacin on ACh-induced dilatation of isolated collateral vessels from animals with unilateral occlusion of the femoral artery and kept Sedentary or Trained for 3 weeks, as compared to the vessels of the Control non-occluded contralateral limbs of the Sedentary group. Arteries were pressurized at 45 cmH₂O (A and B) and 120 cmH₂O (C and D). Values are means ±s.e.m. *Significantly greater than Control non-occluded group (P < 0.05).

Vascular responses to PE and SNP

As illustrated in Fig. 4, all arteries demonstrated a robust vasoconstriction in response to cumulative addition of PE in a dose-dependent manner that was not different among groups (EC₅₀, Control: 1.3 × 10⁻⁶m; Sedentary: 2.5 × 10⁻⁶m; Trained: 1.0 × 10⁻⁶m). Responses to SNP, which are generally but not always related to NO per se, were significantly reduced by occlusion (cf. Sed group; P < 0.05)), as compared to the response of the occluded Trained group which was not different from that observed in the Control group. Further, there were no difference in EC₅₀ values among groups (Control: 1.1 × 10⁻⁶m; Sedentary: 7.2 × 10⁻⁷m; Trained: 4.1 × 10⁻⁷m).

Figure 4. Concentration–response to phenylephrine and sodium nitroprusside.

Concentration–response to phenylephrine (A and B) and sodium nitroprusside (C and D) of isolated collateral vessels from animals with unilateral occlusion of the femoral artery and kept Sedentary or Trained for 3 weeks, as compared to the vessel of the Control non-occluded contralateral limb of the Sedentary group, and evaluated at 120 cmH₂O. Values are means ±s.e.m. *Trained group greater (P < 0.02) than Sedentary group, but not different from Control non-occluded group.

Responses to intraluminal flow

During flow at the low perfusion pressure of 45 cmH₂O, there was a trend for the Trained and Sedentary vessels to vasodilate (P < 0.10) at low shear rates (Fig. 5B); in contrast, at relatively high shear rates the Sed and Control vessels exhibited vasoconstriction (P < 0.05). When the Trained vessels were perfused at the higher pressure of 120 cmH₂O, there was a robust vasodilatation (P < 0.005) over a range of relatively low shear rates (Fig. 6B). High pressure perfusion tended to increase vasodilatation and lessen the vasoconstriction in the Sedentary and Control groups. Interestingly, when the Trained vessels were incubated at low pressure in Indo, the vasodilatation remained (P < 0.05; Fig. 5D), whereas vasodilatation became evident in the Sedentary (P < 0.05) but not Control vessels. Similarly, in the presence of l-NAME vasodilatation remained in the Trained vessels (P < 0.001; Fig. 5F); however, there was no response to flow in the Control vessels, even at fairly high shear rates. The vasoresponsiveness to flow during incubation with both Indo and l-NAME incubation revealed that a non-NOS and non-COX pathway-dependent vasodilatation (P < 0.025) was operating in the Trained vessels when evaluated at 45 cmH₂O (Fig. 5H). While this response was marginally evident in the Sedentary group vessels (P < 0.20), vasodilatation did occur at exceptionally higher shear rates (Fig. 5H), but was not evident at high pressure (Fig. 6H).

Figure 5. Effects of COX inhibition with indomethacin, NOS inhibition with l-NAME, and their combination on flow-induced vasomotion of collateral vessels measured at low pressure.

Effects of COX inhibition with indomethacin (C and D), NOS inhibition with l-NAME (E and F), and their combination (G and H) on flow-induced vasomotion of collateral vessels (A and B) measured at low pressure (45 cmH₂O). Vessels were isolated from animals with unilateral occlusion of the femoral artery and kept Sedentary or Trained for 3 weeks, as compared to the vessel of the Control non-occluded contralateral limb of the Sedentary group. Results are presented as absolute diameter (left column) and change in diameter (right column). Values are means ±s.e.m. †Trend for dilatation (P < 0.10). Significant dilatation/constriction: ‡P < 0.05, §P < 0.025 and ∫P < 0.001. *Response of the vessels remained unchanged with increased shear stress up to 45 dynes cm⁻² (6 additional points not shown); **Response of Sedentary vessels increased to 20–30% vasodilatation (P < 0.05) over shear stress to 45 dynes cm⁻² (3 additional points not shown).

Figure 6. Effects of COX inhibition with indomethacin, NOS inhibition with l-NAME, and their combination on flow-induced vasomotion of collateral vessels measured at normal pressure.

Effects of COX inhibition with indomethacin (C and D), NOS inhibition with l-NAME (E and F), and their combination (G and H) on flow-induced vasomotion of collateral vessels (A and B) measured at normal pressure (120 cmH₂O). Vessels were isolated from animals with unilateral occlusion of the femoral artery and kept Sedentary or Trained for 3 weeks, as compared to the vessel of the Control non-occluded contralateral limb of the Sedentary group. Results are presented as absolute diameter (left column) and change in diameter (right column). Values are means ±s.e.m. †Trend for dilatation (P < 0.10). Significant dilatation/constriction: ‡P < 0.05, §P < 0.025, γP < 0.005 and ∫P < 0.001. *Response of the vessels remained unchanged with increased shear stress up to 45 dynes cm⁻² (6 additional points not shown).

Discussion

The results of these experiments support our hypothesis that the exercise training-induced increase in endothelium-mediated vasodilatory response to ACh (cf. Fig. 2), typical of normal vessels, extends to peripheral collateral arteries that develop following femoral artery occlusion even though they are subjected to a low intraluminal pressure. We confirm the well-recognized response that radial wall stress is an important determinant of vasoresponsiveness (Kuo et al. 1991), as the dulled responses observed at low luminal pressure (45 cmH₂O) were modified when the vessels were taken to 120 mmH₂O. However, even in the presence of the reduced luminal pressure within the collateral vessels, typical of that following occlusion of the femoral artery, vascular remodelling occurs that recovers endothelial-mediated vasodilatation. Curiously, even in the presence of dual blockade of NOS and COX function with l-NAME and indomethacin, a modest but significant dilatation persisted in the Trained vessels (Fig. 3). Similarly, our findings indicate that an enhanced flow-mediated dilatation in Trained arteries remained in the presence of l-NAME, indomethacin or both. These data indicate that the increased ACh- and flow-mediated vasodilatation in Trained arteries involves an alternative endothelium-dependent mechanism, potentially endothelium-derived hyperpolarizing factor (EDHF). Further, the modest reduction in endothelium-independent vasodilatory responsiveness to sodium nitroprusside caused by occlusion (cf. Fig. 4) was reversed by exercise training, suggesting a training-induced influence on smooth muscle responsiveness to NO in collateral arteries. Taken together, these results may provide insight into the training-induced increases in peripheral blood flow that can occur in patients with peripheral arterial disease.

Previous work has established that exercise training increases blood flow to the musculature of the collateral-dependent limb in the rat model of peripheral arterial disease created by ligation of the femoral artery (Mathien & Terjung, 1990; Yang et al. 1990, 1995a, 2002). This adaptation leads to a reduction in vascular resistance within the collateral circuit circumventing the site of obstruction (Lash et al. 1995; Yang et al. 2002) and is consistent with an increase in collateral vessel diameter, as reported in this study (see Table 1) and previously (Yang et al. 1995a,b, 1998; Prior et al. 2004). This structural adaptation is likely to provide the basic foundation for the reduced resistance within the collateral circuit (Taylor et al. 2008). For example, at the low distal pressures created by femoral artery occlusion (e.g. ∼45 cmH₂O) the enlargement of the collateral vessel observed in this study, relative to Control, could support an ∼7-fold increase in conductance, in the absence of any vascular tone, based upon measured passive diameters (see Table 1). The much smaller increase in collateral-dependent blood flow (2- to 3-fold) observed following occlusion and training (Lloyd et al. 2001; Prior et al. 2004) attests to the complex nature of vessels that comprise the collateral circuit and the importance of vascular tone in establishing vessel calibre. However, results of this study indicate that these larger vessels are subject to an enhanced potential for endothelium-mediated reduction of collateral artery resistance. Thus, it is likely that functional, as well as structural, adaptations contribute to the enhanced collateral blood flow capacity of the calf muscle following exercise training. This functional adaptation is not unique to collateral vessels as there are numerous reports of an enhanced endothelium-dependent vasodilatation of normal vessels perfusing trained skeletal muscle (Sun et al. 1994; Delp, 1995; Koller et al. 1995; Lash & Bohlen, 1997; McAllister & Laughlin, 1997; Sun et al. 1998; Sun et al. 2002; Laughlin et al. 2004; McAllister et al. 2005). This is likely to be the result of an enhanced NO production by the vascular endothelium related to the increased endothelial NO synthase gene expression induced by exercise training (Sessa et al. 1994; Woodman et al. 1997; Laughlin et al. 2004; McAllister et al. 2005) and attributed to the elevated intravascular shear stress generated by increases in blood flow associated with exercise (Miller & Vanhoutte, 1988; Sessa et al. 1994; Nadaud et al. 1996). These results imply that the enhanced endothelium-mediated dilatation to ACh observed in collateral vessels is part of a generalized functional remodelling observed in vessels in response to changes to haemodynamic stimuli (i.e. increased blood flow and shear stress) associated with exercise. The significance of this enhanced NO modulation was evident even in normal vessels, as prior exercise training modestly increased collateral-dependent blood flow following acute ligation of the femoral artery in a NO-dependent manner (Yang et al. 2000, 2002). Similarly, a down-regulation of NO production, even in normal vessels by l-NAME, limits blood flow through the collateral circuit in the face of an abrupt occlusion of the femoral artery (Unthank et al. 1994, 1996b) and importantly preempts vascular enlargement with training (Lloyd et al. 2001). Thus, we believe that the altered vasoresoponsiveness observed in vitro in this study can be meaningful in modulating the function of the collateral circuit in vivo.

Endothelium-dependent adaptations

We sought to determine the contribution of NOS-generated NO and COX-generated products in mediating adaptations in the Trained vessels with pharmacological blockade of these signalling pathways. Surprisingly, the enhanced vasodilatory responsiveness to ACh observed in Trained arteries persisted in the presence of Indo +l-NAME, whereas the vasodilatory response to ACh in non-occluded Control arteries (at 120 cmH₂O) was effectively abolished by these inhibitors (Fig. 3D, compared to Fig. 2D). Thus, training-induced adaptations occur in endothelium-mediated vasodilatation within the collateral circulation at least in part via pathways independent of NOS or COX perhaps mediated by the putative EDHF. EDHF has been described as an arachadonic acid metabolite of cytochrome P-450 capable of causing hyperpolarization and relaxation of vascular smooth muscle (Harder et al. 1995; Campbell et al. 1996; Harder et al. 1997). Along with NO and PGI₂, EDHF plays a role in the endothelial regulation of vascular tone. There is evidence indicating a redundancy in the role of these vasodilators in the maintenance of vascular tone, such that suppression of one vasodilatory pathway results in the upregulation of another (Campbell et al. 1996; Beverelli et al. 1997; Nishikawa et al. 2000; Huang et al. 2001; Wu et al. 2001). Interestingly, in the case of Trained arteries, the putative EDHF pathway became evident in the presence of an enhanced NOS pathway. How these enhanced responses for vasodilatation respond in vivo remains to be clarified.

Endothelium-independent adaptations

Our findings indicate that the vascular remodelling of collateral arteries induced by exercise training involves more than endothelium-dependent processes, as the vasodilatory responsiveness to sodium nitroprusside was enhanced. The increase was actually a recovery of the depressed responsiveness of collateral vessels to SNP observed in Control animals following occlusion of the femoral artery (Fig. 4D). Thus, the influence of training is different from the enhanced vasodilatory responses to SNP and nitroglycerin reported by Sellke et al. (1990, 1992) in porcine and canine coronary collateral vessels. These authors suggested that these enhanced vasodilatory responses above normal could be related to a diminished basal release of NO in the collateral vasculature, which could potentially result in an increased vascular smooth muscle NO sensitivity. Mounting evidence indicates the sensitivity of the vascular NO response is modulated in part by the level of endogenously produced NO, such that a diminished basal release of NO results in an enhanced vascular NO response (Moncada et al. 1991; Brandes et al. 2000; Mullershausen et al. 2003) while elevated NO levels leads to NO desensitization (Mullershausen et al. 2001). A diminished basal production of NO, however, is an unlikely explanation for the recovered vasodilatory responsiveness to SNP observed in our work, given the enhancement of endothelial NO synthesis known to accompany exercise training (Sun et al. 1994; Delp, 1995; Koller et al. 1995; Sun et al. 2002) and the enhanced eNOS mRNA abundance in this collateral vessel previously reported (Prior et al. 2004). This implies that the deficiency in SNP responsiveness provided the conditions in which the influence of exercise training could be realized.

Complex shear-stress response

The influence of vascular occlusion and subsequent adaptations with exercise training establishes an intriguing response of collateral vessels, one that is consistent with developing evidence illustrating the interactions among pathways that control vasoresponsiveness to luminal flow. Collateral vessels exhibited a modest vasodilatation at low shear stress that then proceeded to a significant vasoconstriction with increasing shear stress (Fig. 5B). Recent evidence raises the potential for altered vascular responsiveness to flow, depending on the direction of flow (Markos et al. 2002; Lu & Kassab, 2004). The direction of flow, however, does not account for our findings, as preliminary evidence demonstrated that the direction of flow did not appreciably modify the presence of vasoconstriction that occurred with increased shear stress (Yang & Terjung, unpublished observations). Thus, the loss of vasoconstriction in these vessels with inhibition of the COX pathway (Fig. 5, compare panels B and D) represents a distinct physiological response. Interestingly, this loss of vasoconstriction is similar to that demonstrated by Koller and colleagues (Huang et al. 1998; Bagi et al. 2001, 2002; Ungvari et al. 2002, 2003) and could reflect the elimination of a vasoconstrictory product of the COX pathway (e.g. thromboxane A2), thereby permitting the net influence of a vasodilatory agent such as NO. Curiously, inhibition of NOS by l-NAME also eliminated this vasoconstrictory response to increased shear stress (Fig. 5, compare panels B and F), consistent with a NO-dependent product that is in turn dependent upon the COX pathway for its vasoconstriction effect. As convincingly shown by Koller and colleagues (Huang et al. 1998; Bagi et al. 2001, 2002; Ungvari et al. 2002, 2003), interaction of NO with reactive oxygen species (ROS) can produce peroxynitrite that can serve as the stimulus for vasoconstriction via thromboxane A2 production through the COX pathway. The net result of these interactions can be a dominant vasoconstriction influenced by the extent of NO and/or ROS production (Huang et al. 1998; Bagi et al. 2001, 2002; Ungvari et al. 2002, 2003). In contrast, it is apparent that inhibition of the COX pathway resulted in a robust vasodilatation in the collateral vessels of the exercise trained animals that was realized at relatively low rates of shear stress (see Fig. 5D). This more substantial vasodilatory response could reflect an enhanced NOS production of NO typical of trained vessels (Griffin et al. 1999; Heaps et al. 2000; Griffin et al. 2001; Fogarty et al. 2004). Interestingly, the robust vasodilatation in the trained collateral vessels, observed as NO was removed with l-NAME (Fig. 5F), may seem anomalous. However, we interpret this sustained vasodilatation as due to a NOS- and COX-independent vasodilatation (Fig. 5H), possibly related to EDHF, similar to that proposed in response to ACh (see Fig. 3). The absence of a frank vasoconstriction in the trained collateral vessels at 45 cmH₂O, without pharmacological intervention (Fig. 5A), may reflect that we did not raise flow high enough in these larger vessels to produce sufficiently high rates of shear stress to evaluate the full range of response. However, there was a robust dilatation to these same flows when evaluated at the higher luminal pressure of 120 cmH₂O (Fig. 5B), again reiterating that wall tension is a critical determinant of endothelial vasoresponsiveness (Kuo et al. 1991). Alternatively, it may be that the inherent abundance of ROS had been tempered, if these collateral vessels exhibited the typical training response of an upregulation in catalase and a reduction in oxidative stress (Rush et al. 2003). This could have lessened peroxynitrite production and led to a reduced vasoconstrictory influence as observed in our data. We suspect that these alterations in vasoresponsiveness are physiologically relevant, as the remodelling of the endothelium-dependent pathways to flow induced by exercise training is consistent with the observed enhanced collateral-dependent flow capacity that we have reported previously (Yang et al. 1995a, 1998; Lloyd et al. 2001; Prior et al. 2004).

Experimental considerations

There are several experimental aspects that should be considered. First, the collateral vessel (perforating artery; Green, 1935) that we choose to evaluate functions as one of the main collateral vessels that develop upon occlusion of the femoral artery, but it is not the only one. Further, this vessel originates from the distal femoral/popliteal artery to deliver flow to the distal hamstring muscle (Green, 1935). As such, it serves as a re-entry vessel of the collateral circuit that delivers flow, collected from the proximal, smaller microvasculature anastomoses that enlarge deep within the thigh muscles, to the normal vasculature of the distal hindlimb. Thus, its vasoresponsiveness may not be representative of the deeper vessels of the collateral network. Second, it is recognized that initial tone developed by vessels can be an important determinant of vasoresponsiveness (Kuo et al. 1991; Thorin-Trescases & Bevan, 1998; Davis & Davidson, 2002). Thus, our findings of a dulled vasoresponsiveness at the low luminal pressure (45 cmH₂O), relative to 120 cmH₂O, was to be expected even though there were differences in baseline vessel tone between these conditions (see Table 3). However, the important comparisons of this study are found at the low pressure, typical of that measured in collateral vasculature in vivo (Yang et al. 2002; Taylor et al. 2008), where vessel tone was effectively similar across groups. Thus, we believe that our data set for shear stress (see Fig. 5) is meaningful and warrant the conclusions presented. Third, we recognize that there may not be direct application of our findings to the behaviour of collateral vessels in human patients where disease processes may dominate.

In summary

Our data indicate that following surgically induced vascular occlusion, exercise training results in an enhanced endothelium-dependent vasodilatory capacity of collateral arteries observed with cumulative addition of ACh in vitro. While this appears similar to the training-induced adaptations that generally occur in normal vessels (Sun et al. 1994; Delp, 1995; Koller et al. 1995; Lash & Bohlen, 1997; McAllister & Laughlin, 1997; Sun et al. 1998, 2002; Laughlin et al. 2004; McAllister et al. 2005), the signalling pathways involved appear to be more encompassing. The enhanced dilatation persisted in the presence of indomethacin and l-NAME, indicating the observed exercise-induced increase in vasodilatory capacity of collateral arteries can be attributed to an upregulation of the putative endothelium-derived hyperpolarizing factor. The enhanced endothelium-mediated vasodilatory response to ACh is coincident with an enhanced flow-mediated dilatation, which includes an element independent of the endothelium-derived NOS and COX vasodilatory pathways, further implicating the potential role of an endothelium-derived hyperpolarizing factor in this adaptation. Additionally, exercise training induced a recovery from the dulled vasodilatory capacity of peripheral collateral arteries that occurs as a result of altered vascular smooth muscle responsiveness to NO, as evidenced by the enhanced dilatation to sodium nitroprusside (Fig. 4). This observation confirms the notion that exercise training-induced increases of collateral-dependent skeletal muscle blood flow occur, at least in part, due to endothelium-dependent and -independent vasodilatory adaptations occurring within the collateral vasculature. These changes operate on the background of a structurally enlarged collateral vessel. This argues that, in addition to the structural enlargement that develops with exercise training (Prior et al. 2004), an enhanced functional response of collateral arteries could contribute to the control of blood flow through the peripheral collateral circuit of the rat. Similar adaptations in humans could contribute to the improved exercise tolerance observed in patients with peripheral arterial insufficiency that participate in routine physical activity.

Glossary

Abbreviations

ACh

acetylcholine

COX

cyclooxygenase

EDHF

endothelium-derived hyperpolarizing factor

NOS

nitric oxide synthase

PE

phenylephrine

PAI

peripheral arterial insufficiency

SNP

sodium nitroprusside

Author contributions

Conception and design of the experiments: M.H.L., R.L.T., H.T.Y. Collection, analysis and interpretation of data: P.N.C., Z.L., H.T.Y., R.L.T., M.H.L. Drafting the article or revising it critically for important intellectual content: R.L.T., M.H.L., H.T.Y. All authors approved the manuscript.