Limb Blood Flow During Exercise Is Dependent on Nitric Oxide
Background—We have recently reported that hypercholesterolemia reduces aerobic exercise capacity in mice and that this is associated with a reduced endothelium-dependent vasodilator function, endothelium-derived nitric oxide (EDNO) production, and urinary nitrate excretion. These findings led us to test the hypothesis that EDNO production contributes significantly to limb blood flow during exercise and to determine whether loss of EDNO production is responsible for the decline in exercise capacity observed in hypercholesterolemia.
Methods and Results—Twelve-week-old wild-type (E+; n=9) and apoE-deficient (E−; n=9) C57BL/6J mice were treadmill-tested to measure indices defining exercise capacity on a metabolic chamber–enclosed treadmill capable of measuring oxygen uptake and carbon dioxide excretion. Urine was collected before and after treadmill exercise for determination of vascular NO production assessed by urinary nitrate excretion. The wild-type mice were then given nitro-l-arginine (E+LNA) in the drinking water (6 mg/dL) for 4 days before undergoing a second treadmill testing and urinary nitrate measurement. An additional set of 12-week-old wild-type mice was divided into 2 groups: 1 receiving regular water (E+; n=8) and 1 administered LNA for 4 days (E+LNA; n=8). These mice, along with an additional set of E− mice (n=8), underwent treadmill testing to determine maximal oxygen uptake (V̇o2max). The mice were then cannulated such that the tip of the tubing was positioned in the ascending aorta. Fluorescent microspheres (20 000) were infused into the carotid cannula while the mice were sedentary and again while approaching V̇o2max. When the mice were euthanized, the running muscles were collected and fluorescence intensity was measured to determine the peak-exercise redistribution of blood flow to the running muscles (expressed as percentage of total cardiac output, %COrm) during both states. Both E+LNA and E− mice demonstrated a markedly reduced postexercise urinary nitrate excretion, aerobic capacity, and %COrm at V̇o2max compared with E+.
Conclusions—EDNO contributes significantly to limb blood flow during exercise. Conditions that reduce EDNO production disturb the hyperemic response to exercise, resulting in a reduced exercise capacity.
We recently reported that hypercholesterolemic mice manifest a reduction in indices of aerobic exercise capacity during treadmill testing, including a decline in V̇o2max, AT, distance run to exhaustion, aerobic work capacity, and RQ at exhaustion.1 This is true of both diet-induced and genetically prone (caused by an apoE deficiency) hypercholesterolemia. Furthermore, the reduction in exercise capacity is associated with an endothelium-dependent vasodilator dysfunction, attenuated elaboration of aortic EDNO, and a reduced exercise-induced NO production assessed by urinary nitrate excretion. These findings lead us to speculate that EDNO production contributes significantly to exercise-induced hyperemia and that a loss of EDNO-mediated hyperemia results in a decreased oxygen transport capacity of the vasculature and an attenuated aerobic capacity.
The present study was performed to determine whether inhibition of EDNO production, either pharmacologically or by hypercholesterolemia, reduces exercise-induced hyperemia and aerobic capacity.
Eight-week-old female E+ and E− C57BL/6J mice (Jackson Laboratories, Bar Harbor, Me, and DCM) were entered into experimental protocols after a 1-week period of acclimation in the housing facilities of the Stanford DCM. All mice were inspected before the study by the DCM veterinarian and monitored daily by DCM technicians and investigators. All experimental protocols were approved by the Administrative Panel on Laboratory Animal Care of Stanford University and were performed in accordance with the recommendations of the American Association for the Accreditation of Laboratory Animal Care. All mice were housed 3 to 4 per cage. They were maintained on a 12-hour light/dark cycle and given unlimited access to food and water for the duration of the study. All mice were handled daily and taught to run on a treadmill with shock-plate incentive (Exer-4 Treadmill, Columbus Instruments) but were otherwise confined to cages for the duration of the study.
The E− mice were generated from targeted disruption of the apoE gene in the 129 embryonic stem cell line. Germ-line chimeras were mated and backcrossed for 10 generations with C57BL/6J wild-type mice.2
To determine the effects of EDNO inhibition on aerobic capacity, a set of mice underwent the following treadmill studies. Eight-week-old E+ (n=9) and E− (n=9) mice were kept sedentary for 4 weeks. At 12 weeks of age, each mouse was treadmill-tested to measure indices defining exercise capacity. Urine was collected before and after treadmill exercise for determination of urinary nitrate excretion. The E+ mice were then given LNA (Sigma Chemical Co) in the drinking water (6 mg/100 mL; E+LNA). This dose of LNA is similar to that shown to attenuate basal urinary nitrate excretion,3 and similar molar doses of LNA methyl ester have been shown to suppress the release of EDNO and the excretion of cGMP and nitrate in other animal models.4 After 4 days, the mice underwent a second treadmill testing and urinary nitrate measurement. Mice were euthanized in random order after treadmill testing by overdose of methoxyflurane (Pitman-Moore) inhalation anesthesia. Blood was collected from the right atrium for measurement of serum total cholesterol levels.
To determine blood flow distribution during exercise, an additional set of mice underwent a microsphere study. Eight-week-old wild-type mice and E− mice (n=8) were kept sedentary for 4 weeks. At 12 weeks of age, the wild-type mice were divided into 2 groups, one receiving regular water (E+; n=8) and one given LNA in the drinking water (6 mg/100 mL) for 4 days (E+LNA; n=8). Each mouse underwent treadmill testing to determine V̇o2max. The aorta was then surgically cannulated, and after overnight recovery, the mice underwent a microsphere delivery study (described below). Mice were euthanized after microsphere delivery by overdose of methoxyflurane inhalation anesthesia. The hindlimb muscles were collected and weighed for determination of microsphere density.
Indices of Exercise Capacity
V̇o2max is defined as the plateau in V̇o2 despite increasing work intensity. Commonly in exercise testing, the V̇o2max is not measurable because the subject becomes exhausted or feels too much discomfort to continue before a plateau in V̇o2 is reached. The treadmill protocol used in this study was designed so that the mice would quickly attain a plateau, reaching their V̇o2max before exhaustion. In a few cases when a plateau was not reached, the V̇o2max was approximated by the peak V̇o2 attained by the animal before exhaustion.
The AT is an independent measure of aerobic capacity expressed in units of V̇o2. For each mouse, the AT was determined from computer analysis (confirmed by blinded observer) of V̇co2/V̇o2 plots by the V-slope method of Beaver.5 In situations in which the slope of V̇co2/V̇o2 did not increase at higher work rates, the V̇o2max was taken as the AT.
The distance run to exhaustion is taken as an approximate measure of overall work performance and is the total distance run.
Aerobic work capacity was determined by the summation of minute oxygen uptake above basal rate over the course of treadmill running until exhaustion. This was multiplied by the constant 20 J/mL O2 to convert oxygen uptake to aerobic work.6 7
The RQ is the V̇co2/V̇o2 at any given time and at exercise intensities above AT is used as an indirect indicator of anaerobic work performance. At high work rates, anaerobic work supplements aerobic work. Lactic acid, produced from anaerobic metabolism, is buffered by serum bicarbonate, resulting in a stoichiometric increase in CO2 output over O2 uptake.8 9 Thus, RQ begins to rise after AT is attained and continues to rise with increasing anaerobic workload until exhaustion.10
At the time of treadmill testing, each mouse was placed on a treadmill at a constant 8° angle enclosed by a metabolic chamber capable of measuring oxygen and carbon dioxide outflow once every minute (model CT-2, Columbus Instruments). After a 15-minute period of acclimation, basal measurements were obtained over a period of 7 minutes. The treadmill was then started at 10 m/min, and the speed was incrementally increased 1 m/min every minute until the mouse reached exhaustion. Exhaustion was defined as spending time on the shocker plate without attempting to reengage the treadmill. Data on oxygen uptake (V̇o2), carbon dioxide output (V̇co2), RQ, and distance run to exhaustion were collected and stored on hard disk (Oxymax Software, Columbus Instruments).
Measurement of Urinary Nitrogen Oxides
Mice were placed in metabolic chambers for basal and postexercise urinary nitrate collection.11 For the basal state, mice were confined to cages for >24 hours, and for the postexercise state, mice were treadmill-exercised for 22 minutes to a final treadmill speed of 32 m/min. Metabolic chambers were constructed as described previously.1 Urine was collected in test tubes containing 100 μL of isopropyl alcohol submerged in ice water for the duration of the 5-hour collection period. Urine was centrifuged at 4000 rpm for 5 minutes, and the supernatant was collected, diluted 1:9 in distilled water, and stored at −80°C for measurement of nitrogen oxides and creatinine.
Nitrogen oxides in the urine were measured with a commercially available chemiluminescence apparatus (model 2108, Dasibi Corp) as previously described.12 The samples (50 μL) were injected into boiling acidic vanadium (III) chloride. This technique uses acidic vanadium (III) chloride at 98°C to reduce both NO2− and NO3− to NO, which is then detected by the chemiluminescence apparatus after reacting with ozone. Signals from the detector were analyzed by computerized integration of curve areas. Standard curves for NaNO2/NaNO3 were linear over the range of 50 pmol/L to 10 nmol/L.
Urine creatinine was measured by the modified method of Slot developed by Sigma Diagnostics.13
Regional Blood Flow Determination
Mice were anesthetized with isoflurane (Ohmeda Caribe) inhalation. An incision was made in the ventral midline of the neck. After the carotid sheath was exposed, the carotid artery was separated from the neurovascular bundle and secured by two 4-0 silk sutures. An incision was made in the carotid artery, and a 30-cm length of PE10 tubing (Becton Dickinson) tapered at one end by gentle stretch was filled with heparin (100 U/mL, Elkins-Sinn), introduced into the carotid artery, and advanced to the ascending aorta just distal to the aortic valve. The incision was oversewn, and the tubing was tunneled subcutaneously to a pouch under the skin on the back. The mice were then given a single dose of ampicillin (100 mg/kg diluted in saline 10 mg/mL IP). After overnight recovery, the mice were assessed for running ability. Mice that could not attain 80% V̇o2max as determined previously were eliminated from the study.
Treadmill Exercise Protocol
The instrumented mice were placed on a treadmill in a metabolic chamber in random order. The tubing was fed through a hole in the chamber, and the carotid artery tubing was connected to a pressure transducer for continuous heart rate and blood pressure measurement. A 100-μL sample of blood was withdrawn for basal lactate measurement. After 20 minutes of acclimation and oxygen uptake analysis, blue-green microspheres were injected into the carotid tubing and infused with normal saline for regional blood flow determination at rest. An equal volume of blue-green microspheres was injected into a reference vial for a “100% of flow” control. The treadmill was then started at 10 m/min and increased 1 m · min−1 · min−1 while the oxygen uptake curve was followed. As the oxygen uptake curve began to plateau, yellow-green microspheres were injected into the carotid cannula and infused with normal saline. An equal volume of yellow-green microspheres was injected into a reference vial. A second 100-μL sample of blood was withdrawn for peak-exercise lactate measurement. After the completion of the yellow-green microsphere infusion, mice were euthanized by methoxyflurane overdose. The gastrocnemius, quadriceps, and both kidneys were removed, dissected free of fat and connective tissue, blotted dry, weighed, and placed in sample vials for fluorimetric determination.
Blue-green and yellow-green fluorescently labeled microspheres (15-μm diameter, Molecular Probes) were diluted to 20 000 microspheres/20 μL saline. At the time of injection, the microsphere solution was vortexed for 5 minutes. Microspheres (20 μL) were drawn into a microinjector syringe and transferred to the carotid artery tubing. The cannula was connected to a glass syringe mounted in an injector pump and filled with saline, which was used to flush the carotid cannula (100 μL over a period of 30 seconds).
Fluorimetric Determination of Microsphere Number
Tissues and reference samples were digested in 2N KOH in methanol overnight in a shaker bath at 40°C. The samples were centrifuged at 3000g, and the supernatant was removed to the level of the tissue plug. The tissue plug was resuspended twice in distilled water with 0.5% Tween80 (Fisher Scientific) and centrifuged, and the supernatant was decanted. The plug was then resuspended in methanol, centrifuged, and decanted. The remaining methanol was removed by evaporative drying. The remaining microsphere residue was dissolved in 2 mL of 2-ethoxyethylacetate (Arcon Organics) and measured by fluorimetry (model LS50B; Perkin-Elmer) using the recommended extinction and emission frequencies for microsphere fluorescence (λex of 425 nm and λem of 468 nm for blue-green and λex of 490 nm and λem of 505 nm for yellow-green).
Regional blood flow for resting and exercise states was calculated from fluorescent intensity as the percent of cardiac output to the tissue (%COt): %COts,i=fts,i(WTtaverage/WTti)/frss,i, where ƒts,i is the fluorescence intensity of the tissue sample residue from mouse i during state s (resting or exercise) and frss,i is the fluorescence intensity of the reference sample residue. Variability in fluorescence intensity due to variation in tissue weight collected is removed by normalizing the tissue weight, WTt i, to the average tissue weight of all animals, WTtaverage. Regional blood flow to the left and right kidneys was determined to assess adequacy of microsphere mixing. A >10% difference in regional flow between kidneys was a criterion for removing an animal from analysis. Of the 24 mice cannulated, 12 mice met the inclusion criteria for analysis (4 died, 3 did not achieve V̇o2 goal, 1 catheter loss, and 4 with poor microsphere mixing).
Hematology and Biochemistry
Blood samples (100 μL) collected at the time of regional blood flow determination were placed in serum separator tubes containing fluoride/oxalate anticoagulant. The samples were immediately centrifuged at 400g for 10 minutes. The serum was kept on ice until analysis. Serum lactate was analyzed within the hour of collection by an enzymatic method developed by Sigma Diagnostics (procedure 739).
Blood samples were collected at the time the animals were killed. These were immediately centrifuged at 3000 rpm for 15 minutes. The serum was separated and stored at −80°C until analysis. Total serum cholesterol was analyzed by the enzymatic method of Allain et al16 as developed by Sigma Diagnostics.
Data are expressed as mean±SEM. Comparisons of single means from multiple populations were made by one- and two-factor univariate one-way ANOVA followed by Fisher’s protected least significant difference. A value of P<0.05 was accepted as statistically significant.
Effect of LNA or Hypercholesterolemia on NO Production
The body weights of the E+ and E− mice were the same, and the weight of the E+ mice did not change after 4 days of LNA administration (Table⇓). The cholesterol levels of the E− mice were significantly elevated compared with E+ mice (941±114 versus 57±4 mg/dL serum). Basal urinary nitrate values were similar in the E+, E+LNA, and E− mice (350±100, 250±50, and 200±10 pmol/mg creatinine, respectively, Figure 1⇓). Postexercise urinary nitrate levels were increased in the E+ mice (550±40 pmol/mg creatinine); by contrast, exercise did not increase urinary nitrate levels in the E+LNA or E− mice (90±25 and 85±30 pmol/mg creatinine, respectively).
Effect of LNA or Hypercholesterolemia on Exercise Capacity
Suppression of NO by LNA resulted in a decrease in aerobic capacity compared with E+ mice as measured by V̇o2max (E+, 120±2; E+LNA, 107±4 ml/min/kg; P<0.05) and aerobic work capacity (E+, 11.2±0.6; E+LNA, 7±1 J/g; P<0.05). There was also a trend toward a decline in running distance (E+, 475±29; E+LNA, 397±30 m; P=NS) and anaerobic threshold (E+, 84±4; E+LNA, 76±3 mlO2/min/kg; P=NS). The magnitudes of reductions in the indices of aerobic capacity observed in E− mice were similar to those seen in the E+ LNA mice (Table⇑, Figure 2⇓).
Effect of LNA or Hypercholesterolemia on Exercise Hyperemia
The basal heart rates and both basal and peak-exercise mean blood pressures of the three groups were not different from each other, although the mean blood pressure of the E+LNA group tended to be higher than the other two, reflecting an effect of NO suppression (Table⇑). Accurate peak-exercise heart rates were not obtained. Basal serum lactate did not differ between the three groups (13±3, 11±2, and 15±2 mg/dL for E+, E+LNA, and E−, respectively, P=NS). Peak-exercise lactate was significantly higher than basal in all three groups but, again, did not differ between groups (62±12, 68±21, and 77±24 mg/dL for E+, E+LNA, and E−, respectively, P=NS). This rise in serum lactate confirmed that the mice were exercising at an intensity above their anaerobic threshold at the time of the second microsphere injection.
Basal %COrm did not differ between the three groups (6±2%, 5±2%, and 5±2% of total cardiac output for E+, E+LNA, and E−, respectively, P=NS) (Figure 3⇓). Peak-exercise %COrm nearly doubled in the E+ mice (from 6±2% to 11±3%, P<0.05). In striking contrast, the %COrm remained essentially the same as basal in the E+LNA and E− mice (7±1% and 6±3% of total cardiac output for E+LNA and E−, respectively).
The salient findings of this study are that (1) inhibition of NO synthase activity reduces exercise capacity; (2) inhibition of NO synthase activity reduces exercise-induced redistribution of blood flow to skeletal muscles; and (3) similar alterations in exercise capacity and exercise-induced limb blood flow are observed in hypercholesterolemia (which is known to inhibit NO activity).
In this study, we set out to determine whether EDNO contributes significantly to exercise-induced hyperemia. We find that inhibition of EDNO activity markedly inhibits exercise-induced redistribution of blood flow to skeletal muscle. The data reveal the importance of EDNO in determining physiological regional shifts in blood flow. These data confirm our previous observation that hypercholesterolemia perturbs EDNO activity and reduces exercise capacity. The present study suggests a mechanism for this dysfunction.
Critique of Methods
To determine whether EDNO plays a significant role in exercise-induced hyperemia, we conducted microsphere experiments in our mouse model. Ideally, use of a double-cannulation method would have provided the most information about cardiac function and regional blood flow determination. However, there is no precedent for this technique in conscious mice, and studies using double cannulation in our laboratory proved to be impractical. In addition, we chose to position the tip of the catheter distal to the aortic valve rather than the traditional position within the left ventricle. This position can be expected to result in poor microsphere mixing at the level of the proximal branches of the aorta but more even mixing at the more distal branches, which was confirmed by the traditional method of comparing microsphere numbers between the left and right kidneys.17 This sacrifice in microsphere uniformity was made to preserve the competency of the aortic valve, which is almost certainly important to the physiology of the exercising animal.
Despite these limitations of this microsphere technique, information about the distribution of microspheres to the running muscles relative to the entire cardiac output was obtainable. Whereas the portion of the cardiac output to the running muscles during exercise nearly doubled in E+ mice (6±2% to 11±3% of total cardiac output; P<0.05), there was no increase in the portion of cardiac output to the running muscles in hypercholesterolemic mice or mice given LNA. This doubling of the portion of cardiac output to the exercising muscles underestimates the absolute increase in blood flow to these muscles by not taking into account the increase in cardiac output, which, in a normal human subject, can increase 4-fold.18
Relationship of EDNO, Exercise Hyperemia, and Exercise Capacity
Previous work has demonstrated the importance of endothelium-derived mediators for flow-dependent vasodilation in the microcirculation of skeletal muscle. Although earlier studies demonstrated that prostaglandin release mediated flow-dependent vasodilation in isolated skeletal muscle preparations,19 20 more recent studies have demonstrated at least an equal and possibly an even greater role for EDNO.21 22 These findings have been confirmed in vivo by Doppler flow studies of human forearm blood flow in response to hypoxia-induced hyperemia.23 24 However, there are conflicting reports on the role of EDNO in exercise-induced hyperemia after repetitive handgrip and wrist-flexion exercises.25 26 27 28 This study not only supports the importance of EDNO in exercise-induced hyperemia, but it is also the first to suggest that a loss of EDNO can be rate-limiting to oxygen delivery and exercise performance. This finding fits with previous work that supports the concept that vascular transport of oxygen can be rate-limiting to metabolic capacity.29 30 31 32 33 It is quite possible, then, that any disturbance in the rate-limiting mechanism of oxygen delivery, such as a defect in flow-mediated vasodilation, would affect exercise capacity.
It is possible that other mechanisms for decreased exercise capacity exist in the presence of systemic administration of LNA. NO plays a role in cardiac and skeletal myocyte function,34 and the role that neuronal NO plays in metabolic activity is yet to be investigated. Inhibition of NO within these systems may cause adverse consequences for exercise capacity. In myocytes, however, a reduction in NO stimulates mitochondrial respiration.34 The effects of hypercholesterolemia on exercise-induced hyperemia and aerobic capacity were remarkably similar to that of treatment with the NO synthase antagonist. We suspect that this is because hypercholesterolemia shares a common mechanism: perturbation of the endothelial NO synthase pathway.
To conclude, EDNO appears to play a significant role in limb blood flow during exercise. A pathological disturbance in EDNO activity like that observed in hypercholesterolemia results in a loss of blood flow redistribution and a reduction in exercise capacity.
Selected Abbreviations and Acronyms
|%COrm||=||redistribution of blood flow to running muscles as percentage of total cardiac output|
|DCM||=||Department of Comparative Medicine|
|E−||=||apoE-deficient (apoE-knockout) mice|
|E+LNA||=||wild-type mice receiving LNA|
|EDNO||=||endothelium-derived nitric oxide|
|V̇o2max||=||maximal rate of oxygen uptake|
This work was supported in part by a grant from the National Heart, Lung, and Blood Institute (1RO1-HL-58638) and was done during the tenure of a Grant-in-Aid award from the American Heart Association and Sanofi Winthrop. Dr Maxwell is the recipient of a Bugher Foundation Fellowship of the American Heart Association. Dr Cooke an Established Investigator of the American Heart Association. The authors thank Bruce Payne of the University of California, San Francisco, for technical assistance and the use of the fluorimeter.
- Received September 16, 1997.
- Revision received February 5, 1998.
- Accepted February 13, 1998.
- Copyright © 1998 by American Heart Association
Maxwell AJ, Niebauer J, Lin PS, Tsao PS, Bernstein D, Cooke JP. Hypercholesterolemia impairs exercise capacity: role of nitric oxide. J Am Coll Cardiol. 1997;29:2A. Abstract.
Piedrahita JA, Zhang SH, Hagaman JR, Oliver PM, Maeda N. Generation of mice carrying a mutant apolipoprotein E gene inactivated by gene targeting in embryonic stem cells. Proc Natl Acad Sci U S A. 1992;89:4471–4475.
Beaver WL, Wasserman K, Whipp BJ. A new method for detecting the anaerobic threshold by gas exchange. J Appl Physiol. 1986;60:2020–2027.
Harpur RP. The rat as a model for physical fitness. Comp Biochem Physiol. 1980;66A:553–574.
Wasserman K, Hansen JE, Sue DY, Whipp BJ, Casaburi R. Principles of Exercise Testing and Interpretation. 2nd ed. Philadelphia, Pa: Lea & Febiger; 1994:479.
Stringer W, Casaburi R, Wasserman K. Acid-base regulation during exercise and recovery in humans. J Appl Physiol. 1992;72:954–961.
Wasserman K. Coupling of external to cellular respiration during exercise: the wisdom of the body revisited. Am J Physiol. 1994;266:E519–E539.
Tsao PS, McEvoy LM, Drexler HD, Butcher EC, Cooke JP. Enhanced endothelial adhesiveness in hypercholesterolemia is attenuated by l-arginine. Circulation. 1994;89:2176–2182.
Musch TI, Terrell JA, Hilty MR. Effects of high-intensity sprint training on skeletal muscle blood flow in rats. J Appl Physiol. 1991;71:1387–1395.
Barbee RW, Perry BD, Ré RN, Murgo JP. Microsphere and dilution techniques for the determination of blood flows and volumes in conscious mice. Am J Physiol. 1992;263:R728–R733.
Allain CC, Poon LS, Chan CSG, Richmond W, Fu PC. Enzymatic determination of total serum cholesterol. Clin Chem. 1974;20:470–475.
Ishise S, Pegram BL, Yamamoto J, Kitamura Y, Frolich ED. Reference sample microsphere method: cardiac output and blood flows in the conscious rat. Am J Physiol. 1980;239:H443–H449.
Koller A, Dong S, Huang A, Kaley G. Corelease of nitric oxide and prostaglandins mediates flow-dependent dilation of rat gracilis muscle arterioles. Am J Physiol. 1994;267:H326–H332.
Koller A, Kaley G. Prostaglandins mediate arteriolar dilation to increase blood flow velocity in skeletal muscle microcirculation. Circ Res. 1990;67:529–534.
Koller A, Sun DGK. Role of shear stress and endothelial prostaglandins in flow and viscosity-induced dilation of arterioles in vitro. Circ Res. 1993;72:1276–1284.
Koller A, Sun D, Huang A, Kaley G. Exercise training augments EDRF (nitric oxide) synthesis in skeletal muscle arterioles. In: Halpern W, Bevan JA, Brayden J, Dustan H, Nelson M, Osol G, eds. The Resistance Arteries: Integration of the Regulatory Pathways. Totowa, NJ: Humana Press. 1994;153–161.
Blitzer ML, Lee SD, Creager MA. Endothelium-derived nitric oxide mediates hypoxic vasodilation of resistance vessels in humans. Am J Physiol. 1996;271:H1182–H1185.
Meredith IT, Currie KE, Anderson TJ, Roddy MA, Ganz P, Creager MA. Postischemic vasodilation in human forearm is dependent on endothelium-derived nitric oxide. Am J Physiol. 1996;270:H1435–H1440.
Endo T, Imaizumi T, Tagawa T, Shiramoto M, Ando S, Takeshita A. Role of nitric oxide in exercise-induced vasodilation of the forearm. Circulation. 1994;90:2886–2890.
Gilligan DM, Panza JA, Kilcoyne CM, Waclawiw MA, Casino PR, Quyyumi AA. Contribution of endothelium-derived nitric oxide to exercise-induced vasodilation. Circulation. 1994;90:2853–2858.
Wilson JR, Kapoor S. Contribution of endothelium-derived relaxing factor to exercise-induced vasodilation in humans. J Appl Physiol. 1993;75:2740–2744.
Di Prampero PE. An analysis of the factors limiting maximal oxygen consumption in healthy subjects. Chest. 1992;101:188S–191S.
Barclay JK, Stainsby WN. The role of blood flow in limiting maximal metabolic rate in muscle. Med Sci Sports Exerc. 1975;7:116–119.
Di Prampero PE. Metabolic and circulatory limitations to VO2 max at the whole animal level. J Exp Biol. 1985;115:319–331.