Corresponding author J. Hansen: Copenhagen Muscle Research Centre, Rigshospitalet, section 7652, Blegdamsvej 9, DK-2100 Copenhagen, Denmark. Email: email@example.com
1Sympathetically evoked vasoconstriction is modulated by skeletal muscle contraction, but the underlying events are incompletely understood. During contraction, intramuscular oxygenation decreases with increasing exercise intensity. We therefore hypothesized that tissue hypoxia plays a crucial role in the attenuation of sympathetic vasoconstriction in contracting skeletal muscle.
2In 19 subjects, near-infrared spectroscopy was used to measure decreases in muscle oxygenation (ΔtHbO2+MbO2) as an estimate of the vasoconstrictor response to reflex sympathetic activation with lower body negative pressure (LBNP) in the microcirculation of resting and contracting forearm muscles. Oxygen delivery to the muscles was reduced by decreasing (a) arterial O2 content by breathing 10 % O2, or (b) muscle perfusion by applying forearm positive pressure (FPP, +40 mmHg).
3In resting forearm, reflex sympathetic activation decreased muscle oxygenation by 11 ± 1 %. Handgrip alone at 5 and 20 % of maximal voluntary contraction (MVC) decreased muscle oxygenation by 4 ± 1 and 28 ± 4 %, respectively. When superimposed on handgrip, LBNP-induced decreases in muscle oxygenation were preserved during handgrip at 5 % MVC, but were abolished during handgrip at 20 % MVC. Oral administration of aspirin (1 g) did not restore the latter response.
4When the decrease in forearm muscle oxygenation elicited by handgrip at 20 % MVC was mimicked by either (a) systemic hypoxia plus 5 % handgrip (ΔtHbO2+MbO2, −32 ± 3 %), or (b) hypoperfusion of resting muscle by FPP (ΔtHbO2+MbO2, −26 ± 6 %), LBNP-induced decreases in muscle oxygenation were greatly attenuated.
5These data suggest that local tissue hypoxia is involved in the metabolic attenuation of sympathetic vasoconstriction in the microcirculation of exercising human skeletal muscle. The specific underlying mechanism remains to be determined, although products of the cyclo-oxygenase pathway do not appear to be involved.
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
Many of the cardiovascular adjustments to exercise are mediated by activation of the sympathetic nervous system. This sympathetic activation is dependent, in part, on a reflex mechanism, termed the muscle metaboreflex, that is initiated in skeletal muscle when local metabolic products of contraction stimulate thin fibre afferent nerves. These afferents in turn activate cardiovascular centres in the brainstem, resulting in increased efferent sympathetic outflow (Kaufman et al. 1983; Mense & Stahnke, 1983; Mitchell & Schmidt, 1983). The muscle metaboreflex-induced increases in efferent sympathetic nerve activity are targeted to exercising, as well as non-exercising, human skeletal muscle (Savard et al. 1987; Hansen et al. 1994). While sympathetic activation in resting muscle causes vasoconstriction, thereby redistributing cardiac output to the exercising muscles, the functional consequence of sympathetic activation in exercising muscle has been difficult to define (for review see Laughlin et al. 1996; Rowell et al. 1996).
We recently provided evidence that sympathetic vasoconstriction in rat and mouse hindlimb (Thomas et al. 1994, 1997, 1998; Thomas & Victor, 1998) and human forearm circulation (Hansen et al. 1996, 1999) is sensitive to modulation by local metabolic events in exercising skeletal muscle, a phenomenon initially termed functional sympatholysis (Remensnyder et al. 1962). Using near-infrared spectroscopy to measure oxygen availability in the skeletal muscle microcirculation in humans, we demonstrated that reflex activation of sympathetic vasoconstrictor nerves consistently decreased oxygenation in resting forearm muscles, but had no effect on oxygenation when the muscles were exercised (Hansen et al. 1996). However, the underlying mechanisms by which muscle contraction modulates sympathetic vasoconstriction are incompletely understood.
With the increased metabolic demand that accompanies muscle contraction, tissue PO2 decreases in proportion to exercise intensity. We therefore hypothesized that in humans tissue hypoxia is one of the pivotal proximal events involved in the metabolic modulation of sympathetic vasoconstriction in the microcirculation of exercising skeletal muscle. To test this hypothesis, we used near-infrared spectroscopy to estimate the vasoconstrictor response at the level of the microcirculation to reflex sympathetic activation (with lower body negative pressure) in resting and contracting forearm muscles when oxygen delivery was reduced either by systemic hypoxaemia (breathing of 10% O2 in N2) or by a local decrease in muscle perfusion (forearm positive pressure). We also determined whether products of the cyclo-oxygenase pathway are involved in the modulation of sympathetic neural control of skeletal muscle oxygenation in humans by studying subjects before and after cyclo-oxygenase inhibition with aspirin.
All protocols were approved by the Institutional Review Board at the University of Texas Southwestern Medical Center and by the Ethics Committee for Copenhagen and Frederiksberg. Written, informed consent was obtained from each subject prior to study. We studied a total of 19 healthy volunteer subjects, 23–32 years old (mean age, 25). Several of the subjects participated in more than one protocol on separate occasions. The experiments were carried out in accordance with the standards laid down in the Declaration of Helsinki.
Subjects were studied in the supine position. Heart rate was measured continuously by electrocardiography, arterial pressure was measured by finger photoplethysmography (Finapres; Ohmeda, Englewood, CO, USA). All measurements were recorded continuously on an electrostatic recorder (model ES 1000, Gould, Cleveland, OH, USA) and on a computer-based data acquisition system (Powerlab, ADInstruments, Hastings, UK).
In vivo near-infrared multiwavelength spectroscopy
Near-infrared spectroscopy (NIRS) was performed as described in detail previously (Hansen et al. 1996). Briefly, the near-infrared method exploits the principle that light between the wavelengths of 700 and 900 nm penetrates tissues with relative ease, and is absorbed mainly by the iron-porphyrin moieties in oxygenated and deoxygenated haemoglobin and myoglobin with a small contributing absorption by the copper moiety in cytochrome a,a3 (Piantadosi, 1989). Changes in absorption are proportional to changes in the relative concentrations of oxygenated haemoglobin and myoglobin (ΔtHbO2+MbO2), and deoxygenated haemoglobin and myoglobin (ΔtHb+Mb). Thus, the NIRS measurements reflect the local balance between oxygen supply and demand mainly at the level of the microcirculation within the region of illumination (Piantadosi, 1989; Mancini et al. 1994).
Near-infrared signals were obtained by means of two fibre optic bundles (optrodes) placed over the flexor digitorum profundus muscle of the left arm, which is the main muscle recruited during handgrip (Fleckenstein et al. 1992). The optrodes were held in place with a spring-loaded stereotactic device designed to minimize motion artifacts during handgrip. The optrodes were applied directly to the skin in parallel with each other 2.3 cm apart for optimal reflectance measurements. Near-infrared signals were sampled at a rate of 1 Hz and each data point was displayed as the mathematical average of 10 consecutive measurements. The maximal change in the near-infrared signals, termed the total labile signal (TLS), was determined in each experiment as the difference between baseline and complete deoxygenation produced by inflation of a pneumatic cuff at the level of the brachial artery to 280 mmHg for 8 min. Changes in the near-infrared signals are expressed as a percentage of the TLS.
Recording of sympathetic nerve discharge
Recordings of multiunit efferent postganglionic sympathetic nerve activity (SNA) to the skeletal muscle bed were obtained with unipolar tungsten microelectrodes inserted selectively into muscle nerve fascicles of the peroneal nerve posterior to the fibular head by microneurography (Vallbo et al. 1979). The neural signals were amplified ((20–50) × 103), filtered (bandwidth, 700–2000 Hz), rectified, and integrated (time constant, 0.1 s) to obtain a mean voltage display of sympathetic activity. A recording of muscle SNA was considered acceptable when the neurograms revealed spontaneous, pulse-synchronous bursts of neural activity, with a minimum signal-to-noise ratio of 3:1, that increased during phases II and III of the Valsalva manoeuvre, but not during arousal stimuli (loud noise or skin pinch). Sympathetic bursts were detected by inspection of the mean voltage neurograms. SNA was expressed as (a) the number of bursts of sympathetic activity per minute and (b) the number of bursts per minute multiplied by the mean burst amplitude in that minute (total activity).
Handgrip exercise was performed with either a custom-made or a standard (Stoelting, Wood Dale, IL, USA) handgrip dynamometer connected to a force transducer. Force output was recorded on paper and displayed on an oscilloscope to provide the subject with visual feedback. Prior to the experiment, each subject's maximal voluntary contraction (MVC) was determined as the best of four or five trials with verbal encouragement to improve at each trial. The exercise protocols consisted of intermittent isometric exercise in which subjects matched force production to a visual target, to the rhythm of a metronome (40 beats min−1) with a 50 % duty cycle.
Subjects performed rhythmic handgrip exercise at 5 or 20 % MVC for a total of 5 min. These intensities of rhythmic handgrip exercise do not engage the muscle metaboreflex and have been shown not to increase muscle SNA (Batman et al. 1994; Ray et al. 1994; Hansen et al. 1996).
Lower body negative pressure
The subject's lower body was enclosed in a negative pressure chamber to the level of the iliac crest. The pressure inside the chamber was measured using a Statham transducer (Gould, Oxnard, CA, USA). A non-hypotensive level of lower body negative pressure (LBNP) at −20 mmHg for 2 min was used selectively to unload cardiopulmonary baroreceptors, thereby increasing muscle SNA without concomitant changes in systemic arterial pressure (Jacobsen et al. 1993). These LBNP episodes cause highly reproducible reflex increases in muscle SNA before, during and after handgrip exercise (Scherrer et al. 1988; Hansen et al. 1996).
Because the reproducibility of LBNP-induced reflex increases in muscle SNA previously had not been established during systemic hypoxaemia, microneurography was performed through an opening in the side of the chamber. Once a stable recording of SNA was obtained, the opening was closed and sealed during the experiment.
Acute hypoxaemia was induced by breathing a ‘hypoxic’ gas mixture consisting of 10 % O2 in N2. The calibrated gas mixture was continuously fed to a reservoir bag (volume 70 l) and breathed by the subjects through a three-way valve using a nose-clip to ensure mouth breathing only. Arterial O2 saturation was estimated by pulse oximetry (Nellcor, Hayward, CA, USA) with the probe placed on the right index finger.
Forearm positive pressure
Forearm positive pressure (FPP) was produced by the method of Joyner (Joyner, 1991). The subject's left arm and hand were enclosed in a custom-made Plexiglas chamber that also contained the handgrip dynamometer and the stereotactic device with the fibre optic bundles for NIRS. A seal was formed proximal to the elbow by the use of snuggly fitting non-restrictive neoprene cuffs. The pressure inside the box was measured using a Statham transducer (Gould).
By pressurizing the box it is possible to reduce limb perfusion pressure, thereby reducing blood flow and oxygen delivery to the forearm muscles (Joyner, 1991). In the present study, this strategy was used to mimic contraction-induced decreases in muscle oxygenation in resting forearm. In pilot studies we determined that this could be accomplished by raising box pressure by 40 mmHg above atmospheric pressure. In a pilot study on four subjects a 5 min period of FPP at 40 mmHg produced a rapid decrease in muscle oxygenation that was complete within 30 s and remained at a stable steady-state level for the remainder of the 5 min.
Protocol 1. Effect of increasing metabolic demand on reflex sympathetic neural control of oxygenation in skeletal muscle (n= 6). The purpose of this protocol was to test the hypothesis that the ability of sympathetic activation to decrease muscle oxygenation is attenuated as muscle metabolic demand is increased. To test this, blood pressure, heart rate, handgrip force and near-infrared signals were recorded in response to reflex sympathetic activation elicited by LBNP at −20 mmHg before and during 5 min of a mild (5 % MVC) or a moderate (20 % MVC) intensity of rhythmic handgrip. LBNP was applied during the 3rd and 4th minutes of each exercise period. At least 15 min of recovery was allowed between the handgrip periods. We have previously shown that sympathetic vasoconstriction in the muscle is fully restored within 10 min of recovery from contraction (Hansen et al. 1996). Eight minutes of complete circulatory arrest was performed at the end of the protocol.
Protocol 2. Effect of decreasing oxygen supply by hypoxaemia on reflex sympathetic neural control of oxygenation in skeletal muscle (n= 7). In protocol 1, muscle oxygenation was manipulated by increasing metabolic demand. In protocol 2, instead of changing metabolic demand, muscle oxygenation was manipulated by the breathing of a hypoxic gas mixture to reduce oxygen supply. Blood pressure, heart rate, O2 saturation, handgrip force, muscle SNA and near-infrared signals were recorded in response to LBNP (–20 mmHg for 2 min) before and during rhythmic handgrip at 5 % MVC for 5 min alone and in combination with breathing of a hypoxic gas. Subjects breathed the hypoxic gas continuously for 10 min before and throughout the experimental protocol. Because this level of acute hypoxic hypoxaemia previously has been reported to cause sympathetic neural activation (Rowell et al. 1989), muscle SNA was recorded to ascertain that LBNP evoked similar increases in SNA in normoxia and hypoxia. Eight minutes of complete circulatory arrest was performed at the end of the protocol.
Protocol 3. Effect of decreasing oxygen supply by reducing perfusion pressure on reflex sympathetic neural control of oxygenation in skeletal muscle (n= 7). Rather than reduce oxygen supply globally as in protocol 2, the aim of protocol 3 was to reduce muscle oxygenation locally in the forearm by applying positive pressure to decrease forearm perfusion pressure. Blood pressure, heart rate and near-infrared signals were recorded in response to LBNP at −20 mmHg for 2 min before and during FPP at 40 mmHg for 5 min. LBNP was applied during the 3rd and 4th minutes of FPP. After 15 min of recovery, the response to a 1 min period of complete forearm circulatory arrest (pneumatic cuff inflated to 280 mmHg) was recorded before and during FPP at 40 mmHg for 5 min. The forearm circulatory arrest was applied during the 3rd minute of FPP. We previously determined that 1 min of complete circulatory arrest of a resting forearm produced decreases in muscle oxygenation equivalent to those produced by 2 min of LBNP at −20 mmHg (Hansen et al. 1996). Eight minutes of complete circulatory arrest was performed at the end of the protocol.
Protocol 4. Effects of cyclo-oxygenase inhibition on reflex sympathetic neural control of oxygenation in contracting skeletal muscle (n= 5). The aim of this protocol was to test the hypothesis that products of the cyclo-oxygenase cascade play an essential role in the attenuation of sympathetic control of oxygenation in exercising skeletal muscle. If so, administration of a cyclo-oxygenase inhibitor should restore the ability of reflex sympathetic activation to decrease tissue oxygenation in contracting muscle. Blood pressure, heart rate, handgrip force and near-infrared signals were recorded in response to LBNP at −20 mmHg for 2 min before and during a 5 min period of rhythmic handgrip at 20 % MVC. One gram of aspirin was administered orally and the protocol was repeated after 2 h. Eight minutes of complete circulatory arrest was performed at the end of the protocol. The efficacy of this dose of aspirin to inhibit cyclo-oxygenase activity was demonstrated previously in human hand veins, where the large vasodilator response to infused arachidonic acid was completely abolished 2 h after oral administration of 1 g aspirin (Bhagat et al. 1995).
Statistical analysis was performed using repeated measures analysis of variance with Dunnett's post hoc test to detect values that were different from control values. When appropriate, group mean values were compared using Student's t test for paired comparisons. A P value < 0.05 was considered significant. Data are expressed as means ±s.e.m.
Sympathetic neural control of muscle oxygenation in contracting forearm is attenuated as metabolic demand is increased
In resting forearm, reflex sympathetic activation with non-hypotensive LBNP produced a robust and reproducible decrease in muscle oxygenation which returned to baseline values with the offset of LBNP. Rhythmic handgrip alone at 5 or 20 % MVC produced intensity-dependent decreases in muscle oxygenation that reached steady state within the first 2 min of exercise. When superimposed on mild handgrip at 5 % MVC, reflex increases in sympathetic nerve activity induced by LBNP produced decreases in muscle oxygenation that were similar to those in resting muscle. In contrast, when superimposed on moderate handgrip at 20 % MVC, the response to LBNP was completely abolished. See Fig 1, Fig 2, and Table 1.
Table 1. Haemodynamic values and forearm muscle oxygenation during steadystate handgrip and LBNP performed alone and in combination
HR (beats min−1)
ΔtHbO2+MbO2 (% TLS)
ΔtHb+Mb (% TLS)
Data represent means ±s.e.m.in 7 subjects. HR, heart rate; MAP, mean arterial pressure; TLS, total labile signal; LBNP, lower body negative pressure. *P < 0.05vs. Rest; †P < 0.05vs. 5% handgrip.
Values during steady state
60 ± 3
85 ± 7
61 ± 4
91 ± 6
+10 ± 2*
65 ± 5 *
100 ± 6 *
−28 ± 4 *†
+45 ± 10 *†
Responses to LBNP
+5 ± 3
+0 ± 1
−11 ± 1
+8 ± 2
+11 ±3 *
+4 ± 2
−12 ± 1
+11 ± 2
+11 ± 3*
+5 ± 2*
−1 ± 2*†
−6 ± 3*†
Sympathetic neural control of muscle oxygenation in contracting forearm also is attenuated when oxygen delivery is decreased by systemic hypoxia
During breathing of 10 % oxygen, O2 saturation decreased from 97 ± 1 to 74 ± 3 % (P < 0.05) and muscle SNA increased by 105 ± 26 % (P < 0.05). Despite this change in baseline SNA, the LBNP-induced reflex increases in muscle SNA were similar in subjects breathing room air or 10 % O2. During normoxia, LBNP-induced increases in muscle SNA produced robust decreases in muscle oxygenation in the resting forearm. This LBNP-induced decrease in muscle oxygenation was well preserved during mild rhythmic handgrip at 5 % MVC, which by itself decreased muscle oxygenation by 13 ± 3 %. Acute hypoxaemia reduced oxygenation of the resting muscle by 11 ± 3 %. When superimposed on this new baseline, reflex increases in muscle SNA were associated with decreases in oxygenation in resting forearm that were similar to those observed during normoxia. In contrast to normoxia, handgrip at 5 % MVC performed during hypoxia produced a large decrease in muscle oxygenation of 32 ± 3 %. Under these conditions of handgrip plus hypoxia, LBNP-induced increases in SNA were not acompanied by decreases in muscle oxygenation. See Fig. 3 and Table 2.
Table 2. Haemodynamic values, muscle SNA and forearm muscle oxygenation during steadystate handgrip and responses to LBNP performed during normoxia or systemic hypoxia (10% O2)
HR (beats min−1)
Frequency (bursts min−1)
Total activity (units)
Data represent means ±s.e.m.in 7 subjects (muscle SNA,n= 6). *P < 0.05vs. Rest; †P < 0.05vs. normoxia.
Values at rest and during handgrip
63 ± 3
93 ± 5
11 ± 3
69 ± 4
102 ± 6
13 ± 3
155 ± 46
−13 ± 3*
+16 ±2 *
88 ± 7†
105 ± 6
22 ± 4†
259 ± 44†
−11 ± 3†
+11 ± 3†
93 ± 6 †
105 ± 5
26 ± 5 †
359 ± 65 †
−32 ± 3 *†
+35 ± 4 *†
Changes in response to LBNP
+2 ± 1
+4 ± 1
+12 ± 2
+247 ± 122
−10 ± 1
+5 ± 1
+1 ± 1
+9 ± 1
+136 ± 30
−9 ± 1
+7 ± 1
+4 ± 2
−3 ± 3 †
+10 ± 1
+227 ± 124
−11 ± 1
+7 ± 1
+3 ± 2
−4 ± 2
+10 ± 5
+354 ± 194
0 ± 1*†
0 ± 1*†
Sympathetic neural control of muscle oxygenation in resting forearm is attenuated when oxygen delivery is decreased by forearm positive pressure
In resting forearm, similar decreases in muscle oxygenation were produced either by LBNP-induced reflex sympathetic activation or by a brief period of forearm vascular occlusion. Application of FPP at 40 mmHg to reduce perfusion pressure markedly decreased muscle oxygenation by 26 ± 6 %. When applied during FPP, LBNP-induced decreases in muscle oxygenation were greatly attenuated. In contrast, the decrease in muscle oxygenation produced by a brief period of vascular occlusion was fully preserved during FPP, demonstrating that the ability of the method to detect reductions in oxygen delivery is preserved under the conditions of FPP and tissue hypoxia. See Fig. 4.
Pharmacological inhibition of the cyclo-oxygenase pathway with aspirin has no effect on the attenuated sympathetic neural control of muscle oxygenation in contracting forearm
In resting muscle, LBNP caused significant but similar reductions in muscle oxygenation before and after administration of aspirin (ΔtHbO2+MbO2, −17 ± 2 and −14 ± 2 %, respectively). Moderate rhythmic handgrip at 20 % MVC alone produced similar steady-state reductions in muscle oxygenation before and after aspirin (ΔtHbO2+MbO2, −48 ± 8 and −40 ± 7 %, respectively). During handgrip, LBNP-induced decreases in muscle oxygenation were greatly attenuated before, as well as after, administration of aspirin (ΔtHbO2+MbO2, −1 ± 2 and 1 ± 3 %, respectively, not significant).
We previously have provided several lines of evidence in anaesthetized rats and mice and in conscious humans that sympathetic vasoconstriction is greatly attenuated in contracting skeletal muscle (Thomas et al. 1994, 1997, 1998; Hansen et al. 1996, 1999; Thomas & Victor, 1998). In the present study in humans, we now have examined the role played by muscle tissue hypoxia in this situation. We used near-infrared spectroscopy (NIRS) to provide an index of sympathetic vasoconstrictor responses in the microcirculation of skeletal muscle that was made hypoxic either by increasing oxygen utilization (by handgrip) or by decreasing oxygen supply (by systemic hypoxaemia or forearm positive pressure). The major new finding is that the ability of reflex sympathetic activation to evoke vasoconstriction at the microcirculatory level, as estimated by NIRS, is dependent on the incipient level of muscle oxygenation. We found that minor reductions in baseline muscle oxygenation of <15 % did not prevent sympathetic activation from further decreasing muscle oxygenation. In contrast, regardless of the method used to produce tissue hypoxia, reductions in baseline muscle oxygenation of >25 % diminished the ability of sympathetic activation to further decrease muscle oxygenation (Fig. 5). Based on these observations, we suggest that contraction-induced skeletal muscle tissue hypoxia plays an important role in the local metabolic attenuation of sympathetic vasoconstriction in the microcirculation of active muscle. Although the precise mechanism by which muscle hypoxia attenuates sympathetic control of oxygenation remains to be determined, our data suggest that hypoxia-induced activation of the cyclooxygenase pathway is not essential.
In this study, we chose to use NIRS to examine the metabolic modulation of sympathetic vasoconstriction in the microcirculation of exercising human skeletal muscle because the technique provides several advantages over more conventional haemodynamic methodologies. First, NIRS permits the continuous measurement of oxygen availability at the level of the skeletal muscle microcirculation, that portion of the vascular bed most accessible to metabolic products of contraction. Second, the spatial resolution of NIRS is such that measurements can be obtained from the truly active, rather than adjacent inactive, small muscles of the forearm. The use of NIRS as an index of sympathetic vasoconstriction in human forearm has been validated previously by demonstrating that LBNP-induced decreases in blood flow (–39 ± 5 % as measured by plethysmography) and resting muscle oxygenation (–10 ± 2 %) were abolished by local sympathetic blockade (Hansen et al. 1996). Because NIRS reflects the balance between oxygen supply and utilization, we cannot assume that the relationship between changes in blood flow and changes in tissue oxygenation are similar in resting and contracting muscle. However, due to the higher oxygen consumption rate in contracting muscle, a given reduction in muscle blood flow would be expected to produce a greater reduction in oxygenation in contracting than in resting muscle. Thus, the sensitivity of the NIRS measurements to detect vasoconstriction should, if anything, be enhanced in contracting muscle.
Our initial observation that reflex sympathetic neural control of skeletal muscle oxygenation is differentially affected by handgrip at 5 and 20 % MVC confirmed our previous study in which we found that the threshold intensity of contraction needed to attenuate sympathetically mediated decreases in muscle oxygenation was somewhere between 10 and 20 % MVC (Hansen et al. 1996). Although there are numerous potential explanations for the differential effect of mild versus moderate handgrip on the response to reflex sympathetic activation, one particularly notable difference between 5 and 20 % MVC handgrip is the greater contraction-induced decrease in skeletal muscle oxygenation produced by the latter (Fig. 1). In previous animal and human studies, experimentally induced tissue hypoxia has been shown to impair the vasoconstrictor response to sympathetic nerve stimulation as well as to exogenous noradrenaline infusion (Skinner & Costin, 1969; Heistad & Wheeler, 1970; Costin & Skinner, 1971; Heistad et al. 1975; Granger et al. 1976; Boegehold & Johnson, 1988; Tateishi & Faber, 1995). We therefore asked whether contraction-induced tissue hypoxia might modulate the vasoconstrictor response to reflex sympathetic neural activation.
We used two complementary experimental approaches to examine the oxygen dependency of the antagonistic interaction between skeletal muscle contraction and sympathetic vasoconstriction as estimated by NIRS. First, we exposed subjects to acute systemic hypoxia, which previously has been shown to reduce tissue oxygen delivery despite a slight increase in perfusion (Heistad & Wheeler, 1970; Koskolou et al. 1997). Second, we localized the hypoxic stimulus solely to the forearm by applying forearm positive pressure, an intervention which has been shown to reduce tissue oxygen delivery due to decreased perfusion (Joyner, 1991). The combination of these two approaches allowed us to associate the observed impairment in the sympathetic neural control of muscle oxygenation specifically with a local effect of muscle tissue hypoxia, rather than with a generalized humorally mediated effect of systemic hypoxaemia. Furthermore, we also were able to associate the impaired sympathetic neural control of muscle oxygenation with a metabolic effect of muscle contraction rather than with a purely mechanical effect of contraction. The evidence is that the same intensity of muscle contraction (5 % MVC) had different effects on the response to reflex sympathetic activation when oxygen delivery was varied experimentally. Taken together, these results suggest that when muscle contraction produces a mismatch between oxygen supply and demand, the resultant tissue hypoxia, or some factor that changes as a direct result of hypoxia, serves to modulate the microvascular response to sympathetic neural activation.
Prostaglandins initially were attractive candidates to mediate the contraction-induced attenuation of sympathetic vasoconstriction in humans for several reasons. Prostaglandins are produced in contracting skeletal muscle (Young & Sparks, 1980; Wilson & Kapoor, 1993; Lang et al. 1997) and have been implicated in both contraction-induced and hypoxia-induced vasodilatation (Busse et al. 1984; Wilson & Kapoor, 1993; Fredericks et al. 1994; Lang et al. 1997). Furthermore, prostaglandins have been shown to inhibit sympathetic vasoconstriction produced by either α-adrenoceptor agonists or sympathetic nerve stimulation in rabbit hindquarters (Lippton et al. 1983) and in rat cremaster muscle (Faber et al. 1982). In the present study, administration of aspirin, at a dose previously shown to completely abolish the vasodilator response to infused arachidonic acid (Bhagat et al. 1995), had no effect on sympathetic neural control of skeletal muscle oxygenation in resting or contracting forearm. Based on these results, products of the cyclo-oxygenase pathway do not appear to be essential for the attenuation of sympathetically mediated decreases in muscle oxygenation in exercising human forearm. A similar conclusion was reached in a recent study by Thomas & Victor (1998), who found no evidence for a primary role of endogenous prostaglandins in the metabolic attenuation of sympathetic vasoconstriction in contracting rat hindlimb.
Several other highly oxygen-sensitive vasodilator mechanisms such as KATP channels in vascular smooth muscle and nitric oxide recently have been implicated in hypoxia- or contraction-induced inhibition of α-adrenoceptor-mediated vasoconstriction in rat skeletal muscle vasculature (Tateishi & Faber, 1995; Thomas et al. 1997, 1998; Thomas & Victor, 1998). However, in human skeletal muscle the relative importance of these mechanisms as modulators of sympathetic vasoconstriction in contracting muscle has yet to be determined. The precise role of other oxygen-sensitive vasodilator mechanisms including direct effects of hypoxia on vascular smooth muscle (Taggart & Wray, 1998), ATP/adenosine production (Mian & Marshall, 1991; Ellsworth et al. 1995; Skinner & Marshall, 1996), or formation of 20-HETE (Harder et al. 1996), none of which are mutually exclusive, is beyond the scope of the present investigation.
In conclusion, the present data in humans implicate local tissue hypoxia as an important factor in the contraction-induced modulation of sympathetic vasoconstriction at the level of the microcirculation. Critical decreases in muscle oxygenation of >25 % due to increased metabolic demand or decreased oxygen delivery attenuate local sympathetic vasoconstriction in the microcirculation (as estimated by NIRS) by a mechanism which does not seem to depend on products of the cyclo-oxygenase pathway. Such inhibition, in combination with preserved sympathetic vasoconstriction in non-hypoxic tissues, would serve to optimize oxygen delivery to the most hypoxic tissue regions under conditions of sympathetic activation, such as intense muscular activity or systemic hypoxia. Because more proximal segments of the vasculature may be less susceptible to such metabolic modulation these findings do not exclude the possibility that sympathetic control of proximal vascular dimensions and vascular resistance is partially preserved under conditions when the influence on more distal vessels is lost. While these findings in healthy humans emphasize the role of the sympathetic nervous system in the normal matching of regional oxygen supply to metabolic demand, we further speculate that they may have implications for our understanding of blood flow and blood pressure regulation in disease states accompanied by severe tissue hypoxia, such as pulmonary and/or cardiac insufficiency.
We thank Dr Michael J. Joyner for generously providing the forearm positive pressure chamber. J.H. was supported by a Fogarty International Research Fellowship (NIH-1-F05-TW04949-01) and grants from the Danish Heart Foundation. M.S. was supported by a Fogarty International Research Fellowship (NIH-1-F05 TW05085-02) and by the Michaelsen Foundation. G.D.T. was supported by a Muscular Dystrophy Association Postdoctoral Fellowship. This research was supported by grants from the Danish Heart Association, the Danish Research Council, the NOVO Foundation, the Copenhagen Muscle Research Centre and the National Institutes of Health (PO1-HL-06296).