INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

DURING EXERCISE, BLOOD VESSELS in active skeletal muscle vasodilate and the autonomic nervous system is engaged (23). The mechanisms responsible for vasodilatation and sympathoexcitation are complex; however, it is clear that metabolic by-products of muscle contraction contribute to both processes (26). In the case of vasodilatation, metabolic by-products can alter the degree of smooth muscle contraction and, thereby, alter vascular tone (6). In addition, autonomic activation of the exercise pressor reflex occurs in part due to the stimulation of thin fiber muscle afferents (2), whose free nerve endings are located in the interstitium of contracting skeletal muscle (38). Accordingly, to understand the role any given substance might play in either or both processes requires the ability to directly measure their concentrations within the interstitium. Until recently, this has been a major limitation in our understanding of the local characteristics of these processes. However, the microdialysis technique (17) affords the opportunity to directly measure and quantify the concentrations of many substances suggested to play a role in these processes.

Previous investigations of potential metabolites (K⁺, lactate, adenosine, phosphate, and pH) involved in vasodilatation and/or stimulation of muscle afferents have often yielded conflicting findings in animals (22, 24, 29, 30) and humans (5, 7, 18, 30, 37). However, in many of these reports, interstitial concentrations were inferred from measurements of venous plasma concentrations or by ³¹P-nuclear magnetic resonance determinations of the cellular concentrations of various substances during bouts of muscle contraction. There have been only a few studies (9, 11, 14, 17) that examined specific individual interstitial metabolites (e.g., lactate, K⁺, and adenosine), and to our knowledge, there have not been prior human studies in which multiple metabolites were determined by microdialysis in conjunction with blood flow measurements during exercise.

Therefore, the purpose of the present study was to utilize the microdialysis technique to measure interstitial K⁺, lactate, adenosine, phosphate, pH, and H⁺ within the quadriceps skeletal muscle of human subjects during one-legged dynamic knee extension exercise at different relative maximal work capacities (W_max). Simultaneous measurements of heart rate (HR), blood pressure (BP), and femoral mean blood velocity (MBV) were made to draw inferences regarding the role these metabolites may play in neurocirculatory control.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Subjects. Seven healthy subjects (4 male and 3 female) aged 20-26 yr (23 ± 1), 160-183 cm height (168 ± 3), 49.5-96.0 kg wt (70.1 ± 6.5), and body mass index of 19.1-28.8 kg/m² (24.5 ± 1.6) participated in this study. Values in parentheses are in means ± SE. The subjects were sedentary normotensive nonsmokers. In addition, the subjects were without orthopedic limitations, and they were not taking any medications. The Institutional Review Board of the Milton S. Hershey Medical Center approved the experimental protocol. Each subject had the purposes and risks of the protocol explained to them, and written informed consent was obtained before the subjects entered the study.

HR, BP, and blood flow. HR was monitored by electrocardiogram, and systolic BP and diastolic BP were measured using the volume-clamp method (Finapres, Ohmeda; Madison, WI). Finapres readings were adjusted based on resting BP determined using an automated sphygmomanometer (Dinamap, Critikon; Tampa, FL). Femoral artery MBV was measured using a 4-MHz pulsed-wave Doppler ultrasound probe (model 500M, Multigon; Yonkers, NY) with Zero Crossing (Hokanson; Bellevue, WA) taped in a fixed position ~2-3 cm distal to the inguinal ligament at a 45° angle of insonation (20). An index of vascular conductance was obtained by dividing the flow velocity by mean arterial pressure (MAP).

Microdialysis probes. The semipermeable fibers used to construct the microdialysis probes (Spectrum Laboratories; Laguna Hills, CA) had a molecular mass cutoff of 13,000 Da. Each end of a single fiber was inserted ~1 cm into a hollow polyamide tube [0.50 mm inner diameter (ID), 0.63 mm outer diameter (OD)] and glued. The actual probe length (e.g., distance between the two polyamide tubes) was 4 cm (0.20 mm ID, 0.22 mm OD). To withstand the forces generated by muscle contraction, tensile strength was added to the microdialysis probe by gluing a 10-cm piece of 5-0 suture (Ethicon) to the polyamide tubing. The suture was attached so that 3 cm was glued to the polyamide tubing on one side of the probe and 3 cm was glued to the polyamide tubing on the other. Thus the suture was glued only to the polyamide tubing but spanned the distance of the probe. This modification allowed the microdialysis probes to function very well during muscle contractions (14, 16).

Microdialysis probe insertion. Six microdialysis probes were inserted into the vastus lateralis muscle of either the right or the left leg of the subject. The skin and subcutaneous tissue at the probe entrance and exit sites was first anesthetized with a local injection (0.5-1.0 ml) of lidocaine and epinephrine (20 mg/ml + 12.5 µg/ml). The probes were inserted into the muscle via a 20-gauge cannula in a direction parallel to muscle fiber orientation (i.e., 45° moving proximally and laterally). The distance between the entrance and exit sites of the probes was ~9 cm and the distance between each probe was ~1-2 cm. After insertion, the microdialysis probes were attached to a perfusion pump (model 102, CMA) and perfused at a rate of 5 µl/min with a Ringer solution (4.0 meq/l K⁺, pH 7.3, and 0 mM phosphate). In an effort to minimize the possibility of draining the interstitial space (4, 13), the perfusate contained 3.0 mM glucose and 0.5 mM lactate. The outflow tube of two microdialysis probes were attached to flow-through K⁺ microelectrodes, whereas another two probes were attached to flow-through pH microelectrodes (Microelectrodes; Londonderry, NH). The K⁺ and pH microelectrodes were connected to an Orion pH meter with separate channels for ionic determinations and a Fischer Scientific pH meter, respectively, which allowed the manual recording of both K⁺ and pH. It should be noted that there was a 6-min delay between the passage of the perfusate through the microdialysis probe and detection by the microelectrodes. This time delay was taken into consideration when calculating the data, and thus all data are presented in real time. Dialysate was collected in 100-µl microcentrifuge tubes, immediately sealed to prevent evaporation, and stored at 80°C until analysis. The probes were perfused, and the subjects rested supine for 60 min before the experiment was initiated.

Determination of probe recovery. To fully utilize the microdialysis technique, an estimate of the in vivo extraction fraction of the compound being measured from the interstitial space needs to be made which is defined as "probe recovery." This determination is necessary to calculate actual interstitial concentrations and to document any changes in probe recovery associated with muscle contraction because probe recovery can vary as a factor of workload (9, 14, 21). In the present study, the "internal reference" method introduced by Scheller and Kolb (27) was used. The details of this method used in our laboratory have been reported previously (15). The major advantage of this method is that probe recovery can be determined for each collected sample allowing the continual monitoring of probe recovery over time. In the present study, very small amounts of L-[U-¹⁴C]lactate and [2-³H]adenosine (<0.22 µCi/ml) were added to the final perfusion solution as the internal reference markers to determine lactate and adenosine probe recovery.

Experimental protocol. Before the experiment, subjects were familiarized with the Krogh bicycle ergometer modified for one-legged knee extension exercise as previously described (1). Briefly, subjects were seated with their upper body, hips, and knees strapped in to ensure stability and their lower legs flexed at an angle of 90° at the knee. The foot was placed in a boot attached to a rod containing a strain gauge for force measurements and connected to the pedal arm of a cycle ergometer placed behind the subject. Subjects actively extended their leg at the knee to ~170°, while the momentum of the pedal arm passively returned the leg to the starting position. This type of exercise resulted in muscle contractions almost exclusively isolated to the quadriceps muscle mass which accounts for ~3 kg of muscle (1). This allowed for the simultaneous measurement of cardiovascular variables and interstitial metabolites from a specific isolated muscle mass (14, 25).

View larger version (16K): [in this window] [in a new window]   Fig. 1.   Time course for exercise bouts [i.e., 30% and 60% maximal work capacity (Wmax)]. Resting heart rate (HR), blood pressure (BP), and mean blood velocity (MBV) were averaged over 5 min and then continuously measured during exercise and recovery. Interstitial K+ and pH were continuously measured and manually recorded each minute during exercise and recovery. Interstitial adenosine, lactate, and phosphate were measured during the periods of rest, exercise, and recovery and represent a sum value for each of these time periods (, specific time points recorded).

Data analysis. HR, BP, and MBV were calculated during the last 15 s of each minute during rest, exercise, and recovery periods. Resting values represent an average of the entire 5-min baseline for these variables. Because of the technical difficulties obtaining MBV during contractions, we elected to use the MBV obtained immediately after stopping exercise with corresponding HR and MAP values. This allowed us to compare peak exercise flow responses at the different workloads. Recovery values for these variables represent the last minute of recovery.

Calculations. Probe recovery based on the internal reference method was calculated as follows
where P_dpm and D_dpm represent the disintegrations per minute in the perfusate and dialysate, respectively, for lactate and adenosine. The probe recoveries were then used to calculate the actual interstitial concentrations of lactate and adenosine as follows
where D_c and P_c represent the dialysate and perfusate concentrations of lactate and adenosine, respectively.

Statistics. Rest (average of 5 min), end of exercise, and recovery time points were used for HR, MAP, MBV, and conductance. The sum of the effects for over the duration of rest, exercise, and recovery were reported for interstitial adenosine, lactate, and phosphate. Because K⁺ and pH were continuously collected online, the exercise slopes from rest to the end of exercise and recovery slopes from end of exercise to the end of recovery were calculated. If the data was not normally distributed, linear log transformations were used (e.g., K⁺ and pH). Changes in HR, MAP, MBV, conductance, probe recovery, and interstitial metabolites (adenosine, lactate, phosphate, K⁺, and pH) were analyzed using a two-way repeated measures of ANOVA comparing changes in workload (30 vs. 60% W_max) and changes over time (rest vs. end of exercise and rest vs. end of recovery). If significant changes were observed, a Bonferroni post hoc test for multiple comparisons was used to determine where the significant changes occurred. Finally, regression analysis was used in comparing MBV and interstitial variables at rest and during exercise and recovery. Values are means ± SE and significance was accepted at P < 0.05.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Workload and hemodynamics. The mean workload used for exercise at 30% W_max was 14 ± 1 W (range 9 to 18 W) and for exercise at 60% W_max was 27 ± 3 W (range 18-36 W). During exercise, HR and MAP increased during both exercise bouts (P < 0.05). There was a greater change in HR for exercise at 60% W_max (34 ± 9 beats/min) compared with exercise at 30% W_max (18 ± 2 beats/min; P < 0.05), whereas there was no difference between workloads on the MAP response to contraction (13 ± 3 mmHg and 21 ± 6 mmHg for 30 and 60% W_max, respectively). HR and MAP returned to baseline at the end of 5 min of recovery for both exercise intensities.

Mean blood flow and conductance. Resting MBV was similar before exercise at 30 and 60% W_max, 16.0 ± 4.0 and 16.4 ± 2.5 cm/s, respectively (Fig. 2A). With exercise, MBV increased (P < 0.05) to 59.7 ± 5.1 and 93.4 ± 5.0 cm/s for the 30 and 60% W_max bouts, respectively. The change from rest to exercise was greater for 60% (77.0 ± 6.8 cm/s) compared with 30% W_max bout (43.7 ± 3.7 cm/s) (P < 0.05). Conductance also significantly increased during exercise with a greater change occurring during the 60% W_max bout (P < 0.05) (Fig. 2B). During recovery, MBV and conductance returned to baseline after the 30% W_max bout. However, recovery MBV and conductance remained significantly elevated above baseline after the 60% W_max bout despite being significantly lower than exercise.

View larger version (26K): [in this window] [in a new window]   Fig. 2.   MBV (A) and conductance (B) significantly increased during exercise (n = 7 subjects). An intensity effect was observed with higher flows on immediate recovery and recovery for exercise at 60% Wmax compared with exercise at 30% Wmax. *P < 0.05, significantly different from baseline; P < 0.05, change from baseline significantly different between exercise at 30% and 60% Wmax.

Interstitial metabolites. There was a significant difference in probe recovery for adenosine between rest and exercise for both exercise intensities (30% W_max, 25.6 ± 1.5 vs. 32.0 ± 1.5%; 60% W_max, 25.0 ± 1.2 vs. 32.6 ± 1.4%). However, there were no significant differences in probe recovery for adenosine between rest and recovery. We encountered technique difficulties in determining the specific activity of [¹⁴C]-lactate. As a result, we were unable to estimate probe recovery. The determination of probe recovery, especially during exercise, is critical because we have shown (14) that probe recovery increased as a function of exercise intensity. Therefore, without an estimation of probe recovery, one cannot accurately compare dialysate concentrations, as a portion of the change in dialysate lactate-phosphate concentrations will be due to factors other than those associated with exercise. To overcome this drawback, we corrected the changes in exercise dialysate lactate concentrations [lactate = (exercise  base)/correction factor + base] using a correction factor generated from probe recoveries obtained from a previous study (14) utilizing the exact same exercise protocol intensity. It should be noted that this correction actually decreases the overall change observed for lactate during exercise and thus compensates for the possibility of overestimating changes in interstitial lactate concentrations.

View larger version (34K): [in this window] [in a new window]   Fig. 3.   Interstitial metabolite changes from baseline resting conditions to after 5 min of exercise and recovery for adenosine (A), lactate (B), and phosphate (C). *P < 0.05, significantly different from baseline; P < 0.05, change from baseline significantly different between 30 and 60% Wmax.

View larger version (22K): [in this window] [in a new window]   Fig. 4.   Interstitial metabolite changes continually recorded each minute during exercise and recovery for K+ (A), pH (B), and H+ (C). 0, Resting baseline. P < 0.05 change of slopes significantly different between exercise at 30% and 60% Wmax.

Correlations between MBV and interstitial metabolites. To further examine the relationship between MBV and interstitial metabolites, correlations from rest to exercise and recovery were examined. A significant correlation was shown between only MBV and interstitial adenosine (r = 0.69, P < 0.05) and K⁺ (r = 0.53, P < 0.05) but not for interstitial lactate or phosphate.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

This study represents the first time that multiple interstitial metabolites (adenosine, lactate, phosphate, K⁺, pH, and H⁺) were examined by microdialysis in conjunction with MBV determinations during dynamic one-legged knee extension exercise in young men and women. The one-legged knee extension exercise model used in this study provides an opportunity to examine the effects that active work have on cardiovascular responses without the confounding influences associated with the engagement of other muscle groups (e.g., hamstring and gluteus) (1). Our MBV findings at rest and during exercise are consistent with others (1, 20) and supports the contention that Doppler ultrasound can be effectively used to examine blood flow changes during exercise. The major findings of this study are that a difference in exercise intensity resulted in differences in MBV and interstitial K⁺ and adenosine. In addition, MBV and interstitial K⁺ and adenosine followed similar temporal patterns during exercise and recovery and were significantly correlated. These findings suggest that the concurrent responses of K⁺ and adenosine may play an active role in the regulation of blood flow during exercise. This discussion section will focus on the design of the study, main findings, and the possible limitations of our results.

Previous studies have investigated potential metabolites involved with exercise vasodilatation and/or stimulation of muscle afferents using techniques such as blood sampling (28), metabolite infusions (30), and/or muscle biopsies to the vastus lateralis (35). Although these studies have provided important data, they did not allow for the direct assessment of metabolite concentrations within the skeletal muscle interstitium. Microdialysis provides a useful tool to directly measure interstitial metabolites; however, to date, only interstitial adenosine (9) has been determined simultaneously with exercise blood flows. In the following paragraphs, we discuss the time course and the workload effects of altering work intensity on various hemodynamic and interstitial measurements.

This study showed a similar temporal pattern of response for MBV and interstitial adenosine during exercise. The relationship between adenosine and work intensity has been shown in other studies (9, 17). In addition, our findings agree with Hellsten et al. (9), who showed a nonlinear increase in interstitial adenosine, which was a function of exercise intensity and muscle blood flow responses. By further examining the association between interstitial adenosine and blood flow, Hellsten et al. (9) also showed a strong correlation coefficient between adenosine and leg blood flow (r = 0.98). In the present report, we also showed a significant but lower correlation for the relationship between adenosine and MBV (r = 0.69) which accounts for only 48% of the variance. We believe our lower correlation is due to two physiological factors: 1) variability of interstitial adenosine seen in the different probes (i.e., different fiber types surrounding the probes), and 2) blood flow variability in the local region of each probe. In addition, the lower correlation may be due to the fact that only two work bouts were used. Nevertheless, we believe that the significant correlation between MBV and adenosine, the similar temporal patterns seen with rest, contraction, and recovery, as well as the effect that varying workloads had on adenosine, suggests that adenosine may play part of a role in exercise hyperemia.

Previous studies (10, 31) have shown that skeletal muscle releases K⁺ continuously during exercise. In this report, the rate of increase in interstitial K⁺ increased in conjunction with an increase in workload. During recovery, interstitial K⁺ returned to baseline at similar rates for the two workloads. This pattern of increase and decrease in interstitial K⁺ has been shown during dynamic exercise in animals (16) and humans (11) as well as during recovery in humans (11). Interstitial K⁺ has also been shown to increase during static exercise paradigms (8). In addition, our study showed a significant correlation between K⁺ and blood flow (r = 0.53), which only accounts for 28% of the variability of flow. Similar factors accounting for this lower correlation have been described previously with adenosine. However, the similar patterns of rise and fall with exercise and recovery (the effect of varying workloads), combined with a significant correlation between K⁺ and MBV, suggest that K⁺, along with other interstitial variables, may also lead to vasodilatation of the skeletal muscle vasculature during exercise.

This is the first study that has examined interstitial lactate, phosphate, and pH with blood flow during dynamic exercise. Our data showed that interstitial lactate increased during exercise similarly regardless of work intensity (30% vs. 60% W_max). Previous animal (16) and human (14) studies have shown that dynamic exercise is associated with increases in interstitial lactate concentrations. In prior reports (14), interstitial lactate concentrations were shown not to change substantially during exercise at 10-30 W, whereas dramatic increases were observed at higher workloads (i.e., 40 and 50 W). In this study, exercise at 30% and 60% W_max represented a power output of 14 ± 1 and 27 ± 3 W, respectively. The modest changes in lactate levels reported in our study support these previous findings. Interestingly, our study showed that after 5 min of recovery, interstitial lactate continued to increase for exercise at 60% W_max, whereas lactate levels decreased slightly after exercise at 30% W_max. This same pattern of lactate following recovery from exercise has been observed after static exercise in animals (16) and humans (15). The higher lactate levels in recovery may be due to the reduction in blood flow leading to a slower lactate removal from the interstitium. Because the pattern of change in interstitial lactate from rest to recovery does not correspond to the temporal pattern changes observed for MBV as well as the lack of a significant correlation between MBV and interstitial lactate, we do not believe that interstitial lactate plays a role in smooth muscle vasodilatation.

Interstitial phosphate levels increased with exercise with similar values regardless of workload; however, during recovery, phosphate levels continued to rise after exercise at 60% W_max. This is a different response than we previously observed following static quadricep work (15, 16). In is unclear why phosphate levels remained elevated after dynamic exercise in our study, whereas, they returned to baseline levels after static exercise. It may be due to the type of exercise intensity, the degree of blood flow change (reduction), as well as muscle washout. In addition, previous studies (19, 34) have shown that ATP is a cotransmitter at nerve endings. Thus higher phosphate levels in recovery may reflect a slower disengagement of sympathetic nervous system during recovery.

This study showed a decrease in H⁺ (e.g., alkalosis) with exercise regardless of intensity with a return to baseline levels in recovery. The alkalotic interstitial response seen with contraction and the subsequent return to baseline values during recovery has been shown in electrically stimulated muscle in cat studies (16) as well as in humans performing static exercise (15). These findings are consistent with the Stewart hypothesis, which suggests that within a given space, ionic neutrality must be preserved. In other words, H⁺ is a dependent variable and its concentration within a given space, such as the skeletal muscle interstitium, depends on the relative concentration of strong cations such as K⁺ and anions such as lactate and phosphate (33). Thus if the change from rest to exercise for cations such as K⁺ (0.88-1.61 meq/l) is greater than for anions such as lactate (0.32-0.37 mM) and phosphate (0.14-0.28 mM), the H⁺ concentration will be expected to fall to maintain electrical neutrality. Likewise, in recovery, if the lactate and phosphate continued to increase, whereas K⁺ levels fell, one would expect H⁺ concentration to increase.

There are several limitations of this study that should be discussed. First, we were unable to examine the relationship between muscle metabolites and the muscle reflex. This is because the BP response to rhythmic quadriceps contraction is influenced not only by the muscle reflex engagement, but also by the degree of skeletal muscle vasodilation (12). Second, because we did not inflate a cuff just below the knee to supersystolic pressures of ~200 mmHg, it is possible that the blood flow determinations may have been affected by contributions from the lower leg. Our rationale for not using the cuff was that we believed it unlikely that lower limb flow would change as a function of workload and recovery. Moreover, because cuff inflation could not be maintained for the entire study, we would need to inflate the cuff during exercise. Cuff inflation during exercise would reduce leg flow. We were concerned that we would be unable to discriminate between a change in Doppler probe placement (i.e., relative to the femoral artery) and the physiologic "true" fall in flow associated with cuff inflation.

In conclusion, this study represents the first time that multiple interstitial metabolites as well as femoral artery MBV were examined during one-legged knee extension exercise. The major findings of this study were that we observed similar workload and response patterns as well as positive correlations between MBV and interstitial K⁺ and adenosine. Thus the concurrent responses of interstitial K⁺ and adenosine with MBV suggest that K⁺ and/or adenosine may play a role in the regulation of blood flow during exercise.