Corresponding author E. A. Martin: Department of Anesthesiology, Mayo Clinic, 200 First Street SW, Rochester, MN 55905, USA. Email: email@example.com
We previously demonstrated a bimodal distribution of forearm vasodilator responsiveness to adenosine (ADO) infusion in the brachial arteries of human subjects. We also demonstrated that ADO receptor antagonism blunted exercise hyperaemia during heavy rhythmic handgripping, but vasodilator responses to exogenous ADO were only blunted in ADO responders. In this study, we continued investigating the contribution of ADO to exercise hyperaemia and possible differences between responders and non-responders. We hypothesized that ADO transporter antagonism would increase vasodilatation in response to exogenous ADO in responders only, but not effect exercise-mediated vasodilation. To test this hypothesis, we compared forearm vascular conductance (FVC) during infusion of ADO to FVC during handgripping before and after infusion of dipyridamole (DIP) in 20 subjects. In ADO responders, change in FVC above baseline (ml min−1 (100 mmHg)−1) for low, medium and high doses of ADO, respectively, was 58 ± 8, 121 ± 22 and 184 ± 38, and after DIP was 192 ± 32, 238 ± 50 and 310 ± 79. For non-responders, these values were 23 ± 2, 43 ± 5 and 66 ± 9, respectively, before DIP (P < 0.01 versus responders). Contrary to our hypothesis, these values were increased by DIP in non-responders (P < 0.001) and therefore not different from responders (P > 0.20). We found that ADO transporter blockade had no effect on exercise hyperaemia in either subgroup. We conclude that there may be increased ADO transporter activity in non-responders resulting in reduced ADO-mediated vasodilatation. The failure of DIP to augment exercise hyperemia under these conditions suggests that ADO concentrations may not rise enough during rhythmic handgripping to have a major impact on these responses.
Metabolites, which are produced during muscle contraction and exert local vasodilator actions in the muscle vasculature, are thought to play a major role in exercise hyperaemia, although the main contributors and their interactions remain uncertain (Shepherd, 1983; Rowell et al. 1996; Joyner & Proctor, 1999). Adenosine (ADO) may be involved in exercise hyperaemia, although the specific role of ADO in exercise hyperaemia has been debated for several decades.
The ‘adenosine hypothesis’, which suggested that coronary blood flow is controlled by interstitial ADO concentration, was first developed by Robert Berne and colleagues in the early 1960s (Berne, 1963). This hypothesis quickly inspired several studies investigating the role of ADO in the control of blood flow during mismatches between oxygen supply and demand in other vascular beds, including skeletal muscle, brain and kidney. Several studies in animals have failed to demonstrate a significant reduction of exercise hyperaemia during ADO receptor antagonism (Honig & Frierson, 1980; Koch et al. 1990). However, other animal (Tabaie et al. 1977; Metting et al. 1986; Poucher et al. 1990; Persson et al. 1991) and human studies (Radegran & Calbet, 2001) have demonstrated significant reduction of exercise hyperaemia with blockade of ADO receptors.
In previous studies, we investigated the contribution of ADO to exercise hyperemia, as suggested by the adenosine hypothesis, by comparing vasodilator responsiveness to brachial arterial infusion of ADO and handgrip exercise (Martin et al. 2006a,b). In these previous studies, we were surprised to find a consistent and repeatable bimodal distribution of vasodilator responsiveness to intra-arterial infusion of ADO, among human subjects. In these studies, one subgroup of subjects demonstrated significant, dose-dependent and repeatable increases in forearm vascular conductance (FVC) during ADO infusion. This subgroup was categorized as ‘ADO responders’. The remaining subjects demonstrated dramatically blunted vasodilator responsiveness to ADO, in comparison to exercise hyperaemia, and were categorized as ‘ADO non-responders’. We expected to find a difference in the contribution of ADO to exercise hyperaemia between these two subgroups of subjects, but failed to do so. Because there was no difference in exercise hyperaemia between the two subgroups, ADO non-responders were defined as subjects who demonstrated twice the increase in FVC during handgrip exercise, as compared to the increase in FVC during ADO infusion.
Further investigation into mechanisms of exercise hypereamia, as well as the possible differences between ADO responders and non-responders, demonstrated that blockade of nitric oxide synthase (NOS) with NG-monomethyl-l-arginine (l-NMMA) resulted in significant blunting of ADO-mediated vasodilatation only in ADO responders and had no effect on exercise hyperaemia in either subgroup. Furthermore, ADO receptor antagonism blunted ADO vasodilator responses only in ADO responders, and blunted exercise hyperaemia by approximately 15% only at high workloads in both subgroups (Martin et al. 2006b). These data suggest that: (1) a portion of the difference in ADO-mediated vasodilator responsiveness between ADO responders and non-responders may be due to a difference in a nitric oxide component of ADO-mediated vasodilatation, predominantly through a cyclic GMP-mediated pathway; (2) there may be reduced ADO receptor function, density or sensitivity in the ADO non-responders; and (3) ADO may play a minimal role in exercise hyperaemia during rhythmic handgripping at submaximal workloads.
Although these previous findings suggest that differences in the ADO receptor may account for demonstrated differences between ADO responders and non-responders and that the contribution of ADO to exercise hyperaemia may be minimal, especially in non-responders, we could not exclude possible differences in ADO transporter structure, function or density between the ADO responders and non-responders. Differences in the ADO transporter function might manifest as differences in vasodilator responsiveness to ADO infusion and/or reduced contribution of ADO to exercise hyperaemia in the ADO non-responders.
The primary source of ADO in the skeletal muscle interstitium is from the breakdown of AMP by the ecto-form of 5′-nucleotidase, which is bound to the cell membrane by a glycosyl-phospatidylinositol (GPI) anchor (Misumi et al. 1990; Hellsten, 1999). ADO is transported across the cell membrane via equilibrative nucleoside transporters. Blockade of these ADO transporters with infusion of dipyridamole (DIP) results in an increased interstitial concentration and reduced metabolism of ADO. Therefore, more ADO is available to bind to ADO (purinergic P1) receptors on skeletal muscle cells, vascular endothelial cells and vascular smooth muscle cells (Lynge & Hellsten, 2000). Binding of ADO to two of the four ADO receptor subtypes (A2A and A2B) on vascular smooth muscle cells results in vasodilatation.
In a limited number of human studies, intra-arterial infusion of DIP, by increasing the interstitial concentration of ADO, augmented blood flow before and during ADO infusion (van Ginneken et al. 2002; Gamboa et al. 2003; Bijlstra et al. 2004). Furthermore, although DIP infusion augmented exercise hyperaemia at maximal workloads in exercising swine (Laughlin et al. 1989), DIP did not augment exercise hyperaemia at submaximal workloads in dog gracilis muscle (Honig & Frierson, 1980; Klabunde, 1986). To our knowledge, the effect of DIP on skeletal muscle vasodilatation during exercise hyperemia in humans has not been investigated.
Because the amount of ADO available in the interstitium (which can be increased by DIP infusion) may affect the binding of ADO to vasoditatory ADO receptors, the purpose of the present study was two-fold. We wanted to investigate whether differences in ADO transporters may contribute to different (1) ADO vasodilator responsiveness and (2) ADO contribution to exercise hyperemia, between ADO responders and ADO non-responders. We therefore tested whether ADO transporter antagonism with DIP would differentially affect vasodilator responsiveness to ADO or exercise in ADO responders versus non-responders. First, we hypothesized that there would be a bimodal distribution of vasodilator responsiveness to ADO among our subjects, as before. Second, we hypothesized that exercise responses would be similar among all subjects, and that ADO transporter antagonism would not augment exercise hyperaemia in either group. Finally, in our previous study, ADO receptor antagonism with aminophylline blunted the responsiveness to ADO infusion only in the ADO responders, suggesting that ADO non-responders have reduced ADO receptor function (Martin et al. 2006b). Therefore, we hypothesized that DIP would augment vasodilator responses to ADO only in ADO responders.
All protocols and procedures were approved by the Institutional Review Board at Mayo Clinic and were performed according to the Declaration of Helsinki. Twenty young healthy subjects (nine males and 11 females) participated in the study after giving written informed consent. All subjects were moderately active, non-smokers, non-obese, normotensive and not taking any medications other than oral contraceptives. Female subjects were not pregnant, as determined by a pregnancy test less than 24 h prior to the study. Female subjects were studied during the placebo phase of oral contraceptive use, or in the early follicular phase of their menstrual cycle, to minimize possible confounding influences of reproductive hormones on control of blood flow (Minson et al. 2000; Charkoudian, 2001). The subjects fasted overnight and refrained from intake of caffeine for 48 h prior to the study, and intake of alcohol or exercise for 24 h prior to the study.
The brachial artery of the non-dominant forearm was catheterized under aseptic conditions after local anaesthesia (1% lignocaine (lidocaine)). A standard 5 cm, 20 gauge Teflon catheter was inserted and connected to a three-port connector system in series. One port was connected to a pressure transducer for continuous measurements of arterial pressure and was continuously flushed with 3 ml h−1 heparinized saline. The other two ports allowed for local intra-arterial administration of study drugs (Dietz et al. 1994).
Forearm blood flow and vascular conductance
Brachial artery mean blood velocity (MBV) and brachial artery diameter were monitored with a 12 MHz linear-array Doppler probe (Model M12L, Vivid 7, General Electric, Chalfont St Giles, UK) as previously described (Wilkins et al. 2006). The probe insonation angle was 60 deg. Brachial artery diameter measurements were taken at end diastole during baseline and during the last 45 s of each drug infusion dose or exercise workload and used to calculate brachial artery cross-sectional area (π(diameter/2)2). Forearm blood flow (FBF) was calculated as the product of MBV (cm s−1) and brachial artery cross-sectional area (cm2) and multiplied by 60 to present values as ml min−1. Forearm vascular conductance (FVC) was then calculated as FBF/mean arterial pressure (MAP) × 100, and expressed as ml min−1 (100 mmHg)−1.
Rhythmic handgrip exercise
Rhythmic forearm handgrip exercise was performed at three different workloads using a 2.3, 4.6 or 6.9 kg weight for females or a 3.2, 6.4 or 9.6 kg weight for males. The weight was lifted 4–5 cm over a pulley at a duty cycle of 1 s contraction–2 s relaxation (20 contractions min−1) using a metronome to ensure the correct timing. These workloads averaged 7.4 ± 0.3%, 14.9 ± 0.6% and 22.3 ± 0.9% of maximum voluntary contraction (MVC) for the low, medium and high workloads, respectively, for all subjects (P > 0.80 between ADO responders and non-responders and between males and females).
Brachial artery infusions
Drugs were infused on the basis of forearm volume (FAV, water displacement) via the brachial artery catheter. The absolute infusion rate was less than 3 ml min−1 in every trial.
ADO was infused at 3.125, 6.25 and 12.5 μg min−1 (dl FAV)−1. ADO doses and exercise workloads were chosen according to previous studies (Dinenno & Joyner, 2003; Martin et al. 2006a,b). Increases in FVC during infusion of the low, medium and high ADO doses were previously shown to be similar to the increases in FVC during the low, medium and high exercise workloads, respectively, in ADO responders. However, in ADO non-responders, the FVC during ADO infusion was previously shown to be blunted in comparison to FVC during handgripping.
DIP was infused at 20 μg min−1 (dl FAV)−1 with the goal of blocking equilibrative nucleoside transport across cell membranes. Blockade of ADO uptake into skeletal muscle cells and subsequent ADO metabolism results in increased interstitial concentrations of ADO. This leads to increased binding of ADO to ADO receptors and vasodilatation. The DIP dose was chosen after review of the literature, which demonstrated that infusion of DIP into the forearm inhibited nucleoside transport with subsequent increased levels of ADO in skeletal muscle interstitium (van Ginneken et al. 2002; Gamboa et al. 2003; Bijlstra et al. 2004). The final dose of DIP used in this study was based on the forearm distribution and the total infusion duration. These considerations allowed for blockade of nucleoside transport throughout the study, while limiting inherent vasodilator and systemic effects of a continuous DIP infusion.
Figure 1 shows the experimental timeline. Each trial consisted of 1 min of baseline measurement with saline infusion, followed either by sequential intra-arterial infusions of low, medium and high doses of ADO or by rhythmic handgripping at low, medium and high workloads (4 min at each dose or workload). The order of the trials was randomized among subjects. We have previously shown that both exercise and ADO trials demonstrate repeatable vasodilator responsiveness (Martin et al. 2006a,b). After one ADO dose–response trial and one exercise workload–response trial were performed, DIP was infused for 20 min at rest. Infusion of DIP continued for the remainder of the study, which consisted of one more ADO trial and one more exercise trial. Each trial was separated by 20 min of rest to allow FBF to return to baseline values.
Data acquisition and analysis
Data were collected and stored on computer at 250 Hz and analysed off-line with signal processing software (WinDaq, DATAQ Instruments, Akron, OH, USA). MAP was determined from the arterial pressure waveform. Baseline FBF and MAP represent averages taken during the minute of baseline, and the hyperemic values represent 30 s averages taken during the last minute of each dose or workload.
All values are reported as means ±s.e.m. Subject demographics were compared using a rank-sum test, and gender was compared using Fisher's exact test. Repeated measures analysis of variance (ANOVA) was used to assess differences between treatment groups and levels. Two-sample t test was used to compare pairs of group means. When significance was detected, Tukey's post hoc test was used to identify individual differences and to adjust P values to account for multiple comparisons, in order to preserve an overall type-1 error rate of 0.05. Significance was set at P < 0.05.
Estimating from our previous studies that the average standard deviation of FVC during each trial is 50 ml−1 min−1 (100 mmHg)−1 and the correlation of repeated measurements is 0.80, a sample size of 20 will provide statistical power of 95% to detect a main effect of drug or exercise. We studied 20 subjects in the current study. Subjects were not recruited as ADO responders or non-responders, but were categorized as ADO responders or non-responders following the first ADO and first exercise trial. Exactly half of our subjects were categorized as ADO responders and the other half were non-responders, and our calculated statistical power was 80%. Although subjects were not recruited as ADO responders and non-responders, the finding that 50% of subjects studied in the current study were ADO responders and the other 50% were non-responders is in agreement with our previous studies, in which roughly half of subjects were ADO responders (n= 35) and the rest were non-responders (n= 31). Based on criteria outlined below and the approach used in our previous studies (Martin et al. 2006a,b), subjects demonstrating robust vasodilator responses to both ADO infusion and handgripping were identified as ADO responders and those that had blunted vasodilator responses to ADO infusion were identified as ADO non-responders. Specifically, ADO non-responders were defined as subjects in whom the FVC during each exercise workload was more than twice the FVC during the respective ADO dose (for the low, medium and high doses or workloads, respectively).
Mean age, body mass index (BMI) and FAV were 26 ± 1 years, 24 ± 0.6 kg m−2 and 910 ± 50 ml, respectively, for all subjects. Group mean age, BMI and FAV for ADO responders (n= 10; five males and five females) were 26 ± 2 years, 24 ± 1.0 kg m−2 and 920 ± 87 ml and for non-responders (n= 10; four males and six females) were 27 ± 1 years, 23 ± 0.7 kg m−2, and 900 ± 58 ml. There was no statistical difference in subject demographics, including caffeine use and exercise training, between ADO responders and non-responders (P > 0.05).
Baseline FVC and MAP were significantly influenced by DIP infusion (P < 0.05) in both responders and non-responders; however, the effects of DIP did not differ between ADO responders and non-responders (P > 0.05). Baseline FVC values (ml min−1 (100 mmHg)−1) for the first ADO trial (ADO) and first exercise trial (exercise) were 41 ± 7 and 45 ± 9 for ADO responders and 36 ± 4 and 29 ± 3 for the non-responders, respectively. After ADO transporter antagonism for the ADO trial (ADO + DIP) and exercise trial (exercise + DIP), these values (ml min−1 (100 mmHg)−1) were 87 ± 28 and 89 ± 25 for ADO responders and 44 ± 4 and 62 ± 10 for the non-responders, respectively (P < 0.05 for ADO versus ADO + DIP, and exercise versus exercise + DIP for both ADO responders and non-responders. For ADO responders versus non-responders, the P values were 0.56, 0.14, 0.17 and 0.33 for ADO, exercise, ADO + DIP and exercise + DIP, respectively).
Because of large intersubject variability, the means were not different between responders and non-responders. Baseline MAP (mmHg) was 92 ± 2, 95 ± 2, 99 ± 3 and 99 ± 2 during the ADO, exercise, ADO + DIP and exercise + DIP trials, respectively (P < 0.05 for ADO versus ADO + DIP, and exercise versus exercise + DIP). DIP-induced increases in baseline FBF and MAP have been previously described (van Ginneken et al. 2002; Gamboa et al. 2003; Bijlstra et al. 2004). Because of the increase in baseline FVC with DIP infusion, we have reported values as change in FVC above baseline (ml min−1 (100 mmHg)−1).
Adenosine versus exercise vasodilator response
Ten subjects showed robust vasodilator responses to both ADO infusion and handgrip exercise, and were, hence, categorized as ADO responders. The other 10 subjects showed blunted vasodilator responses to ADO infusion in comparison to vasodilator responses to handgrip exercise (FVC during each exercise workload was more than twice the FVC during the corresponding dose of ADO), and were, hence, categorized as ADO non-responders. As seen in Fig. 2, for the low, medium and high doses of ADO, respectively, the change in FVC above baseline (ml min−1 (100 mmHg)−1) before DIP infusion for the 10 ADO responders was 58 ± 8, 121 ± 22 and 184 ± 38, and for the 10 non-responders was 23 ± 2, 43 ± 5 and 66 ± 9 (P < 0.01 at each dose for ADO responders versus non-responders).
As seen in Fig. 3, for the low, medium and high handgrip exercise workloads, respectively, the FVC above baseline (ml min−1 (100 mmHg)−1) before DIP infusion for the ADO responders was 87 ± 13, 162 ± 28 and 213 ± 31, and for the non-responders was 84 ± 11, 166 ± 16 and 228 ± 20 (P > 0.70 at each workload for ADO responders versus non-responders).
Effect of DIP infusion on ADO responsiveness
As seen in Fig. 2, ADO transporter blockade significantly augmented ADO responsiveness at each dose in both ADO responders and non-responders. For the low, medium and high ADO doses, respectively, the FVC above baseline (ml min−1 (100 mmHg)−1) after DIP infusion was 192 ± 32, 238 ± 50 and 310 ± 79 for the ADO responders, and 145 ± 20, 227 ± 34 and 302 ± 46 for the non-responders. Blockade of the ADO transporter resulted in significant augmentation of FVC values at each dose in both responders and non-responders (P < 0.02 for ADO versus ADO + DIP in responders and P < 0.001 for ADO versus ADO + DIP in non-responders). Additionally, the change in FVC values after DIP infusion was not significantly different between responders and non-responders (P > 0.20 at each dose of ADO + DIP for ADO responders versus non-responders), which suggests that DIP had the effect of making vasodilator responses to ADO similar in both subject groups.
Effect of DIP infusion on exercise hyperaemia
As seen in Fig. 3, ADO transporter blockade did not significantly increase exercise hyperaemia in either ADO responders or non-responders (P > 0.30 exercise versus exercise + DIP for both ADO responders and non-responders). For the low, medium and high handgrip exercise workloads, respectively, the change in FVC above baseline (ml min−1 (100 mmHg)−1) after DIP infusion was 79 ± 13, 162 ± 23 and 217 ± 22 for the ADO responders, and 89 ± 9, 162 ± 14 and 243 ± 23 for the non-responders. The change in FVC values at each workload was not significantly different between responders and non-responders (P > 0.40 at each workload for ADO responders versus non-responders).
There were three main findings of the present study. First, as in our previous studies, there was a bimodal distribution of ADO vasodilator responsiveness among subjects, despite similar responses to rhythmic handgrip exercise. Second, ADO transporter antagonism did not augment exercise hyperemia in either ADO responders or non-responders. Finally, ADO transporter antagonism significantly augmented ADO-mediated vasodilatation in both ADO responders and non-responders. This effect was more dramatic in ADO non-responders, such that the difference in ADO-mediated vasodilatation between ADO responders and non-responders was not present after DIP infusion. Together, these findings suggest that rhythmic handgrip exercise may not be associated with a functionally significant increase in ADO in the interstitial space of the exercising skeletal muscle, and that ADO transporter activity may normally be elevated in ADO non-responders (compared with ADO responders), with subsequent increased clearance of ADO from the interstitial space.
Our finding of a lack of effect of DIP infusion on exercise hyperaemia partially contrasts with data from a previous study in which DIP infusion augmented exercise hyperaemia in exercising swine (Laughlin et al. 1989), but is consistent with findings from previous studies in dog gracilis muscle, in which DIP failed to increase muscle blood flow during submaximal contraction (Honig & Frierson, 1980; Klabunde, 1986). In the study by Laughlin et al. (Laughlin et al. 1989), DIP augmented muscle blood flow in all muscles tested (cardiac, respiratory and all limb muscles) during maximal exercise in swine, but only in certain muscles (cardiac, respiratory and slow-twitch limb muscles) during submaximal exercise. This suggests that DIP may have differential effects on vascular beds of different muscle fibre types and at different workloads. To our knowledge, our present study is the first in humans to investigate the effect of ADO transporter antagonism on exercise hyperaemia.
The failure of DIP infusion to augment exercise hyperaemia in the present study contradicted our hypothesis, but is consistent with the idea that ADO may not be obligatory for exercise hyperaemia during submaximal exercise in humans. Findings from our previous study demonstrated that non-specific ADO receptor antagonism failed to blunt exercise hyperaemia at submaximal workloads; however, it did blunt exercise hyperaemia by 15% at the highest workload in all subjects tested (Martin et al. 2006b). This suggested that ADO may contribute to exercise hyperaemia at high workloads when there might be a greater mismatch between oxygen supply and oxygen demand and skeletal muscle is relatively ‘under-perfused’ (Sparks, 1980). Although in the current study, DIP infusion failed to increase exercise hyperaemia, this does not suggest that ADO does not contribute to exercise hyperaemia, as suggested by the adenosine hypothesis, but simply that it may not be obligatory. Therefore, both sets of data support the idea that redundant vasodilator mechanisms contribute to exercise hyperaemia, and that no one vasodilator may be responsible or obligatory for exercise hyperaemia in humans (Delp & Laughlin, 1998; Saltin et al. 1998; Laughlin & Korzick, 2001; Clifford & Hellsten, 2004).
Our finding of a bimodal distribution of vasodilator responsiveness to ADO is consistent with and strongly supports findings from our previous studies (Martin et al. 2006a,b), in which 19 subjects using various protocols demonstrated robust vasodilator responsiveness to ADO infusion and 16 subjects demonstrated blunted ADO responsiveness. As in earlier studies, which used the same criteria for categorizing ADO responders and non-responders, both subgroups in the present study showed similar vasodilator responses to handgripping.
Our finding that DIP infusion augmented ADO-mediated vasodilatation is consistent with studies by both Gamboa et al. (2003) and Bijlstra et al. (2004), in which DIP infusion augmented forearm vasodilatation induced by either low or high ADO doses (0.6 or 6.0 nmol min−1(dl FAV)−1 and 0.125 or 0.5 mg min−1, respectively). Our findings are only partially consistent with the study by van Ginneken et al. (2002) in which DIP augmented vasodilatation induced by a low dose of ADO (0.6 nmol min−1(dl FAV)−1), but not at a higher dose of ADO (6.0 nmol min−1(dl FAV)−1). However, none of the mentioned studies reported on differences in ADO vasodilator responsiveness among subjects, which could contribute to these discrepant findings.
Our results demonstrate that DIP blocks ADO transporters in both ADO responders and non-responders. However, the observed effect of DIP was more dramatic in ADO non-responders whether expressing the data as absolute FVC values or change in FVC above baseline values; the increase in FVC above baseline during infusion of ADO + DIP was large enough that at each dose of ADO + DIP there was no difference between ADO responders and non-responders. In terms of ADO vasodilator responsiveness, DIP has the effect of eliminating the differences between ADO responders and non-responders that are present during ADO infusion alone. This suggests that ADO non-responders may have increased activity of ADO transporters, resulting in increased uptake and clearance of ADO from the interstitial space, which may be eradicated in the presence of the ADO transporter antagonist DIP, in these subjects. Thus, in the non-responders, increased activity of ADO transporters may limit the exposure of ADO receptors to stimulation with endogenous ADO, and therefore may explain the augmented vasodilator responses when more ADO was available in the interstitial space.
Based on our previous results (Martin et al. 2006b), we hypothesized in the present study that ADO non-responders would have blunted ADO vasodilator responsiveness, even in the presence of ADO transporter antagonism, due to decreased density or affinity of the ADO receptor, because blockade of the ADO receptor with aminophylline did not alter ADO-mediated vasodilatation in ADO non-responders, but did block ADO-mediated vasodilatation in the ADO responders. In the current study, ADO non-responders had robust vasodilator responses to ADO during DIP infusion that were not different to responses in ADO responders. Therefore, the present data suggest that ADO non-responders may have normal ADO receptor function but a reduced availability of ADO to bind to ADO receptors, possibly due to increased uptake and metabolism of ADO. Furthermore, we hypothesized that ADO responders may have augmented exercise hyperemia during ADO transporter antagonism, which contrasts with data from the current study.
We chose ADO doses and workloads from previous studies in our laboratory (Martin et al. 2006a,b), which resulted in identification of ADO responders and non-responders. In our previous and ongoing studies, the use of these ADO doses has consistently resulted in a bimodal distribution of ADO vasodilator responsiveness in subjects, while the exercise workloads chosen consistently result in similar vasodilator responses in all subjects studied. We have used the same criteria for categorizing ADO responders and non-responders in each study. Subjects demonstrating twice the increase in FVC during low, medium and high exercise workloads compared to the FVC during the low, medium and high ADO infusion doses, respectively, are categorized as ADO non-responders. Using this criterion we have found that of 66 subjects studied in various protocols, approximately half (n= 35) have been ADO responders and the other half (n= 31) have been non-responders (Martin et al. 2006a,b).
Although this is an indirect method of investigating the role of ADO in exercise hyperaemia, we believe it sheds new light on a controversial topic that has resulted in conflicting results and conclusions over several decades. By categorizing our subjects into ADO responders and non-responders, rather than pooling the data from both groups, this approach may help explain previously conflicting data regarding the role of ADO in exercise hyperaemia in humans.
The DIP dose was chosen on the basis of recent studies in the human forearm (van Ginneken et al. 2002; Gamboa et al. 2003; Bijlstra et al. 2004), with the goal of completely inhibiting ADO transporters. DIP has previously been shown to alter baseline FVC, probably because of its effects on the interstitial concentration of ADO through ADO transport inhibition, as well as non-specific effects of DIP, including induction of release of prostacyclin (Blass et al. 1980), inhibition of phosphodiesterases and inhibition of platelet aggregation (McElroy & Philip, 1975). These non-specific effects of DIP, which result in increases in either intracellular cAMP or cGMP (McElroy & Philip, 1975; Blass et al. 1980), could have resulted in the observed increases in baseline FVC.
In this study, the change in baseline FVC after infusion of DIP exhibited large inter-individual variability, such that the standard deviation was similar to the mean. Therefore, there was no significant difference between this value in the responders versus the non-responders. The P values for the comparison between baseline FVC values (ml min−1 (100 mmHg)−1) values in responders were as follows: (1) ADO, P= 0.56 for responders (41 ± 7) versus non-responders (36 ± 4); (2) exercise, P= 0.14 for responders (45 ± 9) versus non-responders (29 ± 3); (3) ADO + DIP, P= 0.17 for responders (87 ± 28) versus non-responders (44 ± 4); and (4) exercise + DIP, P= 0.33 for responders (89 ± 25) versus non-responders (62 ± 10). A time effect of the trials was eliminated by randomizing the order of trials with regard to adenosine and exercise in order to avoid this potential problem. Regardless, data in the present study were presented as change in FVC above baseline to account for this baseline shift.
It could be argued in our previous studies (Martin et al. 2006a,b)that delivery of exogenously infused ADO into the interstitial space may have been limited by a significant endothelial barrier to exogenous ADO. In the previous studies, only ADO responders demonstrated increases in FVC during ADO infusion. Although both ADO responders and non-responders demonstrated increased baseline FVC values following non-specific ADO receptor blockade with aminophylline infusion, only ADO responders demonstrated blunting of ADO-mediated vasodilatation following aminophylline infusion. This may suggest that the infused ADO increased FVC in ADO responders mostly by stimulating ADO receptors located on the endothelial cells and not the abluminal receptors located on the vascular smooth muscle cells. This may also suggest that the receptors on vascular smooth muscle cells are most probably stimulated by endogenous release of ADO during muscle contraction, because of lack of delivery of the exogenous ADO into the interstitial space. However, both ADO responders and non-responders demonstrated increased baseline FVC values following DIP infusion (increases in baseline FVC was not significantly different between responders and non-responders during each trial, despite large inter-subject variability) and dose-dependent increases in FVC during ADO + DIP infusion. This suggests that there were adequate amounts of endogenous ADO in the interstitial space during DIP infusion, and that there was adequate delivery of exogenous ADO into the interstitial space during ADO + DIP infusion.
Finally, our subjects refrained from caffeine use during the 48 h prior to the study. This was done to prevent confounding effects of methylxanthine antagonism of ADO receptors during the study. We also considered possible differences in chronic caffeine use among subjects, which could contribute to variable responses to ADO, but after querying all subjects about caffeine use, we found no difference between ADO responders and non-responders, eliminating the possibility that sensitivity or tolerance to ADO infusion could be related to excessive caffeine use in either subgroup of subjects.
In summary, we investigated the potential role of ADO in exercise hyperaemia by comparing ADO-mediated vasodilatation to that during voluntary forearm contractions in human subjects before and after ADO transporter antagonism with DIP. We also investigated possible differences between ADO responders and non-responders. In agreement with our two previously published studies (Martin et al. 2006a,b), we report a bimodal distribution of ADO vasodilator responsiveness in the human forearm. This difference in ADO-mediated vasodilatation between ADO responders and non-responders was eradicated following ADO transporter antagonism with DIP infusion. This observation suggests that differences in function or density of ADO transporters may contribute to differences in ADO responsiveness between ADO responders and non-responders. Furthermore, exercise hyperaemia was not augmented by DIP in either subgroup, suggesting that ADO may not be obligatory for exercise hyperaemia at these workloads, due perhaps to ongoing contributions of redundant vasodilator pathways.
We wish to thank the Mayo Clinic Human Integrative Physiology Laboratory, especially Dr Niki Dietz, Branton G. Walker, Lakshmi P. (Madhuri) Somaraju, Maile L. Ceridon, Christopher P. Johnson, Shelly K. Roberts, Pamela Engrav and Karen P. Krucker for their contributions, and the subjects who volunteered for this study. This research was supported by NIH grants HL-46493 (M.J.J.), GM-08685 (W.T.N.) and RR-172520 (J.H.E.) and NIH General Research Center Grant RR-00585 (to the Mayo Clinic, Rochester, MN, USA).