α1-Adrenergic receptor control of the cerebral vasculature in humans at rest and during exercise

Authors

  • Sushmita Purkayastha,

    1. Institute For Aging Research, Hebrew Senior Life, Boston, MA, USA
    2. Harvard Medical School, Boston, MA, USA
    Search for more papers by this author
  • Ashwini Saxena,

    1. Department of Integrative Physiology and the Cardiovascular Research Institute, University of North Texas Health Science Center, Fort Worth, TX, USA
    Search for more papers by this author
  • Wendy L. Eubank,

    1. Department of Integrative Physiology and the Cardiovascular Research Institute, University of North Texas Health Science Center, Fort Worth, TX, USA
    Search for more papers by this author
  • Besim Hoxha,

    1. Department of Integrative Physiology and the Cardiovascular Research Institute, University of North Texas Health Science Center, Fort Worth, TX, USA
    Search for more papers by this author
  • Peter B. Raven

    1. Department of Integrative Physiology and the Cardiovascular Research Institute, University of North Texas Health Science Center, Fort Worth, TX, USA
    Search for more papers by this author

S. Purkayastha: Institute for Aging Research, Hebrew Senior Life, Boston, MA 02131, USA. Email: sushmitapurkayastha@hsl.harvard.edu

New findings

  • • What is the central question of this study?Despite the abundance of sympathetic nerve fibres emanating from the cervical and stellate ganglia that innervate the cerebral arteries, the role of the sympathetic nervous system in regulation of cerebral vasculature in humans remains equivocal.
  • • What is the main finding and its importance?The findings from this study support the role of the sympathetic nervous system, mediated by activation of α1-adrenoreceptors, in dynamic cerebral autoregulation and cerebral vascular tone at rest and during moderate dynamic exercise. Blockade of the α1-adrenoreceptors impaired dynamic cerebral autoregulation and attenuated any increases in cerebral vascular tone during moderate dynamic exercise in healthy humans.

We tested the hypothesis that pharmacological blockade of α1-adrenoreceptors (by prazosin), at rest and during steady-state dynamic exercise, would impair cerebral autoregulation and result in cerebral vasodilatation in healthy humans. In 10 subjects, beat-to-beat mean arterial pressure and mean middle cerebral artery blood velocity were determined at rest and during low (Ex90) and moderate workload (Ex130) on an upright bicycle ergometer without and with prazosin. Plasma noradrenaline concentrations increased significantly from rest to Ex130 during control conditions (from 1.8 ± 0.2 to 3.2 ± 0.3 pmol (ml plasma)−1). In the control conditions, the transfer function gain between mean arterial pressure and mean middle cerebral artery blood velocity in the low-frequency range was decreased at Ex90 (P= 0.035) and Ex130 (P= 0.027) from rest. A significant increase in critical closing pressure (CCP) was also observed in the control conditions from rest to Ex90 to Ex130 (from 18 ± 3 to 24 ± 4 to 31 ± 4 mmHg). An average of 74 ± 2% blockade of blood pressure response was achieved with oral prazosin. Following blockade, plasma noradrenaline concentrations further increased at rest and during Ex130 from the control value (from 2.6 ± 0.3 to 4.4 ± 0.5 pmol (ml plasma)−1). Prazosin also resulted in an increase in low-frequency gain (P < 0.003) compared with the control conditions. Prazosin blockade abolished the increases in CCP during Ex130 and increased the cerebrovascular conductance index (P= 0.018). These data indicate that in the control conditions a strengthening of cerebral autoregulation occurred with moderate dynamic exercise that is associated with an increase in CCP as a result of the exercise-mediated augmentation of sympathetic activity. Given that α1-adrenergic receptor blockade attenuated the increase in dynamic cerebral autoregulation and CCP, we conclude that increases in sympathetic activity have a role in establishing cerebral vascular tone in humans.

Despite the abundance of sympathetic nerve fibres emanating from the cervical and stellate ganglia that innervate the cerebral arteries (Edvinsson, 1975, 1982; Heistad et al. 1978; Hamel, 2006), the role of the sympathetic nervous system in regulation of cerebral vasculature in humans remains equivocal. The neurovascular coupling of the perivascular neurons, astrocytes, endothelial cells and smooth muscle cells is implicated in the regulation of cerebral blood flow (CBF; Iadecola, 2004). Previous studies using isolated human pial arteries have demonstrated weak neurogenic responses to noradrenaline (NA; Bevan et al. 1998). In addition, a study using electromagnetic flowmeters was unable to identify pressure-independent increases in intracarotid artery blood flow with intracarotid artery infusions of NA (Greenfield & Tindall, 1968). However, a recent investigation using the tritiated NA spillover technique identified the presence of functional α1-adrenergic receptors on cerebral microvasculature outside the blood–brain barrier (BBB) and the presence of tritiated NA metabolites of brain neurogenic origin in the internal jugular vein of humans (Mitchell et al. 2009). These findings indicate that infused NA does cross into the brain from the cerebral blood vessels. Investigations into the impermeability of the BBB indicate that although the capillaries have impenetrable tight junctions, the endothelium of arterioles and venules of the cerebral microvasculature are subject to greater modulation (Abbott et al. 2006). Furthermore, reductions in the tightness of the BBB are associated with increases in cerebral perfusion pressure (Bill & Linder, 1976), free radicals (Abbott, 2005) and interleukin-6 concentrations (Mayhan, 2001). All these modulators of BBB impermeability are present during dynamic exercise (Pedersen & Febbraio, 2008). In addition, vasoactive substances, such as bradykinin and nitric oxide, appear to increase the permeability of the BBB via activation of second-messenger pathways (Mayhan, 2001). More recently, high-intensity exercise has been found to increase the permeability of the BBB without causing structural damage (Bailey et al. 2011).

Findings independent of dynamic exercise indicated the involvement of α1-adrenoreceptors in cerebrovascular blood flow regulation during experimentally induced hypotension in humans (Ogoh et al. 2008). During recovery from acute hypotension, the changes in the cerebral vascular conductance index (CVCi) were mediated by increases in arterial blood pressure (myogenic mechanisms) and sympathetically mediated cerebral vasoconstriction, which when blocked with prazosin diminished the rate of regulation, an index of dynamic cerebral autoregulation (CA). Furthermore, additional studies involving dynamic exercise identified sympathetically mediated cerebral vasoconstriction in exercising humans with activation of the exercise pressor reflex following a decrease in cardiac output with β1-adrenergic blockade. This decrease in cerebral blood flow velocity following sympathoexcitation was eliminated by stellate ganglion blockade (Ide et al. 2000). A decrease in cerebral tissue oxygenation, a surrogate measurement of decreased CBF via increased sympathetic activity, was reported at rest and during low-intensity exercise following injection of a bolus of phenylephrine. However, the decrease in CBF was attenuated with increased metabolism associated with moderate-intensity exercise (Brassard et al. 2010), suggesting ‘functional sympatholysis’, as evidenced in the peripheral vasculature (Keller et al. 2004). We hypothesized, therefore, that increased sympathetic activity during moderate dynamic exercise would increase cerebral vascular tone and that pharmacological blockade of the α1-adrenoreceptors with prazosin at rest and during steady-state dynamic exercise would result in decreased cerebral vascular tone and impaired cerebral autoregulation in healthy humans.

Methods

Seven men and three women (age, 28 ± 1 years; height, 176 ± 4 cm; weight, 78 ± 4 kg; means ± SEM), volunteered to participate in the present investigation. All subjects were healthy, free of known cardiovascular and respiratory diseases, and were not using any prescription or over-the-counter medication at the time of participation in the study. The subjects were asked to abstain from drinking alcohol and caffeine and not to exercise for a 24 h period prior to any scheduled experiments. All screening and experimental procedures were approved by the Institutional Review Board at the University of North Texas Health Science Center (IRB # 2009–106) and were in accordance with the guidelines of the Declaration of Helsinki.

Experimental protocols

The protocol required 2 days of participation. On day 1, subjects were informed of the study protocol, signed an informed consent, completed a health history questionnaire and were familiarized with all the testing protocols. They also underwent seated and standing 12-lead ECG measurements for detection of orthostatic intolerance and arrhythmia. A progressive exercise workload (WL) stress test on a stationary electrically braked upright cycle ergometer to volitional exhaustion was performed. The subjects were asked to maintain a pedalling cadence of ∼60 r.p.m. throughout the stress test.

Day 2 was separated from day 1 by at least 2 days. All subjects arrived in the laboratory in the morning, approximately 2 h after having a light breakfast. The subjects were seated on an upright cycle ergometer for 10 min while their resting data were acquired. The subjects exercised at two workloads, and the heart rate (HR) from the stress test recorded on day 1 was used to quantify the workload that each subject would require to achieve a steady-state HR of 130 beats min−1 and was referred to as moderate WL. Each subject pedalled at 10 W for the low WL regardless of his or her baseline HR. Following the rest protocol, each subject pedalled at 60 r.p.m. at a low WL of 10 W for 10 min. The group average HR achieved for the low WL was 88 beats min−1 (Ex90). Without pause, the WL was increased to achieve each subject's respective moderate WL at a steady-state target HR of 130 beats min−1 (Ex130) for the next 10 min.

Prazosin After the control conditions, which incorporated the rest and the two exercise trials, the subjects recovered, seated in an armchair, for 45 min to enable recovery of HR and mean arterial pressure (MAP) from the preceding exercise trials. The same rest and exercise protocol was repeated 2 h after oral ingestion of the α1-adrenoreceptor antagonist prazosin at a dose of 50 μg (kg body weight)−1 (Mylan Pharmaceuticals, Morgantown, WV, USA). The 2 h postingestion duration was chosen to coincide with the peak activity period of the drug (Jaillon, 1980).

Phenylephrine challenge To test the effectiveness of the α1-adrenoreceptor blockade achieved with prazosin, a phenylephrine (PE) challenge was administered following 45 min rest after the control exercise trials. A bolus intravenous injection of PE hydrochloride (Baxter Healthcare Corp., Deerfield, IL, USA) at a dose of 1.0 μg (kg body weight)−1 was given. Increases in blood pressure were seen immediately upon PE injection and ended within 3 min, consistent with other investigations (Goertz et al. 1993). The same dose of PE was repeated after the rest period following the prazosin trial as well as at the end of the study. The difference between the peak pressor response obtained during control conditions and that following prazosin was used to estimate the percentage of activated α1-adrenoreceptor-mediated blood pressure response achieved in the study.

Measurements

Beat-to-beat MAP recordings were obtained for each subject using a finger photoplethysmographic arterial blood pressure cuff (Finometer; Finapres Medical Systems, Amsterdam, The Netherlands). The Finometer transducer was calibrated, and the height sensor was zeroed at heart level and placed on the finger cuff while the subject rested their arm on a raised side-table throughout the study. In six subjects, beat-to-beat blood pressure was simultaneously acquired using a catheter (2.54 cm, 22 gauge catheter; Terumo Corporation, Tokyo, Japan) placed in the radial artery of the non-dominant arm under local anaesthesia with 1 ml of lidocaine using sterile techniques. The catheter was connected to a pressure transducer (Argon Medical Devices, Inc., Plano, TX, USA) positioned at the level of the right atrium in the midaxillary line. The overall correlation between MAP obtained by Finometer (y-axis) and direct radial arterial line (x-axis) was 0.88, and the regression equation between the two values was y= 1.0x– 3. Furthermore, a Bland–Altman analysis identified that the difference between all data points obtained from the two methods was within the 95% confidence interval. In addition, a venous catheter was inserted into the median antecubital vein to enable bolus injection of PE and blood sampling.

Heart was monitored using a three-lead ECG (model 78342A; Hewlett Packard) for continuous recording throughout the rest and exercise trials. Mean middle cerebral artery blood velocity (MCA V) was measured by transcranial Doppler ultrasonography (Multi-Dop X; DWL, Sipplingen, Germany). A 2 MHz Doppler probe was placed over the temporal window and fixed with an adjustable headband and adhesive ultrasonic gel (Tensive; Parker Laboratories, Orange, NJ, USA). The first four subjects tested were also instrumented with a sensor attached to a clip on the earlobe with a drop of electrolyte solution for detection of transcutaneous inline image (TOSCA 500; Radiometer, Copenhagen, Denmark). The inline image measured by this technology reflects arterial inline image (Gancel et al. 2011).

Blood sampling

A 3 ml blood sample was drawn from the antecubital venous catheter at the end of the rest period and the two exercise periods at different workloads during control conditions and following ingestion of prazosin. The blood samples were transferred to tubes pretreated with heparin and glutathione. A total of six blood samples were obtained from each subject throughout the experiment. The blood samples were centrifuged and the plasma samples stored at −70°C for analysis. Samples were later thawed and plasma concentrations of NA and adrenaline were analysed by high-performance liquid chromatography as described previously (Napier et al. 1998).

Data analysis

After the subject reached the steady state for each of the conditions (control and prazosin), 5 min of data were collected for transfer function analysis. The control condition always preceded the prazosin condition. Haemodynamic variables were obtained by averaging 1 min of steady-state data at the end of the rest period and each exercise WL. Analog signals of arterial blood pressure and MCA V were sampled continuously at 1 kHz using an analog-to-digital converter interfaced with a computer. Beat-to-beat arterial blood pressure and MCA V were obtained from the waveform.

Transfer function analysis The relationship between changes in MAP and MCA V was evaluated using transfer function analysis of dynamic CA using the DADisp program (DADisp 4.1; DSP Development, Cambridge, MA, USA). The transfer function phase, gain and coherence were calculated in the very low-frequency (VLF; 0.02–0.07 Hz), low-frequency (LF; 0.07–0.20 Hz) and high-frequency ranges (HF; 0.20–0.30 Hz). These frequency ranges reflect patterns of the dynamic pressure–flow relationships, as identified by transfer function analysis reported earlier (Ogoh et al. 2005). The blood pressure fluctuations in the HF range are induced primarily by respiration, whereas those in the LF range are independent of respiratory frequency and are dampened by CA. The LF fluctuations in blood pressure have been identified as reflecting fluctuations in sympathetic nervous system activity (Saul et al. 1991; Brys et al. 2003; Ogoh et al. 2005). The VLF range of both pressure and flow variability appear to reflect multiple physiological mechanisms that confound interpretations (Ogoh et al. 2005). Thus, the LF range for pressure and flow variables was used for transfer function phase and gain analysis to identify dynamic CA. The transfer function gain (tfg) reflects the relative amplitude between the changes in perfusion pressure and blood flow over the specified frequency range. Effective CA decreases the transmission of pressure fluctuation on flow. Therefore, an increase in tfg can be interpreted as an increase in transmission, suggesting relative impairment in CA between conditions in the present study and vice versa. In addition, normalized LF-tfg was calculated as the 20 log of tfg to express values in decibels (Ogoh et al. 2005). A value of zero indicates that the output varies by the same fraction of the mean value as the input; a negative value indicates that it varies less, and a positive value indicates that it varies more than the input. The transfer function phase shift reflects the time relationship between transmissions of changes in perfusion pressure on cerebral blood flow velocity. A decrease in phase shift indicates impairment in CA and vice versa. The coherence varies from zero to one, similar to the correlation coefficient, expressing the fraction of MCA V signal that is linearly associated with the MAP and was used to reflect the validity of the transfer function analysis between MAP and MCA V. Generally, coherence values of >0.5 are accepted as having statistical significance.

Critical closing pressure (CCP) The apparent CCP of the cerebral circulation was estimated and used as an index of cerebral vascular tone (Panerai, 2003; Panerai et al. 2005). The CCP for each individual was calculated from the last minute of steady-state data obtained at rest and at the end of each exercise trial without and with prazosin. Twenty pairs of systolic and diastolic arterial blood pressures (ABPs) were associated with the systolic and diastolic velocities of their consecutive MCA V waveforms. Linear regression between consecutive pairs of systolic and diastolic values of ABP and MCA V waveforms from the 20 cardiac cycles were determined, and the ABP axis (abscissa) intercept of the extrapolated MCA V/ABP regression line was used to determine the CCP (Panerai, 2003; Ogoh et al. 2010).

Cerebrovascular conductance index (CVCi) The CVCi was estimated by dividing beat-to-beat MCA V by MAP for each last minute of steady-state rest and exercise WLs, and an average was reported. The CVCi was used as an estimate of changes in cerebrovascular conductance resulting from changes in perfusion pressure during exercise conditions (O’Leary, 1991). Furthermore, a change in conductance best reflects a change in vascular tone rather than changes in resistance (Lautt, 1989).

Statistics

Subject comparisons were made across three exercise WLs (rest, Ex90 and Ex130) and across two conditions (control and prazosin). Two (3 × 2) factor ANOVA with repeated measures across each factor was used to assess the differences in haemodynamic variables, transfer function phase, gain and coherence in each frequency (SigmaStat; Jandel Scientific Software, SPSS, Chicago, IL, USA). Significant main effects were analysed using a Student–Newman–Keuls post hoc test. There were no significant interactions. Statistical significance was set at P < 0.05. All data are expressed as means ± SEM.

Results

A bolus dose of PE produced an average increase in MAP of 12 ± 1 mmHg; however, following prazosin, the same bolus dose of PE resulted in an increase in MAP of only 3.2 ± 1 mmHg, indicating a 74 ± 2% blockade of the pressor response mediated by α1-adrenergic receptor blockade. At the end of the experimental protocol, there was a 71 ± 4% blockade of the blood pressure response, confirming that the prazosin blockade of the α1-adrenergic receptor-mediated blood pressure increase was maintained throughout the protocol.

The mean values for the cardiovascular and haemodynamic variables obtained at rest and for the two exercise WLs during control conditions and with prazosin are presented in Table 1. An increase in MAP at Ex130 was observed compared with Ex90 (P= 0.040) and rest (P < 0.001) in the control conditions. The Ex90 workload also resulted in an increase in MAP from rest (P= 0.044) in the control conditions. Following prazosin, however, an overall decrease in MAP from the control conditions was observed at rest and during both the exercise WLs (P < 0.001). Heart rate increased from rest to Ex90 (P < 0.019) and Ex130 (P < 0.001) in the control conditions. Following prazosin, further increases in HR were observed compared with the control conditions (P < 0.001). The MCA V increased by 10% from rest to Ex130 in the control conditions (P= 0.015). Following prazosin, an overall decline in MCA V was observed from the control conditions (P < 0.001), and the increase in MCA V from rest to the two exercise WLs was abolished.

Table 1.  Steady-state haemodynamic variables at rest, Ex90 and Ex130 during control conditions and 2 h following oral prazosin (50 μg (kg body weight)−1)
ParameterUnitsControlPrazosin
RestEx90Ex130RestEx90Ex130
 
  1. Values are means ± SEM. Abbreviations: A, adrenaline; CCP, critical closing pressure; CVCi, cerebrovascular conductance index; Ex90, exercise at low workload; Ex130, exercise at moderate workload; HR, heart rate; MAP, mean arterial pressure; MCA V, middle cerebral artery blood flow velocity; and NA, noradrenaline. * Main effect of dynamic exercise (P < 0.05). † Main effect of prazosin (P < 0.05).

MAP*†mmHg79.9 ± 385.9 ± 394 ± 371 ± 374.6 ± 280 ± 3
MCA V*†cm s−149.8 ± 351 ± 454 ± 445 ± 546 ± 547 ± 5
HR*†beats min−179 ± 488 ± 4127 ± 190 ± 3102 ± 3149 ± 3
CVCi*†cm s−1 mmHg−10.65 ± 0.070.57 ± 0.040.55 ± 0.050.69 ± 0.060.62 ± 0.060.61 ± 0.07
CCP*mmHg18 ± 324 ± 431 ± 421 ± 422 ± 524 ± 4
NA*#pmol (ml plasma)−11.8 ± 0.22.0 ± 0.23.2 ± 0.32.6 ± 0.33.4 ± 0.34.4 ± 0.5
Apmol (ml plasma)−10.4 ± 0.10.4 ± 0.10.5 ± 0.11.0 ± 0.30.6 ± 0.11.1 ± 0.2

The plasma NA concentrations increased during Ex130 compared with rest (P < 0.001) and Ex90 (P < 0.001) in the control conditions. Prazosin resulted in further increases in NA at rest and the two WLs compared with the control conditions (P < 0.001).

Transfer function analysis of MAP and MCA V in the frequency spectra range of 0–3 Hz at rest and during Ex130 without and with prazosin is shown in Fig. 1. A decrease in the LF-tfg (0.07–0.2 Hz) from rest was observed at Ex90 (P= 0.035) and Ex130 (P= 0.027) in the control conditions (Fig. 2). In contrast, prazosin resulted in an increase in the LF-tfg from control values (P= 0.003), regardless of the exercise conditions, indicating a relative impairment in dynamic CA. Similar results were obtained with normalized LF-tfg. The coherence between MAP and MCA V for all the exercise and drug conditions for each subject remained above 0.5 and was accompanied by a decrease in phase from control values in the prazosin conditions (P= 0.017).

Figure 1.

Cross-spectral analysis in the spectra from 0 to 0.3 Hz with no drug and with oral prazosin 
Group averaged phase (top), normalized gain (middle) and coherence (bottom) between mean arterial pressure and middle cerebral artery blood flow velocity at rest (left) and during exercise at moderate workload (Ex130; right) are shown. Values are means; n= 10.

Figure 2.

Group average low-frequency (LF; 0.07–0.2 Hz) phase (A), LF gain (B), normalized LF gain (C) and LF coherence (D) at rest, and during exercise at low (Ex90) and moderate workload (Ex130) during control conditions and with oral prazosin 
Low-frequency gain at Ex90 (P= 0.035) and Ex130 (P= 0.027) and normalized LF gain at Ex90 (P= 0.030) and Ex130 (P= 0.011) were lower than rest in the control conditions. Prazosin increased LF gain (P= 0.003) and normalized LF gain (P < 0.001) in comparison to the control conditions regardless of exercise. There was a decrease in phase (P= 0.017) in the presence of prazosin compared with the control conditions. *Significantly different from rest and control conditions.

An increase in CCP was observed at Ex130 compared with rest from 18 ± 3 to 31 ± 4 mmHg (P < 0.001) and Ex90 from 24 ± 4 to 31 ± 4 mmHg (P < 0.001) in the control conditions. The increase in CCP was erased from rest to Ex90 to Ex130 from 21 ± 4 to 22 ± 5 to 24 ± 4 mmHg) in the prazosin conditions, indicating an overall decline in cerebral vascular tone following prazosin.

The CVCi decreased at Ex130 (P= 0.005) compared with rest in the control conditions, indicating cerebral vasoconstriction. In contrast, following prazosin, CVCi increased compared with the control conditions (P= 0.018) regardless of the WLs, indicating cerebral vasodilatation.

Discussion

The major findings of the present investigation are identified as follows. (i) A decrease in LF-tfg during low and moderate WL supports a stronger dynamic CA associated with exercise. (ii) The increase in LF-tfg following prazosin at rest and during the two WLs indicates an impairment of dynamic CA associated with blockade of the accessible α1-adrenoreceptors present in the cerebral microvasculature (Ogoh et al. 2008; Mitchell et al. 2009). (iii) The increase in CCP from rest to Ex130 in the control conditions was attenuated with prazosin. This finding suggests the presence of an exercise-induced increase in cerebral vascular tone mediated by activation of α1-adrenoreceptors in the autoregulatory response to the exercise-induced increases in perfusion pressure (Panerai, 2003). (iv) Prazosin resulted in an augmentation of CVCi, identifying an increase in CBF per unit change in arterial blood pressure. This finding suggests that cerebral vasodilatation occurred as a result of the α1-adrenergic receptor blockade. (v) The lack of an increase in MCA V despite mild increases in MAP during exercise in the prazosin conditions is attributed to impairment in dynamic CA associated with a decrease in cerebral vascular tone as well as decreases in perfusion pressure with α1-adrenergic receptor blockade. These findings highlight the role of α1-adrenergic receptors in establishing cerebral vascular tone, which when blocked during exercise impairs dynamic CA, decreases cerebral vascular tone and results in cerebral vasodilatation in conjunction with the established myogenic properties of the cerebral blood vessels (Zhang et al. 2009). The decrease in LF-tfg during low- and moderate-intensity upright cycling exercise observed in this study suggests stronger dynamic CA. However, in the past no change in LF-tfg was reported during mild- and moderate-intensity semi-recumbent cycling exercise (Ogoh et al. 2005). Previously, upright exercise, supine exercise and exercise at different pedalling frequencies have demonstrated that increases and decreases in central blood volume decrease and increase the operating point of the arterial baroreflex control of MAP and muscle sympathetic nerve activity, respectively (Volianitis et al. 2004; Ogoh et al. 2007); therefore, we suggest that postural differences between the two studies provide a rationale for the differences in the calculated LF-tfg.

Perivascular nerves have been identified in close proximity to smooth muscles in cerebral vessels, and the density and innervation of cerebral resistance vessels are extensive and similar to those of the mesenteric and the femoral arterial beds (Sandor, 1999). Despite the anatomical existence of a sympathetic neural network associated with the cerebral vessels, identification of a functional role for the sympathetic nervous system in the regulation of CBF has remained elusive (Bevan et al. 1998; Ogoh et al. 2008; Zhang et al. 2009; Hamner et al. 2010). In animal models, sympathetically mediated vasoconstriction during severe hypertension protected the rupture of the BBB (Bill & Linder, 1976). However, there are reports indicating that the sympathetic nervous system does not appear to have any substantial role in CBF regulation in humans (Skinhoj, 1972). On the contrary, there are reports highlighting the role of the sympathetic nervous system as a potential candidate influencing dynamic CBF regulation (Ogoh & Ainslie, 2009). In humans, infusions of tritiated NA at rest resulted in spillover from the cerebral blood vessels (difference between arterial and internal jugular vein NA concentrations) as well as from internal jugular vein samples containing tritiated NA metabolite overflow from behind the BBB, respectively. In addition, ganglionic blockade with trimethaphan infusions in healthy subjects lowered brain NA spillover and increased NA metabolite spillover in the internal jugular vein. In addition, clonidine suppression of central sympathetic outflow lowered brain NA spillover without changing NA metabolite overflow concentrations, while neuronal NA block decreased the transcranial extraction of NA (Mitchell et al. 2009). Furthermore, in humans at rest, stimulation of α1-adrenoreceptors with PE decreased cerebral tissue oxygenation, an indirect measurement of the decrease in CBF, due to activated α1-adrenoceptor vasoconstrictor mechanisms (Brassard et al. 2010). However, the increased cerebral metabolism associated with high-intensity exercise eliminated the effect, suggesting a balance between cerebral metabolism and a functional lysis of sympathetic control of blood flow in the brain at higher exercise intensities, which may be related to sympathetically mediated redistribution of regional cerebral blood flow (Secher et al. 2008). It was also reported that BBB permeability increases during intense exercise in healthy humans without any adverse structural damage to the blood vessels in the brain (Bailey et al. 2011). These findings corroborate that the BBB is permeable to NA and allows binding to the α1-adrenergic receptors of the cerebral vasculature.

The measurement techniques used for assessment of CA during steady-state conditions suggested that the myogenic properties of the smooth muscle are fundamental to CA and the effect of arterial carbon dioxide pressures on the myogenic properties of cerebral vascular tone is a primary modulator of CA. With dynamic measurement techniques, such as the transcranial Doppler ultrasound measuring cerebral blood flow velocity changes and the near-infrared spectroscopic analysis of cerebral tissue oxygenation, the dynamic regulation of CBF by CA, when arterial blood pressure is rapidly changing, can be assessed. With the use of dynamic techniques, there is a growing body of evidence identifying NA activation of the accessible α1-adrenergic receptors on smooth muscles of the cerebral vasculature than previously accepted (Heisted & Kontos, 1983). This increased sympathetic influence on cerebral vascular function appears to be related to its effect on cerebral vascular tone (D’Alecy et al. 1979; Ogoh et al. 2008; Ogoh & Ainslie, 2009).

The critical closing pressure of a blood vessel indicates the value of ABP at which blood flow approaches zero (Burton, 1951). In the cerebral circulation, CCP is equal to the sum of the intracranial pressure and the tension developed in the vascular smooth muscle cells (Dewey et al. 1974). The calculation of apparent CCP of cerebral blood vessels using extrapolation of the regression line of MCA V to the ABP axis intercept is a useful measure of the dynamics of the cerebral circulation (Richards et al. 1999) and a relevant index of cerebrovascular tone (Panerai et al. 1995, 1999). In the present study, dynamic exercise induced an increase in CCP that is interpreted as a sympathetically mediated increase in cerebral vascular tone, which would serve to protect the BBB from exercise-induced hypertension (Ogoh et al. 2010) and the exercise-induced increases in intracranial pressure (Panerai, 2003).

Increases in exercise WLs result in an increase in plasma NA (Kotchen et al. 1971; Hartley et al. 1972) and systemic NA spillover during leg exercise (Savard et al. 1989). In this investigation, there was a stepwise increase in cerebral CCP from rest to Ex90 and Ex130, suggesting an increase in cerebral vascular tone with the increase in sympathetic activity associated with exercise. However, blockade of the α1-adrenergic receptors with prazosin attenuated the increase in CCP during dynamic exercise. We suggest, therefore, that activation of α1-adrenergic receptors via an increase in sympathetic activity contributes to the increase in cerebral vascular tone (Schubert & Mulvany, 1999). This α1-adrenoreceptor activation may contribute to the myogenic stretch–tension mechanism of the cerebral vascular smooth muscles (Panerai et al. 2005; Panerai, 2008). In the present study, the presence of cerebral vasoconstriction is supported by the concomitant decrease in CVCi at Ex130 in the control conditions. However, prazosin blockade resulted in increases in CVCi compared with the control conditions even though the increases in MAP from rest to Ex90 and Ex130 were blunted compared with the control values. Similar decreases in cerebrovascular conductance was also reported following cold pressor test-mediated sympathetic stimulation at rest and during static exercise (Hartwich et al. 2010).

Experimental limitations

In this study, MCA V was estimated using transcranial Doppler ultrasonography, which is non-invasive and has high temporal resolution (Willie et al. 2011). However, the potential limitation of vasoconstriction of the insonated vessel would increase MCA V at any given volume of flow. In humans, however, the MCA diameter is reported to remain relatively constant in the presence of a variety of sympathetic stimuli (Schreiber et al. 2000; Serrador et al. 2000), and the diameter of large cerebral arteries does not change significantly during exercise (Pott et al. 1997). The regulation of CBF takes place in smaller arteries and arterioles (Giller et al. 1993). However, the effects of pharmacological intervention on MCA diameter is unknown and was not accounted for in the present study. Furthermore, findings that changes in MCA V increased in a similar manner to the inflow of the internal carotid artery (Hellström et al. 1996) and the ‘initial slope index’ of the 133Xe clearance-determined CBF (Jørgensen et al. 1992) support our use of transcranial Doppler measurements for identifying changes in cerebral blood flow. The coherence between MCA V and MAP remained above 0.5 in all conditions, suggesting that there was little effect of signal noise on the validity of transfer function analysis during exercise. The CCP was estimated using 20 waveforms of consecutive pairs of systolic and diastolic values of ABP and MCA V during steady-state rest and exercise conditions. The calculation of CCP using only the systolic and diastolic values versus the beat-to-beat data from cardiac cycles was reported not to be different (Ogoh et al. 2010). Given that CCP was calculated as the pressure when flow is zero and is extrapolated from the linear regression between MAP and MCA V, this technique may not be equal to the true CCP of the vessel. This is because the velocity–pressure relationship is highly curvilinear at low blood pressure (Jørgensen et al. 1992). However, the estimated CCP is considered to be an index of changes in cerebral vascular tone (Panerai, 2003). In this study, we achieved an average of 74% of blockade of the blood pressure response with oral prazosin; hence, if we had established a complete blockade, the identified impairment of dynamic CA would have been exacerbated, and the presence of sympathetic activation of α1-adrenoreceptors of the cerebral vessels in the establishment of cerebral vascular tone would have been more marked.

In summary, α1-adrenoreceptors play an important role in dynamic CA and cerebral vasculature tone at rest and during dynamic exercise. Blockade of the α1-adrenoreceptors impaired dynamic CA and attenuated any increase in CCP during moderate dynamic exercise in healthy humans.

Appendix

Acknowledgements

The authors thank Dr James Caffrey, PhD, and Darice Yoshishige for laboratory support, their expertise and assistance in the analyses of plasma catecholamines. The authors also thank the subjects for volunteering for the study and their interest in its outcomes. This study was supported in part by funds provided by the Cardiovascular Research Institute and the Department of Integrative Physiology, University of North Texas Health Science Center.

Ancillary