Cutaneous Active Vasodilation in Humans Is Mediated by Cholinergic Nerve Cotransmission
Abstract During heat stress, increases in blood flow in nonglabrous skin in humans are mediated through active vasodilation by an unknown neurotransmitter mechanism. To investigate this mechanism, a three-part study was performed to determine the following: (1) Is muscarinic receptor activation necessary for active cutaneous vasodilation? We iontophoretically applied atropine to a small area of forearm skin. At that site and an untreated control site, we measured the vasomotor (laser-Doppler blood flow [LDF]) and sudomotor (relative humidity) responses to whole-body heat stress. Blood pressure was monitored. Cutaneous vascular conductance (CVC) was calculated (LDF÷mean arterial pressure). Sweating was blocked at treated sites only. CVC rose at both sites (P<.05 at each site); thus, cutaneous active vasodilation is not effected through muscarinic receptors. (2) Are nonmuscarinic cholinergic receptors present on cutaneous arterioles? Acetylcholine (ACh) was iontophoretically applied to forearm skin at sites pretreated by atropine iontophoresis and at untreated sites. ACh increased CVC at untreated sites (P<.05) but not at atropinized sites. Thus, the only functional cholinergic receptors on cutaneous vessels are muscarinic. (3) Does cutaneous active vasodilation involve cholinergic nerve cotransmission? Botulinum toxin was injected intradermally in the forearm to block release of ACh and any coreleased neurotransmitters. Heat stress was performed as in part 1 of the study. At treated sites, CVC and relative humidity remained at baseline levels during heat stress (P>.05). Active vasodilator and sudomotor responses to heat stress were abolished by botulinum toxin. We conclude that cholinergic nerve activation mediates cutaneous active vasodilation through release of an unknown cotransmitter, not through ACh.
Reflex control of blood flow to nonglabrous skin in humans is accomplished by two branches of the sympathetic nervous system: a noradrenergic active vasoconstrictor system and an active vasodilator system of uncertain neurotransmitter.1 2 3 4 5 6 7 During cold stress, vasoconstrictor nerve activity is increased, and any vasodilator activity is abolished, reducing CVC, decreasing SkBF, and conserving body heat.5 6 7 At the onset of heat stress, as internal temperature rises, an initial abolition of noradrenergic vasoconstrictor nerve activity occurs. With further increases in internal temperature, the active vasodilator system is activated. These changes in nerve activity relax cutaneous resistance vessels, thus increasing CVC and SkBF to aid in the dissipation of body heat. The activation of the vasodilator system is responsible for 80% to 95% of the elevation in SkBF accompanying heat stress.5 6 7 Considering that whole-body SkBF can achieve a level approaching 8 L/min, or 60% of cardiac output during heat stress,5 6 7 this vasodilator system is a critical component of thermoregulatory responses and can have an important role in systemic hemodynamics.
Active vasoconstrictor nerves are known to be adrenergic, secreting norepinephrine, which acts on postsynaptic α1- and perhaps α2-receptors.5 6 7 8 Cotransmission may also be involved in active vasoconstriction.9 10 In contrast, the neurotransmitter and neural mechanism for active vasodilation are unclear.11 The observation that SkBF tends to parallel sudomotor activity during heat stress led to the suggestion that sweat gland activation plays a role in cutaneous active vasodilation. Drawing on earlier work done with salivary glands,12 Fox and Hilton2 proposed that cholinergic activation of sweat glands elicited a glandular production of kininogenase that ultimately resulted in bradykinin production via an enzymatic cascade. Bradykinin was proposed as the effector of active cutaneous vasodilation, although atropine completely blocked sweat production and only slightly delayed active vasodilation during heat stress.2 They also detected a bradykinin-forming enzyme in human sweat.2 Subsequent work by Frewin et al13 failed to confirm this proposal. Although there is agreement that systemic administration of atropine can abolish sweating, active vasodilation is reported to be delayed in onset,2 14 unchanged in onset,15 reduced in magnitude,13 completely inhibited,16 or enhanced15 by atropine. These inconsistencies suggest that cutaneous active vasodilation may not be mediated by activation of muscarinic receptors and does not require the activation of sweat glands. Alternatively, another cholinergic receptor type on the skin vessels could mediate active vasodilation. These conflicting results even suggest that cutaneous active vasodilation may be effected by noncholinergic nerves. However, the postulate that cholinergic-muscarinic sudomotor activity effects active cutaneous vasodilation persists.2 5 6 7 16
The foregoing proposals are based on principles of classic pharmacology. An alternative hypothesis proposed by Hökfelt et al17 derives from studies of cotransmitter systems in the cat paw. According to this hypothesis, a single set of cholinergic nerves could control both active dilation of cutaneous arterioles and sweating through corelease of ACh and VIP. ACh would stimulate sweating, whereas coreleased VIP would cause active vasodilation. This colocalization and corelease mechanism could explain why systemic atropine blocks sweating but has inconsistent effects on active vasodilation. In testing this hypothesis, Savage et al18 found active vasodilation to be normal in patients with cystic fibrosis, a group having very low VIP content in cutaneous postganglionic neurons. Although VIP as a cotransmitter is largely ruled out by this finding, the possibility of cholinergic cotransmission as the mechanism for cutaneous active vasodilation is not excluded.
To elucidate the mechanism of cutaneous active vasodilation in humans, we performed a study with three protocols to examine the potential involvement of muscarinic cholinergic systems, nonmuscarinic cholinergic systems, and cholinergic cotransmitter systems. Thus, the following protocols were designed to test three related hypotheses: (1) Is muscarinic control part of cutaneous active vasodilation? (2) If not, can nonmuscarinic cholinergic receptors mediate cutaneous active vasodilation? (3) If not, is cutaneous active vasodilation a function of release of a cotransmitter from cholinergic terminals, or is it mediated by noncholinergic nerves?
Subjects and Methods
Seven male and two female subjects participated in one or more of the protocols. Their average age (±SEM) was 29±3 years; average height, 172±3 cm; and average weight, 71±3 kg. All subjects were in good health, and their informed consent was obtained before their participation in these institutionally approved studies.
In protocols 1 and 3, whole-body heat stress was produced as follows: Subjects wore a tube-lined suit that was used to control Tsk by perfusion with water of different temperatures.19 20 Over the suit, subjects wore a water-impermeable plastic garment. The suit and garment covered the entire body except for the head, arms, and feet. The suit was perfused with warm water to raise Tsk to 38°C to 39°C during heating periods or perfused with cold water to lower Tsk from a range of 34°C to 34.5°C to a range of 31.5°C to 32°C for cold stress.
Internal temperature was monitored from the esophagus (Tes).19 Tsk was recorded as the weighted electrical average from six thermocouples placed over the body surface.19 20 Heart rate was monitored continuously from the ECG. Mean arterial pressure was recorded continuously from a finger blood pressure monitor (Finapres, Ohmeda).
LDF from skin, assessed by dual-channel flowmeter (MBF3D, Moor Instruments Ltd), and SR, assessed by relative humidity monitors (HX92V, Omega Engineering; IH3602-L humidity sensors, HY-CAL Engineering), were measured simultaneously at each of two forearm sites. CVC was indexed as LDF (in volts)÷mean arterial pressure (in millimeters of mercury).
We constructed special LDF probe holders to permit simultaneous measurement of LDF and SR with concomitant control of local temperature at the site of LDF measurement. The LDF probe head was placed in a central chamber in the probe holder and was slightly raised from the skin surface. Dry air was passed through the central chamber and conducted to a relative humidity detector to measure SR. Local warming was accomplished by heating elements embedded within the probe holder. A thermocouple was placed between the skin surface and the probe holder to provide measurement of local temperature and feedback for temperature control.
In protocol 1, we characterized the role of cholinergic muscarinic receptors in cutaneous active vasodilation and sweating by investigating the effects of locally applied atropine, a cholinergic muscarinic receptor antagonist. Atropine sulfate was iontophoretically applied at a dose of 400 μA per cm2 for 45 seconds to a 0.64-cm2 area of forearm skin to block cutaneous cholinergic receptors in those areas. By local application, we could test the blocking effects of atropine, uncomplicated by any systemic effects. We then simultaneously monitored SkBF by laser-Doppler flowmetry21 22 and SR by relative humidity at the atropine-treated site and at an adjacent untreated site during periods of body heating and cooling. Body heating was used to produce heat stress and thus served to activate the vasodilator system. Body cooling was used to produce cold stress to test for intact noradrenergic vasoconstrictor function and thus to examine for unanticipated effects of atropine iontophoresis.
Data collection began with a 5-minute normothermic control period followed by a 3-minute application of whole-body cold stress to verify that vasoconstrictor nerve function was intact and that atropine pretreatment had not produced unanticipated effects on either nerve or vascular function. After a few minutes of recovery, Tsk was raised to 38°C to 39°C and was maintained at that level for 35 to 45 minutes to induce heat stress. After hyperthermia, subjects were cooled to normothermia. Local temperatures at both LDF sites were raised to 42°C for maximal vasodilation.23 24 For data analysis, values of CVC were normalized to those maximal levels.25
The vasomotor responses to cold stress and heat stress were analyzed by comparing the levels of CVC at treated and untreated sites during the initial control period with the levels achieved during the final minute of cold stress and heat stress. The sudomotor responses to heat stress were analyzed by comparing the level of SR during the initial control period with that achieved during the final minute of heat stress. Internal temperature thresholds (the level of Tes at which CVC or SR began to rise during whole-body heating) were chosen from plots of CVC or SR versus Tes by an investigator blinded to the conditions, subjects, and drug treatments involved. CVC, SR, and Tes responses were analyzed by ANOVA followed by the Student-Newman-Keuls test.
In protocol 2, atropine was applied by iontophoresis at one forearm site at the same current and duration as in protocol 1. Subjects were then placed in a supine position and instrumented at the treated site with a laser-Doppler probe equipped with a probe holder that provided for simultaneous LDF measurement, iontophoresis, and local warming. This probe holder contained a recessed plastic iontophoresis chamber with a platinum electrode. The chamber was filled with a solution of ACh in propylene glycol. The LDF probe head was placed in the holder and was slightly raised from the skin surface. Local warming was accomplished by resistors with thermocouple feedback. This enabled the monitoring of LDF while ACh was being applied. Blood pressure was measured as described above. Eight of the nine subjects participated in this protocol.
The protocol began with a 5-minute control period. ACh was then iontophoretically applied at a dose of 31 μA per cm2 for 45 seconds to the 0.64-cm2 area of forearm skin previously treated with atropine. Responses to ACh were monitored for 10 minutes, and then the local temperature at the forearm site was raised to 42°C for maximal vasodilation. The LDF probe was then moved to an untreated control site on the forearm, and the protocol was repeated. The responses of CVC at the two sites were analyzed by ANOVA followed by the Student-Newman-Keuls test.
Protocol 3 involved presynaptic cholinergic nerve blockade with local intradermal injection of botulinum toxin type A (Allergan Pharmaceuticals). This is an anticholinergic agent that is taken up selectively by presynaptic cholinergic nerve terminals, where it abolishes release of ACh and any colocalized neurotransmitters.26 Each of six subjects had a 0.6-cm2 area of skin on the ventral surface of the forearm injected intradermally with 5 U botulinum toxin. Two to 14 days after injection, during which cholinergic nerve blockade developed, we tested the effects of this blockade on LDF and SR at the area of forearm skin pretreated with botulinum toxin and at an adjacent untreated site. The study design and analysis used in protocol 3 were otherwise identical to those used in protocol 1.
Fig 1⇓ shows the time course of results from protocol 1 in one subject. During cold stress, CVC fell from an average of 18±2% to 12±1% of the maximal level at untreated sites (P<.05) and from 19±2% to 14±2% of the maximal level at atropine-treated sites (P<.05). These responses were not different from each other (P>.10) and demonstrated that atropine iontophoresis had not altered either the adrenergic vasoconstrictor system or the responsiveness of the cutaneous vessels themselves.
During heat stress, no sweating was detected at sites pretreated with atropine. The value of Tes at which vasodilation began at the atropine-treated sites was slightly greater than that at the untreated sites (37.22±0.06°C [atropine-treated sites] and 37.14±0.06°C [untreated sites], P<.05). Despite total abolition of sweating by atropine treatment, CVC rose significantly at treated sites during heat stress. At the peak of heat stress, CVC had increased to 67±6% of the maximal level at untreated sites and to 50±6% of the maximal level at atropine-treated sites (P<.05 both sites). However, this vasodilation at atropine-treated sites was reduced in relation to that at control sites (P<.05 between sites). No differences were found in the absolute LDF values achieved with local warming at the two sites (P>.05 between sites). This indicated that atropine pretreatment had not altered the cutaneous vascular response to local warming and validated the normalization of CVC values to maximal levels.
In the second protocol, iontophoresis of ACh was used to exclude the possibility that the foregoing results were due to an inadequate blockade of muscarinic receptors on the cutaneous arterioles and to test for a nonmuscarinic cholinergic vasodilation. The results from one subject are illustrated in Fig 2⇓. In response to ACh, CVC at untreated control sites rose from 16±1% to 51±4% of the maximal level (P<.05), whereas at sites pretreated with atropine, CVC was unchanged by ACh (14±1% to 17±3% of the maximal level, P>.05). These responses were significantly different (P<.05 between sites).
The third protocol involved presynaptic cholinergic nerve blockade with local intradermal injection of botulinum toxin type A. Typical results are shown in Fig 3⇓. During cold stress, CVC fell from an average of 10±2% to 7±1% of the maximal level at untreated sites and from 10±1% to 7±1% of the maximal level at botulinum toxin–treated sites (P<.05 at each site, P>.10 between sites), demonstrating that botulinum toxin had not altered the noradrenergic mediation of the vasoconstrictor response to whole-body cooling. With heat stress, sweating occurred at untreated sites but not at botulinum toxin–treated sites, verifying cholinergic nerve blockade. Simultaneously, CVC rose to 52±9% of the maximal level at untreated sites (P<.05 versus control) but was not significantly different from control levels at botulinum toxin–treated sites (17±2% of the maximal level, P>.05 versus control). Finally, no differences were found in the absolute LDF values achieved during local warming to 42°C at the two sites (P>.05 between sites). This indicated that botulinum toxin pretreatment had not altered the cutaneous vascular response to local warming and therefore again validated the normalization of CVC values to maximal levels.
Overall responses to hyperthermia and to cold stress in protocols 1 and 3 are summarized in Fig 4⇓. The CVC responses to cold stress were similar between the two protocols and were not affected by atropine or botulinum toxin. However, presynaptic blockade of cholinergic function (botulinum) abolished CVC increases during heat stress, whereas postsynaptic antagonism only slightly reduced the CVC response.
The results of the present study show that cutaneous active vasodilation in humans is mediated by cholinergic nerve cotransmission. This conclusion follows from the following observations: muscarinic blockade does not abolish cutaneous active vasodilation during heat stress, muscarinic receptor blockade abolishes the cutaneous vasodilation induced by exogenous ACh, and selective presynaptic cholinergic nerve blockade with botulinum toxin abolishes the active vasodilator response to hyperthermia.
The results of the first protocol showed that iontophoresis of atropine did not alter the vasoconstrictor responses to cold stress. Therefore, atropine iontophoresis did not measurably alter the cutaneous vasculature through nonmuscarinic effects. Atropine did abolish sweating during heat stress but failed to abolish the cutaneous vasodilation. Thus, most of the vasodilator response in CVC at atropine-treated sites is not mediated by activation of muscarinic receptors either on sweat glands or on the cutaneous vessels themselves.
The foregoing results could have been due to an inadequate blockade of muscarinic receptors on the cutaneous arterioles, despite complete blockade of the sweat glands. Alternatively, nonmuscarinic cholinergic receptors not subject to blockade by atropine, such as nicotinic receptors, could mediate cutaneous active vasodilation. These possibilities were examined in our second protocol. ACh treatment increased CVC at control sites to levels comparable to those achieved during heat stress but had no effect on CVC at sites pretreated with atropine. The failure of ACh iontophoresis to increase CVC at sites pretreated with atropine indicates that muscarinic receptors on the vessels themselves were adequately blocked and thus could not mediate active vasodilation. These results also demonstrate that muscarinic receptors are the only cholinergic receptors of functional importance in the cutaneous vasculature and that, therefore, cutaneous active vasodilation cannot be mediated by activation of nonmuscarinic cholinergic receptors.
As previously mentioned, the neurotransmitter of the vasodilator nerves is unknown, although an indirect relation to cholinergic sudomotor nerve activity has been postulated.2 3 Atropine locally abolished sudomotor but not vasomotor responses to hyperthermia. Thus, our results show that activation of the sweat glands is not necessary for active vasodilation and do not support a role for glandular production of kininogenase leading to bradykinin as the mechanism of active vasodilation.
From the results of protocols 1 and 2, we concluded that ACh is not the major mediator of cutaneous active vasodilation. The remaining issue regarding the role of cholinergic nerves in active vasodilation was the possibility of corelease of another neurotransmitter from cholinergic nerves as proposed by Hökfelt et al17 and Lundberg.27 According to this proposal, a single set of neurons could control both active vasodilation of cutaneous arterioles and sweating by coreleasing both ACh and VIP. In this scheme, ACh would control sweating, and coreleased VIP would cause active vasodilation.17 This colocalization mechanism could explain why atropine blocks sweating but inconsistently affects active vasodilation. In the third protocol, we used presynaptic cholinergic nerve blockade with local intradermal injection of botulinum toxin type A to test the possibility that cotransmitter release was involved in cutaneous active vasodilation. Botulinum toxin is an anticholinergic agent that is taken up selectively by presynaptic cholinergic nerve terminals.26 Within the nerve terminal, the toxin cleaves several proteins required for Ca+2-induced exocytosis, leading to total inhibition of this process.28 Abolition of Ca+2-induced exocytosis blocks release of small synaptic vesicles that contain ACh and large dense-core vesicles that mainly contain peptides.26 28 We reasoned that botulinum toxin would thus concomitantly abolish the release of any cotransmitters from cholinergic nerve endings that are secreted by Ca+2-induced exocytosis.26 28 The fact that vasoconstrictor function in the present study was unimpaired testifies to the selectivity of the action of botulinum toxin. Furthermore, it is not thought to affect the function of nonadrenergic noncholinergic nerves.26
Botulinum toxin locally abolished both the active vasodilator and sudomotor responses to hyperthermia. From this result, we conclude that activation of cholinergic nerves mediates cutaneous active vasodilation. Given the selectivity of botulinum toxin for cholinergic nerve terminals and our observations from protocols 1 and 2 that active cutaneous vasodilation is not mediated by activation of cholinergic receptors, these findings indicate that release of a cotransmitter from cholinergic nerves is the major mechanism of cutaneous active vasodilation.
VIP is considered the likely transmitter responsible for atropine-resistant vasodilation with parasympathetic nerve stimulation in the salivary gland.29 Is VIP the cotransmitter that causes cutaneous active vasodilation in humans? Probably not. Savage et al18 found active vasodilation to be normal in patients with cystic fibrosis, a group with very low VIP levels in postganglionic neurons. This finding suggests that VIP is not the specific cotransmitter of cutaneous active vasodilation in humans, although it appears to have a cotransmitter role elsewhere.17 27 Is nitric oxide the cotransmitter that mediates active vasodilation? This also appears unlikely. Dietz et al11 found that intra-arterial L-NMMA did not abolish cutaneous active vasodilation.11 This finding suggests that nitric oxide does not mediate cutaneous active vasodilation in humans. We have tried to iontophoretically deliver L-NMMA into the skin to investigate the role of nitric oxide but found that this could not be done because of problems with the iontophoretic application itself.
One local factor important to the cutaneous circulation is the temperature of the blood vessels themselves. Botulinum toxin did not alter the vasodilation induced by local warming of the skin to 42°C. This indicates that cholinergic nerves do not mediate the vasodilation induced by increased local temperature. Prior work from our laboratory found that the effects of local warming on cutaneous arterioles do not require intact adrenergic function and may involve an axon reflex.30 The results of the present study suggest that the vascular effects of increased local temperature do not involve cholinergic neurotransmission and thus are mediated by a mechanism separate from active vasodilation.
As mentioned previously, atropine was not without effect on the cutaneous vasodilation induced by heat stress. During hyperthermia, a small delay in the onset of cutaneous vasodilation occurred at atropine-treated sites (P<.05 between sites). In addition, the extent of vasodilation was slightly reduced by atropine (P<.05 between sites). These results are similar to previous findings based on intra-arterial infusions of atropine.14 A delayed and slightly reduced vasodilation at atropine-treated sites suggests that activation of muscarinic receptors participates in the increase of CVC during hyperthermia. It is possible that ACh contributes partially to cutaneous active vasodilation through postsynaptic muscarinic receptors on the skin vessels and that atropine eliminates that component. It is also conceivable that the reduction in vasodilation is limited to those vessels that only subserve sweat glands. Perhaps sweat gland activation by muscarinic receptors effects a local periglandular hyperemia that contributes to the vasodilation of heat stress. According to this hypothesis, increases in SkBF during heat stress are the sum of increased blood flow at cutaneous arterioles by the cotransmitter and the hyperemia of sweat gland activation by ACh. In addition to postsynaptic effects, ACh may have influences through presynaptic muscarinic autoreceptors that exert negative feedback on release of both ACh and the cotransmitter.27 Atropine blockade of such autoreceptors could lead to increased release of the cotransmitter. Increased cotransmitter release could mask the effect of postsynaptic muscarinic blockade and restore most of the vasodilation that was prevented by atropine.
In summary, we found that muscarinic receptor blockade with atropine did not abolish cutaneous active vasodilation during heat stress but did abolish the vasodilation induced by ACh. Presynaptic blockade of cholinergic nerves with botulinum toxin selectively abolished the active vasodilation induced by heat stress. Collectively, these results show that cutaneous active vasodilation is mediated by cholinergic nerves and indicate that the mechanism of cutaneous active vasodilation involves a cotransmitter released from those nerves. Identification of this cotransmitter and establishing whether these cholinergic nerves are identical with sudomotor nerves represent future challenges to understanding the mechanism of cutaneous active vasodilation.
Selected Abbreviations and Acronyms
|CVC||=||cutaneous vascular conductance|
|LDF||=||laser-Doppler blood flow|
|SkBF||=||skin blood flow|
|Tes||=||internal (esophageal) temperature|
|VIP||=||vasoactive intestinal peptide|
This study was supported by an American Federation for Aging Research/Merck Fellowship in Geriatric Clinical Pharmacology (Dr Kellogg); National Heart, Lung, and Blood Institute grants HL-36080 (Dr Johnson) and HL-08861 (Dr Crandall); and a NATO grant through the German Academic Exchange Service (Dr Grossmann). The authors gratefully acknowledge Moor Instruments for their generous loan of an MBF3D laser-Doppler blood flow monitor and the volunteers for their participation and cooperation.
- Received April 27, 1995.
- Accepted August 2, 1995.
- © 1995 American Heart Association, Inc.
Grant RT, Holling HE. Further observations on the vascular responses of the human limb to body warming: evidence for sympathetic vasodilator nerves in the normal subject. Clin Sci. 1938;3:273-285.
Fox RH, Hilton SM. Bradykinin formation in human skin as a factor in heat vasodilation. J Physiol (Lond). 1958;142:219-232.
Brengelmann GL, Freund PR, Rowell LB, Olerud JE, Kraning KK. Absence of active vasodilation associated with congenital absence of sweat glands in humans. Am J Physiol. 1981;240:H571-H575.
Edholm OG, Fox RH, MacPherson RK. Vasomotor control of the cutaneous blood vessels in the human forearm. J Physiol (Lond). 1957;139:455-465.
Johnson JM, Proppe DW. Cardiovascular adjustments to heat stress. In: Blatteis C, Fregly M, eds. Handbook of Physiology: Adaptations to the Environment. Bethesda, Md: American Physiological Society. In press.
Fox RH, Edholm OG. Nervous control of the cutaneous circulation. Br Med Bull. 1963;19:110-114.
Taddei S, Salvetti A, Pedrinelli R. Persistence of sympathetic-mediated forearm vasoconstriction after α-blockade in hypertensive patients. Circulation. 1989;80:485-490.
Dietz NM, Rivera JM, Warner DO, Joyner MJ. Is nitric oxide involved in cutaneous vasodilation during body heating in humans? J Appl Physiol. 1994;76:2047-2953.
Hilton SM, Lewis GP. The mechanism of the functional hyperemia in the submandibular salivary gland. J Physiol (Lond). 1955;129:253-271.
Frewin DB, McConnell DJ, Downey JA. Is a kininogenase necessary for human sweating? Lancet. 1973;2:744.
Roddie IC, Shepherd JT, Whelan RF. The contribution of constrictor and dilator nerves to the skin vasodilation during body heating. J Physiol (Lond). 1957;136:489-497.
Love AHG, Shanks RG. The relationship between the onset of sweating and vasodilation in the forearm during body heating. J Physiol (Lond). 1962;162:121-128.
Savage MV, Brengelmann GL, Buchan AM, Freund PR. Cystic fibrosis, vasoactive intestinal peptide, and active cutaneous vasodilation. J Appl Physiol. 1990;69:2149-2154.
Johnson JM, Park MK. Effect of upright exercise on the threshold for cutaneous vasodilation and sweating. J Appl Physiol. 1981;50:814-818.
Rowell LB, Murray JA, Brengelmann GL, Kraning KK II. Human cardiovascular adjustments to rapid changes in skin temperature during exercise. Circ Res. 1969;24:711-724.
Johnson JM. The cutaneous circulation. In: Shepherd AP, Öberg PÅ, eds. Laser-Doppler Blood Flowmetry. Boston, Mass: Kluwer Academic Publishers; 1990:121-140.
Taylor WF, Johnson JM, O’Leary DS, Park MK. Effect of high local temperature on reflex cutaneous vasodilation. J Appl Physiol. 1984;57:191-196.
Johnson JM, O’Leary DS, Taylor WF, Kosiba WA. Effect of local warming on forearm reactive hyperemia. Clin Physiol. 1986;6:377.
Kellogg DL Jr, Johnson JM, Kenney WL, Pérgola PE, Kosiba WA. Mechanisms of control of skin blood flow during prolonged exercise in humans. Am J Physiol. 1993;265:H562-H568.
Lundberg JM. Evidence for coexistence of vasoactive intestinal polypeptide (VIP) and acetylcholine in neurons of cat exocrine glands. Acta Physiol Scand. 1981;112(S496):1-57.
Thiel G. Recent breakthroughs in neurotransmitter release: paradigm for regulated exocytosis. News Physiol Sci. 1995;10:42-46.
Pérgola PE, Kellogg DL Jr, Johnson JM, Kosiba WA, Solomon DE. Role of sympathetic nerves in the vascular effects of local temperature in human forearm skin. Am J Physiol. 1993;265:H785-792.