Autonomic regulation of cutaneous vascular resistance in the bullfrog Rana catesbeiana

In amphibians, skin is a major respiratory and osmoregulatory organ. The exchange of gases across the skin (Malvin and Hlastala, 1989), as well as ions and water (Christensen, 1974; Mahany and Parsons, 1978), is regulated in part by cutaneous blood flow. The mechanisms controlling cutaneous blood flow, however, are poorly understood. Several studies have shown that cutaneous blood flow can be regulated by sympathetic innervation of the cutaneous vasculature (Langley, 1911; Krogh et al. 1922; Smith, 1976). However, little is known of the neural pathways, transmitters and receptors involved in such regulation, nor has possible hormonal regulation of the cutaneous vasculature by circulating catecholamines been evaluated.

The purpose of this study was to characterize further the vascular innervation, the adrenoceptors and the vascular sensitivity to catecholamines of bullfrog skin. There were four specific goals: (1) to determine whether nerves associated with the vagal ganglion, the first two cervical sympathetic ganglia, cranial nerves IX and X and spinal nerves 1 and 2, can regulate cutaneous vascular resistance (CVR); (2) to determine whether catecholamines, acetylcholine or ATP are among the transmitters used by cutaneous vasomotor nerves; (3) to determine the doseresponse relationships of epinephrine (EPI) and norepinephrine (NE) in the cutaneous vasculature and to assess the possibility of hormonal control of CVR by these substances; and (4) to determine the adrenoceptor types in the cutaneous vasculature. All experiments were performed on a pump-perfused skin preparation in which the cutaneous artery was perfused at constant flow while perfusion pressure was measured.

Bullfrogs, Rana catesbeiana, were obtained from an animal dealer and housed in a large aquarium at 24°C at least 1 week before experimentation. Body mass ranged from 100 to 691g (mean=308g; s.D. = 115g; N=68).

A bullfrog was double-pithed and placed supine on an elevated wire mesh. A midline incision through the pectoral girdle was made to expose the aortic arches and large veins entering the heart. Heparin (50 units; Elkins-Sinn, Inc.) was injected into the ventricle to prevent clotting in the preparation. A cannula (PE 50) was placed in the left anterior vena cava which receives venous effluent from the portions of skin to be perfused. Since venous pressure in frogs is approximately 0kPa (Jones and Shelton, 1972), the distal end of the cannula was positioned 2 cm below the level of the heart. This ensured that outflow pressure in the vein was approximately 0kPa to compensate for the cannula resistance that caused an approximately 0.27 kPa fall in pressure along its length at the flow rates used. The left pulmonary artery just proximal to the lung was then ligated. A cannula was placed in the left pulmocutaneous artery so that the skin supplied by the left cutaneous artery could be selectively perfused at constant flow. A roller pump (Masterflex 7523-10) withdrew perfusate from a reservoir and pumped it at constant flow through the cannula in the pulmocutaneous artery. The perfusata was bubbled with a gas mixture of 98 % O₂ and 2 % CO₂ delivered by a Wösthoff gas-mixing pump. Perfusion pressure was measured with a pressure transducer (Statham) connected to the perfusion line via a T-connection. Just distal to the T-connection was a thick rubber injection port through which small volumes of drugs could be administered (Fig. 1). Flow through the arterial cannula was adjusted so that perfusion pressure was approximately 2 kPa. Mean arterial pressure in bullfrogs at rest is 1.96 kPa (Herman et al. 1986). Flow was determined after the experiments by cutting the pulmocutaneous artery at the cannula tip and then measuring the volume of perfusate pumped over 5 min. Perfusion pressure had no effect on pump flow. The perfusate was a modified Ringer’s solution containing (in mmoll⁻¹): NaCl, 76.1; NaHCO₃, 30.7; NaH₂PO₄, 1.4; KC1, 2.5; MgSO₄, 3.1; CaCh, 4.0; glucose, 5.6; and 30gl⁻¹ dextran (M_r 80000). All experiments were made at room temperature (22–24°C).

The tissue perfused in this preparation was the skin on the sides and back. This was determined by adding Methylene Blue to the perfusate in four preparations and noting the areas of skin that were dyed after 20min. Dye was never observed in areas other than the sides and back.

The names of the nerves follow the nomenclature of Pick (1970; Fig. 2 of this paper). The vagal ganglion, sympathetic ganglia 1 and 2, cranial nerves IX (glossopharyngeal) and X (vagus) and the first two spinal nerves on the left side of the animal were exposed by an incision in the roof of the mouth. The cutaneous artery and vein were cannulated as described and perfusion of the skin begun. 20 min after the initiation of skin perfusion, the nerves were electrically stimulated one at a time. The order of nerve stimulation was randomized. Stimulation was performed with bipolar platinum electrodes placed beneath the nerves or ganglia. 5 mA pulses of 1 ms duration at 5 Hz were delivered in a square-wave pattern for 20 s by a Grass S88 stimulator with a Grass stimulus isolation unit (PSIU6). Preliminary experiments determined that these parameters were well above threshold but below the stimulus required for a maximal response in the sympathetic ganglia and cranial nerve X. If a response was obtained, perfusion pressure was allowed to return to prestimulation levels (approximately 10–20 min) before the next nerve was stimulated.

To determine neural pathways, nerves were stimulated again before and after neural connections had been selectively severed. Specifically, sympathetic ganglion 2 was stimulated, then the connections between the first and second sympathetic ganglia were severed and sympathetic ganglion 2 was stimulated again. Sympathetic ganglion 1 was then stimulated, the connections between the vagal ganglion and the first sympathetic ganglion were cut and the first sympathetic ganglion was then stimulated. Next, the vagal ganglion was stimulated, then cranial nerves IX and X were cut just lateral to the vagal ganglion and then the vagal ganglion was stimulated. Finally, cranial nerve X was stimulated.

To determine whether the nerves innervating the cutaneous vasculature were adrenergic, cholinergic or neither, sympathetic ganglion 1 and cranial nerve X were stimulated before and after adrenergic and cholinergic antagonists had been added to the perfusate. In six frogs, sympathetic ganglion 1 and cranial nerve X were stimulated as previously described with 20min between the two stimulations. Then the a-adrenoceptor antagonist phentolamine (10⁻⁵moll⁻¹; Regitine, CIBA Pharmaceutical Co.) was added to the perfusate. 30 min later the two nerves were stimulated as before. Atropine (10⁻⁵ mol I⁻¹; Sigma Chemical Co.), a muscarinic antagonist, was then added to the perfusate and 30 min later the same two nerves were stimulated. This protocol was repeated on six additional frogs but atropine was added before phentolamine.

Another four frogs were tested to determine if cutaneous vasomotor nerves in the vagus use ATP as a transmitter. Animals were prepared as described above. Cranial nerve X was stimulated, then α, β-methylene ATP (10⁻⁴moll⁻¹; Sigma Chemical Co.) was added to the perfusate. [α,β-methylene ATP is a methylated analogue of ATP which desensitizes P₂-purinoceptors (Kasakov and Bumstock, 1983). P₂-purinoceptors mediate ATP-induced responses in many different types of blood vessels (Bumstock and Brown, 1981; Su, 1981).] 30 min later, the vagus was stimulated again. The vagus was stimulated a third time 30 min after 10⁻³moll⁻¹ α,β-methylene ATP had been added to the perfusate.

The cutaneous vasculature was perfused as described. 20 min after perfusion had been initiated, either EPI (Adrenalin, Parke-Davis; N=7) or NE (Levophed, Winthrop-Breon; N=8) was added cumulatively to the perfusate. Dose-response relationships between 1.8×10⁻⁸ and 1.8×10⁻⁴moll⁻¹ were produced. Perfusate catecholamine concentration was increased by 0.5 log units after CVR became stable (approximately 10–20 min).

Two groups of experiments evaluated the adrenoceptor types in the cutaneous vasculature. In one group, the response of the cutaneous vasculature to EPI and NE was tested before and after α- and β-adrenoceptor antagonists (phentolamine and propranolol, respectively) had been added to the preparation. The other group of experiments tested the effect of selective α- and β-adrenoceptor agonists (phenylephrine and isoproterenol, respectively). This second group also evaluated the effects of acetylcholine on the cutaneous vasculature.

In the first set of experiments, two doses (0.1 and 1.0 nmol) of both EPI and NE were administered in a random order through the injection port by a Hamilton syringe. Injection volume was always 2 μl, which had only a small, brief effect on perfusion pressure. Injections usually caused a transient increase in CVR. Succeeding injections were administered only after CVR had returned to the basal level (approximately 5–20min). After the fourth injection, either propranolol (10⁻⁶moll⁻¹; Inderal, Ayerst Laboratories, Inc.; N=7) or phentolamine (10⁻⁶moll⁻¹; N=8) was added to the perfusate. 30min later, the four different catecholamine injections were administered in a random order. Controls (N=8) were performed over the same period by repeating the above protocol without adding an adrenoceptor antagonist.

In the second set of experiments (N=10), phenylephrine (10.0 and 100 nmol; Neosynephrine, Winthrop Laboratories, Inc.), isoproterenol (0.1, 1.0 and 10.0 nmol; Sigma Chemical Co.) and acetylcholine (0.1,1.0 and 10.0 nmol; Sigma Chemical Co.) were injected into the perfusion line in random order. Drugs were injected only after CVR had returned to the basal level after the preceding injection (approximately 5–20min). Then KC1 was added to the perfusate to produce a KC1 concentration of 40mmoll⁻¹. The added KC1 elevated CVR. This was done to determine if the small decreases in CVR caused by isoproterenol and acetylcholine could be augmented by first increasing CVR. Then the same doses of isoproterenol and acetylcholine were administered as before. Finally, propranolol (10⁻⁶moll⁻¹) was added to the perfusate and 30min later the three doses of isoproterenol were given to determine if the responses to isoproterenol were mediated by β-adrenoceptors.

Cutaneous vascular resistance was determined by dividing perfusion pressure by perfusate flow and then subtracting the resistance of the arterial cannula. Cannula resistance was determined after the experiment by cutting the pulmocutaneous artery at the cannula tip and dividing the resulting perfusion pressure by the flow.

Responses to nerve stimulation and drug injections into the perfusion line were expressed as the maximal percentage change in CVR recorded after the stimulus. Cumulative dose-response curves to EPI and NE were analyzed for (1) the minimum concentration required for a significant increase in CVR, (2) the maximal CVR caused by the catecholamines and (3) the median effective doses (ED50), calculated by expressing CVR as a percentage of the maximal response, performing a logit-log transformation on the data and then fitting a line to the data with a least-squares linear regression.

Paired t-tests were used to evaluate: (1) differences in responsiveness to catecholamine injections caused by adrenoceptor antagonists and by time; (2) the effects of adrenoceptor agonists and acetylcholine; and (3) the difference in basal CVR at the beginning and at the end of the experiment. The maximal response to, and ED50 values of, EPI and NE were compared with an unpaired t-test. One-way ANOVA, followed if necessary by a Newman-Keuls test, evaluated differences in responsiveness to (1) nerve stimulation caused by atropine and/or phentolamine and (2) the two doses of α,β-methylene ATP. Statistical significance was set at a level of P=0.05.

At the beginning of the experiments, mean (±S.D.) perfusion pressure was 2.3±0.6kPa and CVR was 3.1 ±1.5 kPa min kg body mass ml⁻¹. In experiments without sustained increases in CVR due to the addition of KC1 or catecholamines to the reservoir of perfusate there was no change in basal CVR between the beginning and the end of the experiment (P>0.5).

Changes in CVR produced by nerve stimulation are shown in Figs 3 and 4. Stimulation of the vagal ganglion, sympathetic ganglion 1 and cranial nerve X always produced increases in CVR. Sympathetic ganglion 2 stimulation increased CVR in two frogs but had no effect in the other two. Stimulation of cranial nerve IX and spinal nerves 1 and 2 had no effect on CVR in the four frogs tested. The elevation of CVR in response to nerve stimulation lasted several minutes. There was no apparent difference in the time course of the response between vagal and sympathetic ganglia stimulation. Stimulating sympathetic ganglion 2 after cutting between the first and second sympathetic ganglia never altered CVR. This sectioning had no effect on changes in CVR caused by stimulation of sympathetic ganglion 1. Cutting between sympathetic ganglion 1 and the vagal ganglion eliminated the response to stimulation of sympathetic ganglion 1, but had no effect on the response to stimulation of the vagal ganglion. Sectioning cranial nerve X lateral to the vagal ganglion eliminated the response to stimulation of the vagal ganglion but had no effect on the response to stimulation of cranial nerve X.

Addition of phentolamine to the perfusate decreased the response to stimulation of sympathetic ganglion 1 by 78 % (P<0.01), but had no significant effect on the response to stimulation of cranial nerve X (P>0.07; Fig. 5). Subsequent addition of atropine to the perfusate (already containing phentolamine) did not change the response to stimulation of either of these nerves (P>0.1; Fig. 5). Adding atropine to the perfusate before phentolamine had no effect on the response to nerve stimulation (P>0.1; Fig. 5). Adding phentolamine after atropine caused a 68% decrease in the response to stimulation of sympathetic ganglion 1 (P<0.01), but did not alter the response to stimulation of the cranial nerve X (P>0.1; Fig. 5). Neither 10⁻⁴ nor 10⁻³moll⁻¹ α-,β-methylene ATP affected the response to stimulation of sympathetic ganglion 1 and cranial nerve X (P>0.1). Atropine, phentolamine and α-, β-methylene ATP had no effect on basal CVR (P>0.1).

Addition of EPI and NE to the perfusate caused dose-dependent increases in CVR up to a concentration of approximately 10⁻⁵moll⁻¹ (Fig. 6). Higher catecholamine concentrations caused CVR to fall. The lowest concentration of EPI tested which significantly raised CVR was 1.8×10⁻⁷moll⁻¹ (P=0.048), while that for NE was 1.8×10⁻⁶moll⁻¹ (P=0.003). There was no significant difference in the maximal CVR caused by EPI and NE (P>0.5). The EDso for EPI at 2.1±0.7x10⁻⁶moll⁻¹ was significantly different (P=0.035) from the ED₅₀ for NE at 4.8±0.9×10⁻⁶moll⁻¹.

Injections of EPI and NE into the perfusion line caused transient increases in CVR lasting for approximately 5–20 min. In the control experiments, the magnitude of the responses to repeat injections after 30 min was not different from that of the initial injections (P>0.29; Fig. 7). Phentolamine reduced the responses to NE and EPI (P<0.05), except for the response to 0.1 nmol NE (P>0.21; Fig. 7). Propranolol had no effect on the responses (P>0.1; Fig. 7). Neither antagonist had an effect on basal CVR (P>0.4).

1 and 10 nmol of phenylephrine caused significant increases in CVR of 11.2±4.3 and 28.1 ±7.0 %, respectively (P<0.034). Without KC1 (40 mmol I⁻¹) added to the perfusate to raise CVR, isoproterenol decreased CVR at the three doses tested (P<0.03, Fig. 8). However, the mean changes were always less than 9%. Addition of KC1 to the perfusate raised CVR by 76.9±20.2%. Administration of isoproterenol under these conditions of elevated basal CVR also decreased CVR, but the percentage decrease in CVR was not greater than before KC1 was added (Fig. 8). Propranolol abolished the changes in CVR caused by isoproterenol (P>0.4; Fig. 8). Without extra KC1 added to the perfusate, acetylcholine decreased CVR by 7 % (P<0.002) at the lowest dose tested (0.1 nmol; Fig. 8). The two higher doses (1 and 10 nmol) of acetylcholine had no effect on CVR (P>0.09). With CVR elevated by the addition of KC1 to the perfusate, acetylcholine caused significant decreases in CVR at the two lowest doses (P<0.004; Fig. 8). The magnitudes of the decreases were under 13 %. The highest dose of acetylcholine had no effect on CVR (P=0.08).

In the intact bullfrog, most blood flowing through the cutaneous artery goes to the skin on the sides and back, although some of this blood reaches other skin areas (Moalli et al. 1980). In the preparation of this study, only the skin of the sides and back was perfused. It is not known why the cutaneous artery supplies other areas in vivo but not in a pump-perfused preparation.

The preparation remained viable throughout the experimental periods. There was no decrease in responsiveness to EPI and NE injections in the control experiments. Responsiveness to vagal stimulation did not decrease over time and responsiveness to stimulation of sympathetic ganglion 1 decreased only after the addition of phentolamine to the perfusate. In addition, there was no change in basal CVR during the experiments.

It is not known how the basal contractile state of the cutaneous vasculature in this preparation compares to that in the intact animal. However, there was some myogenic tone in the vasculature because isoproterenol and acetylcholine produced decreases in CVR. The basal resistance of the perfused preparation was probably greater than the resistance of the skin vasculature perfused by the cutaneous artery in vivo. Moalli et al. (1980) reported cutaneous arterial flows of 2.6 ml min⁻¹ kg⁻¹ during air-breathing and 1.3 ml min⁻¹ kg⁻¹ during diving in intact animals. In this study, mean basal perfusion flow was 0.74mlmin⁻¹ kg⁻¹. The reason for this difference may, in part, be related to the smaller skin region supplied by the cutaneous artery in this preparation than in vivo.

The magnitude of the CVR increases observed with this preparation may have been less than would have been obtained if a constant-pressure, variable-flow setup had been used. In this preparation, for example, an increase in CVR causes perfusion pressure to increase, and such an increase may partially counteract the increase in CVR by acting to distend vessels. The degree of this possible effect of changing perfusion pressure on CVR is not known. However, a variable-pressure preparation may approximate some physiological situations; for example, lung ventilation in amphibians causes blood pressure to increase (Segura et al. 1981) and cutaneous blood flow to decrease (Moalli et al. 1980; Malvin and Hlastala, 1986b).

This study showed that the bullfrog cutaneous circulation receives sympathetic innervation because stimulation of sympathetic ganglia 1 increased cutaneous vascular resistance. In half of the animals tested, stimulation of sympathetic ganglion 2 produced a response indicating that in some animals cutaneous sympathetic nerves are associated with that ganglion as well. It was shown that these nerves travel anteriorly in the sympathetic chain to join the vagus, because cutting the chain posterior to the stimulation site had no effect on the response to stimulation, but cutting anterior to the stimulation site abolished the response. It is likely that these nerves release EPI, which stimulates (adrenoceptors. Epinephrine is the most common catecholamine released by adrenergic nerves in amphibians (Bumstock, 1969) and most of the response to stimulation of sympathetic ganglion 1 was blocked by phentolamine. It is not known whether the remaining response after blockade was due to incomplete blockage of a- adrenoceptors or to other transmitters.

Sympathetic innervation of the cutaneous circulation exists in other amphibians as well. Early studies on Rana temporaria showed that stimulation of the posterior sympathetic ganglia produced vasoconstriction in the skin of the foot web (Langley, 1911; Krogh et al. 1922). Using a preparation similar to the one of this study, Smith (1976) stimulated the cervical sympathetic connective in the toad Bufo marinus and increased vascular resistance in the skin supplied by the cutaneous artery. In addition, he demonstrated histologically the presence of adrenergic varicosities in the proximal segment of the cutaneous artery.

The lack of an effect of phentolamine on the response to vagal stimulation indicates that non-adrenergic nerves also travel in the vagus and innervate the cutaneous vasculature. These nerves are probably not cholinergic because atropine had no effect on the response to vagal stimulation. In addition, they do not appear to use ATP as a transmitter, since a,/8-methylene ATP had no effect on vagal stimulation. It is possible that the transmitter is a peptide, but that possibility was not evaluated.

Both EPI and NE are pressor agents in the cutaneous vasculature, suggesting that circulating catecholamines may directly influence CVR. However, plasma [EPI] and [NE] in resting bullfrogs are 2.1×10⁻⁹moll⁻¹ and 4.8×10⁻¹⁰moll⁻¹, respectively (Wilson et al. 1988), several orders of magnitude less than the minimum concentrations required for a significant increase in CVR in this present study. Consequently, if the sensitivity of the cutaneous vasculature to catecholamines is the same in vivo as that measured in this study, then these hormones do not directly influence CVR in the resting animal. Various stresses increase plasma [EPI] and [NE], so it is possible that under some circumstances plasma [EPI] and [NE] could rise enough to affect CVR. Unfortunately, the amount that circulating catecholamines levels can rise in the bullfrog is not known. In other amphibian species, the increase in plasma catecholamine levels in response to exercise and handling is less than 10-fold (Wahlqvist and Cambell, 1988; Bourgeois et al. 1978). Thus, a direct role for circulating catecholamines in the control of CVR is doubtful.

The cutaneous vasculature appears to contain both α- and β-adrenoceptors. The presence of α-adrenoceptors is supported by phentolamine antagonizing the responses to sympathetic nerve stimulation and to EPI and NE injection, and by the increase in CVR induced by phenylephrine, β-Adrenoceptors may also be present because isoproterenol produced decreases in CVR which were antagonized by propranolol. However, it is unlikely that β-receptors play a significant role in regulating CVR. Decreases in CVR caused by isoproterenol were small (<10%) and propranolol had no effect on the responses to EPI and NE. α-Adrenoceptors are also probably present in the foot web. Burggren and Moalli (1984) blocked a decrease in the number of perfused capillaries in the bullfrog web caused by air exposure with phenoxybenzamine, an α-adrenoceptor antagonist.

Significant control of CVR by cholinergic nerves is unlikely. Atropine had no effect on the responses to vagal stimulation, indicating the absence of cholinergic vagal control. Direct cholinergic control of CVR by nerves not associated with the vagus is also unlikely. Injections of acetylcholine into the perfusion Une caused significant, but small (<10 %), decreases in CVR. Thus, if cholinergic innervation of the cutaneous vessels exists, it probably has only a minor influence on CVR.

Although this study has described several aspects of possible autonomic regulation of CVR, the conditions under which such regulation may be used are still not known. Stimuli that alter cutaneous blood flow in amphibians include diving (Moalli et al. 1980), environmental hypoxia (Malvin and Hlastala, 1986a, 1989), environmental hypercapnia (Malvin and Hlastala, 1986a), lung ventilation (West and Burggren, 1984; Malvin and Hlastala, 1986b) and exposure to air (Burggren and Moalli, 1984). It is possible that autonomic control over CVR is involved in these changes of cutaneous perfusion but this remains to be tested.

This research was supported by NIH grant HL38942.