A Possible Role of the Glomus Cell in Controlling Vascular Tone of the Carotid Labyrinth of Xenopus Laevis

Tohoku J. exp. Med., 1987, 151, 395-408

A Possible Role of the Glomus Cell in Controlling Vascular Tone of the Carotid Labyrinth of Xenopus laevis

TATSUMIKUSAKABE, * KAZUKOISHII and KOSEI ISHIIt Department of Physiology, Fukushima Medical College, Fukushima 960

KUSAKABE,T., ISHII,K. and Isxu, K. A Possible Role of the Glomus Cell in Controlling Vascular Tone of the Carotid Labyrinth of Xenopus laevis. Tohoku J. exp. Med., 1987, 151(4), 395-408 To clarify the physiological significance of the g-s connection (intimate apposition of the glomus cell to the smooth muscle) in the Xenopus carotid labyrinth, experiments were carried out morphologically and physiologically. Results obtained are as follows. 1. Efferent electrical stimulation of the glossopharyngeal nerve resulted in concentrating dense-cored vesicles on the peripheral region of the glomus cell, and a decrease of vesicles as a whole. 2. In the carotid labyrinth perfused artificially, outflow of the internal and the external carotid arteries decreased with administration of catecholamines (adrenaline, noradrenaline and dopamine). 3. Acetylcholine reduced only the internal outflow. This response was depressed by atropine, hexamethonium and phenotolamine, whereas accelerated by propranolol. 4. Sodium cyanide reduced the internal outflow without affecting the external outflow, and its effect is depressed by phentolamine. From these results, a possibility that the glomus cell participates in controlling the blood flow in the labyrinth through the intervention of the g-s connection was discussed. Xenopus laevis ; carotid labyrinth ; glomus cell ; catecholamine ; arterial chemoreceptor

The carotid labyrinth of amphibia has a chemoreceptor function (Ishii et al. 1966) and contains glomus cells similar to those of many animal species in fine structure (Rogers 1963; Ishii and Oosaki 1969; Poullet-Krieger 1973). Recent- ly, an electron microscopic study of Xenopus laevis carotid labyrinth (Ishii and Kusakabe 1982) revealed that some of glomus cells were intimately connected with the smooth muscle in the labyrinth (g-s connection). Moreover, efferent electrical stimulation of the glossopharyngeal nerve resulted in facilitation of exocytosis of dense-cored vesicles at the g-s connection. These findings suggest that some of glomus cells may affect the vascular tone of the labyrinth. In Rana temporaria, Banister et al. (1967) confirmed fluorescent cells in the carotid labyrinth, and by perfusion experiment Banister et al. (1975) have proposed that these

Received December 3, 1986; accepted for publication February 23, 1987. Present address : Department of Anatomy, Yokohama City University School of Medicine, Yokohama, 236, Japan ; tWatari Nakaecho 9-2, Fukushima, 960, Japan. 395 396 T. Kusakabe et al. fluorescent cells might be responsible for controlling blood flow in its vascular network. Their results have been discussed by Smith et al. (1981) based on the perfusion and tissue culture experiments of the labyrinth of Bufo marinus. In the present study, the distribution of dense-cored vesicles and its change after nerve stimulation were morphologically studied, and vascular responses to some sympathomimetic substances and arterial chemoreceptor stimulants were inves- tigated in the perfusion experiments. Results suggested that the glomus cell participates in controlling the blood flow in the carotid labyrinth.

MATERIALS AND METHODS Electron microscopy. Eight male Xenopus laevis weighing 40-60 g were used for electron microscopic study. The region of the carotid labyrinth was exposed on both sides of pithed animals. The right glossopharyngeal nerve was cut just below the jugular ganglion and isolated from surrounding tissues. Cutting all its branches with the carotid nerve left intact, the nerve was subjected to electrical stimulation. The left labyrinth was kept intact for the control. Electrical stimulation was done by rectangular wave pulses for 1 min in the following parameters ; 2 volts, 1 msec, 20 Hz. A thin nylon tube was inserted into the aortic trunk to perfuse the carotid labyrinth. After stimulation the labyrinths on both sides were washed with heparinized Ringer solution, then perfused for 10 min with 2% glutaraldehyde in 0.1 mole cacodylate buffer (pH L3) at a pressure of 50 cmHZO. Then, the labyrinths were removed from the body and immersed in the fresh fixative for 3 hr. The specimen was divided into small blocks and postfixed in osmium tetroxide buffered with 0.1 mole cacodylate for 3 hr and embedded in Epon 812. Sections of 11am in thickness were stained with toluidine blue for light microscopy. Ultrathin sections were serially cut from the block in which glomus cells were confirmed by light microscope in advance, and stained with 1% uranyl acetate and Reynold's lead solution. Observation was done by a Hitachi

Fig. 1. Diagram of the arrangement for artificial perfusion of the carotid labyrinth. R1 and R2, reservoir bottles ; T1, T2 and T3, three-way cocks ; S, small compartment (0.25 ml) partitioned by TZ and T3 in which stimulants were filled. With turning the three way cock (T1) stimulants reach the labyrinth. labyr, carotid labyrinth ; c.c.a, common carotid artery ; e.c.a, external carotid artery ; i.c.a, internal carotid artery. Xenopus Carotid Labyrinth Controlling Blood Flow 397

HU-11 electron microscope. To measure distribution of dense-cored vesicles in the cytoplasm, photographs of the glomus cell were classified in 3 parts ; the part containing nucleus in its cytoplasm (N), without nucleus (M), and the cell processes (P). The density of vesicles per l ,um2was counted in each regions and the mean densities were compared among regions, and between the control and the stimulated groups. The significance of differences was tested by Student's t-test. Perfusion experiments. Twenty three pithed animals of both sexes weighing 40-60 g were used. As shown in Fig. 1, the experimental arrangement was composed of two reservoir bottles. Both bottles were filled with about 100 ml of modified Ringer solution and connected with a three-way cock (T1) ; the one (R1) directly, the other (R2) with insertion of another 2 three-way cocks (T2, T3), and joined to a fine nylon tube inserted into the common carotid artery. It was designed to perfuse the labyrinth with a fixed volume of stimulants stored in the space between T2 and T3 (ca. 0.25 ml) by turning the cock T1. By flowing Ringer solution in the control bottle (R1) the labyrinth was washed up. In the case of testing effects of some blockers on responses to stimulant substances, both bottles (R1, R2) were filled with the saline containing the blocker in a certain concentration, and the stimulant substance diluted with the blocker containing saline was put in the small compartment between T1 and T2. Thus, a fixed volume of stimulant was given during administration of the blocking agent. Fine stainless steel tubes were inserted into both the internal and the external carotid arteries, respectively, and drops escaping from the tip of them were led to a drop-counter and recorded on magnetic tape for the computer analysis. The pressure of perfusion was selected between 15-20 cmH20. For comparison, responses to stimulation were quantitatively expresed according to the formula R.O.E = (Y-.-X)/X (rario of effect), where X is the mean drop number per min during the control period for 2 min and Y is the mean drops per min for 3-5 min or more from 1 min after starting the stimulation. Thus, R.O.E was positive for vasoconstriction and negative for vasodilatation. All these procedures were processed by a minicomputor (TI-980B). The composition of Ringer solution was as follows : NaCI, 112 mM ; KCI, 3.0 mM ; CaCl2, 2.2 mM ; NaHC03, 1.2 mM ; glucose, 0.6 mM ; albumin (fraction V), 5 g/liter ; heparin, 2,000 u/liter.

RESULTS Electron microscopicobservation Distribution of dense-coredvesicles in the intact glomus cell. In micrographs taken from 8 carotid labyrinths of the control side, the number of pictures of glomus cells were 184 in total including all N-, M- and P- regions. Among these 3 regions there is a dissimilarity in the density of vesicles in the cytoplasm. As shown in Fig. 3A, the mean density was minimum in N(7.43± 2,l0/,u m2, n = 74), maximum in P (21.94 + 8.73/p m2, n = 41), and that in M(9.31+ 3.30/,um2, n = 69) lay between N and P. The ratio of the density in N to that in P was about 3. All of these differencesof mean densities among these regions were significant (p < 0.01). These results suggested that dense-coredvesicles have a tendency to aggre- gate in the glomus cell process. Distribution of dense-coredvesicles after efferentstimulation of the glossopharyngeal nerve. Efferent stimulation of the glossopharyngeal nerve resulted in a significant (p <0.01) decrease of vesicles in the glomus cell (Fig. 2). On 101 micrographic figures of the glomus cell mean densities of vesicles in each regions were measured as 1.99+ 0.50/,um2 (n -42) in N, 3.14± 1.60/,um2 (n = 40) in M and 398 T. Kusakabe et al.

Fig. 2. Electronmicrograms showing the change of the number of dense-cored vesicles in the glomus cell after efferent stimulation of the carotid nerve, a, control ; b, after stimulation.

Fig 3. Densities of dense-cored vesicles in N, M and P. A, control ; B, after simulation.

8.43 + 2.62/1am2 (n =19) in P. As summarized in Fig. 3B, vesicles in all 3 regions decreased with nerve stimulation as a whole, and the inequality existing in regions was augmented ; density ratio N to P increased from 3 to 4. Xenopus Carotid Labyrinth Controlling Blood Flow 399

Perfusion experiments of the carotid labyrinth Effects of catecholamines The glomus cells contain some catecholamines in their granules and are intimately apposed with the smooth muscles which are one of structural elements of the labyrinth. Together with the fact that efferent stimulation of the glossopharyngeal nerve facilitated exocytosis of dense vesicles, the results described above suggest that activity of the smooth muscle in the labyrinth should be influenced by catecholamines released from the glomus cells. Perfusing the labyrinth artificially, we tested the effect of some catecholamines on the outflow from the internal and the external carotid arteries. Adrenaline (AD). Adrenaline produced a marked decrease in the outflow from the internal carotid artery in most labyrinths (25/32) but an increase in some (5/32). As shown in Fig. 4, when adrenaline 2.5><10-6 g was administered , the internal outflow began to decrease with a latency of 2 min, reached a minimum in 5 min, and then gradually returned to the control level in about 30 min. The threshold dose was 2.5 X 10-g g and responses were dose-dependent. In 5 of 32 labyrinths, an increase in the outflow developed more slowly, with the threshold nearly the same as that for a decrease. The outflow from the external carotid artery decreased slightly (Fig. 4), but in most labyrinths it seemed to depend on the response of the internal carotid artery, namely it increased slightly when the outflow from the internal carotid artery decreased strongly and vice versa (Fig. 5). In the remaining 2 labyrinths no response was provoked. On the responses to AD the effect of blocking agents was studied ; by administrating phentolamine (Phentl) to the labyrinth responding vasoconstrictively and propranolol (Prop) to that provoking vasodilatation. After perfusing with Phentl (5 X 10-7 g/ml) for 30 min a decreasing response of the internal flow turned to an increasing one reversibly in all 7 labyrinths tested (Fig. 5A), and the perfusion of Prop (10-6 g/ ml) for 30 min converted reversibly the vasodilatation to the vasoconstriction in all 5 labyrinth tested (Fig. 5B).

Fig .4. Effect of adrenaline (AD) the outflow from the internal and external carotid arteries. xxxxx, drops from the internal carotid artery ; 00000, drops from the external carotid artery. At the arrows AD (2.5 x 10-6 g) was administered into the labyrinth. 400 T. Kusakabe et al.

Fig . 5. Effect of adrenergic blockers on the vascular response of the labyrinth to AD administration and its graphical representation by ROE. A-1: Vasoconstrictor effect of AD in the control. AD (2.5 X 10-' g) provoked strong vasoconstriction on the internal and very slight vasodilatation on the external carotid artery. A-2: Thirty min after turning to saline containing Phentl (5 x 10-' g/ml), AD (1.25 x 10-' g) increased the internal outflow and decreased very slightly the external outflow. B-1: Vasodilatator effect of AD in the control. AD (1.25 x 10-6 g) provoked vasodilatation on the internal and very slight vasoconstriction the external carotid artery. B-2: Thirty min after turning to saline containing Prop (10-6 g/ml), AD (1.25 x 10-6 g) decreased the internal outflow and increased slightly the external outflow. xxxxx, drops from the internal carotid artery ; ooooo, drops from the external carotid artery. At arrows AD was administered. Bottom : Graphic representation of effect of blockers on vascular responses in A and B. Plain columns correspond to 1 in A and B, shadow-columns correspond to 2 in A and B, respectively. Upward columns indicate the vasoconstriction and downward columns indicate the vasodilatation.

Noradrenaline (NA). Both an increase in the internal outflow were provoked by NA administration about in a half, and a decrease in a half of labyrinths tested. In 13 of 29 labyrinths a decrease occurred dose-dependently with the threshold dose of 2.5 x 10_8 g, and was reversed to an increase after perfusion with Phentl (5x 10-' g/ml) for 30 min in all 4 labyrinths tested (Fig. 6A). In 12 other labyrinths, an increase was provoked, which turned to a decrease after Prop (10-6 g/ ml) in all 4 labyrinths tested, (Fig. 6B). In one labyrinth the constriction was followed by the dilatation. In the remaining 3, no response was provoked. The X enopus Carotid Labyrinth Controlling Blood Flow 401

Fig. 6. Vascular responses to NA and the effect of adrenergic blockers, and their graphical representation. A-l : Vasoconstrictor response in the control. NA (2.5 x 10-' g) decreased the outflow of the internal carotid artery without accompanying a definite effect on the external. A-2: thirty min after turning to saline containing Phentl (S x 10-' g/ml), NA (2.5 x 10-' g) increased the internal outflow accompanying a slight decrease in the external outflow. B-1: Vasodilatator response in the control. NA (1.25 x 10-6 g) increased the internal outflow. B-2: Thirty min after turning to saline containing Prop (10-6 g/ml ), NA (1.25>< 10-6 g) decreased the internal outflow. In both cases the external outflow did not change. xxxxx, drops from the internal carotid artery. Cam, drops from the external carotid artery. At arrows NA was administered. Bottom : Graphic representation of results in A and B as the same as in Fig. 5. outflow from the external did not react markedly. Dopamine (DA). In 8 labyrinths obtained from 5 frogs, dopamine was tested. In 3 labyrinths the outflow from the internal decreased nearly in the same way as in the administration of Ad (Fig. 7A), but the threshold concentration was about 10 times of that of AD. In one labyrinth vasodilatation occurred for a short time (Fig. 7B). Two labyrinths responded to dopamine in a diphasic fashion ; the constriction was followed by the dilatation. In another 2, no change occurred. The external outflow did not respond delectably to DA administration in all the labyrinths. Phentl (5><10-' g/ml) converted the decreasing response to the increasing in all 3 labyrinths, and Prop (10-6 g/ml) turned the increasing response observed in one labyrinth into the decreasing (Fig. 7A, B). In these experiments with catecholamine administration the outflow from the 402 T. Kusakabe et al.

Fig. 7. Responses to DA and the effect of adrenergic blockers. A-1: Decreasing effect of DA (2.5 x 10-6 g) on the outflow of the internal carotid artery. A-2: After perfusion of the saline containing Phentl (5 x 10-' g/ml) for 30 min, DA increased outflow of the internal artery. B-1: Increasing effect of DA (2.5 x 10-6 g) on the outflow of the internal artery. B-2: Prop (10-6 g/ml) for 30 min, DA (2.5 x 10-6 g) decreased outflow of the internal artery. No detectable effect of DA was seen in the external outflow. xxxxx, drops from the internal carotid artery. ooooo, drops from the external carotid artery. At arrows DA was perfused. Bottom : Graphic representation as in Fig. 5.

external carotid changed only slightly, but in most cases the changes occurred in an opposite direction to those in the internal carotid, suggesting that they were passive ones. Since the external carotid artery originates directly from the common carotid artery without intervention of vascular maze, responses of the internal outflow to catecholamines described above were attributable to the response of smooth muscles in maze structure where chemoreceptors resided. Therefore, the effects of some chemoreceptor stimulant substances on labyrinth vascularity were tested. Effects of acetylcholine (Ach) and Sodium cyanide (NaCN) It is well known that Ach and NaCN are strong stimulants to the arterial chemoreceptors, though their mechanisms of action are as yet unclear. Ach. In all 22 labyrinths obtained from 13 frogs the outflow from the internal carotid was markedly reduced by perfusion with Ach, 1.25 x 10_9 g, while in the outflow from the external no detectable change was observed (Fig. 8A). Responses occurred in dose-dependency with the very low threshold (2.5 x 10-'s g). Xenopus Carotid Labyrinth Controlling Blood Flow 403

Fig. 8. Effect of Ach on the flow from the internal carotid artery and its modification by various blockers. A-1: Control. Administration of Ach (1.25 x 10_9 g) decreased remarkably the outflow of the internal artery. A-2: After perfusion with Atr (10-6 g/ml) for 30 min, Ach was administered. Vasoconstriction did not occur. A-3: Diagram showing dose-dependency of responses to Ach. B-1: Control. Administration of Ach (1.25 x 10_9g). B-2: After perfusion with C6 (10-5 g/ml) for 30 min. Effect of Ach was almost completely inhibited. B-3: After washing for 60 min. Response recovered. C-1: Control. Administration of Ach (1.25 x 10-9 g). C-2: After perfusion with Phentl (5 x 10-' g/ml) for 30 min. Response to Ach was strongly depressed. C-3: After washing for 60 min. Response recovered again. D-1: Control. Administration of Ach (1.25 x 10_9 g). D-2: After perfusion with Prop (10-6 g/ml) for 30 min. Vasoconstriction was intensified. D-3: After washing for 60 min. Response returned to control level. Bot- tom : Graphic representation of results in above tracings. A, B, C and D correspond, respectively. 404 T. Kusakabe et al.

Fig . 9. Effect of NaCN on the flow in the carotid labyrinth and its disapearance with Phentl. A-1: Control. NaCN (1.25 x 10 g) decreased the flow from the internal carotid remarkably. A-2: After perfusion with Phentl (5 x 10-' g/ml) for 30 min. The response to NaCN in the same dose was completely abolished. A-3: After washing sufficiently, NaCN (1.25 x 10-5 g) was ad- ministrated. Vasoconstriction appeared again. xxxxx, drops from the internal carotid artery. 0000x, drops from the external carotid artery. At arrows NaCN was perfused. B : Diagram showing dose-dependency of the response to NaCN. C : Graphic representation of A. Plain columns correspond to 1 and 3, and the shadowed corresponds to 2, respectively, with upward columns for vasoconstriction and the downward for vasodilatation as the same as in Fig. 5.

Perfusion with atropine (Atr) in 10-6 g/ml in 3 labyrinths, hexamethonium (C6) in 10-' g/ml (in 8 labyrinths) or Phentl in 5 x 10 -' g/ml (in 3 labyrinths) for 30 min always depressed the effect of Ach definitely (Fig. 8A, B, C). To the con- trary, it was augmented by perfusion with Prop (10-' g/ml) for 30 min (Fig. 8D). After washing with Ringer solution for about 1 hr responses to Ach restored nearly the same as in the control. NaCH. When NaCN (1.25 x 10-s g) was administered, the outflow from the internal carotid reduced rapidly in 17 labyrinths of 10 animals (Fig. 9A-1). With the threshold of 2.5 x 10-5 g the response was dose-dependent (Fig. 9B). After perfusion with Phentl (5 x 10-' g/ml) for 30 min NaCN no longer provoked the vasoconstriction in all 4 labyrinths (Fig. 9A-2), while perfusion with C6 exerted no effect on the vascular response to NaCN.

DISCUSSION Secretory function of glomus cell has been suggested in mammals (Biscoe and XenopusCarotid Labyrinth Controlling Blood Flow 405

Stehbens 1966 ; Kobayashi 1968) and in fowls (Hodges et al. 1975 ; King et al. 1975). In a previous paper, Ishii and Kusakabe (1982) reported on a secretory function of the glomus cell of the carotid labyrinthin in Xenopus laevis. Accord- ing to their report the glomus cell was intimately connected to the smooth muscle (g-s connection), and the glossopharyngeal nerve stimulation provoked exocytosis of dense-cored vesicles at g-s connection together with the reduction in number of dense-cored vesicles, suggesting that a target of this secretory function may be smooth muscle in the labyrinth vascularity. In the present study it was quantitatively demonstrated that dense-cored vesicles distributed more densely in the peripheral portion of the cell, and that the glossopharyngeal nerve stimulation gave an impetus to it. Decreasing in number of dense-cored vesicles by nerve stimulation has been reported in the previous paper by Ishii and Kusakabe (1982) and in mammals by Yates et al. (1970). It has been estimated in some species of amphibians that catecholamines involved in dense-cored vesicles in the glomus cell are adrenaline, noradrenaline and dopamine. Nerve stimulation reduced both the outflows from the internal and the external carotid anteries (Banister ; in personal communication). In the present experiments the administration of these catecholamines, Ach and NaCN predominantly affect the outflow from the internal carotid artery, in opposition to the result obtained by Smith et al. (1981) in the toad (Buf o marinus) that the administration of catecholamines or Ach always reduced the flow from the external carotid artery, while increased the flow from the internal carotid in conse- quence of the augmentation of total resistance. Accordingto them the contractile element must be located not at the region associated with the internal carotid artery where the glomus cells are situated, but at the external carotid artery. In the carotid labyrinth of Xenopus laevis the external carotid artery is directly branched from the common carotid artery without constructing maze structure, and only the internal carotid artery is originated from maze structure. Thus responses of outflow from the internal may be attributable to those of maze vascularity. In addition to the vasoconstriction, which was abolished by Phentl, these catecholamines provoked vasodilatation, which was antagonized by Prop. Burn and Robinson (1951) reported in the rabbit ear vessel that the vasoconstrictor response to NA or AD was turned to dilatation by adding histamine to the perfusate. It may be considered that vasodilatatooy response to catecholamines in the present study was carried out in the same mechanism, but the fact that NA provoked more often vasodilatation than AD and DA suggests the peculiarity of smooth muscle in the maze structure of Xenopus carotid labyrinth. DA, which has effects on a-and $-adrenergic receptors and is known as an inhibitor to chemoreceptor excitation in mammalian carotid body, also provoked both the constriction and the dilatation of labyrinth vascularity with higher threshold dose than other catecholamines. Although precise information on membranous mecha- 406 T. Kusakabe et al. nisms in their working site is deficient, at least we can state that catecholamines involved in the glomus cell, AD, NA and DA, have both a- and /3-effects on smooth muscles in the labyrinth. Ach provoked vasoconstriction in very low dose, and it was depressed by Phentl, C6 and Atr, and exagerated by Prop (Fig. 8), suggesting the intervention of some adrenergic mechanisms. Hatakeyama and Kato (1958, 1959, 1960) reported that a great deal of vascularity of amphibians was usually constricted by Ach, and they attributed this to the release of Ach at sympathetic nerve terminals. Ferry (1963) reported that high dose of Ach excited the sympathetic post gang- lionic nerves somewhere near their endings and this was not affected by Atr, but was abolished by C6 administration. In the present experiment, however, the constriction was provoked in very low dose (2.5 X 10-13 g in threshold dose), which seemed to be too low to stimulate smooth muscle or sympthetic nerve terminals directly in view of reports by Hatakeyama and Kato (1958, 1959, 1960) and Ferry (1963). Regarding effects of adrenergic blocking agents and the electron- microscopic observation that the preganglionic efferent terminal contacted to the glomus cell with synaptic specialization (Ishii and Oosaki 1969), it may not be unlikely to suppose that Ach stimulates the nicotinic receptor of glomus cell, consequently catecholamines were released from the glomus cell to provoke vasoconstriction. The degeneration experiment in the toad demonstrated this terminal to be sympathetic (Ishii and Ishii 1973). Therefore, the glomus cell was possibly endowed a function as a booster cell emphasizing sympathetic vasoconstrictive action to the smooth muscle in the labyrinth. On the other hand, there is no doubt on the arterial chemoreceptor function of the carotid labyrinth (Ishii et al. 1966). Although sensory mechanisms are still in enigma, the glomus cell has been now accepted as a local control element for chemoreceptor excitation rather than as receptor cell, though not conclusive. Being stimulated by severe hypoxia and hypercapnia, reciprocal nerve terminals, which were found on the glomus cell in synaptic contact (McDonald and Mitchell 1975), send afferent impulses to the respiratory center, while they release Ach efferently to the glomus cell. Thus excited glomus cell released catecholamines to the nerve terminal to inhibit feedbackly the chemoreceptor excitation. NaCN, commonly used as chemoreceptor stimulant reduced the outflow from the internal carotid. This reaction was not blocked by C6, but blocked by Phentl, suggesting the intervention of the adrenergic mechanism. In our few experiments, the perfusion of N2-saturated saline provoked slight vasoconstriction (unpublished data). Recently Rigual et al. (1986) reported that hypoxic stimulation increased the rate of release of dopamine in cat carotid body. When the chemoreceptor is strongly stimulated, augmentation of the flow in the labyrinth should be provoked by reflex hyperventilation and hypertention. Then, the chemoreceptor excitation would be depressed by local feedback control described above, concurrently the released catecholamines would provoke vasoconstriction of the carotid labyrinth Xenopus Carotid Labyrinth ControllingBlood Flow 407 through the g-s connection, resulting in slowing down the blood flow to the brain to adequate level. The glomus cell is an elemental component of the chemorecep- tion, on the other hand, it participates for an another function of the labyrinth which has been suggested by Carman (1955), Ishida (1954) and Banister et al. (1975), to control the blood flow of the internal carotid artery.

References 1) Banister, R.J., Portig, P.J. & Vogt, M. (1967) The content and localization of catecholaminesin the carotid labyrinths and aortic arches of Rana temporaria. J. Physiol. (Loud.), 192, 529-535. 2) Banister, J., Herpol, C. & Nail, B.S. (1975) Vascularflows in the carotid labyrinth of Rana. In : The Peripheral Arterial Chemoreceptors,edited by M.J. Purves, Cam- bridge University Press, London-NewYork, pp. 315-323. 3) Biscoe, T.J. & Stehbens,WE. (1966) Ultrastructure of the carotid body. J. Cell Biol., 30, 563-578. 4) Burn, J.H. & Robinson,J. (1951) Reversalof the vascularresponse to Acetylcholine and Adrenaline. Brit. J. Pharmacol.,6, 110-119. 5) Carman,J.B. (1955) The carotid labyrainth in Hyla aurea, with a note on that in Leiopelmahochstetteri. J. Anat. (Long.), 89, 503-525. 6) Ferry, C.B. (1963) The sympathomimeticeffect of acetylcholineon the spleenof the cat. J. Physiol. (Loud.), 167,487-504. 7) Hatakeyama, I. & Kato, R. (1958) Vasodilative and vasoconstrictive action of acetylcholineon the hindlimb perfusionpreparation. Yokohamamed. Bull., 9, 391- 408. 8) Hatakeyama,I. & Kato, R. (1959) A comparisonof sensitivityof large blood vessels to adrenaline and acetylcholine. Yokohamamed. Bull., 10, 34-38. 9) Hatakeyama, I. & Kato, R. (1960) The action of adrenaline, noradrenaline and acecylcholineon the pulmonaryblood vessels: Studieson the lung perfusionpreparation of Japanese toads. Yokohamamed. Bull., 11, 259-271. 10) Hodges,RD., King, AS., King, D.Z. & French, El. (1975) The generalurtrastruc- ture of the carotid body of the domesticfowl. Cell Tiss. Res., 162, 483-497. 11) Ishida, S. (1954) So-calledcarotic body of the Amphibia. Igaku Kenkyu (Fukuo- ka), 24, 1024-1050.(Japanese) 12) Ishii, K. & Ishii, K. (1973) Degeneration in the efferent nerve endings on the chemoreceptorcells of the carotid labyrinth after the sympatheticnerve trunk section. TohokuJ. exp. Med., 109, 339-345. 13) Ishii, K. & Kusakabe,T. (1982) The glomuscell of the carotid labyrinth of Xenopus laevis. Cell Tiss. Res., 224, 459-463. 14) Ishii, K. & Oosaki,T. (1969) Fine structure of th chemoreceptorcell in the amphibian carotid labyrinth. J. Anat. (Loud.), 104,263-280. 15) Ishii, K., Honda, K. & Ishii, K. (1966) The function of the carotid labyrinth in the toad. TohokuJ. exp. Med., 88, 103-116. 16) King, AS., King, D.Z., Hodges,RD. & Henry, J. (1975) Synaptic morphologyof the carotid body of the domesticfowl. Cell Tiss. Res., 162, 459-473. 17) Kobayashi,S. (1968) Fine structure of the carotid body of the dog. Arch. histol. jap., 30, 95-120. 18) McDonald,D.M. & Mitchell,R.A. (1975) The innervation of glomuscells, ganglion cells and blood vesselsin the rat carotid body : A quantitative ultrastructural analysis. J. Neurocytol.,4, 177-230. 19) Poullet-Krieger,M. (1973) Innervation du labyrinthe carotidien du crapaud Bufo Bufo: Etude ultrastructurale et histochemique. J. Microscopie,18, 55-64. 408 T. Kusakabe et al.

20) Rigual, R., Gonzalez, E., Gonzalez, C. & Fidone, S. (1986) Synthesis and release of catecholamines by the cat carotid body in vitro : Effects of hypoxic stimulation. Brain Res., 374, 101-109. 21) Rogers, D.C. (1963) Distinctive cell types in the amphibian carotid labyrinth. Nature (Lond. ), 200, 492-493. 22) Smith, D.G., Rogers, D.C., Chamley-Campbell, J. & Campbell, G.R. (1981) The mechanism of blood flow redistribution within the carotid labyrinth of the toad, Bufo marinas. J. exp. Zool., 216, 387-394. 23) Yates, R.D., Chen, I. -LI. & Duncan, D. (1970) Effects of sinus nerve stimulation on carotid body glomus cells. J. Cell Biol., 46, 544-552.