- Email: [email protected]
Sensory receptors in the first gill arch of rainbow trout
Respiration Physiology, 93 (1993) 97-110
97
© 1993 Elsevier Science Publishers B.V. All rights reserved. 0034-5687/93/$06.00
RESP 02031
Sensory receptors in the first gill arch of rainbow trout M a r k L. Burleson and William K. Milsom Department of Zoology, University of British Columbia, Vancouver, B.C., Canada (Accepted 22 February 1993) Abstract. Afferent neural activity was recorded from sensory receptors innervated by the glossopharyngeal nerve (cranial nerve IX) in isolated, perfused first gill arch preparations from rainbow trout. The present study demonstrates the presence of every major type of peripheral cardio-respiratory receptor described in fish in this preparation. Oxygen-sensitive chemoreceptors responsive to internal and/or external hypoxia and cyanide were present. Qualitatively these receptors behaved in an identical fashion which was also similar to that described for mammalian carotid body chemoreceptors. About 5To of the sensory receptors examined were O2-sensitive. Proprioceptors were the most numerous receptor type identified and were sensitive to mechanical stimulation of the arch, rakers or filaments. Finally, baroreceptors, the least numerous class of receptor identified, were also present with activity that was altered in response to changes in perfusion pressure. While the reflex responses elicited by the stimulation of these receptors were not addressed in this study, it is likely that these receptors contribute to the reflex cardio-respiratory responses to changes in gill perfusion, gill deflection and hypoxia (environmental or internal) described in fishes. These data thus support suggestions concerning homologies between the first gill arch of teleosts and the carotid bodies of mammals and the importance of the first gill arch in trout in cardio-respiratory control.
Baroreceptors; Chemoreceptors; Fish, rainbow trout; Hypoxia, peripheral receptors (rainbow trout); Mechanoreceptors; Receptors, sensory
Although the major functions of the gills of fish are gas and ion exchange, these organs are also important sensory structures and many of the sensory receptors found in the gills appear to be involved in the control of cardiovascular and ventilatory reflexes. Mechanoreceptors are present in the arches, rakers, and filaments of all gills. These receptors have been identified in Atlantic salmon (Salmo salar), sea raven sculpin (Hemitripterus americanus) (Sutterlin and Saunders, 1969) and dogfish (Squalus acanthias) (Satchell and Way, 1962) and extensively studied in carp (Cyprinus carpio) (de Graaf et al., 1987; de Graaf and Ballintijn, 1987). Ironically, although much is now known about the discharge characteristics of this receptor group, there are no conclusive studies concerning their reflex roles. They have been implicated in cough and expulsion reflexes, maintenance of the gill curtain, stabilization of the breathing patCorrespondence to: Dr. M.L. Burleson, Section of Comparative Physiology, Dept. of Biology, University of Texas, Arlington, Box 19498 Arlington, TX 76019, USA.
98 tern and regulation of ram ventilation. Mechanoreceptors are also present in the vasculature of the gills and, although the reflex effects of increased blood pressure have been noted in teleosts and elasmobranchs (see Nilsson, 1984 for review), few studies have examined afferent baroreceptor activity directly (Irving et al., 1935). Finally, chemoreceptors sensitive to 02 in blood and water appear to be present in all of the gill arches in actinopterygian fishes and are innervated by branchial branches of cranial nerves IX (glossopharyngeal) and X (vagus). Reflex studies indicate that there are at least two distinct loci for these O2-sensitive chemoreceptors in fishes with different central projections and reflex roles (Burleson et al., 1993). Aquatic (i.e. external) hypoxia stimulates ventilation and elicits bradycardia while cyanide, a potent chemoreceptor stimulant, also elicits bradycardia and stimulates ventilation when added to the inspired water (Burleson et al., 1993). In trout, these receptors appear to reside exclusively in the first gill arch (Smith and Jones, 1978). Altering internal 02 levels, on the other hand, by inducing hypoxemia, injecting hypoxic blood into the ventral aorta or reducing blood flow to the gills, stimulates ventilation but has little or no effect on heart rate (Cameron and Wohlschlag, 1969; Holeton, 1971; Smith and Jones, 1982). Again, there are few studies of the functional characteristics of this receptor group. Laurent and Rouzeau (1972) observed small, slow changes in nerve activity from cranial nerve IX to the pseudobranch of rainbow trout that were correlated with changes in P%. Milsom and Brill (1986), on the other hand, recorded O2-sensitive afferent activity from receptors in the first gill arch of yellowfin tuna (Thunnus albacares) similar to that recorded from mammalian carotid body chemoreceptors (which are derived from the first gill arch and innervated by the glossopharyngeal nerve). No studies have yet compared the discharge characteristics of internal vs external 0 2sensitive chemoreceptors. Given the extensive literature on cardiorespiratory reflexes in rainbow trout, it is curious that the only electrophysiological study of afferent activity arising from the gills of trout is the documentation of the anomalous low-sensitivity, slow, low amplitude voltage changes associated with hypoxia from the glossopharyngeal innervation of the pseudobranch (Laurent and Rouzeau, 1972). Given the extensive use of this species for studies of cardiorespiratory control, and the relative importance of the first gill arch in cardiorespiratory reflexes, the present study set out to examine the presence and discharge characteristics of putative mechano- and chemo-receptors in the glossopharyngeal nerve to the first gill arch of rainbow trout (Oncorhynchus rnykiss). As a basis for future studies of the reflex roles of these various receptor groups in trout, this study examines whether there are mechanoreceptors in the first gill arch of trout (proprioceptors and baroreceptors) similar to those previously described in other species, whether there are O2-sensitive chemoreceptors located in this gill arch similar to those described in tuna and higher vertebrates (and dramatically different from those previously described in the pseudobranch of trout) and whether there are functional differences, as well as differences in anatomical location and central connections, between the internal and external O2-sensitive chemoreceptors identified in reflex studies.
99
Materials and methods Rainbow trout (Oncorhynchus mykiss) were obtained from a local supplier and maintained outdoors in fiberglass tanks of circulating dechlorinated tap water at the University of British Columbia. On the day of an experiment a fish was netted from the tank, checked to confirm that it was in good health and given an intracardiac injection of 5000 units of sodium heparin (Allen and Hanburys, Canada). The fish was then killed by a sharp blow to the head and pithed, thus avoiding the use of anesthetics which could interfere with nerve and receptor function. The fish was placed, right side down, in a dissecting pan containing crushed ice to lower its temperature and presumably its metabolism during the time it took to remove the gill arch. The left operculum was removed, exposing the gills. During the procedure to remove the first gill arch the gills were periodically irrigated with ice cold hyperoxic saline to prevent desiccation and hypoxia. The first gill arch on the left side was cut from its attachment to the floor of the pharynx and the other gill arches. The pretrematic branch of the vagus innervating the first gill arch was located underneath a thin layer of muscle where the gill arch attaches to the roof of the pharynx and sectioned. Fine forceps were used to dissect tissue and expose the glossopharyngeal nerve just anterior to the pretrematic branch of the vagus. The glossopharyngeal nerve was then sectioned central to the petrosal ganglion, where it exits the neurocranium, and was carefully freed from the surrounding tissue. The gill arch was then cut away from its dorsal attachment and placed in a petri dish containing ice cold, hyperoxic, gill perfusate (filtered (Millipore 45 #m) physiological saline with polyvinylpyrrolidone (MW = 40 000, Sigma) added as a colloid osmotic filler (Perry et al., 1984)). Excess tissue was trimmed away and the efferent branchial artery was cannulated (PE-50, Intramedic) and flushed with 3 ml of gill perfusate to clear it of blood. Silk suture (000, Ethicon) was attached to the dorsal and ventral cut portions of the gill for suspension in the gill chamber (Fig. 1). The gill was then placed in a water-jacketed chamber, thermostatically controlled to
/ , ~ ~ |~"." ~
electrode ~/syrlnge ~ .
a,r valve~ ~ - air/nltrogen perfusate
II "" II I rpressure II II ~ transducer
air/n~rogen
waterjacketedchamber Fig. 1. Diagram of the experimental apparatus used to perfuse gills and record afferent neural activity from branchial nerves.
100 10 ° C, and perfused at constant physiological pressure (30-60 cm H 2 0 ). In some early experiments isoproterenol (ISO; 10-8 M, Sigma) was added to the perfusate to keep the gill vasculature open. It was subsequently noted that ISO may have a mild inhibitory effect on gill O2 receptor discharge. If so, however, this would have been constant throughout these early studies and should not affect the interpretation of the data. In later experiments, high initial perfusion pressure (100-120 cm H20 for 5 to 10 rain) was used to open the gill vasculature and ISO was not added to the perfusate. Once the gill was suspended in the gill chamber and attached to the perfusion apparatus, the glossopharyngeal nerve was cleaned of connective tissue and desheathed under a dissecting microscope (Wild). The nerve was placed on a stainless steel platform under mineral oil to prevent desiccation during recording. Small bundles of axons were peeled from the main nerve trunk using fine forceps and placed over bipolar platinum wire (1/1000" diameter) electrodes. The nerve signal was amplified using an A.C. pre-amplifier (Grass P5 model P511K) with associated regulated power supply (Grass model RPS 107E) and high impedance probe (Grass model HIPS11). A window discriminator (WPI model 121) was used to distinguish between units in multiunit preparations. Window discriminator output was averaged over 1 sec time intervals using an integrating amplifier (Gould model 134615-70). Nerve activity was displayed on a storage oscilloscope (Tektronix model 5111A) and also monitored with an audio monitor (Grass model AM 8B). Perfusion pressure was measured using a pressure transducer (Statham model P23 Db) with associated pre-amplifier and amplifier (FRAMP model GPA-2). The window discriminator output (electroneurogram, ENG), integrated window discriminator output and perfusion pressure were displayed on a pen recorder (Beckman Type R Dynograph). These same signals, along with the amplified nerve activity were also recorded on audio tape using an instrumentation tape recorder (Hewlett-Packard model 3968A). These recordings were played back for analysis on a 4-channel pen recorder (Gould model 2400S) and computer data acquisition system; an Olivetti M24 personal computer equipped with an analog to digital converter card (Data Translation model DT2801) using a commercial software package (Labtech Notebook, Laboratory Technologies Corporation, Wilmington, MA). The Po: and pH of the perfusate were measured using a Radiometer acid base analyzer (PHM71 Mk 2) and associated electrodes. The gill chamber and electrodes were maintained at 10 °C using a circulating water bath (Lauda model RM 6). The 02 electrode was zeroed with a solution of 0.1 M sodium borate and sodium sulphite and calibrated with air equilibrated water. The pH electrode was calibrated using Radiometer precision buffers. The oxygen sensitivity of each unit recorded using this preparation was tested by either bubbling the bath with N 2 to lower the external Po2, applying NaCN externally to the gill filaments, perfusing the gill with hypoxic perfusate or injecting NaCN into the perfusion cannula. This also established whether O2-sensitive units responded to changes in internal, external or both bathing solutions. Bubbling the bath with N 2 to achieve various levels of external Po: was a simple procedure. We were not able to alter
101 internal (perfusate) Po2 levels with any precision, however. Thus, stimulus-response curves for changing 02 tensions were obtained for external receptors only. Filament, raker and arch proprioceptors were identified by prodding various areas with a blunt wire probe. Several filaments were subjected to sustained deflections to determine if these receptors were rapidly adapting (RARs) and/or slowly adapting receptors (SARs). Only two fibers were found which responded to changes in perfusion pressure (baroreceptors). The effects of step changes in perfusion pressure on their afferent discharge were examined.
Results
Oxygen-sensitive chemoreceptors. More than 800 fibers were tested for 02 sensitivity during the course of this study. Approximately 5 To of these fibers displayed increased activity in response to hypoxia. Afferent information from a variety of different sensory receptors in the first gill arch is carried in the glossopharyngeal nerve and most of this neural activity was unresponsive to changes in Po~. Thus, gill mechanoreceptors sensitive to filament or raker displacement or changes in perfusion pressure, which were by far the most numerous receptors observed, did not respond to either hypoxia or NaCN. Fibers which did respond to changes in Po2, on the other hand, were insensitive to mechanical stimuli and were considered to be specific Oz-sensitive chemoreceptors. The amplitude of action potentials from O2-sensitive fibers was much smaller than that of mechanoreceptor action potentials suggesting that chemoreceptor afferent fibers are quite small. Of nineteen fibers that exhibited 02 sensitivity and were subjected to both external and internal stimuli, seven were sensitive to external changes in Po2 only, seven were sensitive to changes in internal perfusate composition only and five showed sensitivity to changes in both the internal and external milieu. The inability to control internal Po2 levels accurately with this experimental preparation precluded any quantitative comparison of internal and external receptors. External receptors showed an increase in activity as bath Po2 was decreased to 40 Torr (Figs. 2, 3). Below about 40 torr, nerve activity in most units became depressed (Fig. 3). The depressant effect of hypoxia was reversible and 02 sensitivity returned as O 2 tensions were increased again. Oxygen receptor activity increased dramatically when NaCN was applied externally onto the gill filaments and promptly returned to resting levels when the NaCN was rinsed from the filaments. Subjectively, the internal O2-sensitive chemoreceptors behaved just like the external receptors. Internal O2-sensitive chemoreceptors showed increased discharge rates in response to hypoxic perfusate and bolus injections of NaCN (Fig. 4A,B). The response of internal chemoreceptors to occlusion of the perfusate flow (Fig. 5) was slight in comparison to the responses to hypoxia and NaCN.
Mechanoreceptors. Given the extensive literature on gill proprioceptors in other species, our examination of trout gill proprioceptors was qualitative rather than quantita-
102 I min.
Po= - 140 torr
s.G
Po~ - 107 torr
Po= " 1 ~ ton"
! !!!! .= !! !!!
I ~IIII II I
Po~ = 81 tort
E~__lm_ll--II.l.l!!!l!!!l!
l lllllml II
lilll II
PeR " 39 ton"
Po= = 61 torr
~
lil|ll!_ll!
Fig. 2. Etectroneurograms (ENG) and integrated activity (impulses/sec) showing the response of a single chemoreceptor unit to varying levels of external Po2-
20,
15>
Om
10-
EV 0 J=
L)
5-
1 0
25
I
50
I
75
t
1 O0
I
125
PO 2 (torr)
Fig. 3. Mean changes in discharge frequency of eight 02 sensitive units plotted against external Po2" Dashed line (drawn by eye) shows the hypothetical rate of decrease in discharge frequency as Po2 continues to decrease. Horizontal and Vertical bars are + 1 SEM.
103
~ 20 ,-
1 min
_~
A J I lit
!tiltiiil Illilll i liil llHllllI!ll
ENGll
..... i~ ~j-
till
!1
1" Sodium Cyanide (25 ~ )
r[
tmin.
0
B ..._~,=llillllllllHiiiliU=LiilinP.,-=.
1" ~ n
= ,
I.t~oxic PoduJion
, ,
,,
ll,
liaan
..... I~.
L .........
~. . . . . . . . . . . .
.,11
? End Hypoxlc P ~ o n
Fig. 4. A. Response of a single O2-sensitive chemoreceptor to 25 #g N a C N injected into the perfusion cannula. B. Response of an internal O2-sensitive unit to a switch from a normoxic to a hypoxic perfusate (Poz ~ 20 Tort).
tive. Gill raker mechanoreceptor activity was rapidly adapting and the receptive field appeared to be limited to a single raker. Filament proprioceptor activity was the most common afferent activity observed. The receptive field for these receptors appeared to include more than one filament. Based on the response to sustained deflection, two classes of filament proprioceptors were identified (Fig. 6). The afferent discharge from rapidly adapting receptors (RARs) returned to pre-deflection levels within 1 min.
20 ¸
15 o
10 E
0
50 1' OFF
1O0 150 Time (sec)
200
250
Fig. 5. Mean response (n = 13)ofinternalO2-sensitivechemoreceptorsto occlusion (OFF) ofperfusate flow.
104 SAR
=8 a0
RAR
o-_E
!i=='!!ill
ENG
I
mlgllL!ll I!l!li[
Deflection
II
I
Deflection
Fig. 6. Representative traces showing rapidly and slowly adapting filament receptors in response to sustained deflection.
Slowly adapting receptors (SARs) maintained increased afferent activity until the filament was allowed to move back to its normal position. These mechanoreceptors often exhibited a brief burst of activity when moved back to the normal position. Some filament proprioceptors were tonically active while others were silent until stimulated. Finally, we observed two receptors sensitive to movement of the gill arch at the cartilage connection between the epibranchial and ceratobranchial elements. Activity increased as the two halves of the gill arch were spread apart as during the abduction phase of ventilation. The two baroreceptors examined responded to changes in perfusion pressure within the physiological range of blood pressure in this species. The effects of changing perfusion pressure on the discharge of one baroreceptor is shown in Fig. 7. Increasing
_E 1 mm/sec ,LIBLE[L )1, L J i * E l i t , l [ , i , [ , t [ J , L l i , , i i , , i l l E i i [ i
i ~, I ' ' ~5,, L ' i l ~, l i I sin , s l i l i i Eit1 i l , , i i l ~ l l l L L " i ' ' i S L ' " ' i ' & q , Li El i d i l ' i i l L i i ' , l i l U l
ENG II LII III III I tlmlltil~tnm,ll
i * l & H s l i ' H l l did i i ' L i ' L ' L " , ' i I H ' , , l i & ,
H ii,si'' .................
1551L ' ~ I '
ll/!!liliiU]ill! I I1 III
wE
~" 40
~o D.
Fig. 7. Effects of slow and rapid increases in perfusion pressure on the discharge of a single baroreceptor.
105 perfusion pressure caused an increase in afferent discharge. Slowly increasing pressure slowly increased discharge and rapid increases in perfusion pressure resulted in a rapid burst of afferent discharge. Although these receptors showed a low level of activity at 0 pressure, a rapid decrease in perfusion pressure from 60 cm to 0 cm H 2 0 temporarily silenced baroreceptor activity. Baroreceptor activity was not affected by any other mechanical stimulation.
Discussion The present study demonstrates the presence of every major type of peripheral cardiorespiratory receptor described in fish in the first gill arch of rainbow trout with axons carried in the glossopharyngeal nerve. Oxygen-sensitive chemoreceptors responsive to internal and/or external hypoxia and cyanide were present. Proprioceptors were also identified which were sensitive to mechanical stimulation of the arch, rakers or filaments. Finally, baroreceptors were also present with activity which was altered in response to changes in perfusion pressure. While the reflex responses elicited by the stimulation of these receptors were not addressed in this study, it is likely that these receptors contribute to the reflex cardiorespiratory responses to changes in gill perfusion, gill deflection and hypoxia (environmental or internal) described in fishes.
Oxygen-sensitive chemoreceptors. Recording O2-sensitive afferent activity from the branchial branch of the glossopharyngeal nerve in trout proved to be more difficult than anticipated. Only a low percentage (about 5 ~o) of the active fibers found in all preparations showed 02 sensitivity and these appeared to have a diffuse distribution. In addition, the branchial nerves of the trout gill were not very accessible and, as noted by Milsom and Brill (1986) in tuna, were delicate and easily damaged in comparison to other vertebrate nerves. Despite careful dissections and handling, afferent activity in some fibers simply ceased before a complete protocol could be run on them and there were some preparations in which we were unable to record neural activity of any kind. 02 chemoreceptive activity has been recorded previously from only the pseudobranch of trout (Laurent and Rouzeau, 1972) and the first gill arch of tuna (Milsom and Brill, 1986). The characteristics of the receptors described in these two studies are extremely different. The stimulus-response characteristics of the 02 chemoreceptors reported in the present study are unlike those described from the trout pseudobranch but similar to those reported for tuna as well as for mammalian aortic and carotid body chemoreceptors (Milsom and Brill, 1986). Thus, under resting conditions neural discharge was generally random or erratic and activity increased initially as Po~ fell. The discharge rates for the O2-chemoreceptors in trout were higher than those of other vertebrate 02 receptors over the same Po2 range but the functional significance of this is obscure. Another striking similarity between trout 02 receptors and the 02 receptors described in other vertebrates was their powerful activation by cyanide; the chemoreceptor responses to cyanide were comparable to the responses described for
106 other vertebrate 0 2 receptors (Burleson et al., 1993). They were also consistent with observations that sodium cyanide triggers hypoxic ventilatory reflexes in other fishes (see Burleson et al., 1993 for review). At the lowest level of hypoxia administered in the present study ( ~ 30 Torr Po2), activity was then depressed. Such reversible depression of chemoreceptor activity during long-term severe hypoxia has also been demonstrated in mammalian carotid and aortic receptors and is believed to reflect a dependence of prolonged chemoreceptor activity on oxidative metabolism (Lahiri et al., 1983). The physiological significance of the value of Po~ at which depression of chemoreceptor discharge occurred in the present study (30-40 Torr) is not clear. This species has been shown previously to still possess a brisk hypoxic ventilatory response at 02 tensions in this range (Holeton and Randall, 1967). One possible explanation for this discrepancy between reflex and receptor studies is that the P% at the receptor site could have been lower at the same Po~ of the bathing solution in the receptor study. Although, bubbling the gill chamber with the hypoxic gas mixture during the nerve recording study maintained some convection over the gill surface, a relatively thick boundary layer around the gill might have been present. This should not be the case in vivo where high rates of water flow over the gill would reduce boundary layers and bring the Po~ of externally oriented receptors into closer equilibrium with the external medium. Based on the differential responses offish to aquatic hypoxia and arterial hypoxemia, several studies have concluded that fish have at least two functional groups of 0 2sensitive chemoreceptors (Burleson et al., 1993). Reflex hypoxic bradycardia is mediated exclusively by externally (water) oriented receptors while ventilatory reflexes are mediated by externally and internally (blood/tissue) oriented receptors. Neural recordings from the gills of tuna (Milsom and Brill, 1986) suggested that all O2-sensitive chemoreceptors in the first gill arch were sensitive to internal hypoxia while roughly two thirds were also sensitive to external hypoxia. In the present study, roughly one third of all receptors were sensitive to external hypoxia only, one third to internal hypoxia only and one third to both. Although the number of receptors recorded from are low in both of these studies, the data do suggest that there may be some species variation in the distribution of the O2-sensitive chemoreceptors between the external gill surface and the gill vasculature. Histological studies reveal neuroepithelial cells in the primary gill epithelium in this general location. Anatomically these cells resemble glomus or Type I cells of the mammalian carotid body and Falck fluorescence indicates the presence of monoamines in these cells just as they have been described in mammalian carotid body cells (Dunel-Erb et al., 1982; Donald, 1987). Although we were only able to quantify the responses of the receptors responsive to external hypoxia, the data suggest that, qualitatively, the responses of all O2-sensitive chemoreceptors are the same regardless of their orientation. This would imply that the differential heart rate response to external vs internal hypoxia seen in fish is due to a difference in the central connections of externally vs internally oriented receptors rather than differences in the functional characteristics of the receptors themselves. In conscious fish, decreasing blood 02 delivery by inducing hypoxemia, injecting
107 hypoxic blood or reducing blood flow to the gills stimulates ventilation (Cameron and Wohlschlag, 1969; Holeton, 1971; Smith and Jones, 1982). The activity of internal O2 receptors recorded with this preparation, however, were not very perfusion sensitive. The response to occluding perfusate flow was moderate in comparison to the responses of these receptors to perfusion with hypoxic solutions or NaCN. This would suggest that the metabolic rate of the chemoreceptor cells is quite low. Thus the response to exogenous removal of 02 from the perfusate was fast while the response to endogenous removal of 0 2 from stagnant perfusate by local metabolism during the occlusion procedure was slow. This is in contrast to the rapid change in discharge frequency observed in unperfused 02 chemoreceptors of other vertebrates including tuna (Milsom and Brill, 1986). Afferent neural activity showing regular bursting discharge patterns, as reported by Milsom and Brill (1986) in tuna for some 0 2 receptors, was only seen in two 0 2sensitive fibers during preliminary experiments. Milsom and Brill (1986) suggested that intrinsic vasomotion, as reported in fish gills (Satchell, 1962), may have been the cause of the bursting discharge pattern they observed in tuna receptors. In the present study, isoproterenol was initially added to the gill perfusate to enhance perfusion and subsequently, this was achieved by a high initial perfusion pressure which tended to dilate the gill vasculature and promote perfusion. These procedures may have abolished such intrinsic vasomotion in the trout preparations. Also, as described above, trout 0 2 receptors were much less sensitive to occlusion of the perfusate flow (Fig. 5) than tuna receptors, suggesting that vasomotion would probably have had little effect on afferent neural activity even if it were present.
Mechanoreceptors. Several different classes ofmechanoreceptors sensitive to displacement of various gill elements (arches, filaments and rakers) have been identified in fishes. Piscine gill mechanoreceptors, like other vertebrate mechanoreceptors, appear to be simple free nerve endings situated in muscle or connective tissue (Roberts and Rowell, 1988). It is thought that the anatomical location of a mechanoreceptor determines its sensory modality and response characteristics. Although a number of studies have demonstrated the presence of mechanoreceptors in gills, few have addressed the role that these may play in the control of cardioventilatory reflexes in fishes. Filament mechanoreceptors have been identified in sea raven sculpins (Hemitripterus americanus), Atlantic salmon (Salmo salar) and carp (Cyprinus carpio). Afferent neural information from these receptors travels in external branches of the pre- and posttrematic branchial nerves and activity levels increase in response to filament deflection. Only rapidly adapting receptors were described from this location in previous studies (Sutterlin and Saunders, 1969; de Graafet al., 1987), although slowly adapting filament receptors were also present in trout gills. The receptive field for filament mechanoreceptors was limited to a single filament in carp (de G r a a f e t aL, 1987) but appeared to include one or several filaments in salmon, sculpin and trout (Sutterlin and Saunders, 1969; this study). Mechanical displacement of one filament usually causes adjacent filaments to move, however, making it difficult to determine the limits of receptive fields.
108 Histological studies are needed to precisely determine the extent of filament mechanoreceptor innervation. Because filament receptors were not activated during breathing in lightly curarized carp, de Graaf et al., (1987) suggested that these receptors were not involved in cardioventilatory control. They suggested that filament receptors were more likely involved with feeding and gill defense responses. Stimulation of filament mechanoreceptors elicits filament adduction which should lower the resistance of water flowing over the gills thus facilitating cough and expulsion reflexes (de Graaf et al., 1987). Gill raker mechanoreceptors have previously been identified in dogfish, sea raven sculpin and carp. In dogfish (Squalus acanthias), both rapidly and slowly adapting mechanoreceptors were identified (Satchell and Way, 1962) whereas only rapidly adapting raker mechanoreceptor activity was recorded from carp (de Graafet al., 1987) and trout (this study). Although the experiments on dogfish were performed on isolated gills, Satchell and Way (1962) suggest that the slowly adapting receptors should be stimulated by normal ventilatory movements and could contribute to cardioventilatory control. The raker mechanoreceptors in carp were not active during normal ventilation which suggests that rapidly adapting raker mechanoreceptors do not contribute to cardioventilatory control but may also be primarily involved with feeding and protective reflexes. Gill arch mechanoreceptors located in the cartilaginous strip between the epibranchial and ceratobranchial elements have now been described in carp (de Graaf and Ballintijn, 1987) and trout. These receptors are tonically active and respond to adduction by decreasing activity and abduction by increasing activity. Because these were the only gill mechanoreceptors active during ventilation in carp, de Graaf and Ballintijn (1987) suggest that these are the receptors most likely to be involved with cardioventilatory synchronization. Baroreceptors are simply mechanoreceptors located in the walls of blood vessels. These receptors are stimulated by mechanical distortion of the blood vessel caused by changes in blood pressure. Baroreceptor reflexes have been demonstrated in elasmobranchs and teleost fishes (see Nilsson, 1984 for review). Increasing blood or perfusion pressure to the gills and electrical stimulation of branchial nerves causes a reflexive bradycardia. It should be noted, however, that electrical stimulation of the branchial nerves stimulates chemoreceptor and nociceptor afferent fibers which also mediate reflex bradycardia (Satchell, 1978; Burleson and Smatresk, 1990). Irving et al. (1935) recorded baroreceptor activity from branchial branches of cranial nerves IX and X in dogfish in vivo. They observed that branchial baroreceptor activity was synchronous with systole, was abolished by reducing blood pressure (hemorrhage) and increased in response to increased blood pressure (induced by infusion of adrenaline). Laurent (1967) has also recorded baroreceptor activity from the pseudobranch in tench that increased in response to increased perfusion pressure. The discharge characteristics of branchial baroreceptors in trout were very similar to those recorded in dogfish. Both groups of baroreceptors responded to rapid changes in perfusion pressure with a brief burst in activity followed by adaptation. Although the
109 precise location of branchial baroreceptors is not known, histological studies indicate that the baroreceptive loci in the gills may be at the junction of the afferent and efferent branchial arteries (Boyd, 1936; DeKock, 1963).
Are first gill arch receptors homologous to carotid body receptors?. According to phylogenetic and ontogenetic evidence, the carotid artery of mammals and the first gill arch of teleosts are homologs and this has led to the suggestion that "the gill capillaries of the first gill remain as a tangle of capillaries called the carotid body,..." (Yapp, 1965). Now, the accumulated evidence suggests that the stimulus response characteristics of O2-sensitive chemoreceptors and baroreceptors in fish gills are remarkably similar to those reported for carotid body receptors in mammals. The discharge characteristics of O2-sensitive chemoreceptors at rest, and the effects of hypoxia and cyanide on this discharge are very similar in both groups. So too are the discharge characteristics of the arterial baroreceptors, to the extent that they have now been studied in fish. Afferent information from both structures is carried in the glossopharyngeal nerve. The sum of this evidence would support suggestions that the diffuse network of 02 receptors and baroreceptors found in the first gill arch of fish are homologous to the more localized carotid body receptors found in mammals. It also suggests that the mechanism of 02 chemoreception is similar, if not identical, in all vertebrates and that these mechanisms have been conserved throughout the evolution of the vertebrates.
Acknowledgements.This work was supported by NSERC of Canada (WKM) and a McLean Fraser Research Fellowship (MLB). The authors thank Dr. N.J. Smatresk for his comments on the manuscript.
References Boyd, J.D. (1936). Nerve supply to the branchial arch arteries of vertebrates. J. Anat. Lond. 71: 157-158. Burleson, M.L. and N.J. Smatresk (1990). Effects of sectioning cranial nerves IX and X on cardiovascular and ventilatory reflex responses to hypoxia and NaCN in channel catfish. J. Exp. Biol. 154: 407420. Burleson, M.L., N.J. Smatresk and W.K. Milsom (1993). Afferent inputs associated with cardioventilatory control in fish. In: Fish Physiology, Vol. XIIB, edited by W.S. Hoar, D.J. Randall and A.P. Farrell. Orlando: Academic Press, pp. 389-426. Cameron, J.N and D.E. Wohlschlag (1969). Respiratory response to experimentally induced anaemia in the pinfish (Lagodon rhomboides). J. Exp. Biol. 50: 307-317. de Graaf, P.J.F., C.M. Ballintiin and F.W. Maes (1987). Mechanoreceptor activity in the gills of carp. I. Gill filament and gill raker mechanoreceptors. Respir. Physiol. 69: 173-182. de Graaf, P.J.F. and C.M. Ballintijn (1987). Mechanoreceptor activity in the gills of the carp. I1. Gill arch proprioceptors. Re~pir. Physiol. 69: 183-194. DeKock, L.L. (1963). A histological study of the head region of two salmonids with special reference to pressor- and chemo-receptors. Acta Anat. 55: 39-50. Donald, J. (1987), Comparative study of the adrenergic innervation of the teleost gill. J. Morphol. 193: 63-73. Dunel-Erb, S., Y. Bailly and P. Laurent (1982). Neuroepithelial cells in fish gill primary lamellae. J. Appl, Physiol. 53: R1324-R1353.
110 Holeton, G.F. and D.J. Randall (1967). Changes in blood pressure in the rainbow trout during hypoxia. J. Exp. Biol. 46: 297-305. Holeton, G.F. (1971). Oxygen uptake and transport by the rainbow trout during exposure to carbon monoxide. J. Exp. Biol. 54: 239-254. Irving, L., D.Y. Solandt and O.M. Solandt (1935). Nerve impulses from branchial pressure receptors in the dogfish. J. Physiol. (London), 84: 187-190. Lahiri, S., N.J. Smatresk and E. Mulligan (1983). Responses of peripheral chemoreceptors to natural stimuli. In: Physiology of the Peripheral Arterial Chemoreceptors, edited by H. Acker and R.G. O'Regan, Ch. 10. Amsterdam: Elsevier Science Publishers, pp. 221-256. Laurent, P. (1967). La pseudobranchie des T616osteens preuves electrophysiologiques de ses fonctions ch6mor6ceptrice et baror6ceptrice. C.R. Acad. Sei. Paris 264: 1879-1882. Laurent, P. and J.D. Rouzeau (1972). Afferent neural activity from pseudobranch of teleosts. Effects of PO2, pH, osmotic pressure and Na ÷ ions. Respir. Physiol. 14: 307-33l. Milsom, W.K. and R.W. Brill (1986). Oxygen sensitive afferent information arising from the first gill arch of yellowfin tuna. Respir. Physiol. 66: 193-203. Nilsson, S. (1984). Innervation and pharmacology of the gills. In: Fish Physiology, Vol. XA, edited by W.S. Hoar and D.J. Randall. New York: Academic Press, pp. 185-227. Perry, S.F., P.S. Davie, C. Daxboeck, A.G. Ellis and D.G. Smith (1984). Perfusion methods for the study of gill physiology. In: Fish Physiology, Vol XB, edited by W.S. Hoar and D.J. Randall. New York: Academic Press, pp. 325-388, Roberts, J.L. and D.M. Rowell (1988). Periodic respiration of gill-breathing fishes. Can. J. Zool. 66: 182190. Satchell, G.H. (1962). Intrinsic vasomotion in the dogfish gill. J. Exp. Biol. 39: 503-512. S atchell, G.H. and H.K. Way (1962). Pharyngeal proprioceptors in the dogfish, Squalus acanthias L, J. Exp. Biol. 39: 243-250. Satchell, G.H. (1978). The J reflex in fish. In: Respiratory adaptations, capillary exchange and reflex mechanisms, edited by A.S. Paintal and P. Gill-Kumar. Delhi: Vallabhbhai Patel Chest Institute, University of Delhi, pp. 432-441. Smith, F.M. and D,R. Jones (1978). Localization of receptors causing hypoxic bradycardia in trout (Salmo guirdneri). Can. J. Zool. 56: 1260-1265. Smith, F.M and D.R. Jones (1982). The effect of changes in blood oxygen-carrying capacity on ventilation volume in the rainbow trout (Salmo gairdneri). J. Exp. Biol. 97: 325-334. Sutterlin, A.M. and R.L. Saunders (1969). Proprioceptors in the gills of teleosts. Can. J. Zool. 47: 12091212. Yapp. W.B. (1965), Vertebrates: Their Structure and Life. New York: Oxford University Press, pp. 525.
97
© 1993 Elsevier Science Publishers B.V. All rights reserved. 0034-5687/93/$06.00
RESP 02031
Sensory receptors in the first gill arch of rainbow trout M a r k L. Burleson and William K. Milsom Department of Zoology, University of British Columbia, Vancouver, B.C., Canada (Accepted 22 February 1993) Abstract. Afferent neural activity was recorded from sensory receptors innervated by the glossopharyngeal nerve (cranial nerve IX) in isolated, perfused first gill arch preparations from rainbow trout. The present study demonstrates the presence of every major type of peripheral cardio-respiratory receptor described in fish in this preparation. Oxygen-sensitive chemoreceptors responsive to internal and/or external hypoxia and cyanide were present. Qualitatively these receptors behaved in an identical fashion which was also similar to that described for mammalian carotid body chemoreceptors. About 5To of the sensory receptors examined were O2-sensitive. Proprioceptors were the most numerous receptor type identified and were sensitive to mechanical stimulation of the arch, rakers or filaments. Finally, baroreceptors, the least numerous class of receptor identified, were also present with activity that was altered in response to changes in perfusion pressure. While the reflex responses elicited by the stimulation of these receptors were not addressed in this study, it is likely that these receptors contribute to the reflex cardio-respiratory responses to changes in gill perfusion, gill deflection and hypoxia (environmental or internal) described in fishes. These data thus support suggestions concerning homologies between the first gill arch of teleosts and the carotid bodies of mammals and the importance of the first gill arch in trout in cardio-respiratory control.
Baroreceptors; Chemoreceptors; Fish, rainbow trout; Hypoxia, peripheral receptors (rainbow trout); Mechanoreceptors; Receptors, sensory
Although the major functions of the gills of fish are gas and ion exchange, these organs are also important sensory structures and many of the sensory receptors found in the gills appear to be involved in the control of cardiovascular and ventilatory reflexes. Mechanoreceptors are present in the arches, rakers, and filaments of all gills. These receptors have been identified in Atlantic salmon (Salmo salar), sea raven sculpin (Hemitripterus americanus) (Sutterlin and Saunders, 1969) and dogfish (Squalus acanthias) (Satchell and Way, 1962) and extensively studied in carp (Cyprinus carpio) (de Graaf et al., 1987; de Graaf and Ballintijn, 1987). Ironically, although much is now known about the discharge characteristics of this receptor group, there are no conclusive studies concerning their reflex roles. They have been implicated in cough and expulsion reflexes, maintenance of the gill curtain, stabilization of the breathing patCorrespondence to: Dr. M.L. Burleson, Section of Comparative Physiology, Dept. of Biology, University of Texas, Arlington, Box 19498 Arlington, TX 76019, USA.
98 tern and regulation of ram ventilation. Mechanoreceptors are also present in the vasculature of the gills and, although the reflex effects of increased blood pressure have been noted in teleosts and elasmobranchs (see Nilsson, 1984 for review), few studies have examined afferent baroreceptor activity directly (Irving et al., 1935). Finally, chemoreceptors sensitive to 02 in blood and water appear to be present in all of the gill arches in actinopterygian fishes and are innervated by branchial branches of cranial nerves IX (glossopharyngeal) and X (vagus). Reflex studies indicate that there are at least two distinct loci for these O2-sensitive chemoreceptors in fishes with different central projections and reflex roles (Burleson et al., 1993). Aquatic (i.e. external) hypoxia stimulates ventilation and elicits bradycardia while cyanide, a potent chemoreceptor stimulant, also elicits bradycardia and stimulates ventilation when added to the inspired water (Burleson et al., 1993). In trout, these receptors appear to reside exclusively in the first gill arch (Smith and Jones, 1978). Altering internal 02 levels, on the other hand, by inducing hypoxemia, injecting hypoxic blood into the ventral aorta or reducing blood flow to the gills, stimulates ventilation but has little or no effect on heart rate (Cameron and Wohlschlag, 1969; Holeton, 1971; Smith and Jones, 1982). Again, there are few studies of the functional characteristics of this receptor group. Laurent and Rouzeau (1972) observed small, slow changes in nerve activity from cranial nerve IX to the pseudobranch of rainbow trout that were correlated with changes in P%. Milsom and Brill (1986), on the other hand, recorded O2-sensitive afferent activity from receptors in the first gill arch of yellowfin tuna (Thunnus albacares) similar to that recorded from mammalian carotid body chemoreceptors (which are derived from the first gill arch and innervated by the glossopharyngeal nerve). No studies have yet compared the discharge characteristics of internal vs external 0 2sensitive chemoreceptors. Given the extensive literature on cardiorespiratory reflexes in rainbow trout, it is curious that the only electrophysiological study of afferent activity arising from the gills of trout is the documentation of the anomalous low-sensitivity, slow, low amplitude voltage changes associated with hypoxia from the glossopharyngeal innervation of the pseudobranch (Laurent and Rouzeau, 1972). Given the extensive use of this species for studies of cardiorespiratory control, and the relative importance of the first gill arch in cardiorespiratory reflexes, the present study set out to examine the presence and discharge characteristics of putative mechano- and chemo-receptors in the glossopharyngeal nerve to the first gill arch of rainbow trout (Oncorhynchus rnykiss). As a basis for future studies of the reflex roles of these various receptor groups in trout, this study examines whether there are mechanoreceptors in the first gill arch of trout (proprioceptors and baroreceptors) similar to those previously described in other species, whether there are O2-sensitive chemoreceptors located in this gill arch similar to those described in tuna and higher vertebrates (and dramatically different from those previously described in the pseudobranch of trout) and whether there are functional differences, as well as differences in anatomical location and central connections, between the internal and external O2-sensitive chemoreceptors identified in reflex studies.
99
Materials and methods Rainbow trout (Oncorhynchus mykiss) were obtained from a local supplier and maintained outdoors in fiberglass tanks of circulating dechlorinated tap water at the University of British Columbia. On the day of an experiment a fish was netted from the tank, checked to confirm that it was in good health and given an intracardiac injection of 5000 units of sodium heparin (Allen and Hanburys, Canada). The fish was then killed by a sharp blow to the head and pithed, thus avoiding the use of anesthetics which could interfere with nerve and receptor function. The fish was placed, right side down, in a dissecting pan containing crushed ice to lower its temperature and presumably its metabolism during the time it took to remove the gill arch. The left operculum was removed, exposing the gills. During the procedure to remove the first gill arch the gills were periodically irrigated with ice cold hyperoxic saline to prevent desiccation and hypoxia. The first gill arch on the left side was cut from its attachment to the floor of the pharynx and the other gill arches. The pretrematic branch of the vagus innervating the first gill arch was located underneath a thin layer of muscle where the gill arch attaches to the roof of the pharynx and sectioned. Fine forceps were used to dissect tissue and expose the glossopharyngeal nerve just anterior to the pretrematic branch of the vagus. The glossopharyngeal nerve was then sectioned central to the petrosal ganglion, where it exits the neurocranium, and was carefully freed from the surrounding tissue. The gill arch was then cut away from its dorsal attachment and placed in a petri dish containing ice cold, hyperoxic, gill perfusate (filtered (Millipore 45 #m) physiological saline with polyvinylpyrrolidone (MW = 40 000, Sigma) added as a colloid osmotic filler (Perry et al., 1984)). Excess tissue was trimmed away and the efferent branchial artery was cannulated (PE-50, Intramedic) and flushed with 3 ml of gill perfusate to clear it of blood. Silk suture (000, Ethicon) was attached to the dorsal and ventral cut portions of the gill for suspension in the gill chamber (Fig. 1). The gill was then placed in a water-jacketed chamber, thermostatically controlled to
/ , ~ ~ |~"." ~
electrode ~/syrlnge ~ .
a,r valve~ ~ - air/nltrogen perfusate
II "" II I rpressure II II ~ transducer
air/n~rogen
waterjacketedchamber Fig. 1. Diagram of the experimental apparatus used to perfuse gills and record afferent neural activity from branchial nerves.
100 10 ° C, and perfused at constant physiological pressure (30-60 cm H 2 0 ). In some early experiments isoproterenol (ISO; 10-8 M, Sigma) was added to the perfusate to keep the gill vasculature open. It was subsequently noted that ISO may have a mild inhibitory effect on gill O2 receptor discharge. If so, however, this would have been constant throughout these early studies and should not affect the interpretation of the data. In later experiments, high initial perfusion pressure (100-120 cm H20 for 5 to 10 rain) was used to open the gill vasculature and ISO was not added to the perfusate. Once the gill was suspended in the gill chamber and attached to the perfusion apparatus, the glossopharyngeal nerve was cleaned of connective tissue and desheathed under a dissecting microscope (Wild). The nerve was placed on a stainless steel platform under mineral oil to prevent desiccation during recording. Small bundles of axons were peeled from the main nerve trunk using fine forceps and placed over bipolar platinum wire (1/1000" diameter) electrodes. The nerve signal was amplified using an A.C. pre-amplifier (Grass P5 model P511K) with associated regulated power supply (Grass model RPS 107E) and high impedance probe (Grass model HIPS11). A window discriminator (WPI model 121) was used to distinguish between units in multiunit preparations. Window discriminator output was averaged over 1 sec time intervals using an integrating amplifier (Gould model 134615-70). Nerve activity was displayed on a storage oscilloscope (Tektronix model 5111A) and also monitored with an audio monitor (Grass model AM 8B). Perfusion pressure was measured using a pressure transducer (Statham model P23 Db) with associated pre-amplifier and amplifier (FRAMP model GPA-2). The window discriminator output (electroneurogram, ENG), integrated window discriminator output and perfusion pressure were displayed on a pen recorder (Beckman Type R Dynograph). These same signals, along with the amplified nerve activity were also recorded on audio tape using an instrumentation tape recorder (Hewlett-Packard model 3968A). These recordings were played back for analysis on a 4-channel pen recorder (Gould model 2400S) and computer data acquisition system; an Olivetti M24 personal computer equipped with an analog to digital converter card (Data Translation model DT2801) using a commercial software package (Labtech Notebook, Laboratory Technologies Corporation, Wilmington, MA). The Po: and pH of the perfusate were measured using a Radiometer acid base analyzer (PHM71 Mk 2) and associated electrodes. The gill chamber and electrodes were maintained at 10 °C using a circulating water bath (Lauda model RM 6). The 02 electrode was zeroed with a solution of 0.1 M sodium borate and sodium sulphite and calibrated with air equilibrated water. The pH electrode was calibrated using Radiometer precision buffers. The oxygen sensitivity of each unit recorded using this preparation was tested by either bubbling the bath with N 2 to lower the external Po2, applying NaCN externally to the gill filaments, perfusing the gill with hypoxic perfusate or injecting NaCN into the perfusion cannula. This also established whether O2-sensitive units responded to changes in internal, external or both bathing solutions. Bubbling the bath with N 2 to achieve various levels of external Po: was a simple procedure. We were not able to alter
101 internal (perfusate) Po2 levels with any precision, however. Thus, stimulus-response curves for changing 02 tensions were obtained for external receptors only. Filament, raker and arch proprioceptors were identified by prodding various areas with a blunt wire probe. Several filaments were subjected to sustained deflections to determine if these receptors were rapidly adapting (RARs) and/or slowly adapting receptors (SARs). Only two fibers were found which responded to changes in perfusion pressure (baroreceptors). The effects of step changes in perfusion pressure on their afferent discharge were examined.
Results
Oxygen-sensitive chemoreceptors. More than 800 fibers were tested for 02 sensitivity during the course of this study. Approximately 5 To of these fibers displayed increased activity in response to hypoxia. Afferent information from a variety of different sensory receptors in the first gill arch is carried in the glossopharyngeal nerve and most of this neural activity was unresponsive to changes in Po~. Thus, gill mechanoreceptors sensitive to filament or raker displacement or changes in perfusion pressure, which were by far the most numerous receptors observed, did not respond to either hypoxia or NaCN. Fibers which did respond to changes in Po2, on the other hand, were insensitive to mechanical stimuli and were considered to be specific Oz-sensitive chemoreceptors. The amplitude of action potentials from O2-sensitive fibers was much smaller than that of mechanoreceptor action potentials suggesting that chemoreceptor afferent fibers are quite small. Of nineteen fibers that exhibited 02 sensitivity and were subjected to both external and internal stimuli, seven were sensitive to external changes in Po2 only, seven were sensitive to changes in internal perfusate composition only and five showed sensitivity to changes in both the internal and external milieu. The inability to control internal Po2 levels accurately with this experimental preparation precluded any quantitative comparison of internal and external receptors. External receptors showed an increase in activity as bath Po2 was decreased to 40 Torr (Figs. 2, 3). Below about 40 torr, nerve activity in most units became depressed (Fig. 3). The depressant effect of hypoxia was reversible and 02 sensitivity returned as O 2 tensions were increased again. Oxygen receptor activity increased dramatically when NaCN was applied externally onto the gill filaments and promptly returned to resting levels when the NaCN was rinsed from the filaments. Subjectively, the internal O2-sensitive chemoreceptors behaved just like the external receptors. Internal O2-sensitive chemoreceptors showed increased discharge rates in response to hypoxic perfusate and bolus injections of NaCN (Fig. 4A,B). The response of internal chemoreceptors to occlusion of the perfusate flow (Fig. 5) was slight in comparison to the responses to hypoxia and NaCN.
Mechanoreceptors. Given the extensive literature on gill proprioceptors in other species, our examination of trout gill proprioceptors was qualitative rather than quantita-
102 I min.
Po= - 140 torr
s.G
Po~ - 107 torr
Po= " 1 ~ ton"
! !!!! .= !! !!!
I ~IIII II I
Po~ = 81 tort
E~__lm_ll--II.l.l!!!l!!!l!
l lllllml II
lilll II
PeR " 39 ton"
Po= = 61 torr
~
lil|ll!_ll!
Fig. 2. Etectroneurograms (ENG) and integrated activity (impulses/sec) showing the response of a single chemoreceptor unit to varying levels of external Po2-
20,
15>
Om
10-
EV 0 J=
L)
5-
1 0
25
I
50
I
75
t
1 O0
I
125
PO 2 (torr)
Fig. 3. Mean changes in discharge frequency of eight 02 sensitive units plotted against external Po2" Dashed line (drawn by eye) shows the hypothetical rate of decrease in discharge frequency as Po2 continues to decrease. Horizontal and Vertical bars are + 1 SEM.
103
~ 20 ,-
1 min
_~
A J I lit
!tiltiiil Illilll i liil llHllllI!ll
ENGll
..... i~ ~j-
till
!1
1" Sodium Cyanide (25 ~ )
r[
tmin.
0
B ..._~,=llillllllllHiiiliU=LiilinP.,-=.
1" ~ n
= ,
I.t~oxic PoduJion
, ,
,,
ll,
liaan
..... I~.
L .........
~. . . . . . . . . . . .
.,11
? End Hypoxlc P ~ o n
Fig. 4. A. Response of a single O2-sensitive chemoreceptor to 25 #g N a C N injected into the perfusion cannula. B. Response of an internal O2-sensitive unit to a switch from a normoxic to a hypoxic perfusate (Poz ~ 20 Tort).
tive. Gill raker mechanoreceptor activity was rapidly adapting and the receptive field appeared to be limited to a single raker. Filament proprioceptor activity was the most common afferent activity observed. The receptive field for these receptors appeared to include more than one filament. Based on the response to sustained deflection, two classes of filament proprioceptors were identified (Fig. 6). The afferent discharge from rapidly adapting receptors (RARs) returned to pre-deflection levels within 1 min.
20 ¸
15 o
10 E
0
50 1' OFF
1O0 150 Time (sec)
200
250
Fig. 5. Mean response (n = 13)ofinternalO2-sensitivechemoreceptorsto occlusion (OFF) ofperfusate flow.
104 SAR
=8 a0
RAR
o-_E
!i=='!!ill
ENG
I
mlgllL!ll I!l!li[
Deflection
II
I
Deflection
Fig. 6. Representative traces showing rapidly and slowly adapting filament receptors in response to sustained deflection.
Slowly adapting receptors (SARs) maintained increased afferent activity until the filament was allowed to move back to its normal position. These mechanoreceptors often exhibited a brief burst of activity when moved back to the normal position. Some filament proprioceptors were tonically active while others were silent until stimulated. Finally, we observed two receptors sensitive to movement of the gill arch at the cartilage connection between the epibranchial and ceratobranchial elements. Activity increased as the two halves of the gill arch were spread apart as during the abduction phase of ventilation. The two baroreceptors examined responded to changes in perfusion pressure within the physiological range of blood pressure in this species. The effects of changing perfusion pressure on the discharge of one baroreceptor is shown in Fig. 7. Increasing
_E 1 mm/sec ,LIBLE[L )1, L J i * E l i t , l [ , i , [ , t [ J , L l i , , i i , , i l l E i i [ i
i ~, I ' ' ~5,, L ' i l ~, l i I sin , s l i l i i Eit1 i l , , i i l ~ l l l L L " i ' ' i S L ' " ' i ' & q , Li El i d i l ' i i l L i i ' , l i l U l
ENG II LII III III I tlmlltil~tnm,ll
i * l & H s l i ' H l l did i i ' L i ' L ' L " , ' i I H ' , , l i & ,
H ii,si'' .................
1551L ' ~ I '
ll/!!liliiU]ill! I I1 III
wE
~" 40
~o D.
Fig. 7. Effects of slow and rapid increases in perfusion pressure on the discharge of a single baroreceptor.
105 perfusion pressure caused an increase in afferent discharge. Slowly increasing pressure slowly increased discharge and rapid increases in perfusion pressure resulted in a rapid burst of afferent discharge. Although these receptors showed a low level of activity at 0 pressure, a rapid decrease in perfusion pressure from 60 cm to 0 cm H 2 0 temporarily silenced baroreceptor activity. Baroreceptor activity was not affected by any other mechanical stimulation.
Discussion The present study demonstrates the presence of every major type of peripheral cardiorespiratory receptor described in fish in the first gill arch of rainbow trout with axons carried in the glossopharyngeal nerve. Oxygen-sensitive chemoreceptors responsive to internal and/or external hypoxia and cyanide were present. Proprioceptors were also identified which were sensitive to mechanical stimulation of the arch, rakers or filaments. Finally, baroreceptors were also present with activity which was altered in response to changes in perfusion pressure. While the reflex responses elicited by the stimulation of these receptors were not addressed in this study, it is likely that these receptors contribute to the reflex cardiorespiratory responses to changes in gill perfusion, gill deflection and hypoxia (environmental or internal) described in fishes.
Oxygen-sensitive chemoreceptors. Recording O2-sensitive afferent activity from the branchial branch of the glossopharyngeal nerve in trout proved to be more difficult than anticipated. Only a low percentage (about 5 ~o) of the active fibers found in all preparations showed 02 sensitivity and these appeared to have a diffuse distribution. In addition, the branchial nerves of the trout gill were not very accessible and, as noted by Milsom and Brill (1986) in tuna, were delicate and easily damaged in comparison to other vertebrate nerves. Despite careful dissections and handling, afferent activity in some fibers simply ceased before a complete protocol could be run on them and there were some preparations in which we were unable to record neural activity of any kind. 02 chemoreceptive activity has been recorded previously from only the pseudobranch of trout (Laurent and Rouzeau, 1972) and the first gill arch of tuna (Milsom and Brill, 1986). The characteristics of the receptors described in these two studies are extremely different. The stimulus-response characteristics of the 02 chemoreceptors reported in the present study are unlike those described from the trout pseudobranch but similar to those reported for tuna as well as for mammalian aortic and carotid body chemoreceptors (Milsom and Brill, 1986). Thus, under resting conditions neural discharge was generally random or erratic and activity increased initially as Po~ fell. The discharge rates for the O2-chemoreceptors in trout were higher than those of other vertebrate 02 receptors over the same Po2 range but the functional significance of this is obscure. Another striking similarity between trout 02 receptors and the 02 receptors described in other vertebrates was their powerful activation by cyanide; the chemoreceptor responses to cyanide were comparable to the responses described for
106 other vertebrate 0 2 receptors (Burleson et al., 1993). They were also consistent with observations that sodium cyanide triggers hypoxic ventilatory reflexes in other fishes (see Burleson et al., 1993 for review). At the lowest level of hypoxia administered in the present study ( ~ 30 Torr Po2), activity was then depressed. Such reversible depression of chemoreceptor activity during long-term severe hypoxia has also been demonstrated in mammalian carotid and aortic receptors and is believed to reflect a dependence of prolonged chemoreceptor activity on oxidative metabolism (Lahiri et al., 1983). The physiological significance of the value of Po~ at which depression of chemoreceptor discharge occurred in the present study (30-40 Torr) is not clear. This species has been shown previously to still possess a brisk hypoxic ventilatory response at 02 tensions in this range (Holeton and Randall, 1967). One possible explanation for this discrepancy between reflex and receptor studies is that the P% at the receptor site could have been lower at the same Po~ of the bathing solution in the receptor study. Although, bubbling the gill chamber with the hypoxic gas mixture during the nerve recording study maintained some convection over the gill surface, a relatively thick boundary layer around the gill might have been present. This should not be the case in vivo where high rates of water flow over the gill would reduce boundary layers and bring the Po~ of externally oriented receptors into closer equilibrium with the external medium. Based on the differential responses offish to aquatic hypoxia and arterial hypoxemia, several studies have concluded that fish have at least two functional groups of 0 2sensitive chemoreceptors (Burleson et al., 1993). Reflex hypoxic bradycardia is mediated exclusively by externally (water) oriented receptors while ventilatory reflexes are mediated by externally and internally (blood/tissue) oriented receptors. Neural recordings from the gills of tuna (Milsom and Brill, 1986) suggested that all O2-sensitive chemoreceptors in the first gill arch were sensitive to internal hypoxia while roughly two thirds were also sensitive to external hypoxia. In the present study, roughly one third of all receptors were sensitive to external hypoxia only, one third to internal hypoxia only and one third to both. Although the number of receptors recorded from are low in both of these studies, the data do suggest that there may be some species variation in the distribution of the O2-sensitive chemoreceptors between the external gill surface and the gill vasculature. Histological studies reveal neuroepithelial cells in the primary gill epithelium in this general location. Anatomically these cells resemble glomus or Type I cells of the mammalian carotid body and Falck fluorescence indicates the presence of monoamines in these cells just as they have been described in mammalian carotid body cells (Dunel-Erb et al., 1982; Donald, 1987). Although we were only able to quantify the responses of the receptors responsive to external hypoxia, the data suggest that, qualitatively, the responses of all O2-sensitive chemoreceptors are the same regardless of their orientation. This would imply that the differential heart rate response to external vs internal hypoxia seen in fish is due to a difference in the central connections of externally vs internally oriented receptors rather than differences in the functional characteristics of the receptors themselves. In conscious fish, decreasing blood 02 delivery by inducing hypoxemia, injecting
107 hypoxic blood or reducing blood flow to the gills stimulates ventilation (Cameron and Wohlschlag, 1969; Holeton, 1971; Smith and Jones, 1982). The activity of internal O2 receptors recorded with this preparation, however, were not very perfusion sensitive. The response to occluding perfusate flow was moderate in comparison to the responses of these receptors to perfusion with hypoxic solutions or NaCN. This would suggest that the metabolic rate of the chemoreceptor cells is quite low. Thus the response to exogenous removal of 02 from the perfusate was fast while the response to endogenous removal of 0 2 from stagnant perfusate by local metabolism during the occlusion procedure was slow. This is in contrast to the rapid change in discharge frequency observed in unperfused 02 chemoreceptors of other vertebrates including tuna (Milsom and Brill, 1986). Afferent neural activity showing regular bursting discharge patterns, as reported by Milsom and Brill (1986) in tuna for some 0 2 receptors, was only seen in two 0 2sensitive fibers during preliminary experiments. Milsom and Brill (1986) suggested that intrinsic vasomotion, as reported in fish gills (Satchell, 1962), may have been the cause of the bursting discharge pattern they observed in tuna receptors. In the present study, isoproterenol was initially added to the gill perfusate to enhance perfusion and subsequently, this was achieved by a high initial perfusion pressure which tended to dilate the gill vasculature and promote perfusion. These procedures may have abolished such intrinsic vasomotion in the trout preparations. Also, as described above, trout 0 2 receptors were much less sensitive to occlusion of the perfusate flow (Fig. 5) than tuna receptors, suggesting that vasomotion would probably have had little effect on afferent neural activity even if it were present.
Mechanoreceptors. Several different classes ofmechanoreceptors sensitive to displacement of various gill elements (arches, filaments and rakers) have been identified in fishes. Piscine gill mechanoreceptors, like other vertebrate mechanoreceptors, appear to be simple free nerve endings situated in muscle or connective tissue (Roberts and Rowell, 1988). It is thought that the anatomical location of a mechanoreceptor determines its sensory modality and response characteristics. Although a number of studies have demonstrated the presence of mechanoreceptors in gills, few have addressed the role that these may play in the control of cardioventilatory reflexes in fishes. Filament mechanoreceptors have been identified in sea raven sculpins (Hemitripterus americanus), Atlantic salmon (Salmo salar) and carp (Cyprinus carpio). Afferent neural information from these receptors travels in external branches of the pre- and posttrematic branchial nerves and activity levels increase in response to filament deflection. Only rapidly adapting receptors were described from this location in previous studies (Sutterlin and Saunders, 1969; de Graafet al., 1987), although slowly adapting filament receptors were also present in trout gills. The receptive field for filament mechanoreceptors was limited to a single filament in carp (de G r a a f e t aL, 1987) but appeared to include one or several filaments in salmon, sculpin and trout (Sutterlin and Saunders, 1969; this study). Mechanical displacement of one filament usually causes adjacent filaments to move, however, making it difficult to determine the limits of receptive fields.
108 Histological studies are needed to precisely determine the extent of filament mechanoreceptor innervation. Because filament receptors were not activated during breathing in lightly curarized carp, de Graaf et al., (1987) suggested that these receptors were not involved in cardioventilatory control. They suggested that filament receptors were more likely involved with feeding and gill defense responses. Stimulation of filament mechanoreceptors elicits filament adduction which should lower the resistance of water flowing over the gills thus facilitating cough and expulsion reflexes (de Graaf et al., 1987). Gill raker mechanoreceptors have previously been identified in dogfish, sea raven sculpin and carp. In dogfish (Squalus acanthias), both rapidly and slowly adapting mechanoreceptors were identified (Satchell and Way, 1962) whereas only rapidly adapting raker mechanoreceptor activity was recorded from carp (de Graafet al., 1987) and trout (this study). Although the experiments on dogfish were performed on isolated gills, Satchell and Way (1962) suggest that the slowly adapting receptors should be stimulated by normal ventilatory movements and could contribute to cardioventilatory control. The raker mechanoreceptors in carp were not active during normal ventilation which suggests that rapidly adapting raker mechanoreceptors do not contribute to cardioventilatory control but may also be primarily involved with feeding and protective reflexes. Gill arch mechanoreceptors located in the cartilaginous strip between the epibranchial and ceratobranchial elements have now been described in carp (de Graaf and Ballintijn, 1987) and trout. These receptors are tonically active and respond to adduction by decreasing activity and abduction by increasing activity. Because these were the only gill mechanoreceptors active during ventilation in carp, de Graaf and Ballintijn (1987) suggest that these are the receptors most likely to be involved with cardioventilatory synchronization. Baroreceptors are simply mechanoreceptors located in the walls of blood vessels. These receptors are stimulated by mechanical distortion of the blood vessel caused by changes in blood pressure. Baroreceptor reflexes have been demonstrated in elasmobranchs and teleost fishes (see Nilsson, 1984 for review). Increasing blood or perfusion pressure to the gills and electrical stimulation of branchial nerves causes a reflexive bradycardia. It should be noted, however, that electrical stimulation of the branchial nerves stimulates chemoreceptor and nociceptor afferent fibers which also mediate reflex bradycardia (Satchell, 1978; Burleson and Smatresk, 1990). Irving et al. (1935) recorded baroreceptor activity from branchial branches of cranial nerves IX and X in dogfish in vivo. They observed that branchial baroreceptor activity was synchronous with systole, was abolished by reducing blood pressure (hemorrhage) and increased in response to increased blood pressure (induced by infusion of adrenaline). Laurent (1967) has also recorded baroreceptor activity from the pseudobranch in tench that increased in response to increased perfusion pressure. The discharge characteristics of branchial baroreceptors in trout were very similar to those recorded in dogfish. Both groups of baroreceptors responded to rapid changes in perfusion pressure with a brief burst in activity followed by adaptation. Although the
109 precise location of branchial baroreceptors is not known, histological studies indicate that the baroreceptive loci in the gills may be at the junction of the afferent and efferent branchial arteries (Boyd, 1936; DeKock, 1963).
Are first gill arch receptors homologous to carotid body receptors?. According to phylogenetic and ontogenetic evidence, the carotid artery of mammals and the first gill arch of teleosts are homologs and this has led to the suggestion that "the gill capillaries of the first gill remain as a tangle of capillaries called the carotid body,..." (Yapp, 1965). Now, the accumulated evidence suggests that the stimulus response characteristics of O2-sensitive chemoreceptors and baroreceptors in fish gills are remarkably similar to those reported for carotid body receptors in mammals. The discharge characteristics of O2-sensitive chemoreceptors at rest, and the effects of hypoxia and cyanide on this discharge are very similar in both groups. So too are the discharge characteristics of the arterial baroreceptors, to the extent that they have now been studied in fish. Afferent information from both structures is carried in the glossopharyngeal nerve. The sum of this evidence would support suggestions that the diffuse network of 02 receptors and baroreceptors found in the first gill arch of fish are homologous to the more localized carotid body receptors found in mammals. It also suggests that the mechanism of 02 chemoreception is similar, if not identical, in all vertebrates and that these mechanisms have been conserved throughout the evolution of the vertebrates.
Acknowledgements.This work was supported by NSERC of Canada (WKM) and a McLean Fraser Research Fellowship (MLB). The authors thank Dr. N.J. Smatresk for his comments on the manuscript.
References Boyd, J.D. (1936). Nerve supply to the branchial arch arteries of vertebrates. J. Anat. Lond. 71: 157-158. Burleson, M.L. and N.J. Smatresk (1990). Effects of sectioning cranial nerves IX and X on cardiovascular and ventilatory reflex responses to hypoxia and NaCN in channel catfish. J. Exp. Biol. 154: 407420. Burleson, M.L., N.J. Smatresk and W.K. Milsom (1993). Afferent inputs associated with cardioventilatory control in fish. In: Fish Physiology, Vol. XIIB, edited by W.S. Hoar, D.J. Randall and A.P. Farrell. Orlando: Academic Press, pp. 389-426. Cameron, J.N and D.E. Wohlschlag (1969). Respiratory response to experimentally induced anaemia in the pinfish (Lagodon rhomboides). J. Exp. Biol. 50: 307-317. de Graaf, P.J.F., C.M. Ballintiin and F.W. Maes (1987). Mechanoreceptor activity in the gills of carp. I. Gill filament and gill raker mechanoreceptors. Respir. Physiol. 69: 173-182. de Graaf, P.J.F. and C.M. Ballintijn (1987). Mechanoreceptor activity in the gills of the carp. I1. Gill arch proprioceptors. Re~pir. Physiol. 69: 183-194. DeKock, L.L. (1963). A histological study of the head region of two salmonids with special reference to pressor- and chemo-receptors. Acta Anat. 55: 39-50. Donald, J. (1987), Comparative study of the adrenergic innervation of the teleost gill. J. Morphol. 193: 63-73. Dunel-Erb, S., Y. Bailly and P. Laurent (1982). Neuroepithelial cells in fish gill primary lamellae. J. Appl, Physiol. 53: R1324-R1353.
110 Holeton, G.F. and D.J. Randall (1967). Changes in blood pressure in the rainbow trout during hypoxia. J. Exp. Biol. 46: 297-305. Holeton, G.F. (1971). Oxygen uptake and transport by the rainbow trout during exposure to carbon monoxide. J. Exp. Biol. 54: 239-254. Irving, L., D.Y. Solandt and O.M. Solandt (1935). Nerve impulses from branchial pressure receptors in the dogfish. J. Physiol. (London), 84: 187-190. Lahiri, S., N.J. Smatresk and E. Mulligan (1983). Responses of peripheral chemoreceptors to natural stimuli. In: Physiology of the Peripheral Arterial Chemoreceptors, edited by H. Acker and R.G. O'Regan, Ch. 10. Amsterdam: Elsevier Science Publishers, pp. 221-256. Laurent, P. (1967). La pseudobranchie des T616osteens preuves electrophysiologiques de ses fonctions ch6mor6ceptrice et baror6ceptrice. C.R. Acad. Sei. Paris 264: 1879-1882. Laurent, P. and J.D. Rouzeau (1972). Afferent neural activity from pseudobranch of teleosts. Effects of PO2, pH, osmotic pressure and Na ÷ ions. Respir. Physiol. 14: 307-33l. Milsom, W.K. and R.W. Brill (1986). Oxygen sensitive afferent information arising from the first gill arch of yellowfin tuna. Respir. Physiol. 66: 193-203. Nilsson, S. (1984). Innervation and pharmacology of the gills. In: Fish Physiology, Vol. XA, edited by W.S. Hoar and D.J. Randall. New York: Academic Press, pp. 185-227. Perry, S.F., P.S. Davie, C. Daxboeck, A.G. Ellis and D.G. Smith (1984). Perfusion methods for the study of gill physiology. In: Fish Physiology, Vol XB, edited by W.S. Hoar and D.J. Randall. New York: Academic Press, pp. 325-388, Roberts, J.L. and D.M. Rowell (1988). Periodic respiration of gill-breathing fishes. Can. J. Zool. 66: 182190. Satchell, G.H. (1962). Intrinsic vasomotion in the dogfish gill. J. Exp. Biol. 39: 503-512. S atchell, G.H. and H.K. Way (1962). Pharyngeal proprioceptors in the dogfish, Squalus acanthias L, J. Exp. Biol. 39: 243-250. Satchell, G.H. (1978). The J reflex in fish. In: Respiratory adaptations, capillary exchange and reflex mechanisms, edited by A.S. Paintal and P. Gill-Kumar. Delhi: Vallabhbhai Patel Chest Institute, University of Delhi, pp. 432-441. Smith, F.M. and D,R. Jones (1978). Localization of receptors causing hypoxic bradycardia in trout (Salmo guirdneri). Can. J. Zool. 56: 1260-1265. Smith, F.M and D.R. Jones (1982). The effect of changes in blood oxygen-carrying capacity on ventilation volume in the rainbow trout (Salmo gairdneri). J. Exp. Biol. 97: 325-334. Sutterlin, A.M. and R.L. Saunders (1969). Proprioceptors in the gills of teleosts. Can. J. Zool. 47: 12091212. Yapp. W.B. (1965), Vertebrates: Their Structure and Life. New York: Oxford University Press, pp. 525.