Oxygenandacidchemoreceptknin the carotidbody chemoreceptors C. Gonz~lez, L. Almaraz, A. Obeso and R. Rigual C 6onzJlez,L

Almaraz,A. Obeso and R. Rigualare at the Depa~mento de Oioquimica,Oiologia Mo/ecu/ary Fisiolog[a, Facultadde Medicina, Universidadde Valladolid, 47005 Valladolid, 5pain.

The carotid bodies are arterial chemoreceptors that are sensitive to blood Poe, Pco2 and pH. They are the origin of reflexes that are crucial for maintaining Pco~ and pH in the internal milieu and for adjusting the 02 supply according to the metabolic needs of the organism in situations of increased demand, such as exercise and while breathing at decreased 02 partial pressures during ascent or when living at high altitude. Chemoreceptor cells of the carotid body transduce the bloodborne stimuli into a neurosecretory response that is dependent on external Ca 2+. These cells have an 02-sensitive K + current that is reversibly inhibited by low Poz It is proposed that the depolarization produced by inhibition of this K + current activates Ca 2+ channels; Ca 2+ influx and neurosecretion follow. The cells have also a potent Na+-Ca 2+ antiporter that could be responsible for the intracellular Ca 2+ rise required to trigger the release of neurotransmitters during high Pco2 or low pH stimulation.

static and adaptive roles in maintaining Pco., and pH in the internal milieu and adjusting the 02 supply according to the metabolic needs of the organism. Other targets of chemoreflexes are the heart and the resistance and capacitance vessels, although the physiological significance of these reflexes is ill defined3. T h e c h e m o s e n s o r s t r u c t u r e : c e l l s or n e r v e endings?

In dealing with the physiology of the carotid body at the receptor level, the first aspect to consider is which element within the receptor complex is the primary sensing element. In 1928 De Castro 4 showed that chemoreceptor cells of the carotid body are innervated by sensory neurons, and postulated that the chemoreceptor cells are the primary sensors that, via the products of their metabolism, activate the sensory nerve endings. This view was reinforced by ultrastructural and biochemical findings that recogThe carotid bodies, located near the carotid artery nized the synaptic relationships between chemorecepbifurcations, are chemoreceptor organs sensitive to tor cells and sensory nerve endings, and the presence changes in blood Po~, Pco~ and pH. They have two of synaptic vesicles and putative neurotransmitters in types of parenchymatous cells, chemoreceptor and chemoreceptor cells s'6. In the late 1960s, Biscoe and co-workers 7 repeated sustentacular cells, which form clusters surrounded by a dense net of fenestrated capillaries (Fig. 1). The De Castro's degeneration experiments and reported proximity of the capillaries to chemoreceptor cells that the nerve endings in synaptic apposition to minimizes the diffusion pathway for bloodborne stimuli. chemoreceptor cells were efferent, and therefore Chemoreceptor cells are derived from the neural crest secretomotor. They attributed the primary sensory and exhibit a heterogeneous population of synaptic function to poorly defined free nerve endings that had vesicles in their cytoplasm that contain different been described earlier but were never confirmed. In putative neurotransmitters. Sustentacular cells are Biscoe's conception, chemoreceptor cells were glial in nature and lack specialized organelles. Sensory glandular-like, serving either a hormonal or a paranerve fibres reach the organ via the carotid sinus crine function. This model of chemoreception was nerve, ending in synaptic apposition to the chemo- dismantled when several authors confirmed De receptor cells 1. Castro's findings on the sensory nature of the endings The carotid bodies are tonic receptors with low apposed to chemoreceptor cells 1. Finally, the Po~firing frequency (< 2 impulses per s per fibre) at sensing properties of chemoreceptor cells were normal blood Po~ (100 mmHg), Pco2 (40 mmHg) and demonstrated by showing that carotid bodies from pH (7.4). The action potential frequency is augmented which the sensory afferents were chronically denerwhen blood Po decreases or when Pco~ or [H +] vated responded in vitro to a decrease of Po~ in the increase. Phys~ologacally, the most effective stimulus bathing solution with an increased rate of release of is the decrease in blood Po~. An increase in sensory dopamine (DA)8. discharges is already evident at a Po~ of 70 mmHg, when arterial blood 02 content is around 94% of the C h e m o t r a n s d u c t i o n a s a p r o c e s s c o u p l i n g normal level, and at 50 mmHg (moderate hypoxia) the s t i m u l a t i o n a n d s e c r e t i o n sensory activity is similar to that seen at a Pco~ of Upon stimulation, sensory cells in composite recep75 mmHg and a pH of 7.0, which are indications of a tors should release neurotransmitters in proportion to severe acidosis 1. the intensity of the stimulus and in parallel to the Stimulation of the carotid bodies elicit reflexes, the action potential frequency recorded in the sensory main target of which is the respiratory system. A nerve 9. Therefore, the re-definition of the carotid hyperventilation proportional to the stimulus intensity body as a 'composite receptor' made sensory transensues. The carotid bodies are responsible for about duction in this receptor equivalent to a stimulus90% of the hyperventilation seen in hypoxic hypoxia, secretion coupling process 9. It has been shown that, and for 20-50% of that observed in respiratory and among several putative neurotransmitters, chemometabolic acidosis. The rest of the respiratory drive is receptor cells release DA in proportion to the provided by other arterial chemoreceptors when the intensity of hypoxic and acidic stimulation, and that stimulus is hypoxia and by the central chemoreceptors the release response is paralleled by the electrical in the case of acidosisz'3. Throughout these respir- activity in the carotid sinus nerve' ' . The same atory actions, the carotid bodies play crucial homeo- relationship applies to other chemostimulant agents •

.

2

,

.

8

146

© 1992, ElsevierScience PubfishersLtd, (UK) 0166- 2236/92/$05.00

10

11

TINS, Vol. 15, No. 4, 1992

including cyanide, dinitrophenol and 2-deoxyglucose 12'1z. Responses of the carotid body to B A selected intensities of these stimuli are summarized in Fig. 2. Although the role of DA in the organ is not conclusively established, these data indicate that DA 5 release is a reliable index of chemoreceptor cell activation and represents a final output of sensory transduction in these cells. The possible significance of the released DA in the process of chemoreception has been considered in a recent review ~4. An aspect of the release of DA that has been systematically studied is its dependence on external Ca2+. Table I shows that over 95% of the DA released in response to hypoxia and high concentrations of extracellular K +, [K+]o, and about 80% of that elicited by other stimuli are abolished in Ca2+free medium8'11'15-~7. The table also shows that the dihydropyridines nisoldipine and Bay K 8644, modulators of voltage-sensitive Caz+ channels, modify the release of DA induced by high [K+]o (as expected by their known actions in other systems) as well as the release induced by low Po~. In contrast, these agents do not modify the basal release or that induced by a high Pco2 (low pH) stimulus, dinitrophenol or sodium Fig. 1. Drawing of the carotid artery bifurcation and a cellular cluster of the propionate ~7'~8. Therefore, although the response to carotid body. (,6,) A frontal view of the region of the nght carotid artery all the stimuli tested is highly dependent on external bifurcation in the rabbit. The common carotid artery divides (1) to give the intemal (2) and external (3) carotid arteries. The carotid body (4) is located on Ca2+, a clear difference emerges between the two the internal carotid artery close to the bifurcation. Sensory fibers originating in physiological stimuli: while hypoxia seems to promote the petrosal ganglion (5) reach the carotid body via the carotid sinus nerve (6). the entry of Ca2+ via voltage-sensitive Ca2+ channels, The superior cervical ganglion (7) also innervates the bifurcation area, acidic stimuli should activate alternative pathways for including the carotid body, via the ganglioglomerular nerves (8). The nodose Cae+ entry. ganglion (9) is situated externally to the internal carotid artery. (B) A cluster of parenchymatous cells of the carotid body, which are formed by chemoreceptor cells (1) that are partially surrounded by sustentacular cells (2). The proportion Chemotransduetion of 02 (1) A plasma membrane model. An inference from of chemoreceptor to sustentacular cells is about 3-5 to one. Chemoreceptor the data described above is that hypoxia should cells have in their cytoplasm a heterogeneous population of synaptic vesicles depolarize chemoreceptor cells to activate voltage- (3), which are usually located near the contacts with the sensory nerve endings (4) that originate from branching fibers of the carotid sinus nerve (5). The dependent Ca2+ channels. Therefore, a search for the clusters are surrounded by a dense net of capillaries (6). mechanisms capable of producing such a depolarization was initiated. In whole-cell voltage clamp experiments it was found that freshly dissociated chemo- P% (Fig. 3). In a recent study23 in excised membrane receptor cells from the carotid bodies of the adult patches of chemoreceptor cells isolated from adult rabbit are excitable cells with Na +, K + and Ca2+ rabbit carotid bodies, it was found that the Po2voltage-dependent channels. In this initial study 19 it sensitive K + channels have mean conductances of was also found that lowering P% in the bathing 20 pS, and that their opening probability (Po) desolution reversibly inhibited the outwardly directed creased on lowering Po2 - even when the Ca2+ K ÷ currents without affecting the inwardly directed concentration in the solution bathing the internal face Na + and Ca~+ currents. It was proposed that the of the membrane was maintained below I riM. inhibition of the K ÷ current initiated the depolarization Electrophysiological data obtained from chemoof chemoreceptor cells, in a way similar to that receptor cells that have been isolated from embryos proposed for gustatory and olfactory cells exposed to or neonatal animals are different from those just appropriate stimuli. described. Cells from rabbit embryos lack TTXAnalysis of the whole-cell currents z°'21 indicated sensitive inward currents, but Ca2+ and K + currents that the threshold for activation of the K + currents are mostly similar to those obtained in adult chemowas about - 3 0 mV and that the currents have a receptor cells, including the sensitivity of the K + component that is Caz+ dependent. Na + currents current to low Poo (Ref. 24). In addition, embryonic were inhibited by submicromolar concentrations of chemoreceptor ce~s seem to exhibit an ill-defined tetrodotoxin (TTX) and exhibited typical activation inwardly rectifying K + current that appears to be and inactivation properties. Ca2+ currents were active at negative potentials (in the range of the mediated by L-type channels. In all the cells tested, estimated membrane potential of - 5 0 mV), and that switching from a voltage to a current clamp produced is also reversibly inhibited by low Po2 (Ref. 25). a depolarization and large action potentials (normally In some studies of chemoreceptor cells isolated just a single action potential followed by a steady from neonatal rats, these cells have been found to lack depolarization, but trains of spontaneous spikes were TTX-sensitive transient inward currents 26, while in also recorded). other studies, they appear to exhibit these currents 27. An analysis 22 of the inhibition of K + currents by low Another apparent discrepancy in these studies is that Po~ indicated that only a transient CaZ+-independent while the I-V curves for the K + current showed a component of this type of current was sensitive to the prominent shoulder at test potentials between + 10 and TINS, Vol. 15, No. 4, 1992

147

16 (9

[--]

CSN activity

T

l

DA release

',

12 ._> "5

8

.Q O (9 Q. .m

. _ _

Low Po2

Acid

CO2

CN

DNP

2-DG

Fig. 2. Effects of different stimufi on the release of [3H]DA and carotid sinus nerve activity in an in vitro preparation of the carotid bodies of a cat. Responses of both activities were simultaneously obtained from the same preparation. The carotid bodies were loaded with [3H]tyrosine to label DA deposits prior to the actual expenments. The data are expressed as multiples of basal release or nerve activity, respectively. Stimuli were applied for 10 min in every case by superfusing the preparations with the indicated test solutions. Basal conditions: superfusion with balanced saline equilibrated with 95-100% 02 and adjusted to pH 7.4. Experimental conditions are shown on the x axis. (It should be noted that in the slow superfusion system used in these experiments the threshold response to low Po2 was 40-50% 02 in N2.) Low Po2 is 20% 02/80% N2, the acid solution is pH 6.8, CO2 is a solution equilibrated with 20% C02/80% 02, pH 7.4, DNP is dinitrophenol at 0.25 mM, CN is sodium cyanide at 0.25 mM, and 2-DG is glucose-free saline containing 5 mM pyruvate as an energetic substrate and 4 mM 2-deoxyglucose. In all cases data are means +_SEM of 6--12 individual values. (Taken from Refs 10-13.)

+30 mV in the studies by Peers et al. 26'2s, such a component was not obvious in the I-V curves of the study by Stea and Nurse 27. A detailed characterization of the shoulder of the I-V curves indicated that the K + current was Ca 2+ dependent, and, in addition, it was shown that low Po~ selectively inhibited this cornponent of the K + current 28. The identity of the TABLE I. Ca2+ dependence and modulation by dihydropyridines of [3H]DA-

induced release from carotid bodies in the rabbita. Stimulus

Increase in the amount of DA released under certain conditions

2 mM

C a 2+

0 m M C a 2+

Nisoldipine b

2 mM

2 m M Ca 2+ Bay K8~14 b

Ca 2+

25mMK +

2.50_+0.15

0.20_+0.00

0.04_+0.01

12.20+1.30

Low Po2 (7% 02)

10.25-+2.10

0.40-+0.01

2.32+0.51

30.52+4.10

High Pco2 pH = 6.8

1.61 -+0.17

0.31 +0.11

*

1.50+0.21

Na+-Prop. (15 mM)

1.05 -+ 0.16

0.22 + 0.05

*

1.10 -+ 0.17

DNP (75 W~)

2.44 -+ 0.22

0.56 + 0.05

2.31 _+0.21

*

aData are expressed as multiples of basal release and are the means_+sEMof 6--10

measurements (Refs 17, 18 and Gonz~lez, C., Almaraz, L., Obeso, A. and Rigual, R., unpublished observations). bNisoldipine(625 riM) and Bay K 8644 (1 i~) were presented 10 min prior to and during stimulusapplication. * Not studied. Abbreviations:DA, dopamine; DNP,dinitrophenol; Na+-Prop. sodiumpropionate. 148

component of the K + current inhibited by low P02 was not established in the study by Stea and Nurse 27. A conclusion emerging from all of these results is that chemoreceptor cells, which are sensitive to low P o throughout the life of the animals studied, exhibit currents that are inhibited by low Po~. There are age-specific (and species-specific) differences in the type of K + current inhibited, as well as in other basic electrical properties of the cells. Developmental differences of some kind in the basic functioning of chemoreceptor cells are expected since the carotid bodies of fetal animals exhibit stimulus-response curves, which relate Po~ to firing frequency in the carotid sinus nerve, that are displaced far to the left compared with similar curves in the adult29. Important differences have also been observed in the content and turnover of neurotransmitters perinataUy3°. Within the first two weeks of postnatal life 29'31'32 all the parameters studied are reset in such a way that the chemoreceptor function tends towards that observed in the adult. The mechanisms involved in this resetting are u n k n o w n 32. With the information provided in this and the preceding section, and restricting our consideration to adult carotid body chemoreceptor cells, the following model for low Po2 transduction can be proposed. Low Po2, by reducing the Po of K + channels, produces the initial depolarization required to activate the voltagedependent Ca 2+ channels, and entry of Ca 2+ then follows. Simultaneous activation of Na + channels provides a fast recruitment of Ca 2+ channels, potentiating the entry of Ca 2+ and the release of neurotransmitters. Ca'~+-dependent K + channels contribute to cell repolarization (Fig. 4A). The operativity of this model rests on the premise that inhibition of the K + channel is capable of triggering the initial depolarization. Owing to the voltage dependency of the inhibited current (the threshold activation of which is - 3 0 mV), this would be possible if, at more negative potentials, in the range of the membrane potential of - 5 0 mV (Ref. 33), the 02-sensitive K + channels are effectively contributing to the total resting K + conductance of the cells. Taking into account the small size of the cells and the high resistance of their membranes 2°, the closure of only a few K + channels can modify significantly the membrane potential of the cells and trigger action potentials 34. This behavior is consistent with the observation made in current-clamp recordings, where an increase in the firing frequency is observed on switching to low Pox soluti°ns22If the proposed model for the transduction of the hypoxic stimulus is to be accepted, it must be demonstrated that the inhibition of the K + current by low Po, is fast enough to explain the changes occurring in the intact carotid body. Since in the intact animal an increase in carotid sinus nerve action potential frequency is observed within a few seconds after application of the hypoxic stimulus 1, the.inhibition of the K + current by low P% levels should develop almost instantaneously. The timecourse of Na + and K + current inhibition has been studied after superfusing the cells with a TTX-containing hypoxic solution. It was found that the tl/2 of the effect of low P% on the K + current is less than that of TTX blockade of Na + channels (L6pez-L6pez, J. R. and Gonz~lez, C., unpublished observations). Since TTX TINS, Vol. 15, No. 4, 1992

A

control

B

control

I°wl Oo2 ~ ~

low Po~

1 nA

0.5 nA 2 ms

C

control

recovery

Dr

co tro

A

3 nA

r 5 ms

50 ms

Fig. 3. Effects of low Po2 on voltage-dependent ionic currents recorded in whole-cell configurations from chemoreceptor cells of the carotid bodies in adult rabbits. (A) Inward (Na + + C,.12+) currents recorded during voltage steps to O mV from a holding potential of - 8 0 mV in the control solution (Po2 = 150 mmHg), and3 min after exposure to low Po2 (10 mmHg). Solutions (in mM) were: 140 Na + and 2.5 Ca 2+ in the bath; 130 Cs +, 3 Mg2+-A TP and 10 EGTA in the pipette. (B) Na + currents recorded in the control solution and 5 min after switching to the low Po2 solution. The test pulse was to + 10 mV with a return to a holding potential of - 8 0 mV. Solutions (in mM): 140 Na + and 0.5 Cd 2 + in the bath; 130 Cs +, 3 ME2+-ATP and 10 EGTA in the pipette. (C) K + currents recorded during pulses to + 4 0 m V with a return to - 8 0 mV in the control solution and during exposure to low Po2. The recovery trace was obtained 3 min after the solution with normal Po2 was introduced again into the chamber. Solutions (in mM): 140 Na +, 5 Ca2+ and 0.001 TTXinthebath; 130K + , 3 / v l g 2 + -ATPandO.OOO5ionicCa 2 + (4.44Ca 2 + a n d 5 E G T A ) i n t h e p i p e t t e . ( D ) K + currents elicited by 200 ms pulses to +40 mV with a return to a holding potential of - 8 0 mV in the control solution and during exposure to low Po2. Solutions (in mM): 140 Na +, 5 Ca2 + and 0.001 TTX in the bath; 130 K + , 3 ME 2+ -ATP and 10 EGTA in the pipette. (Taken, with permission, from Ref. 22.)

blockade of Na + channels develops within a few Working with an intact carotid body preparation in hundred milliseconds 35, this finding indicates that the vitro, it has been found39 that the relationship between inhibition of the K + current is indeed instantaneous. P02 and the increase in cyclic AMP (cAMP) observed A major criticism 36 concerning the significance of during hypoxic stimulation is similar in shape to that the inhibition of the 02-dependent K + current relates for the K + current inhibition, but is displaced to the to the nature of the relationship between Po2 and the left by about 40 mmHg. In preliminary whole-cell inhibition itself. In whole-cell recordings, a maximum recordings it has been .observed that addition of inhibition was found at a Po~ near 85 mmHg, and the dibutyryl-cAMP (a permanent analogue of the natural same degree of inhibition was observed at Po2 values nucleotide) to the bathing solution also inhibits the of 10, 65 and 110 mmHg 22. In single channel record- transient component of the K + current (L6pez-L6pez, ings from isolated patches the maximum decrease in J. R., unpublished observations). Therefore, it is conthe Po of the 02-sensitive K + channel was observed at ceivable that, in physiological conditions, cAMP (and about 100 mmHg; again, the decrease in the Po was probably other second messengers) modulate the 02about the same at 50 and 130 mmHgz3. There are two sensitive K + channel to make it sensitive to the problems in interpreting these results. effective 02 pressures. In this context is should also be On the one hand, the resting Po,, in carotid body mentioned that addition of forskolin (an activator of tissue is about 75 mmHg, accordifig to Whalen et adenylate cyclase), isobutylmethyLxantine (an inhibitor al. 37'38 (lower according to other authors, see below), of phosphodiesterase) or dibutyryl-cAMP to the and yet maximum inhibition of the K + current is bathing solution potentiates the release of DA in obtained at 85 mmHg in whole-ceU recordings. On the response to different stimuli, the potentiation being other hand, carotid sinus nerve action potential maximum with the hypoxic stimulus 4°. frequency starts to rise at a tissue Po2 of about Another important aspect to be considered in the 50 mmHg and continues to increase as tissue Po~ membrane model of P02 transduction is the identity of decreases 1, while the inhibition of the K + current by the 02 sensor. Although it has been suggested many low Po follows the complex pattern described above. times that it is a hemoprotein 1, nothing is known Thus, t~hephysiological significance of the K + current about its chemical nature. Whatever its chemical inhibition as an initial event in the hypoxic transductive identity, it seems to be located in the plasma cascade is questionable, unless a lack of modulation of membrane, because low Po~ decreases the eo of K + the K + channels, due to the recording conditions, is channels in excised membrane patches. In fact, it has responsible for the high sensitivity of the channel and been postulated that the channel itself may be the the peculiar shape of the relationship between Po~ and sensor ~3. Recently the suggestion has been made 41'42 the K + current inhibition. that a heine-linked NADPH oxidase could be the 02 TINS, Vol. 15, No. 4, 1992

149

A

K+

02

_\

\ \ \

Ca2+

,f

! ~

~

~__

3

\

\ cAMP / 5/

C a 2+

/

//+ 1

ATP

DA

8

1

DNP

co:

AH

I

H + ~ / Cl-

\ C a 2+

-

cI-

Fig. 4. Models for sensory transduction in chemoreceptor cells of carotid bodies. (A) The proposed cascade for low Poe transduction. On lowering P02 (I) a signal originates at the 02 sensor, inhibiting the K + current and activating adenylate cyclase (2). A drop in membrane potential follows the inhibition of the K + current, which in turn activates voltage-dependent Ca2+ channels leading to Ca2+ influx and an increase in the intracellular concentration of Ca2+, [Ca2+]i (3). Increased [Ca2+]i results in the activation of the exocytotic machinery (4). Increased cyclic AMP levels modulate the K + current and the exocytotic machinery, making low Po2 stimulation more effective (5). The model does not exclude the participation of other second messengers systems as well as an interplay between influx of extracellular Ca2 + and Ca2 + from intracellular stores. (B) Proposed model for the transduction of acidic stimuli. Some possible ways of increasing the intracellular H + concentration [H+]~ are indicated (I), such as by diffusion of protonated weak acids (AH), C02 or H + ions, or by the operation of protonophores like dinitrophenol (DNP). The increase in [H+]i stimulates (dashed arrows) the operation of Na +-coupled, H+-extruding systems that bring Na + into the cell. Two of these systems are represented: the Na +--H + exchanger (2) and a Na +-dependent HC03--CIexchanger (3). The increase in [Na +]i drives the entry of Ca2+ through the Na+-Ca 2+ exchanger (4), resulting in the activation of the exocytotic machinery (5). Evidence for the presence of a Na +-independent HC03 - - C I exchanger has recently been provided 64. Since experimental data show that 20% of the acid-evoked release of DA is not dependent on external Ca2+, this part of the response could be attributed to Ca2+ derived from intracellular stores. [Part (B) taken, with permission, from Ref. 17.] 150

sensor in chemoreceptor cells. According to this proposal, a reduction in hydrogen peroxide produced by the oxidase during low Po2 stimulation changes the glutathione redox state, which in turn alters the configuration of membrane proteins and hence the conductivity in ionic channels. The participation of this system in the modulation of ionic conductances during hypoxic stimulation is conceivable, but it remains to be demonstrated which conductances are affected. On the other hand, the isolated membrane patch recordings show that low Po~ is able to modify K + conductance even when this mechanism is not operative. (2) The metabolic hypothesis: a mitochondrial model. A different view of low Po~ transduction in the chemoreceptor cells of the carotid body has evolved in recent years from the metabolic hypothesis of chemoreception put forward in 1963 by Anichkov and Belenkii 43. They proposed that hypoxia, as well as metabolic poisons, activate the chemoreceptors by decreasing ATP levels. No link was provided between the presumed decrease in ATP and the release of the 'hypoxic' neurotransmitter. Criticisms of the metabolic hypothesis came from the consideration that cytochrome oxidase has a high affinity for Oz, while the carotid body chemoreceptors have a very high blood flow and a hypoxic threshold at arterial blood Po~ of about 70 mmHg 44. The description of a cytochrome oxidase with a very low 02 affinity (Km~90 mmHg) within the carotid body tissue 45, in addition to the normal high-affinity one (Km~ 5 mmHg), seemed to provide one of the missing pieces in the metabolic hypothesis. The existence of such a cytochrome oxidase has been questioned on the basis of direct spectrophotometric measurements 46, but supported on the basis of good fitting of O2-consumption data to a simulated model comprising two cytochrome oxidases 37'3s. More recently, however, Wilson et al. 47, using a new optical method for measuring 02 concentrations, found that the O2 dependence of oxidative phosphorylation in isolated liver mitochondria extends well into the physiological range of Po~. They concluded that in tissues such as the carotid body there is no need for a low-affinity cytochrome oxidase to generate a decrease in the [ATP]/[ADP][Pi] quotient (phosphate potential), even at moderate hypoxic stimulation. The oxygen partial pressures reported for the carotid body tissue are not helpful in making any inference on the existence in chemoreceptor cells of a cytochrome oxidase with a low affinity for oxygen. With the mean oxygen partial pressures reported by Acker and co-workers 70 in normal resting conditions 25 and 7 mmHg for the cat and the rabbit carotid body, respectively), the chemoreceptors of the carotid body would be firing at an almost maximal rate ff it is assumed that the Po~ sensor is a low-affinity cytochrome oxidase. On the contrary, the tissue P Q values reported by Whalen and co-workers 37 ' 38 would in fact require the existence of a low-affinity enzyme for the mitochondria to be able to trigger chemoreception in the first place. No carotid body tissue Po2 values were reported by Wilson. Whatever the case regarding the cytochrome oxidases, a different question concerns the nature of the signal for chemoreception that results from the detection of low Pox at the mitochondria. According to TINS, VoL 15, No. 4, 1992

some authors 48 (see below and Refs 36 and 49), the signal is a decrease in the mitochondrial electrochemical gradient, which then results in an impairment of the mitochondrial uptake of Ca2+, a rise in the intracellular Ca2+ concentration [Ca2+]i, and the release of neurotransmitters. According to other authors, the important signal is a decrease in the phosphate potential in the cell cytoplasm47'5°'~1. Although a decrease in the mitochondrial electrochemical gradient also produces a decrease in the phosphate potential, the latest evidence argues that the intrinsic mitochondrial electrochemical gradient is not paramount to chemoreception, because oligomycin (a blocker of the ATP-synthesizing mitochondrial ATPase) preserves such a gradient and excites the chemoreceptors, while hypoxia and cyanide, which decrease the gradient, also excite the chemoreceptors. The signal triggering the chemoreception would be, according to these authors 47,s°,5~, the decrease in the phosphate potential, although no explanations of how it might occur have been put forward. As pointed out by Acker et al. 41, it is difficult to envision a decrease in the phosphate potential being a signal for low Po~ chemoreception, since the carotid body is in fact requtred to function maximally at low P%. A restatement of the Ca2÷ version of the metabolic hypothesis has been put forward quite recently36'49. According to this new theory, low P o slows the electron transfer in the resptratory chain of chemoreceptor cells, leading to a decrease in the mitochondrial proton electrochemical gradient. As a consequence, mitochondria release Ca2÷ and increase [Ca2+]i, which then triggers neurotransmitter release. The presence of a low-affinity cytochrome oxidase is included in the model to explain the full range of carotid body response to hypoxia, although the experimental data presented do not support the premise (see Ref. 49). The assumption is made that mitochondria have high enough levels of Ca2÷ to produce significant rises in [Ca2+]i. In the model, the voltage-operated Ca2÷ channels are activated in normal conditions to provide the Ca2+ needed to load the mitochondria. The crucial experimental observation for the model is that the rise in [Ca2+]i observed both during anoxia (achieved by N2 equilibration and addition of dithionite) or cyanide application persists, although reduced in amplitude, in bathing solution that is virtually free of Ca2+ (Refs 52, 53), and it is not modified by caffein (an agent that releases Ca2+ from endoplasmic reticulum). The effects of dithionite, a very powerful and unspecific reducing agent, on intracellular Ca2÷ deposits are unknown (see below). It is difficult to reconcile all the mitochondrial Ca2+ versions of the metabolic hypothesis with some wellestablished facts such as the dependence on extracellular Ca 2+ and the dihydropyridine sensitivity of the low P%-induced release of DA8' 16,18 (see Table I). It is also difficult to reconcile these proposals with the current ideas about the Ca2÷ content of mitochondria in healthy resting cells~4's5. Mitochondria have levels of Ca2+ that are not dissimilar to those found in the cell cytosol; therefore, they would play a minor role, or no role at all, in the transient elevation of cytosolic Ca2+ (Refs 54, 55). Contrary to the proposals of Biscoe and Duchen 36'49, if mitochondria accumulate Ca2+ it must be during enhanced physiological activity, not at rest, because it is only during cell activation that cytoplasmic Ca2+ would reach the concentrations •

TINS, VoL 15, No. 4, 1992

,2

required for operation of the low-affinity, Ca2+-uptake systems in mitochondria S4'55. This section on low P% (hypoxic hypoxia) transduction should not be concluded without some considerations on the mechanism(s) of action of cyanide, a metabolic poison that produces histotoxic hypoxia and is a powerful chemostimulant t. Because it is easy to work with, cyanide has often been used in arterial chemoreception experiments with the tacit or explicit assumption that it excites the arterial chemoreceptors in the same way that low P02 does. However, in 1968 Coxon$6 had already pointed out that cyanide inhibits, in addition to cytochromes, some 40 different enzymes (see also Ref. 57); that is, cyanide not only blocks 02 consumption to produce histotoxic hypoxia but also affects many other cellular processes. Therefore, it would be an oversimplification to ascribe the chemostimulant power of cyanide solely to its capability to produce histotoxic hypoxia. In fact, it has been reported 41'46 that cyanide (10 mM) and anoxia induce different light absorbance spectra in the carotid body tissue, indicating that different pigments have been affected by both stimuli. In recent electrophysiological experiments 5s in chemoreceptor cells of carotid body tissue that were isolated from adult rabbits, but were not exposed to culture medium, it was found that cyanide (2 mM) increased a Ca2+-dependent component of the whole K + currents elicited by brief depolarizing steps, and at the same time decreased Ca2+ currents. At holding potentials more positive than - 3 0 mV and in the presence of extracellular Ca 2+, prolonged exposures to cyanide also produced an increase in outward currents. This was interpreted to be due to a steady increase in Ca2+-dependent K + currents produced by a steady increase in cytoplasmic [Ca2+] 5s. These observations, added to the fact that the inclusion of ATP in the recording pipette did not modify the electrical responses to the poison58, led to the proposal that the presumed increase in cytoplasmic Ca2+ must be due to the release of Ca2+ from mitochondria 36'49"58. It was assumed that cyanide acted by the same mechanisms as low Po~, and so these findings were pivotal for the restatement of the metabolic hypothesis36, 49. The-~ta of Biscoe and Duchen 58, however, can be interpreted in another way. On the one hand, it is questionable that mitochondria can release Ca 2+ (see above) and in fact Biscoe et aL 52 found that 'in only six out of eleven preparations, a response (an increase in [Ca2+]i) was still seen soon after removal of [Ca2+]o, but was smaller than the control. In the remainder, [Ca2+]i fell rapidly on removal of [Ca2+]o, and no increase could be seen in response to cyanide'. These findings evoke two considerations. (1) If the previously mentioned electrical response to prolonged exposure to cyanide was in fact due to an increase in [Ca2+]i, the intracellular origin of this ion is uncertain. It seems plausible that at holding potentials more positive than - 3 0 mV the Na+-Ca 2+ antiporter is producing a net Ca2+ influx59. (2) The contention in subsequent papers 36'49 that cyanide increases [Ca2+]i in the absence of extracellular Ca 2+ seems questionable, unless [Ca2+]i deposits involved in the response are depleted over an unusually fast timecourse. On the other hand, the net effects of cyanide on chemoreceptor cells from the carotid bodies of adult 151

rabbits contrast with those of low Po... The natural stimulus produces a reduction m a transient Ca-~-independent component of the K + current 19'22'2a and does not modify Ca2+ currents 19'22. However, it should be noted that in these latest studies chemoreceptor cells were exposed to culture medium for between 4 and 48 hours, and, although unlikely (the same responses were obtained in all the cells, independent of their age), some differences between preparations cannot be excluded. In chemoreceptor cells from carotid bodies of rabbit embryos that had been cultured for three days, low Pot reduced K ÷ currents and did not affect Ca2÷ currents, while cyanide (1 mM) reduced Ca2+ currents 24. In chemoreceptor cells from carotid bodies of neonatal rats, cyanide (2 mM) reduced an uncharacterized component of the whole K ÷ current and low P% reduced the Ca2+-dependent component of the K ÷ current 28'57. Even when the apparent differences in the electrophysiological responses to cyanide between preparations of chemoreceptor cells are taken into account, we would suggest, in agreement with other authors 24'41'46'56'57, great caution in equating the mechanisms of action of low P% and cyanide. ,

.

2

,

Transduction of acidic stimuli The Ca2+ dependence of the release of DA induced by acidic stimuli and the insensitivity of DA release to dihydropyridines (see Table I) prompted a search for Caz÷ influx pathways into chemoreceptor cells other than voltage-dependent Ca2+ channels. Since the parameter sensed by these cells is phi 11'6°'61, the increase in [H+]i and the influx of Ca2+ must be coupled in some way. On intracellular acidification, cells export protons in exchange for Na +, leading to an increase in [Na÷]i. If the cells have a Na+-Ca 2÷ antiporter in their plasma membrane, an increase in [Na+]i would reverse the normal functioning of this antiporter, producing net Ca~+ influx and activation of Ca2+-dependent responses 62. These predictions have been demonstrated in mammalian cardiac muscle6a. Evidence has been obtained indicating that chemoreceptor cells indeed possess a strong Na÷-Ca 2+ antiporter, a Na+-H + exchanger, and a Na +dependent anion exchanger17. For example, blockade of the Na ÷ pump with ouabain or removal of extracellular K ÷ induces DA release that is dependent on extracellular levels of Na + and Ca2+, and that is insensitive to dihydropyridines. Release of DA induced by acidic stimuli was totally dependent on the presence of Na ÷ in the medium and partially inhibited by ethylisopropylamiloride, a blocker of the Na+-H + antiporter. The release was further decreased by reduction of HCO3- ions and elimination of C1- from the medium. These findings led to the proposal of the model for transduction of acidic stimuli17shown in Fig. 4B. With minor variations, all the mechanisms involved in this model have been corroborated by direct measurements of [Ca2+]i and pHi52'64. The interactions between hypoxic and acidic stimulia can be explained adequately in the transduction models in Fig. 4. Thus, the increased chemoreceptor response in carotid bodies observed during moderate hypoxia on lowering pH or increasing Pc% could be the result of a Bohr-like effect at the O2 sensor, with the subsequent decrease in the affinity for O2 (Ref. 65) and a greater response by chemo152

receptor cells to a given hypoxic level, or, as is the case in the heart 6s, it could be due to Ca2+ entering via the Na*-Ca 2÷ exchanger as a result of acidosis. The influx of Ca~+ via the antiporter for a given level of acidosis would be augmented during hypoxia because the cells are depolarized for a longer period of time in this condition and would add to the Ca~÷ that is entering the cells via the Ca2* channels activated by the hypoxia. The relationship between [Ca2+]i and the exocytotic release of neurotransmitters for the different ranges of intracellular Ca2+ levels66 would determine the potentiating, additive (or less than additive) interactions between acidic and hypoxic stimuli3. Finally, with very intense hypoxia, when the release of neurotransmitter reaches high levels, it is conceivable that postsynaptic events could also contribute to the ineffectiveness of acidic stimuli to increase the hypoxic response. The kinetics of the model in Fig. 4B need to be worked out to see if they fit the rapid response of the intact carotid body chemoreceptors to the acidic stimuli. The steep dependence of pHi on pHo seen in chemoreceptor cells64'67 and the speed and magnitude of the increase in [Na+]i observed in other systems in response to an acidic challenge 68'69are compatible with the prompt response of the carotid body to acid. The model does not consider the participation of ionic channels to be relevant because, on lowering pHo, a parallel inhibition of Na ÷, K ÷ and Ca2+ conductances was found (L6pez-L6pez, J. R. and Ganfornina, M. D., unpublished observations); also, the release of DA elicited by acidic stimuli was not sensitive to dihydropyridines (see Table I). However, it should be noted that in studies performed in neonatal rats 26, a Ca2÷dependent K ÷ current was inhibited and Ca2÷ currents were unchanged when the cells were perfused with weak acid-containing media or when pHo was lowered. These findings seem to indicate a previously undiscovered difference between chemoreceptor cells of adult and neonatal animals and suggest that membrane potential and then probably Ca2÷ channels play a role in acid transduction. Concluding r e m a r k s The carotid body is an arterial chemoreceptor activated by decreases in blood P% and pH or increases in blood Pco2- From the point of view of a systemic physiologist, the carotid bodies represent the origin of reflexes with homeostatic and adaptive roles, which maintain blood gases and pH within normal ranges. For cellular physiologists, the structures of the carotid body that sense changes in blood gases and pH - the chemoreceptor cells - represent a sensory receptor in which the variations in blood gases and pH are transduced into a neurosecretory response. In this review, we have attempted to dissect and understand the ensemble of the basic mechanisms operating in the cells, from the arrival of the bloodborne stimuli to the exocytotic release of neurotransmitters. We have discussed current (and sometimes opposite) views on these aspects, but many basic mechanisms that surely operate in chemoreceptor cells have not yet been explored. As a consequence, the models proposed must be seen as incomplete. Further characterization of ionic channels and ion transporters, as well as their modulation by second messengers, and the definition of the nature of TINS, VoL 15, No. 4, 1992

the changes that occur in the receptors in perinatal periods and during acclimatization at high altitude are, among many others, aspects of chemoreceptor cells that need to be worked out. Such knowledge will then give a more complete picture of the transductive cascade in the chemoreceptor cells of the carotid bodies.

Selected references 1 Fidone, S. J. and Gonz&lez, C. (1986) in Handbook of

Physiology (Sect. 3) (The Respiratory System, Vol. 2) (Fishman, A. P., ed.), pp. 247-312, American Physiological Society 2 Honda, Y. (1985) Jpn. J. Physiol. 35, 535-544 3 Fitzgerald, R. S. and Lahiri, S. (1986) in Handbook of

Physiology (Sect. 3) (The Respiratory System, Vol. 2) 4 5 6 7 8 9 10 11 12 13 14

15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36

(Fishman, A. P., ed.), pp. 313-362, American Physiological Society De Castro, F. (1928) Trab. Lab. Invest. Biol. Univ. Madrid25, 331-380 Lever, J. D. and Boyd, J. D. (1957) Nature 179, 1082-1083 Chiochio, S. R., Biscardi, A. M. and Tramezzani, J. H. (1966) Nature 212, 834-835 Biscoe, T. J. (1971) Physiol. Rev. 51,427-495 Fidone, S. J., Gonz&lez, C. and Yoshizaki, K. (1982) J. Physiol. 333, 93-110 Grundfest, H. (1971) in Handbook of Sensory Physiology (Principles of Receptor Physiology, Vol. 1) (Loewenstein, W. R., ed.), pp. 135-165, Springer-Verlag Rigual, R., Gonz&lez, E., Gonz&lez, C. and Fidone, S. J. (1986) Brain Res. 374, 101-109 Rigual, R., L6pez-L6pez, J. R. and Gonz&lez, C. (1991) J. Physiol. 433, 519-531 Obeso, A., Almaraz, L. and Gonz&lez, C. (1986) Brain Res. 371, 25-26 Obeso, A., AImaraz, L. and Gonz,~lez, C. (1989) Brain Res. 481,250-257 Fidone, S. J., Gonz,~lez, C., Obeso, A., Gbmez-Nifio, A. and Dinger, B. (1990) in Hypoxia: The Adaptations (Sutton, J. R., Coattes, G. and Remmers, J. E., eds), pp. 116-126, MarcelDecker Almaraz, L., Obeso, A. and Gonz&lez, C. (1986) J. Physiol. 379, 293-307 Shaw, K., Montagne, W. and Pallot, D. J. (1989) Biochim. Biophys. Acta 1013, 42-46 Rocher, A., Obeso, A., Gonz&lez, C. and Herreros, B. (1991) J. Physiol. 433, 533-548 Obeso, A., Rocher, A., Fidone, S. and Gonz,ilez, C. Neuroscience (in press) Lbpez-Barneo, J., Lbpez-Lbpez, J. R., Urefia, J. and Gonz,~lez, C. (1988) Science 241,580-582 Duchen, M. R. et al. (1988) Neuroscience 26, 291-311 UreSa, J., L6pez-Lbpez, J. R., Gonz&lez, C. and LbpezBarneo, J. (1989) J. Gen. Physiol. 93,979-1001 L6pez-L6pez, J. R., Gonz&lez, C., Urefia, J. and L6pezBarneo, J. (1989) J. Gen. Physiol. 93, 1001-1004 Ganfornina, M. D. and Lbpez-Barneo, J. (1991) Proc. Natl Acad. Sci. USA 88, 2927-2930 Hescheler, J., Delpiano, M. A., Acker, H. and Pietruchka, F. (1989) Brain Res. 486, 79-88 Delpiano, M. A. and Hescheler, H. (1989) FEB5 Lett. 249, 195-198 Peers, C. and Green, F. K. (1991) J. Physiol. 437, 589-602 Stea, A. and Nurse, C. A. (1991) PfliJgersArch. 418, 93-101 Peers, C. (1990) FEBS Lett. 271, 37-40 Blanco, C. E., Dawes, G. S., Hanson, M. A. and McCooke, H. B. (1984) J. PhysioL 351, 25-37 Hertzberg, T., Hellstr6m, S., Lagercrantz, H. and Pequignot, J. M. (1990) J. PhysioL 425, 211-225 Eden, G. H. and Hanson, M. A. (1987) J. Physiol. 392, 1-9 Hanson, M. A., Kumar, P. and Williams, B. A. (1989) J. PhysioL 411,563-574 Hayashida, Y. and Eyzaguirre, C. (1979) Brain Res. 167, 189-194 Lynch, J. W. and Barry, P. H. (1989) Biophys. J. 55, 755-768 Ulbricht, H., Wagner, H. and Schmidtmayer, J. (1986) Ann. N. Y. Acad. 5ci. 479, 68-83 Biscoe, T. J. and Duchen, M, R. (1990) Am. J. Physiol. 258, 271-278

TINS, Vol. 15, No. 4, 1992

37 Nair, P. K., Buerk, D. G. and Whalen, W. J. (1986) Am. J. PhysioL 250, 202-207 38 Buerk, D. G., Nair, P. K. and Whalen, W. J. (1989) J. AppL Physiol. 67, 1578-1584 39 P~rez-Garcia, M. T., Almaraz, L. and Gonz~lez, C. (1990) J. Neurochem. 55, 1287-1293 40 P6rez-Garcia, M. T., Almaraz, L. and Gonz~Jez, C. (1991) J. Neurochem. 57, 1992-2000 41 Acker, H., Dufau, E., Huber, J. and Sylvester, D. (1989) FEBS Lett. 256, 75-78 42 Cross, A. R. etal. (1990) Biochem. J. 272, 743-747 43 Anichkov, S. V. and Belenkii, M. L. (1963) Pharmacology of the Carotid Body Chemoreceptors, Pergamon Press 44 Forster, R. E. (1968) in Arterial Chemoreceptors (Torrance, R. W., ed.), pp. 115-132, Blackwell Scientific Publications 45 Mills, E. and J6bsis, F. F. (1972) J. Neurophysiol. 35,405-428 46 Acker, H. and Eyzaguirre, C. (1987) Brain Res. 409, 380--385 47 Wilson, D. F., Rumsey, W. L., Green, T. J. and Vanderkooi, J. M. (1988) J. Biol. Chem. 263, 2712-2718 48 Bernon, R., Leitner, L-M., Roumy, M. and Verna, A. (1983) NeuroscL Lett. 35, 289-295 49 Biscoe, T. J. and Duchen, M. R. (1990) News Physiol. Sci. 5, 229-233 50 Mulligan, E. and Lahiri, S. (1982) Am. J. Physiol. 242, 200-206 51 Mulligan, E., Lahiri, S. and Storey, B. T. (1981) J. Appl. Physiol. 51,438--446 52 Biscoe, T. J., Duchen, M. R., Eisner, D. A., O'Neill, S. C. and Valdeolmillos, M. (1989) J. Physiol. 416, 421-434 53 Biscoe, T. J. and Duchen, M. R. (1990) J. Physiol. 428, 39-59 54 Carafoli, E. (1987) Annu. Rev. Biochem. 56, 395-433 55 McCormack, J. G., Halestrap, A. P. and Denton, R. M. (1990) PhysioL Rev. 70, 391-425 56 Coxon, R. V. (1968) in Arterial Chemoreceptors (Torrance, R. W., ed.), pp. 91-102, Blackwell Scientific Publications 57 Peers, C. and O'Donnell, J. (1990) Brain Res. 522,259-266 58 Biscoe, T. J. and Duchen, M. R. (1989) J. PhysioL 413, 447-468 59 Blaustein, M. P. (1988) Trends Neurosci. 11,438-443 60 Hanson, M. A., Nye, P. C. G. and Torrance, W. (1981) in Arterial Chemoreceptors (Belmonte, C., Pallot, D., Acker, H. and Fidone, S., eds), pp. 403-416, Leicester University Press 61 Rigual, R., Ifiiguez, C., Carreres, J. and Gonz&lez, C. (1985) Histochemistry 82, 577-580 62 Grinstein, S. and Rothstein, A. (1986) J. IVlembr. BioL 90, 1-12 63 Bountra, C. and Vaughan-Jones, R. D, (1989) J. Physiol. 418, 163-187 64 Buckler, K. J., Vaughan-Jones, R. D. and Peers, C. (1991) J. Physiol. 436, 107-129 65 Lahiri, S. and Delaney, R. G. (1976) in Morphology and Mechanisms of Chemoreceptors (Paintal, A. S., ed.), pp. 18-26, Navchetan Press, New Delhi 66 Baker, P. F. (1987) in Structure and Physiology of the Slow Inward Calcium Channel (Venter, J. C. and Triggle, C., eds), pp. 247-265, Alan R. Liss 67 He, S. F., Wei, J. Y. and Eyzaguirre, C. (1990) in Arterial Chemoreception (Eyzaguirre, C., Fidone, S., Fitzgerald, R. S. and McDonald, D. M., eds), pp. 18-23, Springer-Verlag 68 Boron, W. F., Boyarsky, G. and Ganz, M. (1989) Ann. N. Y. Acad. Sci. 574, 321-332 69 Saito, Y., Ozawa, T. and Nishiyama, A. (1990) PfliJgers Arch. 417, 382-390 70 Acker, H., Delpiano, M. and Degner, F. (1983) in Physiology of the Peripheral Arterial Chemoreceptors (Acker, H. and O'Regan, R, G., eds), pp. 89-115, Elsevier

Acknowledgement~ Theauthorswould liketo thankProf's B.Herreros and J. 6atria-Sancho for critical discussionsin preparingthe manuscript. Supportedby DGICYT(PB89/0358) of Spain.

Writing for TINS Most of the articles published in Trends in Neurosciences are specially invited by the editor. However, unsolicited articles will also be considered for publication. If you wish to write for TINS, please contact the editor, or a member of the Editorial Advisory Board first, with an outline of your intended article.

153