INTRODUCTION

Top Abstract Introduction Methods Results Discussion References

INTERACTION BETWEEN CO₂ and O₂ in carotid body (CB) chemosensory discharge is well known (10-12, 16, 19, 20, 28-30). Within limits, the steady-state response to PCO₂ is linear at a given PO₂ (20). During normoxia [arterial PO₂ (Pa_O2) = 100 Torr], lowering arterial PCO₂ (Pa_CO2) decreases chemosensory activity toward zero, and neuronal discharge (ND) increases linearly with hypercapnia. During hypoxia, Pa_CO2 has to be even lower to achieve the same chemosensory activity, and ND increases with hypercapnia at a greater rate than during normoxia. This relationship between ND response and CO₂-O₂ is known as stimulus interaction. One presumed mechanism of chemosensory discharge is that it depends on proportional secretion of an excitatory neurotransmitter. The amount of neurosecretion in response to the stimulus should then correspond to the change in ND. One putative neurotransmitter, dopamine (DA), is generally thought to be inhibitory (see Ref. 12 for review), although there is also evidence for an excitatory role (11). The purpose of this study was to examine the relationship between DA release and ND and to determine whether there was any interaction between CO₂ and O₂ by using electrochemical microsensors to measure tissue DA.

Electrochemical measurements have confirmed that glomus cells release DA with hypoxia. Using carbon-fiber microelectrodes, Ureña et al. (35) have shown that DA is released from cultured rabbit glomus cells in response to hypoxia. Montoro et al. (25) used carbon-fiber microelectrodes to monitor DA secretion from rabbit glomus cells while measuring intracellular Ca²⁺ concentration ([Ca²⁺]_i) by fura 2 microfluorimetry. They found little change in [Ca²⁺]_i and DA secretion until bath PO₂ was reduced to ~50 Torr. With further reduction in PO₂, [Ca²⁺]_i abruptly rose, with a parallel increase in DA secretion. Findings from isolated glomus cells, however, do not provide any direct information on whether the subsequent neurotransmitter action of DA is excitatory or inhibitory or whether it is a neuromodulator of the chemosensory response.

There have been relatively few studies that used electrochemical techniques to follow the time course of DA changes in the tissue of intact CBs to investigate its putative role as a neurotransmitter or neuromodulator. Donnelly (6) and Doyle and Donnelly (8) found that, when rat CBs were superfused with hypoxic solution (PO₂ <10 Torr for 3 min), there was an increase in chemosensory discharge with a slow increase in DA measured with 10-µm-diameter carbon-fiber microelectrodes. Using a smaller (~5-µm tip) DA microsensor in perfused cat CBs, Buerk et al. (4) found very rapid temporal responses and large increases in DA (average ~20 µM) during the development of hypoxia after flow interruption.

In the present study, we simultaneously measured chemosensory discharge and tissue DA in an in vitro perfused cat CB preparation by using the smaller DA microsensor. We confirmed that there was CO₂-O₂ stimulus interaction for steady-state ND responses, with similar temporal responses and proportional relationships between tissue DA and ND during flow interruption and during normocapnic and hypocapnic hypoxia. However, the tissue DA response to normoxic hypercapnia differed from the hypoxic responses. This suggests that hypercapnic excitation does not act through the same mechanism.

	METHODS

Top Abstract Introduction Methods Results Discussion References

DA-sensitive electrochemical sensors (DA microsensors) were fabricated from single-barrel, metal-filled, glass micropipettes, gold plated inside a shallow recess, as described by Buerk et al. (4). Typical DA microsensors were fabricated with tip diameters ranging from 3 to 10 µm and with recesses ranging from 3- to 10-µm deep, and the microsensors were dip coated in liquid Nafion polymer (Aldrich Chemical). Calibrations were made in deoxygenated phosphate buffer at 37°C in a recirculating system controlled by a pump. DA sensitivities were determined by adding precise amounts of a concentrated stock solution. Amperometric currents were measured with a picoammeter (Keithley Instruments, model 610C) at constant voltage (150 mV) relative to a Ag/AgCl reference and were digitized by computer at 2- to 5-Hz sampling rates with a low-pass filter (5-Hz cutoff). There can be minor contributions to the current from other catecholamines (CAs) (epinephrine, norepinephrine) which require higher oxidation potentials for amperometric measurements (4). At 150 mV, the sensitivity to DA is at least 10-fold higher than sensitivity to epinephrine or norepinephrine.

CBs were sequentially removed from cats anesthetized with pentobarbital sodium (initial dose, 35 mg/kg ip) for in vitro studies with the use of a perfusion-superfusion system (3, 4, 16, 19). A section of the carotid artery bifurcation with the intact CB was mounted in a temperature-regulated chamber. The carotid sinus nerve was isolated, placed on a platinum electrode, and lifted up into a layer of paraffin oil for bipolar ND recordings. ND signals were passed through a 60-Hz notch filter and electronic amplitude discriminator with impulses counted every second. An analog ND signal was also tape recorded. CBs were perfused with solutions from a constant-pressure (80 Torr), gravity-driven reservoir. Effluent from the preparation was removed from the perfusion chamber by continuous suction. The control perfusate was a modified Tyrode solution containing (in mM) 112 NaCl, 4.7 KCl, 2.2 CaCl₂, 1.1 MgCl₂, 22 sodium glutamate, 21.4 NaHCO₃, 5.0 HEPES, and 5.0 glucose, with pH ~ 7.4 at 37°C, and equilibrated by bubbling with compressed gas containing 5% CO₂ and 20% O₂. DA microsensors were positioned into CB tissue by a motorized microdrive with 0.5-µm step resolution (World Precision Instruments, model PM-10). Simultaneous ND recordings and DA microsensor measurements were made during normoxic hypercapnia (20% O₂, PCO₂ 55-57 Torr, pH 7.1-7.2) and for 2-3 min during hypoxia (5% O₂), with either normocapnia (PCO₂ 40 Torr, pH 7.4) or hypocapnia (PCO₂ 10-15 Torr, pH 7.6-7.8). The PO₂, PCO₂, and pH levels in different perfusates were checked on a blood-gas analyzer (Radiometer). Tissue DA and ND responses for each experimental maneuver were compared with respective baseline values during control perfusion with normoxic normocapnic bicarbonate buffer.

In seven experiments, simultaneous measurements of ND, tissue DA, and tissue PO₂ were made by using two single-barrel microsensors. Recessed gold microelectrodes (36) polarized at 0.7 V were used to measure tissue PO₂. Experimental details for tissue PO₂ measurements are described by Buerk et al. (3). Tape-recorded signals were digitized by computer (either 2- or 5-Hz sampling rates, 12-bit resolution) for subsequent data analysis.

Both ND and DA responses to hypercapnia were fit to a single exponential decay with scaling constants (c₁, c₂) and time constant () estimated by nonlinear least squares regression (Sigmaplot, Jandel). The half-life [time for 50% change (t_50%) in ND or DA] was calculated from t_50% = log_e(0.5).

The mean time course for each type of stimulus was determined by averaging repeated trials for each experiment. Statistical comparisons were made by ANOVA (SigmaStat, Jandel) by using equal weighting for mean responses from each experiment, and P < 0.05 was considered as significant.

	RESULTS

Top Abstract Introduction Methods Results Discussion References

Effects of hypercapnia. Induction of hypercapnia (PCO₂ = 55-57 Torr) from normocapnic (PCO₂ = 25-30 Torr) perfusate with the same PO₂ (>120 Torr) caused a rapid increase in chemosensory discharge. A total of 80 measurements were made in 13 cat CBs, with repeated trials ranging between three and eight hypercapnic stimuli per experiment. Average values were determined every 0.5 s, and equal weighting was used for each experiment. Figure 1 A shows the resulting values (means ± SE) for ND (top) and tissue DA (bottom) before and after switching from normocapnic to hypercapnic perfusion at t = 0. Hypercapnia caused an increase in ND () from a baseline of 175 ± 12 impulses/s to a peak value of 593 ± 20 impulses/s (P < 0.05, ANOVA) within 4.6 ± 0.9 s. Chemosensory activity declined to ~59% of the peak ND response and then to a final steady-state value of 423 ± 20 impulses/s. The amount of adaptation (difference between peak and steady-state ND) was 170 ± 20 impulses/s. Baseline tissue DA () averaged 17.9 ± 1.1 µM (range 10-25 µM). Hypercapnia caused an initial increase in tissue DA to a peak value of 23.2 ± 1.2 µM (P < 0.05) within 3.0 ± 0.7 s. The time to reach peak tissue DA was not significantly different from the time to reach peak ND. Tissue DA then declined to 2.6 ± 1.0 µM, which was significantly below baseline (P < 0.05). The decrease in tissue DA to steady-state levels below baseline was seen in all 13 experiments.

View larger version (24K): [in this window] [in a new window]   Fig. 1.   A: averaged results from 13 cat carotid body (CB) experiments, showing time courses for neuronal discharge (ND; , top) and tissue dopamine (DA; , bottom) after switching from control perfusate to normoxic hypercapnic solution at t = 0. Mean baseline, peak, and steady-state values (dashed lines) and SE bars are shown. Imps/sec, impulses/s. B: overlay of ND and DA data during adaptation to hypercapnia. ND (, scale at left) decreased from peak chemosensory excitation, while tissue DA (, scale at right) decreased from peak value to below control baseline. Both time courses could be fit by an exponential decay with similar time constants.

Effect of normocapnic hypoxia. When the CB was perfused with hypoxic bicarbonate buffer (PO₂ = 30-36 Torr) with normocapnia (PCO₂ = 25-30 Torr), there was a rapid increase in chemosensory activity, as observed in earlier experiments (4). As shown in Fig. 2 A, responses from one CB experiment (average of 8 measurements) show the increases in ND (top) and tissue DA (bottom) during normocapnic hypoxia. The average control baseline tissue DA (dashed lines, bottom) was ~25 µM for this CB and increased to ~37 µM with normocapnic hypoxia. In Fig. 2B, the similar time courses for the rise in tissue DA (, scale on right) and ND (, scale on left) during normocapnic hypoxia are shown by overlaying the two responses. Data were sampled at 5 Hz in this example, allowing points every 0.2 s to be illustrated.

View larger version (22K): [in this window] [in a new window]   Fig. 2.   A: example of increased (ND, top) and tissue DA (bottom) after switching from control to normocapnic hypoxia (average of 8 measurements from 1 CB). Dashed line, mean baseline tissue DA for control perfusate. B: overlay of ND (, scale at left) and DA (, scale at right) data, after switching to normocapnic hypoxia, shows that time courses are similar. , change.

Effect of hypocapnic hypoxia. When the CB was perfused with hypoxic bicarbonate buffer (PO₂ = 30-36 Torr) with hypocapnia (PCO₂ = 10-13 Torr), there was a rapid decrease in chemosensory activity, which remained depressed during the entire period of perfusion. Results from one CB experiment (average of 5 trials) are shown in Fig. 3A, indicating a decrease in ND (top), with a parallel drop in tissue DA from ~14 to ~5 µM (bottom), with the final steady-state value below baseline. In Fig. 3B, the similar time courses for the fall in ND (, scale at left) and tissue DA (, scale at right) are shown by the overlay of the data for every 0.5-s interval.

View larger version (19K): [in this window] [in a new window]   Fig. 3.   A: example of decreased ND (top) and tissue DA (bottom) after switching from control to hypoxic, hypocapnic perfusate at t = 0 (average of 5 measurements from 1 CB). B: overlay of ND (, scale at left) and DA (, scale at right) data, after switching to normocapnic hypoxia, shows that time courses are similar.

Effect of flow interruption. The largest ND response and increase in tissue DA were observed after stopping perfusion, as observed in earlier experiments (4). Figure 4 A illustrates simultaneous changes in ND (top), tissue DA (middle), and PO₂ (bottom) after interrupting flow at t = 0. The initial rate of PO₂ disappearance was 3.1 Torr/s (dashed line, bottom). After 35 s, tissue PO₂ fell to near zero, and the maximal ND was reached. Tissue DA (middle) began to rise immediately after interruption of flow in parallel with the increase in ND (top). In this example, the rise in tissue DA was 0.8 µM/s, and the final value reached a maximum ~36 µM, which was 22 µM above the control baseline. In Fig. 4 B, the time courses for ND (, scale to left) and tissue DA (, scale to right) are overlaid with data points every 0.5 s, clearly illustrating that the rise in DA preceded the rise in ND at this location.

View larger version (23K): [in this window] [in a new window]   Fig. 4.   A: example of simultaneous measurements of ND (top), tissue DA (middle), and tissue PO2 (bottom) after interrupting perfusion (stop flow). Maximal excitatory response and increase in DA occurs when tissue PO2 reaches zero, ~35 s after perfusion is interrupted. Dashed lines, initial rates for increase in DA and fall in PO2. B:. overlay of ND (, scale at left) and DA (, scale at right) data, after flow interruption, shows that rise in tissue DA preceded ND response.

Interactions with CO₂. In Fig. 5 A, values (means ± SE) for ND (top) and DA (bottom) are plotted with respect to the perfusate PCO₂ for the two hypoxic conditions (left) along with the baseline, peak, and steady-state values for the normoxic hypercapnic responses (right). The stimulus interaction with PCO₂ for hypoxia (ND/PCO₂ slope) was 27.5 impulses · s¹ · Torr¹ (solid line, top left). The peak ND/PCO₂ slope for normoxic hypercapnia was 13.1 impulses · s¹ · Torr¹ (dotted-dashed line, top right), and decreased to 7.6 · s¹ · Torr¹ (solid line, top right) at steady state. The peak ND/PCO₂ slope was 1.7 times higher than the steady-state slope after adaptation. The DA/PCO₂ slope was positive for hypoxia (1.58 µM/Torr; solid line, bottom). The peak DA/PCO₂ slope for normoxic hypercapnia was also positive, but much smaller (0.17 µM/Torr, dotted-dashed line, bottom), and the final slope after adaptation was negative (0.5 µM/Torr; solid line, bottom).

View larger version (25K): [in this window] [in a new window]   Fig. 5.   A: ND (top) and tissue DA (bottom) with different PO2 and PCO2 levels. Respective slopes (ND/PCO2 and DA/PCO2) for hypoxic conditions (left) are indicated by solid lines between steady-state values for hypocapnic hypoxia () and normocapnic hypoxia (). Interactions were different for normoxic hypercapnia (right). Baseline (dashed lines), peak (), and steady-state () ND and DA values are shown. Lower steady-state slopes (solid lines) are compared with peak (dotted-dashed lines). B: linear relationship between tissue DA and ND was found for control, hypoxia, and stop-flow measurements with normocapnic perfusates (dotted-dashed line), with smaller tissue DA responses for other measurements.

1

1

	DISCUSSION

Top Abstract Introduction Methods Results Discussion References

In all four experimental maneuvers, there were similar temporal changes in tissue DA and ND. Whenever ND increased, tissue DA increased and vice versa. However, there were differences in proportional relationships between ND and tissue DA, as shown in Fig. 5B. The highest ND and tissue DA level occurred during flow interruption, presumably due to the severe hypoxic conditions (near zero PO₂). Although there will be some increase in tissue PCO₂ during flow interruption, we believe that this has a minor effect. Less severe hypoxia, during perfusion with 5% O₂, caused an intermediate increase in ND and tissue DA. The lowest ND and DA were measured during hypocapnic hypoxia. This is consistent with CO₂-O₂ stimulus interaction reported earlier from in vivo cat CB studies by Lahiri et al. (20), in which chemosensory activity decreased as Pa_CO2 was reduced. The reduction in chemosensory activity under hypocapnic conditions inhibits ventilation, as shown by Smith et al. (31) in studies in awake dogs in which respiratory control feedback was altered by denervating one CB and perfusing the second, vascularly isolated CB with hypoxic saline and blood solutions with either normocapnic or hypocapnic conditions.

Steady-state responses to normoxic hypercapnia differed from the other stimuli; this suggests that the mechanism of hypercapnic excitation could be different from hypoxic mechanisms. We did observe an initial increase in DA during the peak ND response to hypercapnia, although the increase was less than might be expected if DA were proportional to the ND response. However, tissue DA then decreased substantially below baseline during adaptation of the ND response, even though chemosensory activity remained elevated above baseline after reaching steady state. This change in tissue DA was opposite to that found for the other responses, in which relative changes in tissue DA were proportional to the relative changes in ND.

We recently reported that carbon monoxide (CO) and CO₂ demonstrate stimulus interaction (28). Because CO binds with cytochrome oxidase, the interaction with CO₂ may provide additional evidence that cytochrome oxidase is the primary O₂ sensor in hypoxic chemoreception. These observations, including results with hypoxic stimuli in the present study, might be interpreted as being consistent with the hypothesis that DA is an excitatory neurotransmitter, because DA release and chemosensory excitation are congruent. This is also consistent with our previous finding that DA increases rapidly during flow interruption (4). González-Guerrero et al. (13) have presented evidence in favor of an excitatory role for DA (also see Refs. 10-12).

However, there is overwhelming evidence that exogenous DA administration inhibits excitatory discharge, as first observed by Zapata (37). Furthermore, blocking D₂ receptors can be excitatory. Lahiri et al. (21) found augmented hypoxic excitation after blocking DA receptors, and Iturriaga et al. (17) found that intravenous domperidone (an antagonist of D₂ receptors) enhanced hypoxic responses. However, in another domperidone study, Tomares et al. (34) reported an increase in chemosensory activity in adult cats, but the researchers saw no noticeable effect on hypoxic sensitivity. There were much larger effects in neonatal cats. Dual effects were seen in an earlier study in our laboratory (26) with a perfused cat CB preparation. Morelli et al. (26) reported that bolus injections of DA initially caused ND inhibition, followed by a transient period of ND excitation. The DA antagonist haloperidol abolished ND responses to the DA bolus, as well as abolishing the ND responses to cyanide, nicotine, flow interruption, and hypercapnia. Responses to these stimuli fully recovered after CBs were returned to perfusate without haloperidol. The excitatory response to the DA bolus also returned, but the inhibitory response did not recover. The response to hypercapnia, as found in the present study, could be consistent with an inhibitory role for DA, if the marked decrease in tissue DA at steady state removed the inhibition and allowed much greater chemosensory activity.

There is also evidence that DA is unrelated to chemosensory activity. In a recent rat CB study, Donnelly (5) reported that reserpine caused a reduction in CA secretion with hypoxia. This was expected, because of depletion of CAs, as confirmed from measurements with carbon-fiber microelectrodes. However, reserpine had little or no effect on hypoxic responses. Sun and Reis (32) concluded that DA is not essential for hypoxic chemotransduction, since hypoxic responses, responses to cyanide, and injections of tyramine to stimulate DA release from rat CB in vivo were unchanged after systemic delivery of chlorpromazine, an antagonist of DA receptors and -adrenoreceptors. Iturriaga et al. (15) reported that there was some correlation between CA release and excitatory stimuli. However, they found that CA release was progressively diminished with repeated flow interruptions, whereas ND responses were unchanged. Also, exogenous application of DA did not change hypoxic ND responses. They concluded that CA release is not essential for hypoxic chemoreception. More recently, Iturriaga and Alcayaga (14) measured CA release [reported as efflux or change in CA (CA)] from superfused cat CBs during hypoxia with and without hypercapnia. They reported that hypercapnia, although evoking chemosensory excitation, did not consistently increase CA. Neither the baseline CA nor the CA with hypercapnia was reported. They also observed that repeated hypoxic and hypercapnic challenges caused a progressive decrease in CA, whereas the ND remained essentially the same. Their largest mean increase in CA with the first hypoxic challenge was smaller (2.9 ± 0.9 µM) than we found in the present study (8.3 ± 1.0 µM above control). We did not observe any systematic decline in tissue DA responses with repeated stimuli. The most recent study by Iturriaga and Alcayaga (14) also concludes that DA release is not essential for chemosensory excitation.

In our view, interpretations based on the time course of CA release from CBs measured with carbon-fiber microelectrodes remains problematic, especially when compared with the rapid temporal changes in tissue DA measured with shallow, recessed DA microsensors for all of the different stimuli reported in the present study. We have found differences between time courses for measurements with the two electrodes that were described previously (4). CA measurements in the CB with carbon-fiber microelectrodes (5, 6, 14, 17) tend to be much slower and considerably delayed compared with the chemosensory responses. Perhaps local damage with the larger sensors might be a confounding factor, or perhaps there are different reactions to the stimuli applied to superfused CBs compared with perfused CBs.

Our measurements of increased DA during hypoxia are consistent with results from hypoxic rat CBs by Donnelly (6), and in agreement with measurements of DA release from single glomus cells (25, 35), as well as studies that used radiolabeling techniques (10-12). Similarly, the rise in glomus cell [Ca²⁺]_i (2, 23) due to hypoxia supports the concept that hypoxia depolarizes the cells, followed by influx of Ca²⁺ through voltage-gated Ca²⁺ channels. The role of ion channels in hypoxic chemoreception has been reviewed by López-Barneo (23). The proposed mechanism is that Ca²⁺ channels are normally inhibited until the K⁺ conductance is reduced by low PO₂ (24). This initiates a burst of action potentials, lowering glomus cell membrane potential, removing inhibition of Ca²⁺ channels, and thus allowing Ca²⁺ to enter the cell. The elevation in [Ca²⁺]_i leads to release of CA. The similarity between O₂-dependent Ca²⁺ influx, CA secretion, and sensory activity was pointed out by López-Barneo (23). Although this idea is still a tentative hypothesis, it is supported by the fact that both DA release and chemosensory discharge are inhibited by decreased [Ca²⁺]_i. Also, Bay K 8644 (a dihydropyridine which increases [Ca²⁺]_i) triples the DA release with hypoxia in the presence of normal Ca²⁺ (11). We have shown (22) that thapsigargin, which releases Ca²⁺ from intracellular stores, can partially restore chemosensory activity during flow interruption after CBs have been perfused with zero-Ca²⁺ solutions.

A rise in [Ca²⁺]_i is known to be associated with neurosecretion and sensory discharge. Hypercapnia produces immediate intracellular acidosis and a rise in [Ca²⁺]_i (1), which is consistent with the transient increase in DA that we measured immediately after the hypercapnic stimulus. But with time and in the presence of CO₂-HCO₃, H⁺ is extruded by the Na⁺-dependent Cl/HCO₃ exchanger, thereby decreasing [Ca²⁺]_i and neurosecretion. This decrease in [Ca²⁺]_i may explain the time course of ND adaptation and decrease in DA that we observed with normoxic hypercapnia. The influence of HCO₃ in the perfusate on intracellular pH (pH_i) through differential effects on acid extrusion has been discussed by Thomas (33) with regard to a study by Ganz et al. (9). Another possibility is that, under acidic conditions, Ca²⁺ becomes bound, thereby decreasing [Ca²⁺]_i, as observed by Donnelly and Kholwadwala (7).

Radiolabeling studies by Rigual et al. (29) and Rocher et al. (30) found that a 5-min period of more severe hypercapnia (20% CO₂) caused a much smaller increase in cumulative DA release compared with hypoxia. It is possible that the small efflux of DA that they measured might reflect a transient period of DA release similar to that observed in the present study with a more modest degree of hypercapnia. Using time-resolved electrochemical measurements of tissue DA, we found that the time courses for the decline in ND and tissue DA during adaptation were nearly identical, as shown in Fig. 1B, but that steady-state sensory discharge remained high despite the decrease in DA. Why ND remains elevated is an interesting question. A conventional explanation is that intracellular H⁺ (and pH_i) is responsible for the excitation. However, Iturriaga et al. (18) have shown, using a pH_i dye in the intact perfused cat CB, that hypoxic perfusion or short periods of flow interruption do not cause significant changes in pH_i. Lahiri et al. (19) suggested that the initial ND response to hypercapnia is due to rapid, transient intracellular acidification, which initially exceeds the final steady-state change in pH_i. The peak and adaptation of the ND response could be eliminated by inhibiting carbonic anhydrase. Active pH_i regulation is the most likely explanation for the adaptation of the ND response with hypercapnia. Another possibility is that ionic changes with hypercapnic stimuli alter the sensitivity of D₂ receptors. Neve (27) has shown that the conformation of D₂ receptors in the brain is regulated by both pH and Na⁺.

Another factor could be increased washout, if there were appreciable vasodilation during hypercapnia. Vascular effects that cause changes in flow could complicate interpretations of experimental tissue DA measurements. A generalized mathematical expression for the time rate of DA change in perfused tissue might be Convective transport can be positive or negative, reflecting either delivery from upstream tissue regions with high DA concentrations or removal by washout. Similarly, the diffusion term can also be either positive or negative, depending on DA concentrations in nearby tissue regions. However, in the seven experiments in which tissue DA and tissue PO₂ were measured simultaneously, we did not see, on average, an increase in tissue PO₂ which would reflect an increase in perfusion.

In summary, stimulus interaction with CO₂ was observed with ND, but this interaction did not appear to be tightly coupled with DA neurosecretion. Responses of ND and DA release during hypoxia could be interpreted as being consistent with DA playing a role as an excitatory neurotransmitter since the phenomena were concomitant. However, an inhibitory role for DA might explain the high chemosensory activity during hypercapnia, since DA decreased to very low levels. Further experimental work is needed to elucidate the role of DA in chemosensory transduction.