- Email: [email protected]
Role of carotid and central chemoreceptors in the CO2 response of sympathetic preganglionic neurons
Journal of the Autonomic Nervous System, 3 (1981) 421--435
421
© Elsevier/North-Holland Biomedical Press
ROLE OF CAROTID AND C E N T R A L CHEMORECEPTORS IN THE CO2 RESPONSE OF SYMPATHETIC P R E G A N G L I O N I C N E U R O N S
BRIAN D. HANNA, F R A N C O LIOY I and CANIO POLOSA 2 Department of Physiology, McGill University,Mclntyre Medical Building, 3655 D r u m m o n d Street,Montreal, Que. H 3 G 1Y6 (Canada)
K e y w o r d s : chemoreceptors -- chemoreflexes -- sympathetic tone --
cardiovascular control
ABSTRACT
In anesthetized, vagotomized, paralyzed, artificially ventilated cats with aortic nerves cut, we recorded the response of 28 sympathetic preganglionic neurons (SPNs) of the cervical sympathetic trunk to changes in arterial pCO2. We observed the effects on these responses of: (i) surgical denervation of carotid sinus chemoreceptors in normoxia (PaO2 110 mm Hg); and (ii) hyperoxia (PaO2 > 350 mm Hg) which is known to depress peripheral chemoreceptor sensitivity to CO2. Stimulus--response curves, obtained b y rebreathing at constant PaO2, were used to detect the effects of these manoeuvres. The present experiments have confirmed previous observations demonstrating the CO2-sensitivity of this neuron population. The population average firing rate, as a function o f paCO2, describes a sigmoid curve, increasing continuously between 20 and 90 mm Hg and asymptotically approaching plateaus at the highest and lowest paCO2 values. Carotid sinus nerve section caused a decrease of the average response of the population at all paCO2 values, resulting in a displacement to the right of the response curve, in a decrease in slope and m a x i m u m values. O n the assumption that the CO2 response curve after carotid sinus nerve section is due to central chemoreceptor input, and that there is a simple addition between the effects of central and carotid chemoreceptors, the difference between CO2 response curves
1 Present address: Department o f Physiology, University of British Columbia Medical School, Vancouver, B.C. V6T l W 5 Canada. 2 To whom reprint requests should be addressed.
422
("difference curves") before and after denervation represents the contribution of the carotid chemoreceptors. A comparison of this "difference curve" with the curve obtained after denervation reveals that the contribution of the carotid chemoreceptors is of the same magnitude as that of the central chemoreceptors up to a paCO2 value of 60--70 mm Hg. Beyond this value, the carotid contribution declines and becomes a smaller c o m p o n e n t of the total response, wherea.s the contribution of the central chemoreceptors continues to increase. Similar results were obtained with rebreathing in hyperoxia, after correction for the central excitatory effect of hyperoxia. Hyperoxia never caused a depression of the CO2 response of units after section of the carotid sinus nerve. Observation of the effects of the two manoeuvres on individual SPNs leads to the conclusion that in approximately half of the COs-sensitive units there is an overlap of central and peripheral chemoreceptor input. The remainder of the CO2-sensitive units receive input only from the central chemoreceptors.
INTRODUCTION
In the course of previous studies on the role of CO2 in the generation of the background firing of sympathetic preganglionic neurons (SPNs) and of the basal sympathetic tone of effectors, several facts have been established. First, the o u t p u t of an SPN population is a continuous function of arterial pCO~ within the range 15--50 mm Hg [12]. The sensitivity to CO2, however, varies greatly among members of the neuron population, the most sensitive neurons being those with a respiratory-modulated firing pattern [ 11 ]. Some SPNs are depressed by CO2. Second, neurogenic vasoconstrictor tone of the hind limb is also a continuous function of arterial pCO2 [7]. This relationship persists after peripheral chemoreceptor denervation b u t is nearly completely abolished when, in addition, the ventral surface of the medulla is cooled t o 12°C [3]. The latter observation suggested that chemoreceptors, superficially located on the ventral surface of the medulla [9], may be connected to the SPNs and may be responsible for a significant fraction of the CO2-sensitivity of the SPN population. The organization of the connections of SPNs with CO2-sensitive receptors appears therefore, at least qualitatively, analogous to that of respiratory neurons. There is, in fact, evidence for connections with peripheral chemoreceptors as well. Experiments of carotid chemoreceptor perfusion in dogs [10] "have shown a direct relationship between systemic vascular resistance and carotid chemoreceptor pCO2. In addition, De Groat and Lalley [2] found electrophysiological evidence that 60% of upper thoracic SPNs are synaptically connected to carotid b o d y chemoreceptors. The purpose of this investigation was to clarify h o w these two sets of C02~sensitive receptors, central and peripheral, share control o f an SPN population. We therefore studied the effects of peripheral chemoreceptor inactivation, either surgical (by carotid sinus nerve section) or func-
423 tional (by 100% O2 breathing) [5] on the CO2 responses o f SPNs of the cervical trunk. MATERIALS AND METHODS Data were obtained from 25 cats of either sex and of b o d y weight between 3.0 and 4.0 kg, anesthetized with sodium pentobarbital (30 mg/kg i.p. Nembutal, Abbott). Catheters were placed in a femoral vein and artery for i.v. administration of drugs and for monitoring arterial blood pressure, respectively. The contralateral femoral artery was also cannulated in 3 experiments and used for infusing a noradrenaline solution (5 × 10 -7 g/ml in saline, Levophed, Wintrop), from a container pressurized at 90 mm Hg, if blood pressure fell below this level. Infusion rates were determined b y the difference between desired and actual pressure. If blood pressure was at/or above the desired level, the infusion would stop. The urinary bladder was continuously evacuated through an indwelling catheter in the urethra. Through a t r a c h e o t o m y the cats were ventilated with humidified r o o m air or 100% O5 (Harvard respirator model 614) after paralysis with Pancuronium bromide (0.02 mg/kg i.v., Pavulon, Organon). End-tidal pO2 and pCO2, henceforth designated as paO2 and paCO~, were continuously monitored using, respectively, a Beckman OM-11 and LB-2 gas analyzer connected in series, with a sample rate of 200 ml/min. During rebreathing (see below) the sample was returned to the rebreathing system. Rectal temperature was measured and maintained at 37 + I°C with heating lamps. One hour after induction of anesthesia a continuous i.v. infusion of 10% dextrose in saline (15 ml/h) was started, containing Nembutal (3 mg/kg/h, less 10%/h), Pavulon (0.2 mg/kg/h) and sodium bicarbonate (1 mEq/lzcF/h). Larynx and esophagus were sectioned and retracted through the buccal cavity to expose the carotid sinus nerves (CSN), which were isolated and prepared for later section in 7 experiments. The vago-sympathetic trunk and the aortic depressor nerve were sectioned bilaterally just caudal to the nodose ganglion. One spinal r o o t of the left phrenic nerve was cut, and the central end desheathed and placed on a pair of silver electrodes for recording. Small strands containing one active unit or a few units were dissected from the cervical sympathetic trunk under a microscope and placed on another pair of silver electrodes. All nerves were kept covered with warm mineral oil in a pool made of skin flaps. The discharge rate of the sympathetic units was continuously measured with a gated pulse counter. Since the experiment lasted hours, it was important to verify that the recorded unit was the same throughout the experin~ent. This was done by examining the unit spike wave shape at frequent intervals during the experiment. The electrical activity of the phrenic nerve was amplified, half-wave rectified and integrated using a " l e a k y " RC circuit with a time constant of 100 msec. Arterial blood pressure, paO2, PaCO2, rectal temperature, SPN firing rate and integrated phrenic nerve discharge were recorded on a Grass Polygraph. The raw SPN and phrenic nerve discharge, together with the arterial blood pressure and paCO2
424 signals after FM modulation, were also recorded on magnetic tape. PaCO2 was changed using a rebreathing method. The expired gas went into a small rebreathing bag connected to the inlet of the respiration pump. Total volume of the rebreathing system was 2 liters approximately. 02 was added to the system at a steady rate, adjusted to keep paCaO: constant. The system was filled with either r o o m air or 100% 0 : . A rebreathing run was preceded b y a 10--15 min equilibration period in open circuit with the animal ventilated with either room air or 100% 02, followed b y a short (1--2 min) period of hyperventilation (obtained, at constant tidal volume, b y increasing pump rate from 12 to 50 strokes/min). The hyperventilation lowered paCO: below the threshold value for phrenic nerve burst activity. Then the circuit was closed and the actual rebreathing run began. A rebreathing run lasted an average 17.0 + 0.6 (mean + S.E.M., n = 64) min, during which paCO2 changed by 68.7 -+ 1.6 mm Hg, at an average rate of 4.0 + 0.2 mm Hg/min. The average rate. of change of p~CO2 was the same in all 4 experimental groups studied (see below). On average, three 0.2 ml arterial blood samples were taken at regular intervals during a rebreathing run and analyzed for pH, pO: and pCO2 (Radiometer, model 72). At the end of a run the animal was returned to the open circuit situation and the p u m p rate to 12 strokes/rain. Before beginning a new rebreathing run, any base deficit was corrected b y giving additional sodium bicarbonate i.v. The reason for choosing a rebreathing over a steady-state method for changing arterial pCO: was the requirement to obtain serial stimulus response curves. In each of the 7 animals in which the CSN was prepared response curves were obtained in at least 4 experimental conditions: in normoxia or hyperoxia, with intact or sectioned CSN. In the remaining animals with intact CSN, response curves were obtained in at least two experimental situations, normoxia and hyperoxia. Moreover, for each experimental situation, several points or measurements were required to define the shape of the stimulus--response relationship. It would have been impossible to obtain all these points with the steady-state technique without running into the problem of a changing state of the experimental animal. As a test for the effectiveness of peripheral chemoreceptor denervation a 50 lag~m1 KCN solution was injected i.v. at doses of 1 ml/kg. Each animal received 2 or 3 injections before and the same number after denervation. The typical normal response consisted of one or t w o phrenic bursts with 1.5--2.0 times the amplitude of the control burst. Carotid sinus chemoreceptor inactivation was accomplished b y 100% O2 breathing or by carotid sinus nerve section. The former procedure is known to decrease the sensitivity to CO: of carotid sinus chemoreceptors [5]. Both procedures have drawbacks. With nerve section, in addition to chemoreceptors, baroreceptor afferent fibers are also interrupted. Their tonic inhibitory action is eliminated and the SPNs will be more excitable. To keep constant this source of error in the experiments of denervation, hexamethonium (10 mg/kg/h) was infused i.v. in 3 cats. This lowered arterial blood pressure to a value o f 90 -+ 5 mm Hg. Since CO: has a direct inhibitory effect on vascular
425
smooth muscle [14] which could lead to a decrease in total peripheral resistance and arterial pressure when unopposed by the simultaneously occurring reflex excitatory effect [15], after hexamethonium the pressure~stabilizing device described above was used during the rebreathing experiments. In the remaining 4 cats phentolamine mesylate (2 mg/kg/h) was infused i.v. and lowered blood pressure to 90 + 7 mm Hg. In the cats so treated no decrease in arterial pressure occurred during rebreathing. With 100% O2 breathing, in addition to the effects caused by the decrease in peripheral chemoreceptor activity [5], respiratory effects have been described due to a central nervous system stimulating action [8]. In order to see whether there are any effects of O2 on SPNs, independent of peripheral chemoreceptors, rebreathing in 100% O2 was repeated after denervation. As a measure of the response to CO2 the mean frequency of the unit discharge was used. For each animal, comparisons between response curves in air before and after the carotid chemoreceptor denervation and between response curves with intact peripheral chemoreceptor innervation in air and in 100% O2 (after correction for the excitatory effect o f 02, see above) were performed. The Wilcoxon sign rank test [13] was used to test the significance o f differences. Differences were considered significant if P ~< 0.05. RESULTS
C02 responses in normoxia in cats with intact carotid sinus nerves
A total of 28 units were studied. Of these, 24 increased their firing rate with an increase in paCO2, within at least part of the range of CO2 levels studied, one did not change and 3 decreased. On the basis of a previously oOOOeoA
2, F
oo
..~.2.0
2.0
•
B
2.5
1.5
4"0
1 • ¢) " ' t E
~oO
I
I
I
I
c
•r 2"r
i:I
E
5.0
oo°°o oeeOee
I
i
i
"" I
3-0 I
D
2.0
"O"
i7. 2,0~ •O
1.o 00
00 0
IO eqp 20 40
61 i 0 810 100
, 20
., 40
6'0
Pac02
I 0
100
(mm
I
Hg)
Fig. 1. CO2 responses in normoxia of sympathetic preganglionic neurons. A--D: types 1--4 response patterns. E: average CO2 response (n = 28). Carotid sinus nerves intact.
426
proposed classification [12], 13 (46%) units had a type 1 response, i.e. were silent at low CO2 levels, were recruited at a critical CO2 level, and increased their firing rate as CO2 increased above threshold. One unit (4%) had a type 2 response, i.e. insignificant changes in firing rate as CO2 levels changed. Eleven units (39%) had type 3 response, i.e. the unit firing rate was relatively independent of CO2 within a certain range of low CO2 levels, and was CO2dependent within a range of higher CO2 levels. Finally, 3 units (11%) had type 4 response, i.e. their firing rate was highest at low CO2 levels and decreased with increasing CO2 levels. Fig. 1A--D shows examples of these 4 response types. These different unit responses to CO2 were recorded in preparations with typical phrenic nerve response to CO2, therefore the differences could not be attributed to abnormality of the chemoreceptor reflexes in these preparations. Moreover, units with different response types could be recorded in the same preparation at different times. Thus, the different response types were not related to differences between animals, but reflected a heterogeneity of the sympathetic neuron population with respect to chemoreceptor input distribution and/or intrinsic neuron properties. Fig. 1E shows h o w the average discharge rate o f this SPN population varies with paCO2. The curve was obtained by averaging the mean firing rates of the 28 units at each 5 mm Hg increment of paCO2 between 20 and 90. The curve has a sigmoid shape, with m a x i m u m slope in the 50--70 mm Hg PaCO2 range, and asymptotically approaching plateaus at the lowest and highest paCO2 values. C02 responses in normoxia after carotid sinus nerve section
The effect of bilateral CSN section was studied in 7 of the 28 units. The average response curves for these 7 units, before and after section, are shown
oOO D
~3
o
{g
o o
==
2
o
a L~
0
20
POCO2 [mm Hg)
Fig. 2. Depressant e f f e c t o f bilateral carotid sinus nerve ( C S N ) s e c t i o n o n the average CO2 response in n o r m o x i a o f s y m l m t h e t i c preganglionic n e u r o n s (n = 7). e , CSN intact; D, C S N s e c t i o n e d ; ×, intact m i n u s s e c t i o n e d - - t h e " d i f f e r e n c e curve".
427
in Fig. 2. Denervation resulted in a displacement of the curve to the fight and a decrease in maximum slope and in the high paCO2 plateau. These changes were statistically significant. On the assumption that the CO2 response after denervation is due to central chemoreceptors [3] and that there is simple addition of the effects of central and carotid sinus (CS) chemoreceptors on the SPN, the difference between the curves before and after denervation should represent the contribution of the CS chemoreceptors to the response recorded with intact CSN (difference curve in Fig. 2). Comparison of the response curve after denervation with the difference curve suggests that the CS contribution is comparable to the central in the range of paCO2 from 45 to 65 mm Hg. Above 65 mm Hg, the CS chemoreceptor contribution declines while that of the central chemoreceptors continues to rise. In detail, CSN section abolished the CO2 response of one unit, decreased the response of 3 and had no effect on the response of two others. For the remaining unit the CO2 response not only persisted, but the curve was somewhat displaced to the left. Fig. 3A shows the data for the unit which lost its CO2 response after CSN section. Thus, this unit, which had a type 3 CO2 response, derived its CO2 sensitivity from the peripheral chemoreceptor input only. CSN section did not change significantly the unit firing rate around normocapnia, suggesting the absence of a tonic input from CS afferents. As expected from this inferred connection to CS chemoreceptor, the sensitivity to CO2 of this unit was increased by hypoxia. An example of a unit in which the CO2 response was significantly reduced, but not abolished by CSN section, is shown in Fig. 3B. Two units with a type 3 CO2 response had behavior similar to that of the unit shown in Fig. 3B. This behavior suggests that they received both central
2.5
A
1"6
B QO cO0
1.4 ~2.0
1.2
@ =~
o o
O~ 1,5
o oll
o
o
o
e@
o
o o ~
I
•
o
°
o
o
1.0
o
@@
'8
O
O
O
~ 1.0 C
-6
D
._~ tL
.4
C O.5
•
Q
o o
.2 o
0
0 i
i
i
i
n
0
20
40
60
eo
PoCO
2 (mm Hg)
o
eoeeo~
)
L
i
100
0
20
n 40
Poco
,
i
i
n
60
80
100
2 (mmHg)
Fig. 3. E f f e c t o f bilateral carotid sinus nerve ( C S N ) s e c t i o n o n CO2 response in n o r m o x i a o f t w o units, o, C S N intact; D, CSN s e c t i o n e d .
428 and peripheral CO2-sensitive chemoreceptor input. Of the two units, whose CO2 response was unchanged after CSN section, one had a t y p e 1 and the other a t y p e 4 CO2 response. This lack of effect suggests that the CO2 sensitivity of these two units was due only to connections with central chemoreceptors, unless a displacement to the right of the curve, due to the elimination of the CS chemoreceptors, was exactly matched b y a displacement to left, due to the elimination of the CS baroreceptors. This exact matching seems unlikely. Finally, in one unit, with a t y p e 3 response to CO2, the response was displaced to the left after CSN section. Thus, the COs sensitivity of this unit was due mainly to input from central chemoreceptors. While it is possible that the displacement to the left was due to removal of an inhibitory input originating from CS chemoreceptors, it is more likely that it was due to the elimination of baroreceptor afferents. The fact that the firing rate of this unit in the CO2 insensitive range nearly doubled after CSN section supports this view. The obvious drawback with CSN section as a m e t h o d for estimating the carotid chemoreceptor contribution to the CO2 response of SPNs is the concomitant elimination of the carotid baroreceptors. This could lead to an underestimate of the CS chemoreceptor contribution. As stated in the Methods section, we tried to minimize this source of variability by performing the section at similar arterial blood pressure values, obtained b y i.v. infusion of hexamethonium or phentolamine, in all cats. Treatment with these drugs also lowered arterial blood pressure so that the level of baroreceptor input, prior to section, was also low. This probably explains w h y the average COs response curves before and after CSN section, in the CO: range from 20 to 40 mm Hg, are practically superimposable (Fig. 2). In this range the CO: sensitivity of the SPNs is low, and therefore we would expect to see best the effects of changes in the level of other inputs, e.g. baroreceptors. Since we do not see significant changes, we infer that the effect of baroreceptor removal must have been negligible, with the one exception mentioned above. C02 response curves in hyperoxia
As an additional, seemingly independent procedure for estimating carotid chemoreceptor contribution to the CO2 response of SPNs, we increased p,O2 to values greater than 350 mm Hg on all 28 units. This is known to reduce the responsiveness to CO2 of peripheral chemoreceptors [5]. We assumed therefore that a decrease i n chemoreceptor activity would be reflected in a displacement of the CO2 response curve to the right and/or in a decrease of its slope, and/or in a change of the curve's plateau at high p,CO2, when the experiment was conducted in hyperoxia as compared to experiments conducted in normoxia. Fig. 4 shows the average CO2 response curve obtained in hyperoxia in the whole sample o f 28 units. For comparison the CO2 response curve in normoxia, already shown in Fig. 1E, is plotted. Also plotted is the difference between the t w o curves. Notice that, qualitatively,
429 4
5
~I00
5
.E_ v)
4 o
(n o A o o
~
4O
o o
u.
A o
u.
2O
ooo x
20
40
x
x
~
0
Poco 2 (mm Hg)
x
80
x
x ~'~
ooo
x
I
100
210
I
40
I
60
~
810
1~)0
Paco 2 (mm Hg)
Fig. 4. Depressant effect of hyperoxia on the average CO2 response of sympathetic preganglionic neurons (n = 28). Intact carotid sinus nerves. $, normoxia; A, hyperoxia; ×, normoxia minus hyperoxia. Curves normalized by expressing firing rates as percentage of that at 90 mm Hg PaCO2 in normoxia. Fig. 5. Excitatory effect of hyperoxia on the average CO2 response of sympathetic preganglionic neurons (n = 7) in cats without peripheral chemoreceptors. D, normoxia; A, hyperoxia; ×, hyperoxia minus normoxia.
hyperoxia has an effect similar to CS denervation, namely it causes a displacement to the right, decreased slope and decreased plateaus of the response curve (P < 0.05). Notice also the similarity of the difference curve (normoxia minus hyperoxia) of Fig. 4 with the difference curve of Fig. 2. The displacement to the right and downward, however, is less with hyperoxia than with CSN section. When the effect of hyperoxia on the CO2 response o f individual units is examined, the response of 14 units (50%) was shifted down, that of 8 (28%) was shifted up, and that of 6 (22%) did not change. The increase in CO2 responsiveness in these 8 units could not be attributed to changes in the level of baroreceptor input, since there were no changes in arterial blood pressure. It could have been due either to an excitatory action of hyperoxia, independent of peripheral chemoreceptors, analogous to that described for respiratory neurons [8], or to the fact that CS chemoreceptor input to some SPNs may be inhibitory. The test for the latter possibility has already been performed with negative results, by examining the effects of CSN section. As a test for the first possibility in a series of experiments on 7 o f the 28 units, CO2 response curves in hyperoxia were obtained and compared with CO2 response curves in normoxia, in CS d enervated preparations. This procedure would detect any excitant or depressant effect of hyperoxia, unrelated to CS chemoreceptors. The results of this experiment are shown in Fig. 5. It can be seen that hyperoxia displaces to the left the average CO2 response curve of these SPNs (P < 0.05). Moreover, the effect is CO2-dependent (see difference curve). This experiment demonstrates that hyperoxia can excite SPNs by an
430
0
O O
0 0 0
0 J¢ Q. U~
o
D 0
o
•
o
0
D 0
g°
o
r,-
•
g
o o
o o
C I,I. 1
0 0
I 0
I 20
I 40
I 60
P°c02 (mm Hg)
] 80
I 100
I 0
I 20
l 40
i 60
i 80
I 100
Poco 2 (mmHg)
Fig. 6. A : depressant effect of hyperoxia on the average CO2 response of sympathetic preganglionic neurons (n = 7 ) in cats with intact carotid sinus nerves, after graphical correction for the excitatory effect of hyperoxia, o, normoxia; ©, corrected hyperoxia; ×, normoxia minus corrected hyperoxia. B: similarity of CO2 response curves in normoxia, after carotid sinus nerve (CSN) section, and in hyperoxia, with intact CSN after correction for the excitatory effect of hyperoxia, n = 7. Average CO2 response in normoxia, with intact C S N , is given for comparison. $, normoxia, CSN intact; m, normoxia, CSN sectioned; ©, corrected hyperoxia, CSN intact.
action independent of CS chemoreceptors. This action can account for the left-hand shifts of CO2 response curves observed in hyperoxia in cats with CSN intact. In detail, the CO2 response curve after CSN section was unchanged by hyperoxia in 2 of the 7 units and displaced to the left in the remaining 5. The fact that, in the absence of CS chemoreceptors, hyperoxia never caused a displacement of the right to the curve, can be considered evidence that, in the CSN-intact cat, when hyperoxia causes a displacement to the right of the curve, it must be due to a decrease in CSN chemoreceptor input. From these experiments it can be concluded that the CO2 response curve in hyperoxia of SPNs in preparations with intact CSN results from the algebraic summation of two effects onto the (~O2 response curve in normoxia: a depressant effect of CS chemoreceptor activity [5] a n d a n excitatory effect on other structures, perhaps on ceni~aI nervous system structures [8]. Since any displacement of the CO2 r e s p o ~ curve in hyperoxia from that in normoxia, in preparations with sectione(}Carotid sinus nerves {Fig. 5) is due only to the excitatory effect, it can b~ ~~ e n as a measure of this effect. When this measure of the excitatory effec~ o f O~ is subtracted from the CO2 response curve in hyperoxia in the same preparations, but with intact CS nerves, a new curve is obtained which represent s the effect of the CS chemoreceptor depression only. This procedure ~ a s applied to the data from the same group of 7 units and the resultant curVes are shown in Fig.
431
6A. It can be seen in Fig. 6B that after corrections for the excitatory effect of O~, the hyperoxic CO2 response curve is indistinguishable from the normoxic denervated curve, i.e. both denervation and hyperoxia, after correction for O2 excitation, give similar estimates for the CS contribution to the CO2 response curve of the SPNs. The :,elation between sympathetic preganglionic neuron and phrenic motoneuron C02 responses under various experimental conditions The CO2 response of this SPN population is similar to that of phrenic motoneurons, recently described [4]. This similarity is due to the great preponderance of type 1 and 3 responses in the population and is brought out by the graphs of Fig. 7. In Fig. 7A the CO2 response curves in normoxia of the two neuron populations are shown. In Fig. 7B the responses in normoxia and hyperoxia, before and after CSN section, of the phrenic nerve (peak amplitude) and of the SPN population (mean firing rate) are plotted against each other. This plot shows that the similarity in CO2 response o f the two neuron populations persists after CSN section and in hyperoxia. From Fig. 7A it is clear that both neuron populations show a strong relation between activity levels and CO2 levels. For both neuron systems the
B
A 100
/
O@ • @
~db
@o
0
C~ 8O (el 0 Q. h-
6o D
m
i
@
0 ®
60
Q
E =J
._~
E ®
E4o
o
~o
0 o
x
IE
x
o
*
o e
~O
@oO@
20
o@
z a,
® ®
I
A
0
20
I
|
40 ,
Pot- 0 "
60 2
(mmHg)
.... I
@0
I
I
i
L
I
[
I
100
0
20
40
60
SO
100
Phrenic~ of Maximum Response
Fig. 7. Similarity of CO2 responses of phrenic motoneurons and sympathetic preganglionic neurons. A: average normalized CO2 response (n = 7) of sympathetic preganglionic neurons (mean firing rate, l ) and of phrenic motoneurons (mean peak amplitude of integrated phrenic nerve burst, o) in normoxia, in cats with intact carotid sinus nerves (CSN). B: average normalized CO2 responses of phrenic motoneurons (horizontal axis) plotted against average normalized CO2 responses of sympathetic preganglionic neurons (vertical axis) in normoxia with CSN intact (e) and after CSN section (o), and in hyperoxia, with CSN intact after correction for excitatory effect of 02 Co) and after CSN section (~). n=7.
432
relation is sigmoidal. In Fig. 7B it can be seen that for phrenic responses greater than 40% of maximum {corresponding to paCO2 greater than 50 mm Hg from Fig. 7A) the experimental points fall very close together, i.e. the two systems have the same sensitivity to COs. At phrenic response amplitudes lower than 40% of maximum, however, the sensitivity of the SPN population gradually declines and approaches zero, i.e. SPN activity becomes independent o f phrenic amplitude and of PaCO~ for phrenic amplitudes less than 20% of maximum (corresponding to PaCO2 below 40 mm Hg). The data obtained after CSN section fall along the same path as those before, showing that CS "chemodenervation" has a similar influence on the COs response o f the two neuron populations (Fig. 7B). The same conclusion can be drawn regarding the effect of hyperoxia, before and after CSN section, since these data also indicate close similarity of response.
Role o f central and carotid sinus chemoreceptors in the generation o f unit responses to COs Type 1 and 3 responses in normoxia could be shown in 11 of 24 units to be caused by excitation from both sets of chemoreceptors, in 1 by excitation from central chemoreceptors only, in 1 by excitation from CS chemoreceptors only, and in 2 by excitation from CS chemoreceptors and inhibition from central chemoreceptors. In the latter cases, hyperoxia resulted in a change from either t y p e 1 or 3 to t y p e 4 response. For the remaining 9 units with t y p e 1 or 3 response there was clearly excitation from central chemoreceptors, but the role of CS chemoreceptors could not be assessed, because hyperoxia only was used and its effect was a displacement to the left or no change in the response curve. These results, as discussed above, cannot be interpreted because of the demonstrated excitatory effect of 02. All 3 type 4 responses in normoxia were caused b y inhibition from central chemoreceptors, associated in 2 cases with excitation from CS chemoreceptors. The CS chemoreceptor influence was of such magnitude as to be overcome b y the inhibitory influence of the central chemoreceptors. The one unit with t y p e 2 response did n o t have either central o r CS chemoreceptor input. DISCUSSION
The present results confirm previous observations demonstrating the sensitivity to CO2 of this population of SPNs [12]. The population o u t p u t has been studied in the present investigation over a much wider range of paCO2 values {20--90 mm Hg) than in the previous study. Section of the CS nerve or hyperoxia resulted in a significant decrease in the CO2 response o f the population. This decrease w a s attributed to the removal of a CO2
433
ments [3] and, likewise, taken as a measure of the contribution of this input to the overall CO2 response of the population. Whereas practically all units received input from central chemoreceptors, a smaller number received input from the CS chemoreceptors (4/7 or 56% on the basis of CSN section data). No estimate could be made from the hyperoxia data, due to the already discussed uncertainties in the interpretation of the cases of no change or displacement to the left of the CO2 response curves in hyperoxia. The contribution of the CS chemoreceptors to the CO2 response of this SPN population, estimated from the CSN section data and from the hyperoxia data after correction for the 02 excitation, is comparable in magnitude to that of the central chemoreceptors up to a paCO2 of 60--70 mm Hg. Beyond these values, the CS contribution declines and becomes a progressively smaller component of the total response, whereas the contribution of the central chemoreceptors continues to increase. The similarity of the effects of 100% 02 (after correction) and denervation suggests that the reflex effects of CS chemoreceptor stimulation by CO2 are completely abolished as a result of hyperoxia. A possible interference by baroreceptor reflexes (or their removal) in these experiments has been dealt with by the procedure of ganglion block and noradrenaline infusion. By these means changes in arterial blood pressure, evoked by CO2 either reflexly or by a direct action on vascular smooth muscle [14] have been avoided. Since arterial blood pressure was stable during rebreathing, so must also have been baroreceptor activity levels. Therefore the various CO: response patterns of units cannot be attributed to differences in baroreceptor activity levels. The present experiments do not provide data concerning the sources of the CO2 sensitivity of SPNs in hyperoxia or after CSN section. Previous experiments [3], however, have demonstrated that the CO2 sensitivity of hind limb vasomotor tone, in peripherally "chemodenervated" cats, is nearly completely abolished by cooling the exposed ventral surface of the medulla. This finding suggests that the main source of the CO2 sensitivity of SPNs in hyperoxia or after CSN section is represented by chemoreceptors on the ventral medullary surface. Chemosensitivity of the spinal cord has been suggested by Szulczyk and Trzebski [16], but there is no evidence that this leads, in the spinal animal, to a significant CO2-responsiveness of hind limb vascular tone [7] or of cervical trunk SPNs (Rohliceck and Polosa, unpublished observations). The similarity of responses to CO2 of SPNs and phrenic motoneurons, as well as the analogies in the modification of their responses caused by CSN section or hyperoxia have been pointed out (Fig. 7). These analogies may simply be the result of the fact that the same (or similar) sensory systems are responsible for the CO2 responses of the two neuron populations. An additional possibility, suggested previously [12], is that the CO2 sensitivity of some SPNs does not result from synaptic connections of these neurons with the CO2 sensitive receptor, but from synaptic connections with inspiratory neurons [11], which would make the SPN a "follower" of inspiratory neu-
434
ron activity. These t w o possibilities are n o t mutually exclusive, and it is possible that the observed SPN--phrenic correlation results from both causes. We estimated the CS chemoreceptor contribution to the CO2 response curve of SPNs as the difference between the curves before and after CSN section. This "difference curve" has a characteristic shape: it climbs, as COs increases, up to a peak at between 60 and 70 mm Hg, and b e y o n d this value it declines. The "difference curve" obtained from the hyperoxia data had a similar shape. This decline of the CS chemoreceptor c o m p o n e n t cannot be attributed to a decline in sensory input from CS chemoreceptors, since this increases as CO2 increases and then plateaus, but does not decline [1 ]. More likely explanations are either that at high COs there is occlusion (i.e. hypoadditive interaction) between central and CS chemoreceptor evoked reflexes, with the result that in the absence of the CS c o m p o n e n t the central component is overestimated, or that as COs increases there is some inhibitory interaction between the two inputs. Hyperoxia, in peripheral chemodenervated cats, caused a displacement to the left of the CO2 response curve of the population, i.e. caused an excitation of SPNs. The excitatory effect of hyperoxia on SPN was unknown before this work. However, an excitatory effect o f hyperoxia on respiration [8] was known. A similar mechanism may account for the effects of hyperoxia on the two neuron systems. This excitation of SPNs, or phrenic motoneurons, may result from a decrease in: (i) cerebral blood flow; and/or (ii) carbamino transport and buffer action of hemoglobin with increased oxygenation [6]. An increase in pCO2 and a decrease in pH at the level of the central chemoreceptors could result from either mechanism. Finally, these observations have suggested that central chemoreceptors can exert an inhibitory action on SPNs. Thus, units which were inhibited b y high pCO~ (type 4) were not affected by CSN section or hyperoxia. Moreover, abolishing peripheral chemoreceptor input by hyperoxia transformed the original excitatory effects of COs upon 2 units (type 1 and type 3) into an inhibitory one (type 4). This shows that the excitatory action of the CS chemoreceptors can mask the inhibitory action of the central chemoreceptors. ACKNOWLEDGEMENTS
This work was supported by the Medical Research Council of Canada and the Quebec Heart Foundation. REFERENCES 1 Biscoe, T.J., Purves, M.J. and Sampson, S.R., The frequency of nerve impulses in single carotid body chemoreceptor afferent fibres recorded in vivo with intact circulation, J. Physiol. (Lond.), 208 (1970) 121--131. 2 de Groat, W.C. and Lalley, P.M., Reflex sympathetic firing in response to electrical stimulation of the carotid sinus nerve in the cat, Brain Research, 80 (1974) 17--40.
435 3 Hanna, B.D., Lioy, F. and Polosa, C., The effect of cold blockade of the medullary chemoreceptors on the CO2 modulation of vascular tone and heart rate, Canad. J. Physiol. Pharmacol., 57 (1979) 461--468. 4 Hanna, B.D. and Polosa, C., Carotid and central chemoreceptor components of the phrenic nerve response to CO2 in the anesthetized cat, Canad. Physiol., 10 (1979) 164. 5 Lahiri, S. and DeLaney, R.G., Stimulus interaction in the responses of carotid body chemoreceptor single afferent fibers, Resp. Physiol., 24 (1975) 249--266. 6 Lambertsen, C.J., Kough, R.H., Cooper, D.Y., Emmel, G.L., Loeschcke, H.H. and Schmidt, C.G., Comparison of relationship of respiratory minute volume to pCO2 and pH of arterial and internal jugular blood in normal man during hyperventilation produced by low concentrations of CO2 at 1 atmosphere and by 02 at 3.0 atmospheres, J. appl. Physiol., 5 (1953) 803--813. 7 Lioy, F., Hanna, B.D. and Polosa, C., CO2-dependent component of the neurogenic vascular tone in the cat, Pfliigers Arch. ges. Physiol., 374 (1978) 187--191. 8 Miller, M.J. and Tenney, S.M., Hyperoxic hyperventilation in carotid-deafferented cats, Resp. Physiol., 23 (1975) 23--30. 9 Mitchell, R.A., Loeschcke, H.H., Massion, W.H. and Severinghaus, J.W., Respiratory responses mediated through superficial chemosensitive areas on the medulla, J. appl. Physiol., 18 (1963) 523--533. 10 Pelletier, C.L., Circulatory responses to graded stimulation of the carotid chemoreceptors in the dog, Circulat. Res., 31 (1972) 431--443. 11 Preiss, G., Kirchner, F. and Polosa, C., Patterning of sympathetic preganglionic neuron firing by the central respiratory drive, Brain Research, 87 (1975) 363--374. 12 Preiss, G. and Polosa, C., The relation between end-tidal CO2 and discharge patterns of sympathetic preganglionic neurons, Brain Research, 122 (1977) 255--269. 13 Rumke, C.L. and De Jonge, H., In D.R. Laurence and A.L. Bacharach (Eds.), Evaluation of Drug Activities: Pharmacometrics, Academic Press, N.Y., 1964, pp. 47--100. 14 Somlyo, A.P. and Somlyo, A.V., Vascular smooth muscle. II Pharmacology of normal and hypertensive vessels, Pharmacol. Rev., 22 (1970) 249--353. 15 Suutarinen, T., Cardiovascular response to changes in arterial carbon dioxide tension, Acta physiol, scand., 67, Suppl. 266 (1966) 1--76. 16 Szulczyk, P. and Trzebski, A., The local effect of pH changes in the cerebrospinal fluid off the ventrolateral areas of medulla oblongata and on spinal cord surface on the activity of cardiac and vertebral sympathetic nerves, Acta physiol, pol., 27 (1976) 9--17.
421
© Elsevier/North-Holland Biomedical Press
ROLE OF CAROTID AND C E N T R A L CHEMORECEPTORS IN THE CO2 RESPONSE OF SYMPATHETIC P R E G A N G L I O N I C N E U R O N S
BRIAN D. HANNA, F R A N C O LIOY I and CANIO POLOSA 2 Department of Physiology, McGill University,Mclntyre Medical Building, 3655 D r u m m o n d Street,Montreal, Que. H 3 G 1Y6 (Canada)
K e y w o r d s : chemoreceptors -- chemoreflexes -- sympathetic tone --
cardiovascular control
ABSTRACT
In anesthetized, vagotomized, paralyzed, artificially ventilated cats with aortic nerves cut, we recorded the response of 28 sympathetic preganglionic neurons (SPNs) of the cervical sympathetic trunk to changes in arterial pCO2. We observed the effects on these responses of: (i) surgical denervation of carotid sinus chemoreceptors in normoxia (PaO2 110 mm Hg); and (ii) hyperoxia (PaO2 > 350 mm Hg) which is known to depress peripheral chemoreceptor sensitivity to CO2. Stimulus--response curves, obtained b y rebreathing at constant PaO2, were used to detect the effects of these manoeuvres. The present experiments have confirmed previous observations demonstrating the CO2-sensitivity of this neuron population. The population average firing rate, as a function o f paCO2, describes a sigmoid curve, increasing continuously between 20 and 90 mm Hg and asymptotically approaching plateaus at the highest and lowest paCO2 values. Carotid sinus nerve section caused a decrease of the average response of the population at all paCO2 values, resulting in a displacement to the right of the response curve, in a decrease in slope and m a x i m u m values. O n the assumption that the CO2 response curve after carotid sinus nerve section is due to central chemoreceptor input, and that there is a simple addition between the effects of central and carotid chemoreceptors, the difference between CO2 response curves
1 Present address: Department o f Physiology, University of British Columbia Medical School, Vancouver, B.C. V6T l W 5 Canada. 2 To whom reprint requests should be addressed.
422
("difference curves") before and after denervation represents the contribution of the carotid chemoreceptors. A comparison of this "difference curve" with the curve obtained after denervation reveals that the contribution of the carotid chemoreceptors is of the same magnitude as that of the central chemoreceptors up to a paCO2 value of 60--70 mm Hg. Beyond this value, the carotid contribution declines and becomes a smaller c o m p o n e n t of the total response, wherea.s the contribution of the central chemoreceptors continues to increase. Similar results were obtained with rebreathing in hyperoxia, after correction for the central excitatory effect of hyperoxia. Hyperoxia never caused a depression of the CO2 response of units after section of the carotid sinus nerve. Observation of the effects of the two manoeuvres on individual SPNs leads to the conclusion that in approximately half of the COs-sensitive units there is an overlap of central and peripheral chemoreceptor input. The remainder of the CO2-sensitive units receive input only from the central chemoreceptors.
INTRODUCTION
In the course of previous studies on the role of CO2 in the generation of the background firing of sympathetic preganglionic neurons (SPNs) and of the basal sympathetic tone of effectors, several facts have been established. First, the o u t p u t of an SPN population is a continuous function of arterial pCO~ within the range 15--50 mm Hg [12]. The sensitivity to CO2, however, varies greatly among members of the neuron population, the most sensitive neurons being those with a respiratory-modulated firing pattern [ 11 ]. Some SPNs are depressed by CO2. Second, neurogenic vasoconstrictor tone of the hind limb is also a continuous function of arterial pCO2 [7]. This relationship persists after peripheral chemoreceptor denervation b u t is nearly completely abolished when, in addition, the ventral surface of the medulla is cooled t o 12°C [3]. The latter observation suggested that chemoreceptors, superficially located on the ventral surface of the medulla [9], may be connected to the SPNs and may be responsible for a significant fraction of the CO2-sensitivity of the SPN population. The organization of the connections of SPNs with CO2-sensitive receptors appears therefore, at least qualitatively, analogous to that of respiratory neurons. There is, in fact, evidence for connections with peripheral chemoreceptors as well. Experiments of carotid chemoreceptor perfusion in dogs [10] "have shown a direct relationship between systemic vascular resistance and carotid chemoreceptor pCO2. In addition, De Groat and Lalley [2] found electrophysiological evidence that 60% of upper thoracic SPNs are synaptically connected to carotid b o d y chemoreceptors. The purpose of this investigation was to clarify h o w these two sets of C02~sensitive receptors, central and peripheral, share control o f an SPN population. We therefore studied the effects of peripheral chemoreceptor inactivation, either surgical (by carotid sinus nerve section) or func-
423 tional (by 100% O2 breathing) [5] on the CO2 responses o f SPNs of the cervical trunk. MATERIALS AND METHODS Data were obtained from 25 cats of either sex and of b o d y weight between 3.0 and 4.0 kg, anesthetized with sodium pentobarbital (30 mg/kg i.p. Nembutal, Abbott). Catheters were placed in a femoral vein and artery for i.v. administration of drugs and for monitoring arterial blood pressure, respectively. The contralateral femoral artery was also cannulated in 3 experiments and used for infusing a noradrenaline solution (5 × 10 -7 g/ml in saline, Levophed, Wintrop), from a container pressurized at 90 mm Hg, if blood pressure fell below this level. Infusion rates were determined b y the difference between desired and actual pressure. If blood pressure was at/or above the desired level, the infusion would stop. The urinary bladder was continuously evacuated through an indwelling catheter in the urethra. Through a t r a c h e o t o m y the cats were ventilated with humidified r o o m air or 100% O5 (Harvard respirator model 614) after paralysis with Pancuronium bromide (0.02 mg/kg i.v., Pavulon, Organon). End-tidal pO2 and pCO2, henceforth designated as paO2 and paCO~, were continuously monitored using, respectively, a Beckman OM-11 and LB-2 gas analyzer connected in series, with a sample rate of 200 ml/min. During rebreathing (see below) the sample was returned to the rebreathing system. Rectal temperature was measured and maintained at 37 + I°C with heating lamps. One hour after induction of anesthesia a continuous i.v. infusion of 10% dextrose in saline (15 ml/h) was started, containing Nembutal (3 mg/kg/h, less 10%/h), Pavulon (0.2 mg/kg/h) and sodium bicarbonate (1 mEq/lzcF/h). Larynx and esophagus were sectioned and retracted through the buccal cavity to expose the carotid sinus nerves (CSN), which were isolated and prepared for later section in 7 experiments. The vago-sympathetic trunk and the aortic depressor nerve were sectioned bilaterally just caudal to the nodose ganglion. One spinal r o o t of the left phrenic nerve was cut, and the central end desheathed and placed on a pair of silver electrodes for recording. Small strands containing one active unit or a few units were dissected from the cervical sympathetic trunk under a microscope and placed on another pair of silver electrodes. All nerves were kept covered with warm mineral oil in a pool made of skin flaps. The discharge rate of the sympathetic units was continuously measured with a gated pulse counter. Since the experiment lasted hours, it was important to verify that the recorded unit was the same throughout the experin~ent. This was done by examining the unit spike wave shape at frequent intervals during the experiment. The electrical activity of the phrenic nerve was amplified, half-wave rectified and integrated using a " l e a k y " RC circuit with a time constant of 100 msec. Arterial blood pressure, paO2, PaCO2, rectal temperature, SPN firing rate and integrated phrenic nerve discharge were recorded on a Grass Polygraph. The raw SPN and phrenic nerve discharge, together with the arterial blood pressure and paCO2
424 signals after FM modulation, were also recorded on magnetic tape. PaCO2 was changed using a rebreathing method. The expired gas went into a small rebreathing bag connected to the inlet of the respiration pump. Total volume of the rebreathing system was 2 liters approximately. 02 was added to the system at a steady rate, adjusted to keep paCaO: constant. The system was filled with either r o o m air or 100% 0 : . A rebreathing run was preceded b y a 10--15 min equilibration period in open circuit with the animal ventilated with either room air or 100% 02, followed b y a short (1--2 min) period of hyperventilation (obtained, at constant tidal volume, b y increasing pump rate from 12 to 50 strokes/min). The hyperventilation lowered paCO: below the threshold value for phrenic nerve burst activity. Then the circuit was closed and the actual rebreathing run began. A rebreathing run lasted an average 17.0 + 0.6 (mean + S.E.M., n = 64) min, during which paCO2 changed by 68.7 -+ 1.6 mm Hg, at an average rate of 4.0 + 0.2 mm Hg/min. The average rate. of change of p~CO2 was the same in all 4 experimental groups studied (see below). On average, three 0.2 ml arterial blood samples were taken at regular intervals during a rebreathing run and analyzed for pH, pO: and pCO2 (Radiometer, model 72). At the end of a run the animal was returned to the open circuit situation and the p u m p rate to 12 strokes/rain. Before beginning a new rebreathing run, any base deficit was corrected b y giving additional sodium bicarbonate i.v. The reason for choosing a rebreathing over a steady-state method for changing arterial pCO: was the requirement to obtain serial stimulus response curves. In each of the 7 animals in which the CSN was prepared response curves were obtained in at least 4 experimental conditions: in normoxia or hyperoxia, with intact or sectioned CSN. In the remaining animals with intact CSN, response curves were obtained in at least two experimental situations, normoxia and hyperoxia. Moreover, for each experimental situation, several points or measurements were required to define the shape of the stimulus--response relationship. It would have been impossible to obtain all these points with the steady-state technique without running into the problem of a changing state of the experimental animal. As a test for the effectiveness of peripheral chemoreceptor denervation a 50 lag~m1 KCN solution was injected i.v. at doses of 1 ml/kg. Each animal received 2 or 3 injections before and the same number after denervation. The typical normal response consisted of one or t w o phrenic bursts with 1.5--2.0 times the amplitude of the control burst. Carotid sinus chemoreceptor inactivation was accomplished b y 100% O2 breathing or by carotid sinus nerve section. The former procedure is known to decrease the sensitivity to CO: of carotid sinus chemoreceptors [5]. Both procedures have drawbacks. With nerve section, in addition to chemoreceptors, baroreceptor afferent fibers are also interrupted. Their tonic inhibitory action is eliminated and the SPNs will be more excitable. To keep constant this source of error in the experiments of denervation, hexamethonium (10 mg/kg/h) was infused i.v. in 3 cats. This lowered arterial blood pressure to a value o f 90 -+ 5 mm Hg. Since CO: has a direct inhibitory effect on vascular
425
smooth muscle [14] which could lead to a decrease in total peripheral resistance and arterial pressure when unopposed by the simultaneously occurring reflex excitatory effect [15], after hexamethonium the pressure~stabilizing device described above was used during the rebreathing experiments. In the remaining 4 cats phentolamine mesylate (2 mg/kg/h) was infused i.v. and lowered blood pressure to 90 + 7 mm Hg. In the cats so treated no decrease in arterial pressure occurred during rebreathing. With 100% O2 breathing, in addition to the effects caused by the decrease in peripheral chemoreceptor activity [5], respiratory effects have been described due to a central nervous system stimulating action [8]. In order to see whether there are any effects of O2 on SPNs, independent of peripheral chemoreceptors, rebreathing in 100% O2 was repeated after denervation. As a measure of the response to CO2 the mean frequency of the unit discharge was used. For each animal, comparisons between response curves in air before and after the carotid chemoreceptor denervation and between response curves with intact peripheral chemoreceptor innervation in air and in 100% O2 (after correction for the excitatory effect o f 02, see above) were performed. The Wilcoxon sign rank test [13] was used to test the significance o f differences. Differences were considered significant if P ~< 0.05. RESULTS
C02 responses in normoxia in cats with intact carotid sinus nerves
A total of 28 units were studied. Of these, 24 increased their firing rate with an increase in paCO2, within at least part of the range of CO2 levels studied, one did not change and 3 decreased. On the basis of a previously oOOOeoA
2, F
oo
..~.2.0
2.0
•
B
2.5
1.5
4"0
1 • ¢) " ' t E
~oO
I
I
I
I
c
•r 2"r
i:I
E
5.0
oo°°o oeeOee
I
i
i
"" I
3-0 I
D
2.0
"O"
i7. 2,0~ •O
1.o 00
00 0
IO eqp 20 40
61 i 0 810 100
, 20
., 40
6'0
Pac02
I 0
100
(mm
I
Hg)
Fig. 1. CO2 responses in normoxia of sympathetic preganglionic neurons. A--D: types 1--4 response patterns. E: average CO2 response (n = 28). Carotid sinus nerves intact.
426
proposed classification [12], 13 (46%) units had a type 1 response, i.e. were silent at low CO2 levels, were recruited at a critical CO2 level, and increased their firing rate as CO2 increased above threshold. One unit (4%) had a type 2 response, i.e. insignificant changes in firing rate as CO2 levels changed. Eleven units (39%) had type 3 response, i.e. the unit firing rate was relatively independent of CO2 within a certain range of low CO2 levels, and was CO2dependent within a range of higher CO2 levels. Finally, 3 units (11%) had type 4 response, i.e. their firing rate was highest at low CO2 levels and decreased with increasing CO2 levels. Fig. 1A--D shows examples of these 4 response types. These different unit responses to CO2 were recorded in preparations with typical phrenic nerve response to CO2, therefore the differences could not be attributed to abnormality of the chemoreceptor reflexes in these preparations. Moreover, units with different response types could be recorded in the same preparation at different times. Thus, the different response types were not related to differences between animals, but reflected a heterogeneity of the sympathetic neuron population with respect to chemoreceptor input distribution and/or intrinsic neuron properties. Fig. 1E shows h o w the average discharge rate o f this SPN population varies with paCO2. The curve was obtained by averaging the mean firing rates of the 28 units at each 5 mm Hg increment of paCO2 between 20 and 90. The curve has a sigmoid shape, with m a x i m u m slope in the 50--70 mm Hg PaCO2 range, and asymptotically approaching plateaus at the lowest and highest paCO2 values. C02 responses in normoxia after carotid sinus nerve section
The effect of bilateral CSN section was studied in 7 of the 28 units. The average response curves for these 7 units, before and after section, are shown
oOO D
~3
o
{g
o o
==
2
o
a L~
0
20
POCO2 [mm Hg)
Fig. 2. Depressant e f f e c t o f bilateral carotid sinus nerve ( C S N ) s e c t i o n o n the average CO2 response in n o r m o x i a o f s y m l m t h e t i c preganglionic n e u r o n s (n = 7). e , CSN intact; D, C S N s e c t i o n e d ; ×, intact m i n u s s e c t i o n e d - - t h e " d i f f e r e n c e curve".
427
in Fig. 2. Denervation resulted in a displacement of the curve to the fight and a decrease in maximum slope and in the high paCO2 plateau. These changes were statistically significant. On the assumption that the CO2 response after denervation is due to central chemoreceptors [3] and that there is simple addition of the effects of central and carotid sinus (CS) chemoreceptors on the SPN, the difference between the curves before and after denervation should represent the contribution of the CS chemoreceptors to the response recorded with intact CSN (difference curve in Fig. 2). Comparison of the response curve after denervation with the difference curve suggests that the CS contribution is comparable to the central in the range of paCO2 from 45 to 65 mm Hg. Above 65 mm Hg, the CS chemoreceptor contribution declines while that of the central chemoreceptors continues to rise. In detail, CSN section abolished the CO2 response of one unit, decreased the response of 3 and had no effect on the response of two others. For the remaining unit the CO2 response not only persisted, but the curve was somewhat displaced to the left. Fig. 3A shows the data for the unit which lost its CO2 response after CSN section. Thus, this unit, which had a type 3 CO2 response, derived its CO2 sensitivity from the peripheral chemoreceptor input only. CSN section did not change significantly the unit firing rate around normocapnia, suggesting the absence of a tonic input from CS afferents. As expected from this inferred connection to CS chemoreceptor, the sensitivity to CO2 of this unit was increased by hypoxia. An example of a unit in which the CO2 response was significantly reduced, but not abolished by CSN section, is shown in Fig. 3B. Two units with a type 3 CO2 response had behavior similar to that of the unit shown in Fig. 3B. This behavior suggests that they received both central
2.5
A
1"6
B QO cO0
1.4 ~2.0
1.2
@ =~
o o
O~ 1,5
o oll
o
o
o
e@
o
o o ~
I
•
o
°
o
o
1.0
o
@@
'8
O
O
O
~ 1.0 C
-6
D
._~ tL
.4
C O.5
•
Q
o o
.2 o
0
0 i
i
i
i
n
0
20
40
60
eo
PoCO
2 (mm Hg)
o
eoeeo~
)
L
i
100
0
20
n 40
Poco
,
i
i
n
60
80
100
2 (mmHg)
Fig. 3. E f f e c t o f bilateral carotid sinus nerve ( C S N ) s e c t i o n o n CO2 response in n o r m o x i a o f t w o units, o, C S N intact; D, CSN s e c t i o n e d .
428 and peripheral CO2-sensitive chemoreceptor input. Of the two units, whose CO2 response was unchanged after CSN section, one had a t y p e 1 and the other a t y p e 4 CO2 response. This lack of effect suggests that the CO2 sensitivity of these two units was due only to connections with central chemoreceptors, unless a displacement to the right of the curve, due to the elimination of the CS chemoreceptors, was exactly matched b y a displacement to left, due to the elimination of the CS baroreceptors. This exact matching seems unlikely. Finally, in one unit, with a t y p e 3 response to CO2, the response was displaced to the left after CSN section. Thus, the COs sensitivity of this unit was due mainly to input from central chemoreceptors. While it is possible that the displacement to the left was due to removal of an inhibitory input originating from CS chemoreceptors, it is more likely that it was due to the elimination of baroreceptor afferents. The fact that the firing rate of this unit in the CO2 insensitive range nearly doubled after CSN section supports this view. The obvious drawback with CSN section as a m e t h o d for estimating the carotid chemoreceptor contribution to the CO2 response of SPNs is the concomitant elimination of the carotid baroreceptors. This could lead to an underestimate of the CS chemoreceptor contribution. As stated in the Methods section, we tried to minimize this source of variability by performing the section at similar arterial blood pressure values, obtained b y i.v. infusion of hexamethonium or phentolamine, in all cats. Treatment with these drugs also lowered arterial blood pressure so that the level of baroreceptor input, prior to section, was also low. This probably explains w h y the average COs response curves before and after CSN section, in the CO: range from 20 to 40 mm Hg, are practically superimposable (Fig. 2). In this range the CO: sensitivity of the SPNs is low, and therefore we would expect to see best the effects of changes in the level of other inputs, e.g. baroreceptors. Since we do not see significant changes, we infer that the effect of baroreceptor removal must have been negligible, with the one exception mentioned above. C02 response curves in hyperoxia
As an additional, seemingly independent procedure for estimating carotid chemoreceptor contribution to the CO2 response of SPNs, we increased p,O2 to values greater than 350 mm Hg on all 28 units. This is known to reduce the responsiveness to CO2 of peripheral chemoreceptors [5]. We assumed therefore that a decrease i n chemoreceptor activity would be reflected in a displacement of the CO2 response curve to the right and/or in a decrease of its slope, and/or in a change of the curve's plateau at high p,CO2, when the experiment was conducted in hyperoxia as compared to experiments conducted in normoxia. Fig. 4 shows the average CO2 response curve obtained in hyperoxia in the whole sample o f 28 units. For comparison the CO2 response curve in normoxia, already shown in Fig. 1E, is plotted. Also plotted is the difference between the t w o curves. Notice that, qualitatively,
429 4
5
~I00
5
.E_ v)
4 o
(n o A o o
~
4O
o o
u.
A o
u.
2O
ooo x
20
40
x
x
~
0
Poco 2 (mm Hg)
x
80
x
x ~'~
ooo
x
I
100
210
I
40
I
60
~
810
1~)0
Paco 2 (mm Hg)
Fig. 4. Depressant effect of hyperoxia on the average CO2 response of sympathetic preganglionic neurons (n = 28). Intact carotid sinus nerves. $, normoxia; A, hyperoxia; ×, normoxia minus hyperoxia. Curves normalized by expressing firing rates as percentage of that at 90 mm Hg PaCO2 in normoxia. Fig. 5. Excitatory effect of hyperoxia on the average CO2 response of sympathetic preganglionic neurons (n = 7) in cats without peripheral chemoreceptors. D, normoxia; A, hyperoxia; ×, hyperoxia minus normoxia.
hyperoxia has an effect similar to CS denervation, namely it causes a displacement to the right, decreased slope and decreased plateaus of the response curve (P < 0.05). Notice also the similarity of the difference curve (normoxia minus hyperoxia) of Fig. 4 with the difference curve of Fig. 2. The displacement to the right and downward, however, is less with hyperoxia than with CSN section. When the effect of hyperoxia on the CO2 response o f individual units is examined, the response of 14 units (50%) was shifted down, that of 8 (28%) was shifted up, and that of 6 (22%) did not change. The increase in CO2 responsiveness in these 8 units could not be attributed to changes in the level of baroreceptor input, since there were no changes in arterial blood pressure. It could have been due either to an excitatory action of hyperoxia, independent of peripheral chemoreceptors, analogous to that described for respiratory neurons [8], or to the fact that CS chemoreceptor input to some SPNs may be inhibitory. The test for the latter possibility has already been performed with negative results, by examining the effects of CSN section. As a test for the first possibility in a series of experiments on 7 o f the 28 units, CO2 response curves in hyperoxia were obtained and compared with CO2 response curves in normoxia, in CS d enervated preparations. This procedure would detect any excitant or depressant effect of hyperoxia, unrelated to CS chemoreceptors. The results of this experiment are shown in Fig. 5. It can be seen that hyperoxia displaces to the left the average CO2 response curve of these SPNs (P < 0.05). Moreover, the effect is CO2-dependent (see difference curve). This experiment demonstrates that hyperoxia can excite SPNs by an
430
0
O O
0 0 0
0 J¢ Q. U~
o
D 0
o
•
o
0
D 0
g°
o
r,-
•
g
o o
o o
C I,I. 1
0 0
I 0
I 20
I 40
I 60
P°c02 (mm Hg)
] 80
I 100
I 0
I 20
l 40
i 60
i 80
I 100
Poco 2 (mmHg)
Fig. 6. A : depressant effect of hyperoxia on the average CO2 response of sympathetic preganglionic neurons (n = 7 ) in cats with intact carotid sinus nerves, after graphical correction for the excitatory effect of hyperoxia, o, normoxia; ©, corrected hyperoxia; ×, normoxia minus corrected hyperoxia. B: similarity of CO2 response curves in normoxia, after carotid sinus nerve (CSN) section, and in hyperoxia, with intact CSN after correction for the excitatory effect of hyperoxia, n = 7. Average CO2 response in normoxia, with intact C S N , is given for comparison. $, normoxia, CSN intact; m, normoxia, CSN sectioned; ©, corrected hyperoxia, CSN intact.
action independent of CS chemoreceptors. This action can account for the left-hand shifts of CO2 response curves observed in hyperoxia in cats with CSN intact. In detail, the CO2 response curve after CSN section was unchanged by hyperoxia in 2 of the 7 units and displaced to the left in the remaining 5. The fact that, in the absence of CS chemoreceptors, hyperoxia never caused a displacement of the right to the curve, can be considered evidence that, in the CSN-intact cat, when hyperoxia causes a displacement to the right of the curve, it must be due to a decrease in CSN chemoreceptor input. From these experiments it can be concluded that the CO2 response curve in hyperoxia of SPNs in preparations with intact CSN results from the algebraic summation of two effects onto the (~O2 response curve in normoxia: a depressant effect of CS chemoreceptor activity [5] a n d a n excitatory effect on other structures, perhaps on ceni~aI nervous system structures [8]. Since any displacement of the CO2 r e s p o ~ curve in hyperoxia from that in normoxia, in preparations with sectione(}Carotid sinus nerves {Fig. 5) is due only to the excitatory effect, it can b~ ~~ e n as a measure of this effect. When this measure of the excitatory effec~ o f O~ is subtracted from the CO2 response curve in hyperoxia in the same preparations, but with intact CS nerves, a new curve is obtained which represent s the effect of the CS chemoreceptor depression only. This procedure ~ a s applied to the data from the same group of 7 units and the resultant curVes are shown in Fig.
431
6A. It can be seen in Fig. 6B that after corrections for the excitatory effect of O~, the hyperoxic CO2 response curve is indistinguishable from the normoxic denervated curve, i.e. both denervation and hyperoxia, after correction for O2 excitation, give similar estimates for the CS contribution to the CO2 response curve of the SPNs. The :,elation between sympathetic preganglionic neuron and phrenic motoneuron C02 responses under various experimental conditions The CO2 response of this SPN population is similar to that of phrenic motoneurons, recently described [4]. This similarity is due to the great preponderance of type 1 and 3 responses in the population and is brought out by the graphs of Fig. 7. In Fig. 7A the CO2 response curves in normoxia of the two neuron populations are shown. In Fig. 7B the responses in normoxia and hyperoxia, before and after CSN section, of the phrenic nerve (peak amplitude) and of the SPN population (mean firing rate) are plotted against each other. This plot shows that the similarity in CO2 response o f the two neuron populations persists after CSN section and in hyperoxia. From Fig. 7A it is clear that both neuron populations show a strong relation between activity levels and CO2 levels. For both neuron systems the
B
A 100
/
O@ • @
~db
@o
0
C~ 8O (el 0 Q. h-
6o D
m
i
@
0 ®
60
Q
E =J
._~
E ®
E4o
o
~o
0 o
x
IE
x
o
*
o e
~O
@oO@
20
o@
z a,
® ®
I
A
0
20
I
|
40 ,
Pot- 0 "
60 2
(mmHg)
.... I
@0
I
I
i
L
I
[
I
100
0
20
40
60
SO
100
Phrenic~ of Maximum Response
Fig. 7. Similarity of CO2 responses of phrenic motoneurons and sympathetic preganglionic neurons. A: average normalized CO2 response (n = 7) of sympathetic preganglionic neurons (mean firing rate, l ) and of phrenic motoneurons (mean peak amplitude of integrated phrenic nerve burst, o) in normoxia, in cats with intact carotid sinus nerves (CSN). B: average normalized CO2 responses of phrenic motoneurons (horizontal axis) plotted against average normalized CO2 responses of sympathetic preganglionic neurons (vertical axis) in normoxia with CSN intact (e) and after CSN section (o), and in hyperoxia, with CSN intact after correction for excitatory effect of 02 Co) and after CSN section (~). n=7.
432
relation is sigmoidal. In Fig. 7B it can be seen that for phrenic responses greater than 40% of maximum {corresponding to paCO2 greater than 50 mm Hg from Fig. 7A) the experimental points fall very close together, i.e. the two systems have the same sensitivity to COs. At phrenic response amplitudes lower than 40% of maximum, however, the sensitivity of the SPN population gradually declines and approaches zero, i.e. SPN activity becomes independent o f phrenic amplitude and of PaCO~ for phrenic amplitudes less than 20% of maximum (corresponding to PaCO2 below 40 mm Hg). The data obtained after CSN section fall along the same path as those before, showing that CS "chemodenervation" has a similar influence on the COs response o f the two neuron populations (Fig. 7B). The same conclusion can be drawn regarding the effect of hyperoxia, before and after CSN section, since these data also indicate close similarity of response.
Role o f central and carotid sinus chemoreceptors in the generation o f unit responses to COs Type 1 and 3 responses in normoxia could be shown in 11 of 24 units to be caused by excitation from both sets of chemoreceptors, in 1 by excitation from central chemoreceptors only, in 1 by excitation from CS chemoreceptors only, and in 2 by excitation from CS chemoreceptors and inhibition from central chemoreceptors. In the latter cases, hyperoxia resulted in a change from either t y p e 1 or 3 to t y p e 4 response. For the remaining 9 units with t y p e 1 or 3 response there was clearly excitation from central chemoreceptors, but the role of CS chemoreceptors could not be assessed, because hyperoxia only was used and its effect was a displacement to the left or no change in the response curve. These results, as discussed above, cannot be interpreted because of the demonstrated excitatory effect of 02. All 3 type 4 responses in normoxia were caused b y inhibition from central chemoreceptors, associated in 2 cases with excitation from CS chemoreceptors. The CS chemoreceptor influence was of such magnitude as to be overcome b y the inhibitory influence of the central chemoreceptors. The one unit with t y p e 2 response did n o t have either central o r CS chemoreceptor input. DISCUSSION
The present results confirm previous observations demonstrating the sensitivity to CO2 of this population of SPNs [12]. The population o u t p u t has been studied in the present investigation over a much wider range of paCO2 values {20--90 mm Hg) than in the previous study. Section of the CS nerve or hyperoxia resulted in a significant decrease in the CO2 response o f the population. This decrease w a s attributed to the removal of a CO2
433
ments [3] and, likewise, taken as a measure of the contribution of this input to the overall CO2 response of the population. Whereas practically all units received input from central chemoreceptors, a smaller number received input from the CS chemoreceptors (4/7 or 56% on the basis of CSN section data). No estimate could be made from the hyperoxia data, due to the already discussed uncertainties in the interpretation of the cases of no change or displacement to the left of the CO2 response curves in hyperoxia. The contribution of the CS chemoreceptors to the CO2 response of this SPN population, estimated from the CSN section data and from the hyperoxia data after correction for the 02 excitation, is comparable in magnitude to that of the central chemoreceptors up to a paCO2 of 60--70 mm Hg. Beyond these values, the CS contribution declines and becomes a progressively smaller component of the total response, whereas the contribution of the central chemoreceptors continues to increase. The similarity of the effects of 100% 02 (after correction) and denervation suggests that the reflex effects of CS chemoreceptor stimulation by CO2 are completely abolished as a result of hyperoxia. A possible interference by baroreceptor reflexes (or their removal) in these experiments has been dealt with by the procedure of ganglion block and noradrenaline infusion. By these means changes in arterial blood pressure, evoked by CO2 either reflexly or by a direct action on vascular smooth muscle [14] have been avoided. Since arterial blood pressure was stable during rebreathing, so must also have been baroreceptor activity levels. Therefore the various CO: response patterns of units cannot be attributed to differences in baroreceptor activity levels. The present experiments do not provide data concerning the sources of the CO2 sensitivity of SPNs in hyperoxia or after CSN section. Previous experiments [3], however, have demonstrated that the CO2 sensitivity of hind limb vasomotor tone, in peripherally "chemodenervated" cats, is nearly completely abolished by cooling the exposed ventral surface of the medulla. This finding suggests that the main source of the CO2 sensitivity of SPNs in hyperoxia or after CSN section is represented by chemoreceptors on the ventral medullary surface. Chemosensitivity of the spinal cord has been suggested by Szulczyk and Trzebski [16], but there is no evidence that this leads, in the spinal animal, to a significant CO2-responsiveness of hind limb vascular tone [7] or of cervical trunk SPNs (Rohliceck and Polosa, unpublished observations). The similarity of responses to CO2 of SPNs and phrenic motoneurons, as well as the analogies in the modification of their responses caused by CSN section or hyperoxia have been pointed out (Fig. 7). These analogies may simply be the result of the fact that the same (or similar) sensory systems are responsible for the CO2 responses of the two neuron populations. An additional possibility, suggested previously [12], is that the CO2 sensitivity of some SPNs does not result from synaptic connections of these neurons with the CO2 sensitive receptor, but from synaptic connections with inspiratory neurons [11], which would make the SPN a "follower" of inspiratory neu-
434
ron activity. These t w o possibilities are n o t mutually exclusive, and it is possible that the observed SPN--phrenic correlation results from both causes. We estimated the CS chemoreceptor contribution to the CO2 response curve of SPNs as the difference between the curves before and after CSN section. This "difference curve" has a characteristic shape: it climbs, as COs increases, up to a peak at between 60 and 70 mm Hg, and b e y o n d this value it declines. The "difference curve" obtained from the hyperoxia data had a similar shape. This decline of the CS chemoreceptor c o m p o n e n t cannot be attributed to a decline in sensory input from CS chemoreceptors, since this increases as CO2 increases and then plateaus, but does not decline [1 ]. More likely explanations are either that at high COs there is occlusion (i.e. hypoadditive interaction) between central and CS chemoreceptor evoked reflexes, with the result that in the absence of the CS c o m p o n e n t the central component is overestimated, or that as COs increases there is some inhibitory interaction between the two inputs. Hyperoxia, in peripheral chemodenervated cats, caused a displacement to the left of the CO2 response curve of the population, i.e. caused an excitation of SPNs. The excitatory effect of hyperoxia on SPN was unknown before this work. However, an excitatory effect o f hyperoxia on respiration [8] was known. A similar mechanism may account for the effects of hyperoxia on the two neuron systems. This excitation of SPNs, or phrenic motoneurons, may result from a decrease in: (i) cerebral blood flow; and/or (ii) carbamino transport and buffer action of hemoglobin with increased oxygenation [6]. An increase in pCO2 and a decrease in pH at the level of the central chemoreceptors could result from either mechanism. Finally, these observations have suggested that central chemoreceptors can exert an inhibitory action on SPNs. Thus, units which were inhibited b y high pCO~ (type 4) were not affected by CSN section or hyperoxia. Moreover, abolishing peripheral chemoreceptor input by hyperoxia transformed the original excitatory effects of COs upon 2 units (type 1 and type 3) into an inhibitory one (type 4). This shows that the excitatory action of the CS chemoreceptors can mask the inhibitory action of the central chemoreceptors. ACKNOWLEDGEMENTS
This work was supported by the Medical Research Council of Canada and the Quebec Heart Foundation. REFERENCES 1 Biscoe, T.J., Purves, M.J. and Sampson, S.R., The frequency of nerve impulses in single carotid body chemoreceptor afferent fibres recorded in vivo with intact circulation, J. Physiol. (Lond.), 208 (1970) 121--131. 2 de Groat, W.C. and Lalley, P.M., Reflex sympathetic firing in response to electrical stimulation of the carotid sinus nerve in the cat, Brain Research, 80 (1974) 17--40.
435 3 Hanna, B.D., Lioy, F. and Polosa, C., The effect of cold blockade of the medullary chemoreceptors on the CO2 modulation of vascular tone and heart rate, Canad. J. Physiol. Pharmacol., 57 (1979) 461--468. 4 Hanna, B.D. and Polosa, C., Carotid and central chemoreceptor components of the phrenic nerve response to CO2 in the anesthetized cat, Canad. Physiol., 10 (1979) 164. 5 Lahiri, S. and DeLaney, R.G., Stimulus interaction in the responses of carotid body chemoreceptor single afferent fibers, Resp. Physiol., 24 (1975) 249--266. 6 Lambertsen, C.J., Kough, R.H., Cooper, D.Y., Emmel, G.L., Loeschcke, H.H. and Schmidt, C.G., Comparison of relationship of respiratory minute volume to pCO2 and pH of arterial and internal jugular blood in normal man during hyperventilation produced by low concentrations of CO2 at 1 atmosphere and by 02 at 3.0 atmospheres, J. appl. Physiol., 5 (1953) 803--813. 7 Lioy, F., Hanna, B.D. and Polosa, C., CO2-dependent component of the neurogenic vascular tone in the cat, Pfliigers Arch. ges. Physiol., 374 (1978) 187--191. 8 Miller, M.J. and Tenney, S.M., Hyperoxic hyperventilation in carotid-deafferented cats, Resp. Physiol., 23 (1975) 23--30. 9 Mitchell, R.A., Loeschcke, H.H., Massion, W.H. and Severinghaus, J.W., Respiratory responses mediated through superficial chemosensitive areas on the medulla, J. appl. Physiol., 18 (1963) 523--533. 10 Pelletier, C.L., Circulatory responses to graded stimulation of the carotid chemoreceptors in the dog, Circulat. Res., 31 (1972) 431--443. 11 Preiss, G., Kirchner, F. and Polosa, C., Patterning of sympathetic preganglionic neuron firing by the central respiratory drive, Brain Research, 87 (1975) 363--374. 12 Preiss, G. and Polosa, C., The relation between end-tidal CO2 and discharge patterns of sympathetic preganglionic neurons, Brain Research, 122 (1977) 255--269. 13 Rumke, C.L. and De Jonge, H., In D.R. Laurence and A.L. Bacharach (Eds.), Evaluation of Drug Activities: Pharmacometrics, Academic Press, N.Y., 1964, pp. 47--100. 14 Somlyo, A.P. and Somlyo, A.V., Vascular smooth muscle. II Pharmacology of normal and hypertensive vessels, Pharmacol. Rev., 22 (1970) 249--353. 15 Suutarinen, T., Cardiovascular response to changes in arterial carbon dioxide tension, Acta physiol, scand., 67, Suppl. 266 (1966) 1--76. 16 Szulczyk, P. and Trzebski, A., The local effect of pH changes in the cerebrospinal fluid off the ventrolateral areas of medulla oblongata and on spinal cord surface on the activity of cardiac and vertebral sympathetic nerves, Acta physiol, pol., 27 (1976) 9--17.