Ventrolateral medulla and sympathetic chemoreflex in the rat

174

Brain Research, ~)0t) (1993) 174- I~4 1993 Elsevier Science Publishers B.V. All rights reserved 0006-8993/93/$06.01)

BRES 18745

Ventrolateral medulla and sympathetic chemoreflex in the rat Naohiro

Koshiya,

Donghai

Huangfu

and Patrice

G. Guyenet

University of Virginia Health Sciences Center, Department of Pharmacology, Charlottesville, VA 22908 (USA)

(Accepted 17 November 1992)

Key words: Carotid chemoreceptor; Sympathetic chemoreflex; Caudal ventrolateral medulla; Rostral ventrolateral medulla; Sympathetic baroreflex; Hypoxia

Splanchnic sympathetic nerve discharge (SND), phrenic nerve activity (PND) and putative sympathetic premotor neurons of the rostral ventrolateral medulla (RVL) were recorded in urethane-anesthetized vagotomized rats without aortic baroreceptor afferents. Carotid chemoreceptor stimulation with brief N 2 inhalation increased SND by 1015: 7%, raised mean arterial pressure (MAP) and increased the discharge rate of RVL premotor neurons by 46 + 12% (N = 32). During chemoreceptor activation, SND and most RVL neurons displayed pronounced central respiratory rhythmicity with maximal firing probability immediately after cessation of the PND (postinspiratory phase) and lowest probability during PND (inspiratory phase). Bilateral microinjection of the broad spectrum glutamate receptor antagonist kynurenic acid (Kyn, 5 nmol in 100 nl) into RVL blocked the sympathetic chemoreflex but left the sympathetic baroreflex intact. In contrast, bilateral microinjection of the same dose of Kyn into the caudal ventrolateral medulla (at obex level, CVL) blocked the baroreflex but left the sympathetic chemoreflex intact. Bilateral microinjection of the GABA A agonist muscimol (87.5 pmol in 50 nl) into CVL produced effects identical to those of Kyn. These results confirm that the caudal ventrolateral medulla contains an essential relay of the sympathetic baroreflex and demonstrate that the same area plays no role in the sympathetic chemoreflex. The data suggests that these two reflexes could have a largely independent course through the medulla oblongata and that integration between the baroreceptor and chemoreceptor information used for sympathetic vasomotor control may occur as late as the premotor neuronal stage in RVL.

INTRODUCTION T h e activation o f t h e s y m p a t h e t i c nerve d i s c h a r g e ( S N D ) c a u s e d by s t i m u l a t i o n o f c a r o t i d c h e m o r e c e p tors ( s y m p a t h e t i c c h e m o r e f l e x ) is m a r k e d l y sync h r o n o u s with t h e c e n t r a l r e s p i r a t o r y rhythm, suggesting t h a t it is m e d i a t e d , at least in part, via the activation of s o m e y e t u n i d e n t i f i e d c o m p o n e n t of the c e n t r a l r e s p i r a t o r y r h y t h m g e n e r a t i n g n e t w o r k 3'~8'2°-22. D e t a i l e d i n f o r m a t i o n on c h e m o r e c e p t o r p r i m a r y afferents a n d synaptic p r o c e s s i n g in t h e nucleus of t h e solitary tract is available 8'33'34 b u t t h e rest of the central circuitry o f t h e s y m p a t h e t i c c h e m o r e f l e x is virtually unknown, in c o n t r a s t to t h a t of t h e b a r o r e f l e x 23. T h e c a u d a l p o r t i o n of t h e v e n t r o l a t e r a l m e d u l l a o b l o n g a t a ( C V L ) c o n t a i n s p r o p r i o b u l b a r inhibitory int e r n e u r o n s ( G A B A e r g i c ) which a p p e a r to constitute t h e last m e d u l l a r y link o f t h e polysynaptic p a t h w a y o f the s y m p a t h e t i c b a r o r e f l e x p r i o r to the b u l b o s p i n a l p r e m o t o r n e u r o n a l stage 24'38. A c c o r d i n g l y , b l o c k a d e o f

synaptic t r a n s m i s s i o n in this r e s t r i c t e d a r e a o f the m e d u l l a e l i m i n a t e s the s y m p a t h e t i c baroreflex, pres u m a b l y by p r e v e n t i n g t h e inhibition of the tonic excit a t o r y drive c o n t r i b u t e d by b u l b o s p i n a l s y m p a t h o e x c i t a t o r y p r e m o t o r n e u r o n s l o c a t e d in t h e rostral v e n t r o l a t e r a l m e d u l l a ( R V L ) 1°A3'24. T h e m a i n goals of the p r e s e n t e x p e r i m e n t s are: (i) to verify t h a t the s y m p a t h e t i c p r e m o t o r n e u r o n s o f R V L receive c o n v e r g e n t a n d o p p o s i n g influences from b a r o a n d c h e m o r e c e p t o r s a n d (ii) to d e t e r m i n e w h e t h e r o r n o t C V L also plays a role in the s y m p a t h e t i c c h e m o r e flex. MATERIALS AND METHODS General procedures These have been described in several prior publications ~s'jg'36. Male Sprague-Dawley rats (320-380 g) were intubated and respirated with 100% 0 2. Surgical preparation was done under halothane anesthesia (1.3-1.5%). For recording, intravenous urethane (1 g/kg, initial dose) was substituted for halothane because of the well-known

Correspondence: P.G. Guyenet, University of Virginia Health Sciences Center, Box 448, Department of Pharmacology, Charlottesville, VA 22908, USA. Fax: (1) (804) 982-3878.

175 Baroreceptor activation or unloading was done with i.v. injections of phenylephrine (PE, 5-10 ~g/kg) or sodium nitroprusside (SNP, 10-20 p.g/kg), respectively. At the end of the experiment a high dose of clonidine was administered (see below) and the animals were deeply anesthetized with halothane (2-5%) or additional urethane (0.5-1 g/kg bolus). They were then perfused transcardially with buffered saline (pH 7.4) followed by 4% paraformaldehyde (pH 7.0).

depressant effect of the latter on carotid body function 3°. After equilibration (15 min) and verification of the adequacy of the level of anesthesia (see criteria below), paralysis was achieved with i.v. pancuronium (1.5 mg/kg, 1 mg/kg thereafter as needed). The dose of i.v. urethane used eliminates all noxious responses (corneal reflex, withdrawal reflex and AP increase upon strong nociceptive stimulation of the hind paw) for at least 1 h after the halothane crossover when tested in the absence of the paralyzing agent. In the present case (paralysis with pancuronium) the adequacy of the anesthesia was evaluated by the absence of increase in AP and PND during strong nociceptive stimulation by the hind paw 31. Supplemental doses of urethane (0.1 g/kg i.v.) were administered whenever any of the above responses to nociceptive stimuli was observed. Total dosage of i.v. urethane throughout the experiment (duration 2.5-4 h under urethane anesthesia) ranged from 1.2-1.6 g/kg. Carotid chemoreceptor stimulation was done by ventilating the rats with N 2 for 8-12 s instead of 100% 0 2 (timed with electronic valves) or with 12% 0 2 in N 2 for 2-4 min. As demonstrated before 1'14 PND and SND activation produced by both protocols are completely eliminated by bilateral sections of carotid sinus nerves.

A

Control

(mmHg) 1 0 0 ~ , ~

RVL Kyn

Electrical recordings Phrenic and sympathetic nerve discharge (PND and SND) were recorded with bipolar electrodes (bandpass of 30-3000 Hz) and stored on FM tape (bandpass 0-1150 Hz) along with arterial pressure, end-expiratory CO 2 concentration (eeCO 2) and the concentration of 0 2 in the breathing mixture 5A2. For low time resolution of the effect of chemoreceptor activation (e.g., Fig. 1A), SND and PND were integrated over successive 1 s periods 19. SND was expressed in units, zero being the electrical noise (determined after i.v. injection of 200 /zg/kg of the sympatholytic drug clonidine at the end of the experiments) and 100 units being the

Recovery

~

~1~100"~ ) ,

~~

0

2 min

B

Control

RVL Kyn

SND 200

(un#J)

100

~

0 (u~) 100'

PND

I

0 Fig. 1. Effect of kynurenic acid microinjection into rostral ventrolateral medulla (RVL) on sympathetic baroreflex and chemoreflex. A: low time resolution excerpts (1 s bins). Mean arterial pressure (MAP), splanchnic sympathetic nerve discharge (SND) and phrenic nerve discharge (PND) are shown before (control), 10 min (RVL Kyn) and 60 min (Recovery) after bilateral injection of Kynurenic acid into RVL (5 nmol, 100 nl). Baroreflex was triggered by bolus i.v. injection of phenylephrine (filled arrowheads) followed by i.v. injection of sodium nitroprusside (open arrowheads). Chemoreflex was elicited by inhalation of N 2 for 10 s (filled rectangles). B: higher time resolution (30 ms sample-hold integration) of chemoreflex during control period (control) and 2 min after Kyn injection (RVL Kyn). Chemoreceptor stimulation by 10 s N 2 inhalation at bar. C: Kyn injection sites in RVL plotted on a standard coronal section (interaural -2.80 mm) borrowed from the atlas of Paxinos and Watson 28. Abbreviations: Ac, nucleus ambiguus, compact part; GiV, gigantocellular nucleus, ventral part; IO, inferior olive; NTS, solitary tract nucleus; Pyr, pyramidal tract; Sp5, spinal trigeminal nucleus.

176 resting level at the beginning of the recording session. PND was expressed also in arbitrary units, zero representing the electrical noise recording during central apnea produced by hyperventilation and 100 as the resting level at eeCO 2 of 4-4.5%. For higher time resolution (e.g., Fig. 1B), the rectified SND and PND were subjected to analog integration (time constant 1 s) with 30 ms sample hold resetting (Gould, 13-G4615-70). Unit recording and antidromic activation of presumed sympathetic premotor neurons of the rostral ventrolateral medulla and PND-triggered histograms of unit discharge were done as previously described14.ts. 35.36.

Drug microinjections Electrophysiological mapping was done to target drug microinjections accurately into the rostral or caudal ventrolateral medulla, as described before 12A3. Kynurenic acid (Sigma, 50 mM) or muscimol (Sigma, 1.75 mM) were dissolved in artificial cerebrospinal fluid adjusted at pH 7.30-7.35. Fluorescent latex beads (Lumafluor) were incorporated in the solutions for subsequent histological identification of injection sites 12. Injections were performed in a volume of 100 nl (Kyn) or 50 nl (muscimol). All injection sites were identified in coronal sections from aldehyde fixed brains. The center of the injections was recorded on standardized sections from the atlas of Paxinos and Watson 2s.

and increase in phrenic nerve discharge, typical effects of peripheral chemoreceptor activation. The increase in SND ranged from 29% to 194% of baseline and averaged 101 _+ 7% over 23 rats. These effects are totally abolished by bilateral sectioning of the carotid sinus nerves ~4. Expanded time scale illustrations of the effect of N 2 inhalation on SND and PND are represented in Fig. lB. Note that sympathoactivation is markedly synchronized with the phrenic outflow and that, as shown before in rodents 5'263~, peak activation occurs during the post-inspiratory phase and minima coincide with the peak of the phrenic discharge. The central respiratory phases are defined here after

RVL b

b

6

b

•

O

b

b

b

MAP

•

Resting

[]

N2

200

Statistics One-way ANOVA for repeated measures was used to determine the effect of intraparenchymal drug injection on MAP at rest, at the peak after PE injection or the nadir following SNP injection. Differences in AP between control, drug treatment and recovery were evaluated by Fischer PLSD test (circles in upper panels of Figs. 2, 4 and 6). The effect of intraparenchymal drug injection on chemoreflex activation of MAP was examined by 2-way ANOVA of resting MAP and peak MAP during N 2 inhalation. If the F value indicated a significant interaction between drug treatment and chemoreceptor activation, the effect of chemoreceptor activation (difference score between resting and peak MAP) was evaluated with the Fischer PLSD test (diamonds in upper panels of Figs. 2, 4 and 6). Effects of PE, SNP injection or N 2 inhalation on resting MAP were evaluated by the PLSD test (asterisks in upper panels of Figs. 2, 4 and 6). Since SND was normalized to 100% during the control period, the non-parametric Kruskal-Watlis test (3 time-point experiment with Kyn: control, treatment, recovery) or Mann-Whitney U-test (2 time-point experiments with muscimol: control, treatment) was used to determine the overall effect of drug injection on resting SND. Differences between control, drug treatment and recovery were then evaluated by the Mann-Whitney U-test for three time-point measurements (circles in bottom panel of Figs. 2, 4 and 6). The effect of intraparenchymal drug injection on the SND response to chemoreceptor activation was analyzed by 1-way ANOVA of the differences between resting level and peak response, since these differences were normally distributed. The difference scores (difference between resting values and values at peak reflex activation) were finally compared by Fischer PLSD test (diamonds in bottom panels of Figs. 2, 4 and 6). The response to baroreceptor activation was similarly analyzed. Effects of PE, SNP and N 2 on resting SND were evaluated by Mann-Whitney U-test (asterisks in bottom panels of Figs. 2, 4 and 6). Significance level was set at 0.05. All results are expressed as mean + S.E.M.

RESULTS

Sympathetic chemoreflex A s i l l u s t r a t e d in Fig. 1A, b r i e f N 2 i n h a l a t i o n p r o duced arterial pressure elevation, sympathoactivation

E E 100

Control

Kyn

~'

40O

O

Recovery b

b

b

O

SND

• [] []

300 .

Resting N2 PE

0200/ 100 0

Control

Kyn

Recovery

Fig. 2. Effect of Kyn injections into RVL on sympathetic chemoreflex and baroreflex: summary. Upper panel, mean arterial pressure; lower panel, splanchnic SND. Filled columns represent resting conditions before drug, 5-10 min after bilateral microinjection of Kyn into RVL (5 nmol, 100 nl, injection sites in Fig. 3A) and after recovery. Clear columns represent maximum increase in MAP and SND caused by 10 s N 2 inhalation. Striped columns represent MAP reached after i.v. bolus injection of phenylephrine and lowest SND recorded at this MAP. Asterisks indicate that the parameters (MAP or SND) determined after N 2 inhalation or PE injection were significantly different from resting state. The ANOVA results are represented by the symbols above the graphs. Circles with different fill (open or closed) indicate that the values represented by the underlying columns are significantly different. Diamonds refer to effect of Kyn injection on chemoreflex and baroreflex (ANOVA and post hoc test on difference scores between resting values and values at peak of reflex). Diamonds with different fill indicate significant differences. For details on statistics, see methods.

177 Schwarzacher et al. 31 The postinspiratory phase represents the very early part of the expiratory phase. Phrenic discharge during this period in the rat is very small and not always observable (see, however, Figs. 9, 10) 31.

sympathetic baroreflex. Half recovery time was ranged from 20-40 min. Complete recovery from the effects of Kyn occurred within 60 min (Fig. 1A). The effect of Kyn on the chemoreflex was extremely reproducible and highly significant as indicated in the summary figure (Fig. 2). The results of Fig. 2 represent thre determinations: the first, before interparenchymal injection; the second, 5-10 min after injection; and the third, after full recovery. Chemoreceptor and baroreceptor activations were performed in random order during the experiment. The effect of the drug on PND was time dependent. At first (0.5-5 min), central tachypnea and/or irregularities in respiratory rhythm were observed. This phase was followed by elimination of phrenic activity. Yet tonic activation of PND by chemoreceptor stimulation was commonly observed be-

Effect o f bilateral kynurenic acid microinjection into the ventrolateral medulla on the sympathetic chemoreflex

Bilateral microinjections of the broad spectrum excitatory amino acid receptor antagonist kynurenic acid (Kyn) were successfully placed in the RVL in six animals (sites illustrated in Fig. 1C). An example of one such experiment is illustrated in Fig. 1A. Note that Kyn virtually abolished resting PND and its activation by chemoreceptor stimulation. Kyn also abolished the effect of N 2 inhalation on SND. In contrast, Kyn produced little effect on resting SND and no effect on the

Control

A MAP

CVL Kyn

Recovery

rv

~

200

(rnmHg)

o ----T---g

:7

3OO 8ND (units)

200

\\

100 0 PND (units)

/

,/ //

\ \ ~ ,

J

\~ \\

i/L.~j;, \

;._i/// I

~.. . . . ,',--r//

lOO

2 mln

B

Control

CVL Kyn

600

SND 400-

(units) 2000 PND

(.n~) ! 00 -

or

.

AAA__d .

.

.

51N!¢ Fig. 3. Effect of kynurenic acid microinjection into caudal ventrolateral medulla on sympathetic baroreflex and chemoreflex. A: low time resolution excerpts before (control), 10 min (CVL Kyn) and 60 min (recovery) after intraparenchymal injection of Kyn into CVL. B: higher time resolution of chemoreflex before (control) and 10 min after Kyn injection into CVL (CVL Kyn). C: injection sites (interaural - 4.30 mm from 28). Abbreviations: 12, hypoglossal nucleus; A, ambiguus nucleus; LRN, lateral reticular nucleus; others, see Fig. 1C. For more details, see legend of Fig. 1.

178 fore the full recovery from Kyn. A higher time resolution example of such a case is shown in Fig. lB. Note that despite the persistence of a phrenic outflow (of the tonic type), the sympathetic chemoreflex was nevertheless completely abolished. Bilateral microinjections of Kyn centered in the ventrolateral medulla at or slightly rostral to obex level (CVL, sites illustrated in Fig. 3C) produced dramatically different results with regard to both the baroreflex and the chemoreflex. An example of one animal is illustrated in Fig. 3A. Note that Kyn eliminated PND, raised resting SND, blocked the sympathetic baroreflex but produced no modification of the sympathetic chemoreflex. These results were extremely reproducible and highly significant, as indicated by the summary data of Fig. 4. A higher time resolution of the effect of Kyn injection in CVL is illustrated in Fig. 3B. Note the persistence of the chemoreflex activation of SND. Note also that SND activation during N 2 administration remained rhythmic in the face of a disappearance of the phrenic output. An identical pattern of response was observed in all five cases.

Effect of bilateral microinjection of muscimol into the caudal ventrolateral medulla The preceding results suggested that glutamate neurotransmission in CVL plays no role in the sympathetic chemoreflex. In order to eliminate the possibility that CVL might actually play a role in the chemoreflex but that other transmitters than excitatory amino acids might be involved, injections of the GABAA-receptor agonist muscimol were performed in CVL in five rats with the intent of inhibiting every cell in the area regardless of transmitter content (injection sites represented in Fig. 5C). The effects of muscimol were identical to those produced by kynurenic acid injections into the same region. Muscimol produced the following effects: disappearance of phrenic nerve discharge, elevation of mean arterial pressure, elimination of the sympathetic baroreflex (Fig. 5A), preservation of the sympathetic chemoreflex. The phasic, respiratory-like modulation of SND typical of chemoreceptor activation was also preserved as in the case of Kyn injections into CVL (Fig. 5B). A summary of the results is shown in Fig. 6. In this figure the baroreflex is illustrated by the maximum range of SND Observed during elevation of AP with i.v. phenylephrine (PE) and hypotension with i.v. sodium nitroprusside (SNP). Note that after muscimol injection in CVL, very large swings in AP produced no significant change in SND (e.g., Fig. 5A) but the amplitude of the chemoreceptor activation of SND was unchanged.

Activation of rostral ventrolateral medulla sympathetic premotor neurons by chemoreceptor actication A total of 32 RVL pulse-synchronous and barosensitive cells (i.e., neurons inhibited by baroreceptor activation and at least 75% inhibited at MAP of 150-165 mmHg), were recorded in RVL just caudal to the facial motor nucleus (Fig. 7). Twelve of 19 cells tested, 63%, were also characterized as bulbospinal by antidromic activation (antidromic latency: 18 _+4 ms, n = 12). We have previously demonstrated that a still higher proportion (76%, n = 14016) of these highly characteristic pulse-synchronous ceils of the rat rostral RVL actually project to the spinal cord. The slight difference in the probabilities of antidromic activation in the two studies might be simply explained by the smaller sampling in the present study or by the fact that the preparation was partially debuffered (vagus and aortic depressor nerves cut). In the rat, carotid baroreceptor activation alone is occasionally unable to sufficiently attenuate the cells' activity to prevent frequent collisions of antidromic spikes with spontaneous spikes. We have previously shown that antidromic activation of these highly

CVL 6 0

•

b

•

O

MAP

• Resting [] []

2OO

Control

Kyn

6 400.

•

Recovery

b

~'

b

o

SND 300.

N2 PE

• Resting []

N2

~_.. I~1 PE

2OO

0

Control

Kyn

Recovery

Fig. 4. Effect of microinjcction of kynurenic acid into the caudal ventrolateral medulla on sympathetic baroreflcx and chemorcflex: summary. This summary figure ( N = 5) is presented exactly as Fig. 2. Kyn (5 nmol, 100 nl) was bilaterally injected into the caudal ventro-

lateral medulla(for sites, see Fig.3C).

179 activation. Note that as this cell is activated by chemoreceptor stimulation, its discharge becomes considerably more synchronized with the central respiratory cycle than at rest. At the end of the chemoreceptor stimulation (panel B), the cell stopped firing briefly presumably because of instantaneous interruption of the chemoreceptor-related drive and persistence of the baroreceptor feed-back inhibition due to the chemoreceptor-mediated rise in AP. After computer averaging this cell exhibited a firing probability similar to that described in Fig. 10B, i.e., expiratory phase stimulation with maximum during the postinspiratory phase. An increase in the respiratory rhythmicity of RVL

active cells typically fails if the firing rate of the cell is not decreased to a few spikes/s by baroreceptor activation or iontophoresis of a hyperpolarizing agent (e.g., GABA). The mean discharge rate of all RVL barosensitive cells was 19 + 4 spikes/s. A majority of the cells was activated during N 2 inhalation but a proportion was not activated or was slightly inhibited. The average activation for the entire population was 46 + 12%. The two distribution histograms of Fig. 8 describe the degree of activation of RVL barosensitive cells and that of the splanchnic nerve during chemoreceptor activation. Fig. 9 is an oscilloscopic representation of the response of this type of RVL neuron to chemoreceptor

A

B

200 Control (units) SND

I0

it

l

Control

SND (units)

21100

200.] MAP

PND (unite)

150,

(mmHg)

2O0

100.

5O CVL Musc

CVL Musc

200

SND (~) 1oo, o

200.

MAP (mmHg)

SND (w~s) V

o

¥

15o.

200.

PND (unit.)

100. 50.

1 min

51m¢

Fig. 5. Effect of muscimol microinjection into caudal ventrolateral medulla on sympathetic baroreflex and chemoreflex. A: baroreflex. Low time resolution excerpts before (control), 10 min (CVL Musc) after bilateral intraparenchymal injection of muscimol into CVL (50 nl, 87.5 pmol). B: high time resolution of chemoreflex before (control) and 10 min after muscimol injection into CVL (CVL Musc). C: injection sites (interaural - 4.30 mm). Abbreviations: see Fig. 3C. For more details, refer to Fig. 1.

180 barosensitive cells during peripheral chemoreceptor activation was consistently found when computer averaging was used. In order to be able to perform PND-triggered averaging of unit discharges, chemoreceptor stimulation of longer duration had to be used. Therefore moderate (12% 0 2) hypoxia was used in this case. Under these conditions the central respiratory rhythmicity of all RVL cells was found to increase with chemoreceptor stimulation. Several examples are shown in Fig. 10. At rest, the majority of these units exhibited a mild rhythm with post-inspiratory peak. The respiratory modulation of all cells was increased during hypoxia, although to a vastly different degree

CVL

o

o

°

°

ca

o

•

MAP

• Resting []

20o

N2

!~ PE [ ] SNP

~ I''1~

o= -r-

E E 100

Control

Musc

b o

400

o

,5 •

b

• Ro=inQ ~[] N2 | [] PE r-~ I111SNP I I

SND

300

lO0

0

Control

Fig. 7. Location of units recorded in the rostral ventrolateral medulla. Barosensitive units (filled circle) and postinspiratory on-off units (open circle) plotted on coronal planes sections, 175 /zm apart. Abbreviations: 7, facial nucleus; others, see Fig. 1C.

Mu$c

Fig. 6. Effect of muscimol injection into CVL on sympathetic baroreflex and chemoreflex: summary. Muscimol (87.5 pmol, 50 nl) was bilaterally injected into CVL (injection sites, Fig. 5C). Upper panel: resting MAP, maximum MAP during chemoreceptor 'activation by Nz, MAP maximum recorded after i.v. injection of phenylephrine, MAP minimum after i.v. injection of sodium nitroprusside. Lower panel: corresponding values of SND. Asterisks indicate that the parameters (MAP or SND) determined after (i) N 2 inhalation and after (ii) PE or SNP injection were different from resting state (e.g., after muscimol injection, SND was significantly different from resting only during chemoreceptor activation by N2). Circles indicate comparisons between pre-drug and post-drug values for a given variable (different fills indicate P < 0.05; same fill, no difference). Diamonds indicate effect of muscimol injection on difference scores (e.g., difference between resting and N 2, i.e., chemoreceptor activation) before and after drug. Diamonds with different fill, P < 0.05; same fill, no significance. For example, increase in SND by N 2 inhalation was the same before and after muscimol.

(see Fig. 10A,B). Finally, a minority of the cells (9 out of 32) had an inspiration related discharge (minima in postinspiratory phase) which was also accentuated by mild hypoxia (e.g., Fig. 10C,D). Among or adjacent to these RVL barosensitive units, multiple types of 'on-off respiratory units were recorded (cells exhibiting a period of complete silence during some defined phase of every central respiratory cycle). These included postinspiratory units 29'31, (recording sites, Fig. 7), which were all greatly activated by brief hypoxia (427 + 20% from resting firing rate of 9 + 2 spikes/s, n = 6). DISCUSSION The present data confirms prior evidence that integration between baroreceptor and peripheral chemoreceptor inputs to the sympathetic vasomotor outflow occurs at or prior to the sympathetic premotor neuronal stage in RVL, since a large proportion of these cells are affected by both stimuli 37'39. The present results add to prior work by analyzing the respiratory synchronization of RVL premotor neurons 15'z5 during

181 peripheral chemoreceptor stimulation. During chemoreceptor stimulation, the central respiratory pattern of R V L p r e m o t o r neurons was an exaggeration of the one detected at rest. The dominant pattern was what we have previously called 'postinspiratory' to denote the fact that the nadir coincides with the peak of the inspiratory phase and the apex usually corresponds to the earliest part of the central expiratory phase (Fig. 10A,B). I t should be quite clear that the relationship between the firing rate of R V L sympathetic premotor neurons and the central respiratory cycle is probabilistic, i.e., that these cells may fire at any time during the respiratory cycle, even during peripheral chemoreceptor activation (see Fig. 10). These cells are therefore clearly distinct from the 'postinspiratory' cells identified by respiratory physiologists and also recorded in the present study in the ventrolateral medulla, which are consistently silent during late expiration and inspiration (e.g., ref. 31). The postinspiratory pattern of R V L barosensitive neurons (e.g., Fig. 10B) was also the one observed in the splanchnic sympathetic nerves during peripheral chemoreceptor stimulation (e.g., Fig. 1B). This similarity supports the notion that R V L p r e m o t o r neurons play a role in the chemosympathetic r e f e x but evidence already exists that chemoreceptor input also activates other putative sympathetic premotor neurons, e.g., the A5 neurons 14. In agreement with

A

12 10 8

N 6

0 -100

-50

0

50

100

SND Change

150

200

250

200

250

(%)

B iiiiii~ii!!i .iii!i!!ii!iii:!i~. N

I

4

0

-100

lii!iii!iiiiilii!iiii~!iii!l

~ iiiir:iii~ iii!!i~iiiiiiiiiiiiiiii~iiii~ iiiii ~i

i!!iiiiiiii!iiiiiii!iii!i!iiiii! !!ili!ii i!ii:~ii{{ii~i! -50

0

50

100

150

Firing Rate Change (%) Fig. 8. Effect of chemoreceptor activation on SND and discharge of RVL premotor neurons. A: distribution histogram of maximum increase in splanchnic SND during N 2 inhalation (8-12 s) in 23 rats. B: distribution histogram of maximum activation of RVL premotor neurons by the same procedure (25 cells in 14 rats).

A

!

B

C

I A Fig. 9. Effect of chemoreceptor stimulation on activity of RVL sympathetic premotor neuron: oscillographic record. Unit activity of RVL cell antidromically activated from segment T3 of spinal cord (latency 16 ms) and discharging at 4 Hz at resting AP. A, before N2 inhalation (control). B, last 7 s of a 10 s Nz inhalation (at bar) followed by first three seconds of recovery with 100% 0 2. C: inhibition of the same cell during increase in arterial pressure (lower trace) caused by bolus i.v. injection of phenylephrine (at arrow head). Sweep duration: 10 s in A & B, 20 s in C; 100/zV vertical calibration for spikes, upper traces; calibration arbitrary for integrated PND, lower traces in A & B; 100-200 mmHg for arterial pressure, lower trace in C.

this possibility, activation of R V L premotor neurons by chemoreceptor stimulation was on average smaller, percentage-wise, than that of the splanchnic sympathetic output (Fig. 8). The second new observation of the present study is that blockade of excitatory amino acid transmission in R V L eliminates the sympathetic chemoreflex but has no effect on the baroreflex while the converse occurs when the same agent is introduced in the ventrolateral medulla at obex level (CVL). The increase in resting SND and the elimination of the sympathetic baroreflex caused by local injection of an antagonist of excitatory amino acid receptors in CVL (Kyn) confirm prior observations 1°,13. Kyn antagonizes most glutamate receptor operated channels including the N M D A , kainate and A M P A 2. The pharmacological specificity of intraparenchymal Kyn adminis-

182 tration in the dose used in the present study relies on the following criteria of specificity: (i) the drug produces no effect on GABAergic transmission (refs. 13 and 36 and present results demonstrating lack of effect of Kyn injection into RVL o n GABA-mediated baroreflex); (ii) resting sympathetic tone is unaffected when the agent is introduced into RVL (refs. 13 and 36 and present results); and (iii) the inactive structural analog xanthurenic acid produces no effect when injected into RVL or CVL 13. We also showed previously that the effect of Kyn was regionally specific, since injections centered 1 mm away from CVL were ineffective with regard to the baroreflex 13. This regional specificity is further confirmed in the present study, since the rostrai and caudal medullary injections were separated by merely 1.5 mm but produced totally different effects on two distinct reflexes (chemoreflex and baroreflex). The blockade of the baroreflex by Kyn injections into CVL is consistent with the notion that propriomedullary GABAergic neurons located in the CVL area and projecting to RVL premotor neurons constitute the main inhibitory link of the sympathetic baroreflex 24 (see Fig. 11; for review see ref. 23). In agreement with

A: PUTATIVE SYMPATHETIC CHEMOREFLEX 0 Chemoreceptors

B: PUTATIVE SYMPATHETIC BAROREFLEX

Obex

• epto~...~

/TO IML .--

/

L

i

Pre-Botz RVL

~ GABA • EAA

B

A

500400-

160 120.

300Aotlvlly (IplikllrOln) Unit

A~?~/ 80(l~lku~n) . 40-

200 -

1O0" O" PND 200(unlt=) 100~05

. ~

0

?

0.5 1.0 (we)

(unlt=)PNO 200"100":Lt" 0" ~t ~L,_~''~,,~, .~ o -0.5

1.5

0

0.5 1'0 (~)

15

Time

Time

D

C

Fig. 11. Putative baro- and chemo-sympathetic pathways. Sympathetic baroreflex is redrawn on parasagittal section through medulla oblongata24. Prebotzinger area (presumed 'kernel' of respiratory rhythm generating network) is located after Smith et al. 32 CVL (caudal ventrolateral medulla 13"24'25) is assumed to be roughly coextensive with the rostral ventral respiratory group (rVRG 32) and is depicted as containing the bulk of phrenic premotor neurons in rats 27. All connections with a question mark are anatomically plausible but hypothetical. In particular, the respiratory input to RVL sympathetic premotor neuron could originate elsewhere in the medulla or even in the pons. Abbreviations: CVL, caudal ventrolateral medulla; EAA, excitatory aminoacid; IML, intermediolateral cell column of spinal cord; NTS, nucleus of the solitary tract; Lat.RN, lateral reticular nucleus; PMN, phrenic motor nucleus; pre-Botz, pre-Botzinger area; RVL, rostral ventrolateral medulla.

100.

300. 80

60 ~-t~vity (~el=14iln) Unit

Unit 200" ACUvny (IpikN,t~ln) 100'

20

PND 3o0 (UNit)200 100 0 -0.5

PND 2o0

0

0.5 Time ( ~ )

1.0

1.5

-0.5

0

0.5 1.0 (~:}

15

Time

Fig. 10. Central respiratory modulation of RVL barosensitive units during chemoreceptor stimulation with 12% O 2. A-D: top traces represent PND-triggered peri-event histograms of the unit activity of individual RVL barosensitive neurons at rest (light traces) and during 12% O 2 inhalation (maintained for 2-4 min, dark traces). Bottom traces (PND) represent the corresponding average phrenic discharge. PND and perievent histograms were triggered by the onset of the PND (calibration arbitrary). Each histogram/average was obtained by 200 successive sweeps.

this theory, the baroreceptor input to RVL sympathetic neurons is chloride mediated 11 and bicuculline sensitive 35. Moreover, the presence of neurons with the expected anatomical projections to RVL and the expected physiological properties (excited by baroreceptor stimulation and pulse-synchronous) have now been identified in CVL 38. Their excitation by baroreceptor stimulation is assumed to be mediated by the release of an excitatory amino acid, hence the effectiveness of Kyn in CVL 13. The key observation of the present study is that injections into CVL of either Kyn or muscimol produced no effect on the chemosympathetic reflex. Our interpretation is that while baroreflex inhibition of the

183 sympathetic outflow is synaptically relayed in CVL, chemoreflex activation of the sympathetic vasomotor discharge uses an other anatomical route (see schematic in Fig. 11). A very important clue in support of this scheme is that Kyn or muscimol injections into CVL eliminated the phrenic outflow but did not affect the oscillatory nature of the sympathetic response to chemoreceptor activation (see panels B in Figs. 3 and 5). This suggests that these pharmacological treatments eliminated PND but did not eliminate central respiratory rhythm generation nor the effect of the respiratory rhythm generator network on sympathetic efferents. The explanation of this dissociation may rest in the peculiar anatomy of the rat respiratory network. There appears to be no dorsal respiratory group (ventrolateral nucleus of the solitary tract) in the rat 6'17'31 and the vast majority of phrenic bulbospinal premotor neurons are located in the ventrolateral medulla as opposed to the NTS 27. These cells according to Onai et a127 lie in the area defined by respiratory physiologists as the 'rostral ventral respiratory group (rVRG)' which appears to coincide with the 'CVL' defined in the field of vasomotor control 4'7'1°'13'32 (see Fig. 11). Therefore, it is conceivable that Kyn injections into C V L / r V R G does not affect respiratory rhythm generation but merely uncouples the central respiratory oscillator from the phrenic inspiratory premotor neuronal pool without uncoupling it from the rostrally located sympathetic premotor neuronal pool. This interpretation is also consistent with the theory that the critical area for respiratory rhythm generation, the 'pre-Botzinger complex' lies rostral to the rVRG/CVL, about halfway between obex and the caudal end of the facial motor nucleus 32 (see Fig. 11). Kynurenic acid blocks all classic excitatory amino acid receptor subtypes (kainate, AMPA and NMDA) but has no effect on the 'metabotropic' or trans-ACPD selective subtype 9. It could also be argued that, in the dose used, kynurenic acid might preferentially block NMDA receptors 1°. Thus the lack of effect on the sympathetic chemoreflex of kyn injection into CVL is in itself insufficient evidence to eliminate the alternative explanation, i.e., that CVL might actually be an important area for the sympathetic chemoreflex but that local synaptic processing might not involve kynurenate-sensitive receptors. This remaining interpretation is rendered extremely unlikely by the fact that administration of muscimol into the same region produced effects strictly identical to those of kynurenic acid (compare Figs. 3 and 5, Figs. 4 and 6). Muscimol would be expected to clamp the membrane potential of all neurons reached by the drug at the chloride equilibrium potential, thereby preventing any firing regardless

of what transmitter is normally involved in their discharges. In conclusion, the present results are compatible with the circuitry illustrated in Fig. 11 representing (i) independent pathways for the baro- and chemosympathetic reflexes converging on RVL premotor neurons and (ii) a very rostral location (rostral to CVL) in the ventrolateral medulla of some basic components of the respiratory rhythm generating network and their connection with the vasomotor sympathetic network. Acknowledgements. This work was supported by grants from the National Institutes of Health to P.G.G. (HL 39841 and HL 28785).

REFERENCES 1 Biesold, D., Kurosawa, M., Sato, A. and Trzebski, A., Hypoxia and hypercapnia increase the sympathoadrenal medullary functions in anesthetized, artificially ventilated rats, Jpn. J. Physiol., 39 (1989) 511-522. 2 Collingridge, G.L. and Lester, R.A., Excitatory amino acid receptors in the vertebrate central nervous system, PharmacoL Rev., 40 (1989) 143-210. 3 Coote, J.H., The organisation of cardiovascular neurones in the spinal cord, Rev. Physiol. Biochem. Pharmacol., 110 (1988) 145285. 4 Cravo, S.L., Morrison, S.F. and Reis, D.J., Differentiation of 2 cardiovascular regions within caudal ventrolateral medulla, Am. J. Physiol., 261 (1991) R985-R994. 5 Darnall, R.A. and Guyenet, P.G., Respiratory modulation of preand postganglionic lumbar vasomotor sympathetic neurons in the rat, Neurosci. Lett., 119 (1990) 148-152. 6 Ezure, K., Manabe, M. and Yamada, H., Distribution of medullary respiratory neurons in the rat, Brain Res., 455 (1988) 262-270. 7 Feldman, J.L. and Ellenberger, H.H., Central coordination of respiratory and cardiovascular control in mammals, Annu. Rev. Physiol., 50 (1988) 593-606. 8 Finley, J.C.W and Katz, D.M., The central organization of carotid-body afferent-projections to the brainstem of the rat, Brain Res., 572 (1992) 108-116. 9 Glaum, S.R. and Miller, R.J., Metabotropic glutamate receptors mediate excitatory transmission in the nucleus of the solitary tract, Z Neurosci., 12 (1992) 2251-2258. 10 Gordon, F.J., Aortic baroreceptor reflexes are mediated by NMDA receptors in caudal ventrolateral medulla, Am. J. Physiol., 252 (1987) R628-R633. 11 Granata, A.R. and Kitai, S.T., Intracellular analysis in vivo of different barosensitive bulbospinal neurons in the rat rostral ventrolateral medulla, J. Neurosci., 12 (1992) 1-20. 12 Guyenet, P.G., Darnall, R.A. and Riley, T.A., Rostral ventrolateral medulla and sympathorespiratory integration in rats, Am. J. Physiol., 259 (1990) R1063-R1074. 13 Guyenet, P.G., Filtz, T.M. and Donaldson, S.R., Role of excitatory amino acids in rat vagal and sympathetic baroreflexes, Brain Res., 407 (1987) 272-284. 14 Guyenet, P.G., Koshiya, N., Huangfu, D., Verberne, A.J.M. and Riley, T., Central respiratory control of A5 and A6 pontine noradrenergic neurons, Am. J. Physiol., in press. 15 Haselton, J.R. and Guyenet, P.G., Central respiratory modulation of medullary sympathoexcitatory neurons in rat, Am. J. Physiol., 256 (1989) R739-R750. 16 Haselton, J.R. and Guyenet, P.G., Ascending collaterals of medullary barosensitive neurons and C1 cells in rats, Am. J. Physiol., 258 (1990) R1051-R1063. 17 Hilaire, G., Monteau, R., Gauthier, P., Rega, P. and Morin, D., Functional significance of the dorsal respiratory group in adults and newborn rats: in vivo and in vitro studies, Neurosci. Lett., 111 (1990) 133-138.

184 18 Huang, W.-X., Lahiri, S., Mokashi A. and Sherpa, A.K., Relationship between sympathetic and phrenic nerve responses to peripheral chemoreflex in the cat, J. the Autonomic Nert,ous System, 25 (1988) 95-106. 19 Huangfu, D., Koshiya, N. and Guyenet, P.G., A5 noradrenergic unit activity and sympathetic nerve discharge in rats, Am. Z Physiol., 261 (1991) R393-R402. 20 Janig, W. Pre- and postganglionic vasoconstrictor neurons: differentiation, types and discharge properties, Annu. ReL'. PhysioL, 50 (1988) 525-539. 21 Katona, P.G., Dembowski, K., Czachurski J. and Seller, H., Chemoreceptor stimulation on sympathetic activity: dependence on respiratory phase, Am. J. Physiol., 257 (1989) R1027-1033. 22 Kollai, M. and Koizumi, K., Differential responses in sympathetic outflow evoked by chemoreceptor activation, Brain Res., 138 (1977) 159-165. 23 Kumada, M., Terui, N. and Kuwaki, T., Arterial baroreceptor reflex: its central and peripheral neural mechanisms, Progr. Neurobiol., 35 (1990) 331-361. 24 Masuda, N., Terui, N., Koshiya, N. and Kumada, M., Neurons in the caudal ventrolateral medulla mediate the arterial baroreceptor reflex by inhibiting barosensitive reticulospinal neurons in the rostral ventrolateral medulla in rabbits, J. Auton. Nero'. System, 34 (1991) 103-118. 25 McAllen, R.M., Central respiratory modulation of subretrofacial bulbospinal neurones in the cat, J. Physiol., 388 (1987) 533-545. 26 Numao, Y., Koshiya, N., Gilbey, M.P. and Spyer, K.M., Central respiratory drive-related in symathetic nerves of the rat: the regional differences, Neurosci. Lett. 81 (1987) 279-284. 27 Onai, T., Saji, M. and Miura, M., Projections of supraspinal structures to the phrenic motor nucleus in rats studied by a • horseradis peroxidase microinjection method, J. Auton. Nert,. System, 21 (1987) 233-239. 28 Paxinos, G. and Watson, C., The Rat Brain in Stereotaxic Coordinates, 2nd edn., Academic Press, Sydney, 1986. 29 Richter, D.W., Bischoff, A., Anders, K., Bellingham, M. and Windhorst, U., Response of the medullary respiratory network of the cat to hypoxia, J. Physiol., 443 (1991) 231-256.

30 Seagard, J.L., Hopp, F.A., Donegan, J.H., Kalbfleisch, J.it. and Kampine, J.P., Halothane and the carotid sinus reflex: evidence for multiple sites of action, Anesthesiology, 57 (1982) 191-202. 31 Schwarzacher, S.W., Wilhelm, A., Anders, K. and Richter, D.W, The medullary respiratory network in the rat..L Physiol., 435 (1991) 631-644. 32 Smith, J.C., Ellenberger, H.H., Ballanyi, K., Richter, D.W. and Feldman, J.L., Pre-Botzinger complex - a brainstem region that may generate respiratory rhythm in mammals, Science 254 (1991) 726-729. 33 Spyer, K.M., The central nervous organization of reflex circulatory control. In A.D. Loewy and K.M. Spyer, Central Regulation of Autonomic Functions, Oxford University Press, New York, 1990, pp. 168-188. 34 Spyer, K.M., Izzo, P.N., Lin, R.J., Paton, J.F.R, Silva Carvalho, L.F. and Richter, D.W., The Central Nervous Organization of the Carotid Body Chemoreceptor Reflex. In H. Acker, A. Trzebski and R.G. O'Regan, Chemoreceptors and Chemoreceptor Reflexes, Plenum Press, New York, 1990, pp. 317-321. 35 Sun, M.-K. and Guyenet, P.G., GABA-mediated baroreceptor inhibition of reticulospinal neurons, Am. J. Physiol., 249 (1985) R672-R680. 36 Sun, M.-K. and Guyenet, P.G., Hypothalamic glutamatergic input to medullary sympathoexcitatory neurons in rats, Am. J. Physiol.. 251 (1986) R798-R810. 37 Sun, M.-K. and Spyer, K.M., Response of rostroventrolateral medulla spinal vasomotor neurones to chemoreceptor stimulation in rats, J. Auton. Nert. System, 33 (1991) 79-84. 38 Terui, N., Masuda, N., Saeki, Y. and Kumada, M., Activity of barosensitive neurons in the caudal ventrolateral medulla that send axonal projections to the rostral ventrolateral medulla in rabbits, Neurosci. Lett., 118 (1990) 211-214. 39 Terui, N., Saeki, Y. and Kumada, M., Barosensory neurons in the ventrolateral medulla in rabbits and their responses to various afferent inputs from peripheral and central sources, Jpn. J. PhysioL, 36 (1986) 1141-1161.