facilitation of the arterial baroreflex by the ventrolateral part of the

Journal of Physiology (1994), 477.1

Facilitation of the arterial baroreflex by the ventrolateral partof the midbrain periaqueductal grey matter in rats

Koji Inui, Sumio Murase* and Shoichiro Nosaka*

Departments of Psychiatry and *Physiology, Mie University School of Medicine, Tsu,Mie 514, Japan

1. The effects of stimulation of the ventrolateral part of the midbrain periaqueductal greymatter (PAG) on the arterial baroreflex were investigated in urethane-chloraloseanaesthetized and artificially ventilated rats.

2. Both electrical and chemical stimulation of the ventrolateral PAG provokedhypotension, vagal bradycardia and marked facilitation of baroreflex vagal bradycardia(BVB), which was induced by stimulation of the aortic depressor nerve. The magnitudeof ventrolateral PAG facilitation of BVB was 328 + 193% (n = 34) for electricalstimulation and 243 + 224 % (n = 13) for chemical stimulation. Baroreflex hypotensionwas slightly augmented during either electrical or chemical stimulation of theventrolateral PAG in vagotomized rats.

3. Stimulation of the nucleus raphe magnus (NRM) also provoked hypotension, vagalbradycardia and facilitation of BVB. The magnitude of BVB facilitation was 234 + 132%(n = 8) for electrical stimulation and 328 + 170 % (n = 7) for chemical stimulation. Aftermicroinjection of kainic acid into the NRM region, baroreflex facilitation, as well ashypotension and vagal bradycardia, produced by ventrolateral PAG stimulation, wasalmost abolished.

4. In conclusion, the ventrolateral PAG, besides producing hypotension and bradyeardia,facilitates arterial baroreflexes. These effects are exerted via the NRM, sharplycontrasting with effects of the dorsal PAG.

The midbrain periaqueductal grey matter (PAG) isinvolved in a variety of functions that include defensivebehaviour (Hilton & Redfern, 1986; Bandler & Carrive,1988), vocalization (Jiirgens, 1991), autonomic functions(Carrive, Bandler & Dampney, 1989), sexual behaviour(Sakuma & Pfaff, 1979) and pain control (Lovick, 1993a).Some recent studies have indicated, however, that thePAG is not a functionally homogeneous structure and thatthe dorso- and ventrolateral parts of the PAG producedifferent or even opposing effects. For example,stimulation of neurones in the dorsolateral part of the PAGprovokes defensive behaviour that is accompanied bymarked sympathoexcitation, whereas stimulation of theventrolateral part produces immobility and sympatho-inhibition (Bandler, Carrive & Zhang, 1991). Both thedorso- and ventrolateral portions of the PAG subserveantinociception, but the analgesic effects are of differingtypes with different descending pathways (Lovick, 1993a).Anatomically, it has been demonstrated that there areapparently different patterns of afferent and efferentprojections to/from the dorso and ventrolateral parts ofthe PAG (Veening, Buma, Horst, Roeling, Luiten &Nieuwenhuys, 1991; Holstege, 1991).

Among the different or opposing effects produced bythe dorso- and ventrolateral parts of the PAG, thecardiovascular changes that accompany behaviouralreactions have been studied most intensively by Bandler &Carrive (1988), Carrive et at. (1989), Bandler et al. (1991) andLovick (1992). Stimulation of the dorsolateral part of thePAG produces an increase in blood pressure, tachycardia,mesenteric vasoconstriction and vasodilation in theskeletal musculature in the hindlimbs, which representpart of a complex co-ordinated defence reaction. As aunique cardiovascular involvement, arterial baroreflexesare suppressed during dorsolateral PAG stimulation(Jones, Kirkman & Little, 1990; Nosaka, Murata, Inui &Murase, 1993). To warrant a sufficient blood supply to thedilated vessels in the skeletal musculature during defensivebehaviour, the arterial baroreflex must be activelysuppressed in the face of concomitant hypertension(Nosaka et at. 1993). In contrast to such dynamiccardiovascular changes evoked by the dorsolateral PAG,stimulation of the ventrolateral PAG produces hypo-tension and bradycardia (Lovick, 1992), accompanied byimmobility (Bandler et al. 1991). Furthermore, it has beenreported by Lovick (1992) that activation of neurones in the

WES 2594, pp. 89-101 89

K. Inui, S. Murase and S. Nosaka

ventrolateral PAG inhibits cardiovascular defenceresponses elicited by dorsal PAG stimulation. In view ofthese findings, it is relevant to examine whether theventrolateral PAG affects arterial baroreflexes, a subjectthat has not previously been investigated.

This study was designed to determine whether and howthe ventrolateral part of the PAG modulates the arterialbaroreflex. Furthermore, we sought to identify thedescending pathway of the modulation of baroreflexes bythe ventrolateral PAG. In our study, the aortic depressornerve (ADN) was electrically stimulated to provoke acontrol baroreflex. The major advantage of ADNstimulation to provoke baroreflexes in rats is the fact thatthe ADN in the rat contains only baroreceptor afferents(Sapru & Krieger, 1977).

METHODS

Male Wistar rats, weighing 350-400 g, were anaesthetized byintraperitoneal injection of a-chloralose (60 mg kg-') andurethane (600 mg kg'). One-sixth of the initial dose wasadministered at 3 h intervals to maintain anaesthesia. Thetrachea was cannulated and polyethylene tubes were insertedinto the left femoral vein and artery for administration ofdrugs and measurement of blood pressure, respectively.

bA

Animals were immobilized by intravenous injection ofsuccinylcholine (10 mg kg-') and ventilated artificially. Thedepth of the anaesthesia under paralysis was assessed bystability of arterial blood pressure and heart rate. f-Blockadewas achieved by intravenous injection of propranolol (bolusinjection of 04 mg kg-' followed by 02 mg kg-1 at 2 hintervals). Under such conditions, changes in heart rate couldbe attributed to changes in vagus nerve activities. The aorticdepressor nerves (ADNs) on both sides were prepared forelectrical stimulation as described previously (Nosaka, Murase& Murata, 1991). Carotid sinus nerves (CSNs) on both sides werecut in each rat to eliminate any possible interactions betweeninputs from the ADNs and the CSNs. The ADN on a given sidewas electrically stimulated and responses in terms of bloodpressure and heart rate were recorded. Parameters for thestimulation were 8 V and 0 5 ms duration. By adjusting thestimulation frequency, we reduced the bradycardiac responseto ADN stimulation to less than one-third of the maximumresponse, so that both facilitatory and inhibitory effects ofPAG stimulation could be examined. Once the appropriatestimulation frequency had been determined, it was heldconstant throughout the experiment.

Stimulation of the PAGThe PAG ipsilateral to the ADN used was stimulated eitherchemically or electrically. Chemical stimulation was achievedby pressure injection of DL-homocysteic acid (DLH, 6 jug in

C

150

cmABP -0

400

IC1

E200 ,,

ADN

Ba b c 150

ABP L_0° E

- ~~~~~50

400

HR - ~ 1r< E -~~~~~~~200 c,

ADN ADNADN

vPAG

D

OPAGX

10 s

Figure 1. Effects of electrical stimulation of the dorsal PAG and the ventrolateral PAG on thearterial baroreflexA a and A c, control baroreflex provoked by aortic depressor nerve (ADN) stimulation (8 V, 0 5 ms and30 Hz); Ab, ADN response during dorsal PAG (dPAG) stimulation (30 1uA, 0 5 ms, 50 Hz). Ba and Bc,control baroreflex; Bb, ADN response during ventrolateral PAG (vPAG) stimulation. C and D,stimulated sites. Periods of the respective stimulations are indicated by horizontal bars. ABP, arterialblood pressure; MBP, mean blood pressure; HR, heart rate. A-D, data from one and the same rat.

HR

ADN

ADN

dPAG

J. Phy-siol. 477.190

If I

Barorefiex facilitation by the ventrolateral PAG

100 nl over 15 s) through a tip-blunted glass capillary pipette(tip diameter, 50 ,um). Stimulated sites were confirmedhistologically with Fast Green that had been dissolved in thesolution of DLH. Electrical stimulation was conducted through amonopolar stainless-steel electrode (100 ,um in diameter, fullyinsulated except for the tip) by passing cathodal square-wavecurrent pulses (20-35 ,uA, 0 5 ms and 50 Hz). The sites of thestimulation were confirmed by histochemical demonstration of ironions that had been deposited by delivery of an anodal direct current(DC, 20 ,sA for 25 s) at the end of each experiment.

Brain lesionsIn order to localize the descending pathway of ventrolateralPAG modulation of the arterial baroreflex, lesions in theventromedial medulla were made either electrolytically orchemically. Electrolytic lesions were made by passing acathodal DC (10 mA for 60 s) through a monopolar stainless-steel electrode (150 ,um in diameter). For a chemical lesion,microinjection of kainic acid (KA, 1 jug in 500 nl over 10 min)was used. The extent of the lesion was confirmed byhistochemical demonstration of HRP, dissolved in KAsolution, by the DAB (diaminobenzidine) method. Effects ofventrolateral PAG stimulation were compared before andafter the lesions had been made.

A 1 5

Analysis of data

Changes in blood pressure were assessed by measurement ofmean blood pressure in each experiment. The magnitudes ofbaroreflex vagal bradyeardia (BVB) were evaluated bycalculating the reduction in terms of heart beats during ADNstimulation, from measurements of the response area, the timeintegral of the heart rate. Although the beat to beat interval(R-R interval) has been known to provide a sensitive methodfor analyses of chronotrophic action of the vagus nerve on theheart, this parameter often becomes so large as to scale out theconventional recording range when the ADN is maximallystimulated or when simultaneously stimulated with the PAG,making it impossible to assess the data. Therefore, we usedchanges in heart rate in preference to changes of the R-Rinterval to evaluate the bradycardiac responses to ADNstimulation in the present study. Percentage facilitation ofBVB was defined as

actual response - control responsex

control response

Data were expressed as means + standard deviations (S.D.)and analysed statistically by Student's t test. Pairs of datawere regarded as significantly different when P< 0 05.

A 1.0

A 2-0

A 0-5A 00 P 0-5

Figure 2. Midbrain sites that modulate baroreflex vagal bradycardia when electricallystimulatedThe midbrain was electrically stimulated with 30 #tA for 0 5 ms at 50 Hz. Sites inhibiting baroreflexvagal bradycardia (BVB) by more than 50 % and sites facilitating BVB by more than 100% areindicated by open and filled circles, respectively. Ineffective sites are indicated by dots. CS, superiorcolliculus; GM, medial geniculate body; IP, interpeduncular nucleus; NR, red nucleus; PAG,periaqueductal grey matter; RF, reticular formation.

J. Physiol. 477.1 91


RESULTSEffects of electrical stimulation of theventrolateral PAG on baroreflexesWhen the ventrolateral PAG was electrically stimulated,blood pressure fell (from 108-0 + 19 2 to 88'5 + 20-9 mmHg,P < 0 001, n = 34 rats) with a slight decrease in heart rate(from 307 4 + 30 4 to 287-9 + 32-2 beats min-1). Duringventrolateral PAG stimulation, ADN-induced baroreflexvagal bradyeardia (BVB) was markedly augmented(Fig. IB). The magnitude of BVB facilitation due toventrolateral PAG stimulation ranged from 100 to 858 %,the mean value being 328-7 % above the control response.The effects of ventrolateral PAG stimulation on bloodpressure, heart rate and BVB contrasted with those ofdorsolateral PAG stimulation (Fig. 1A).

In order to localize precisely the midbrain sites thatmodulate BVB on stimulation, tracking studies wereconducted in ten rats. The midbrain was systematicallytracked and stimulated electrically (30 ,uA, 0 5 ms, 50 Hz)at 0 25-0 5 mm consecutive steps. In agreement with ourprevious report (Nosaka et al. 1993), stimulation of thedorsolateral PAG inhibited BVB induced by ADN

A

stimulation. In addition, these BVB inhibitory sites werefound to be located in the rostral two-thirds of thedorsolateral PAG. By contrast, stimulation of the caudalhalf of the ventrolateral PAG facilitated BVB (Fig. 2).Facilitation of BVB was also obtained from two otherareas; the central tegmental field and the tegmentum nearthe decussation of the brachium conjunctivum.

Effects of chemical stimulation of theventrolateral PAG on baroreflexesIn order to examine whether or not the facilitatory effect ofthe ventrolateral PAG on BVB is due to activation ofneuronal cell bodies, chemical stimulation with DLH of themidbrain of seventeen rats was performed. Followingadministration of DLH into the ventrolateral PAG therewas a fall in blood pressure (from 98-7 + 15-4 to863 + 13-1 mmHg, P< 0-001, n =13 rats), a decrease inheart rate (from 320-4 + 24-5 to 310-2 + 32-5 beats min')and BVB was significantly facilitated (percentagefacilitation, 243 + 224 %; Fig. 3B). Figure 3 shows thesharply contrasting effects of chemical stimulation of thedorso- and ventrolateral PAG of a single rat, in terms ofblood pressure, heart rate and BVB. Locations of sites that

b

ABP >_ _W E150F )

-100 EIABP ~~~~~~~~~~~~~~~~~~EABP

- ~~~~~~~~~~~~~~~~50E

HR

ADN

axitC

ADN

PAG, DLH

C

Cs

1

PAG , 1,

*i: ,I

1:l' _,

:.B

RF

Ba b 150

ABP _ __ t1F EM*000 E~~~~~~~~~~E5

HR

ADN

ADNPAG, DLH

400

FE_ -200 EDADN

CZ

10 s

Figure 3. Effects of chemical stimulation of the dorsal and ventrolateral PAG on the baroreflexA a, control baroreflex provoked by aortic depressor nerve (ADN) stimulation (8 V, 0 5 ms, 50 Hz); A b,ADN response following DLH microinjection (6 jug in 100 nl) into the dorsal PAG; A c, controlbaroreflex 15 min after the injection. Ba, control baroreflex provoked by ADN stimulation (8 V, 0.5 ms,25 Hz); Bb, ADN response following DLH microinjection (6 jug in 100 nl) into the ventrolateral PAG;Bc, control ADN response 15 min after the injection. C, sites of injection. A-C, data from one and thesame rat. Periods of microinjection and ADN stimulation are indicated by horizontal bars.

92 J. Physiol. 477.1

Baroreflex facilitation by the ventrolateral PAG

A 1-5 A 1-0

A 2-0

A 0.5 A 0-0 P 0-5

Figure 4. Midbrain sites that modulate baroreflex vagal bradyeardia when chemicallystimulatedThe midbrain was stimulated by microinjection of DLH (6 jug in 100 nl). Sites inhibiting baroreflexvagal bradyeardia (BVB) by more than 50% and sites facilitating BVB by more than 50 % are

indicated by open and filled circles, respectively. Dots denote ineffective sites.

8 200

cJ0

Ca

min 100co

0

0 20 40 60Control BVB (%)

80 100

Figure 5. Relationship between the magnitude of control baroreflex vagal bradyeardia and thepotency of ventrolateral PAG facilitationThe magnitude of control baroreflex vagal bradycardia (BVB) is expressed as a percentage of themaximum response; the potency of ventrolateral PAG facilitation of BVB is expressed as a percentageabove the control BVB.



modulated BVB on chemical stimulation are depicted inFig. 4. Similar to the result of the tracking study describedabove, BVB inhibitory and facilitatory sites were shown tobe located in the rostral part of the dorsolateral PAG andthe caudal part of the ventrolateral PAG, respectively.Those two regions in the midbrain outside the PAG, whichwere shown to produce BVB facilitation upon electricalstimulation, had no effect when chemically stimulated.

Relationship between the magnitude of thecontrol BVB and the facilitatory effect of theventrolateral PAGAs we reported previously (Nosaka et al. 1993), the BVB ishardly influenced by ventrolateral PAG stimulation when

A B

a b

HR

ADNADN

PAG

C

a maximum ADN response is used as the control. In orderto examine the relationship between the facilitatory effectof the ventrolateral PAG and the magnitude of controlbradycardiac responses, the effects of ventrolateral PAGstimulation were examined on control responses ofdifferent magnitudes in four rats. For this purpose, theADN was stimulated at 8 V with frequencies that rangedfrom 7 to 50 Hz. Typical data from a single rat are depictedin Fig. 5. The facilitatory effect of the ventrolateral PAGon BVB was most evident when the control BVB wasreduced to about one-fifth of the maximum response. Bycontrast, the BVB was no longer potentiated when themaximum response was used as the control. This resultindicates that there is a convergence between ADN-

a

ADN

b

ADN

PAG

10 s

150

_-100I

-50 E

0

-400

-200 E,

- A

-0

a 30IEE

Ico 20

70 80 90 100 110

Blood pressure (mmHg)

Figure 6. Effects of electrical stimulation of the ventrolateral PAG on aortic depressor nerve-induced hypotension in vagotomized ratsA a and Ba, control baroreflex before and after vagotomy, respectively; panels A b and Bb, effects ofventrolateral PAG stimulation (35 ,uA) on baroreflex before and after vagotomy, respectively. Periodsof aortic depressor nerve (ADN) stimulation and ventrolateral PAG stimulation are indicated byhorizontal bars. C, relationship between the magnitude of baroreflex hypotension (BHT) and the initialblood pressure after vagotomy. 0, BHT during ventrolateral PAG stimulation; 0, control BHT atdifferent initial blood pressures.

kmwo

J. Physiol. 477.194

ABP

Barorefex facilitation by the ventrolateral PAG

activated and ventrolateral PAG-activated neuronessomewhere in the baroreflex arc.

Effects of ventrolateral PAG stimulation onbaroreflex hypotension in vagotomized ratsIn order to determine whether the ventrolateral PAGmodulates the vascular component of the arterialbaroreflex, animals were bilaterally vagotomized so thatchanges in blood pressure could be attributed entirely tochanges in sympathetic nerve activities. Since the baselineblood pressure was changed during PAG stimulation,baroreflex hypotension before and during PAG stimulationcould not be compared simply by their magnitudes. Forprecise evaluation, a method that we described previously(Nosaka et al. 1993) was employed. In each rat, themagnitude of the decrease in blood pressure due to ADNstimulation was measured at several different initialresting blood pressure levels. The magnitude was plottedagainst the resting blood pressure and a straight line wasobtained (Fig. 6C). Thus, the expected magnitude of ADN-induced reflex hypotension could be obtained from thestraight line for any given blood pressure (for details of thismethod, see Nosaka et al. 1993).

In seven rats, a stimulating electrode was positioned inthe ventrolateral PAG in which stimulation provoked

Ab

marked facilitation of BVB. After bilateral vagotomy,stimulation at the same site produced hypotension (from102-9 + 11-3 to 87-1 + 14 6 mmHg, P< 005) with a minimalchange in heart rate (from 328-6 + 22-2 to 326-9 + 21-8beats min'). During the stimulation, reflex hypotensionwas slightly but significantly facilitated (12-9 + 9 7 % abovethe expected magnitude, P< 005, Fig. 6). Likewise, aslight facilitation of baroreflex hypotension was observedfollowing DLH injection into the ventrolateral PAG(percentage facilitation, 90+ 13-7 %, n =6; Fig. 7). Themagnitude of this facilitation of reflex hypotension bychemical stimulation of the ventrolateral PAG was not,however, statistically significant.

Descending pathway of ventrolateral PAGfacilitation ofBVBThe next experiments were conducted to localize thedescending pathway of ventrolateral PAG effects on thecardiovascular functions. Efferent projections from theventrolateral PAG to the ventromedial medulla includingthe nucleus raphe magnus (NRM) have been establishedanatomically (Beitz, Shepard & Wells, 1983; Fardin,Oliveras & Besson, 1984; Meller & Dennis, 1991). It seemspossible that the cardiovascular effects of the ventrolateralPAG are mediated by the NRM since stimulation of this

BaaAB

ABP _ _

b[150 cm

L E50

PAG

ADN

80 90 100Blood pressure (mmHg)

ADN PAG

D

400'FCE

ADN - 200 E

10os I

110

Figure 7. Effects of chemical stimulation of the ventrolateral PAG on aortic depressor nerve-

induced baroreflex hypotension in vagotomized ratsAa and Ba, responses to aortic depressor nerve (ADN) stimulation before and after vagotomy,respectively; Ab and Bb, ADN responses following DLH (6 #sg in 100 nl) injection into the ventrolateralPAG. Periods of microinjection and ADN stimulation are indicated by horizontal bars. C, relationshipbetween the magnitude of baroreflex hypotension (BHT) and the initial blood pressure level. 0, BHTfollowing chemical stimulation of the ventrolateral PAG; 0, control BHT at different initial bloodpressures. D, site of DLH injection.

HR X

ADN

C

a 30-IE

- 20]m

10 4-_70



P 2-8P20 P2-4....

TSC

P3-2 P3-6

Figure 8. Sites in the ventromedial medulla and the pons that modulate baroreflex vagalbradyeardia when electrically stimulatedThe ventromedial medulla was electrically stimulated with 30 AA for 0 5 ms at 50 Hz. Sites thatinhibited baroreflex vagal bradyeardia (BVB) by more than 50% are indicated by 0. Sites thatfacilitated BVB by more than 100 % are indicated by *. Ineffective sites are indicated by dots. BC,brachium conjunctivum; FLM, medial longitudinal fasciculus; LM, medial lemniscus; NRM, nucleusraphe magnus; NTST, nucleus of the spinal tract of the trigeminal nerve; V, trigeminal nerve; VII,facial nucleus.

a b c d 150

ABP ,_0L E

L 50

150

MBP s100 EE

400

IC:200

NRM

ADN ADN ADN

NRM10 s

Figure 9. Effects of electrical stimulation of the nucleus raphe magnus on the arterial baroreflexa and d, control baroreflexes provoked by aortic depressor nerve (ADN) stimulation (8V, 0 5 ms, 17 Hz).b, effects of electrical (30 ptA, 0'5 ms, 50 Hz) stimulation of the nucleus raphe magnus (NRM) on bloodpressure and heart rate. c, effects of NRM stimulation on baroreflex. Periods of ADN and NRMstimulation are indicated by horizontal bars.

J. Physiol. 477.196


nucleus has been reported to produce hypotension andbradyeardia in a similar manner (Henry & Calaresu, 1974;Adair, Hamilton, Scappaticci, Helke & Gillis, 1977; Yen,Blum & Spath, 1983). To examine this hypothesis, we firstconducted a tracking study in this region, with four rats.The ventromedial medulla was systematically exploredwith electrical stimulation (30 1sA, 0 5 ms and 50 Hz) andBVB modulatory sites were mapped. Results from a singlerat are shown in Fig. 8. Sites that facilitated the BVB were

found in and around the NRM. When the ventral part ofthe NRM was electrically stimulated, blood pressure fell(from 1117 + 94 to 82 5 + 19 9 mmHg, P< 0 05, n =8rats), there was a decrease in heart rate (from 309 4 + 31-1to 289-0 + 37-0 beats min') and BVB was facilitated(percentage facilitation, 234 + 132 %, P< 0 005; Fig. 9).This BVB facilitation by the NRM was reproduced byselective cell body activation in this region. DLH (6 jug in100 nl) administered into the NRM produced vagalbradycardia (from 325-7 + 24 3 to 312-4 + 26-1 beats min',n =7 rats) and BVB facilitation (percentage facilitation,328 + 170 % P< 0 005; Fig. 10). Blood pressure responses

to chemical stimulation of the NRM were inconsistent: a

small increase in three rats, a small decrease in two and no

change in two.In spite of the inconsistent blood pressure changes, these

findings indicated that there are cell bodies in the NRMthat are responsible for cardiovascular effects similar tothose of the ventrolateral PAG stimulation. Therefore, we

examined whether or not the NRM is involved in thecardiovascular effects of ventrolateral PAG stimulation inour subsequent experiments.

The tip of the electrode was positioned in the NRM at a

site at which stimulation provoked marked BVBfacilitation, and an electrolytic lesion was made in thisregion by passing a cathodal direct current (DC) throughthe same electrode. The cardiovascular effects ofventrolateral PAG stimulation were compared before andafter the destruction. We found that BVB facilitation bythe ventrolateral PAG was almost abolished after thedestruction (percentage facilitation, 350 + 196% before;1-7 + 4 0% after, n = 6, Fig. 11). Moreover, the decrease inblood pressure (14-5 + 12-5 mmHg) elicited by ventrolateralPAG stimulation was changed into a small increase(3-3 + 6-0 mmHg) after the destruction. Likewise, thebradycardiac response to ventrolateral PAG stimulation(17-5 + 111 beats min') was almost abolished after thedestruction (0-8 + 4*9 beats min').

Selective destruction of cell bodies in the NRM regionwas also effective in eliminating cardiovascular effects dueto the ventrolateral PAG. Following microinjection of KAinto the NRM region, BVB facilitation was reduced from383 + 226 % to 21 + 35 %, n= 9 (Fig. 12). Depressor andbradycardiac responses to ventrolateral PAG stimulationwere also almost abolished after the destruction of the cellbodies (from 24-7 + 14-1 to -1 1 + 7-4 mmHg and from23-9 + 15-0 to 4'0 + 7-3 beats min'1, respectively). Resting

Aa

ABP

MBP

HR

ADN

b

B

Figure 10. Effects of chemical stimulation of the nucleus raphe magnus on arterial baroreflexA a, control responses of the aortic depressor nerve (ADN); A b, ADN response following DLH (6 ,ug in100 nl) injection into the nucleus raphe magnus (NRM); A c, control ADN response 20 min after theinjection. Periods of microinjection and ADN stimulation are indicated by horizontal bars. B, site ofDLH injection. BC, brachium conjunctivum; CIF, inferior colliculus; CS, superior colliculus; FLM,medial longitudinal fasciculus; LM, medial lemniscus; PAG, midbrain periaqueductal grey matter;VII, facial nucleus.

-150

~~ 9 0~~50 E

-150

\ 100 I)

Koo E

r400

__ I~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~- 200E_,~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~.--

NRM, DLH ADN (a

lOsADN 10 s


.


blood pressure elevated after KA injection into the NRMregion as shown in Fig. 12 (110-2 + 15-7 before and136-1 + 14-1 mmHg after the injection, P< 0 005). Lesionsthat were effective in attenuating the cardiovasculareffects of the ventrolateral PAG extended from the rostralpart to the caudal part of the NRM, but lesions outside theNRM lateral or dorsal to it were without effect (n = 5).

Additionally, we examined whether the inhibitory effectof the dorsal PAG on BVB is mediated by the NRM regionas well. In four rats, stimulating electrodes were positionedseparately in the dorso- and ventrolateral PAG, in whichstimulation provoked inhibition and facilitation of BVB,respectively. After KA injection into the NRM region,BVB facilitation due to the ventrolateral PAG wasremarkably reduced (from 332 + 184 to 25 + 11 %), whereasBVB inhibition by the dorsal PAG was only slightlyaffected (from 67f5 + 29 to 676 + 23 %). Likewise,electrolytic lesions of the NRM region were almost withouteffect on dorsal PAG-induced BVB inhibition. The extentof the inhibition was 56f3 + 11P9 % before, and 58-8 + 5-4 %after the destruction (n = 4).

Aa

MBP

HR

ADN -ADN

PAG

C

b

DISCUSSION

The present study confirmed the previous findings that thedorso- and ventrolateral PAG have opposing effects onblood pressure and heart rate. In addition, this studyrevealed for the first time that these two regions also havesharply contrasting effects on the arterial baroreflex, thatis, the dorsolateral PAG inhibits the baroreflex while theventrolateral PAG facilitates it. As we discussed previouslyelsewhere (Nosaka et al. 1993), the physiological significanceof arterial baroreflex suppression during dorsal PAGactivation is probably related to full-scale expression of theconcomitant defence reaction. Cardiovascular componentsof the defence reaction include hypertension, tachycardiaand redistribution of cardiac output from the viscera to theskeletal muscles, which support augmented behaviouralactivities of animals. If the baroreflex functions to its fullextent during the defence reaction, hypertension shouldcause a decrease rather than an increase in cardiac outputwhich is needed to supply the dilated vessels in thehindlimbs. Furthermore, the arterial baroreflex provokes,

Ba b 150

-100 I

-50 E

50

-400

- 200 co,i:.

ADN PAG10 s

D

Figure 11. Effects of electrolytic destruction of the nucleus raphe magnus on ventrolateral PAGfacilitation of baroreflexA a and Ba, aortic depressor nerve (ADN) responses before and after electrolytic destruction of thenucleus raphe magnus (NRM), respectively; A b and B b, effects of ventrolateral PAG stimulation onthe baroreflex before and after the electrolytic destruction, respectively. Periods of ADN stimulationand ventrolateral PAG stimulation are indicated by horizontal bars. C, site of stimulation. D, extent ofthe lesion in the NRM. FLM, medial longitudinal fasciculus; LM, medial lemniscus; NTST, nucleus ofthe spinal tract of the trigeminal nerve; VII, facial nucleus.

J. Physiol. 477.198


in addition to the cardiovascular effects, inhibition ofrespiration, a decrease in muscle tone and synchronizationof the EEG (Heymans & Neil, 1958), all of which mayinterfere with the full expression of the defence reaction.Thus, arterial baroreflex suppression is a prerequisite forthe defence reaction to develop to the required extent.

In contrast to the dynamic cardiovascular changesproduced by the dorsolateral PAG, ventrolateral PAGstimulation elicited hypotension, bradycardia and markedbaroreflex facilitation. Although the physiologicalsignificance of the cardiovascular effects of the ventro-lateral PAG has not been established, it should beemphasized that all the effects are the opposite of thoseassociated with the cardiovascular defence reactionproduced by the dorsal PAG. It is possible that theventrolateral PAG plays a role in preventing excessivecardiovascular changes due to activation of the dorsalPAG, or in facilitating recovery from energy-consuming,ergotrophic states. Indeed, ventrolateral PAG stimulation

A

a b

has been demonstrated to inhibit the cardiovasculardefence response provoked by the dorsal PAG (Lovick,1992). With respect to baroreflex facilitation, many of itsconsequences, including sympathoinhibition, para-sympathoexcitation, reduction of muscle tone andsynchronization of the EEG (Heymans & Neil, 1958) areconsidered to be mostly advantageous in terms ofrecuperation after the defence reaction. Alternatively, it ispossible to speculate that the ventrolateral PAGconstitutes a crucial site in the forebrain mechanismreinforcing the basic medullary baroreflex function. Such aview has been proposed for the preoptic-anteriorhypothalamic area (Hilton & Spyer, 1971; Spyer, 1972). It isnoteworthy that anatomical connection between thepreoptic-anterior hypothalamic area and the ventrolateralPAG has been delineated (Swanson, 1976).The present study reveals that the baroreflex

facilitation, as well as the hypotension and bradycardia,produced by ventrolateral PAG stimulation is mediated by

Ba

MBP

b-150

100 EE

HR

ADN ADN

PAG

C

ADNADN

PAG

r400

- 200 E

-0

10 s

Figure 12. Effects of selective destruction of neurones in the nucleus raphe magnus onventrolateral PAG facilitation of baroreflexA a and Ba, control responses of aortic depressor nerve (ADN) stimulation before and after an injectionof kainic acid (KA, 1 jug in 500 nl) into the nucleus raphe magnus (NRM) region, respectively; A b andBb, ADN responses during ventrolateral PAG stimulation before and after KA injection, respectively.Periods of ADN and ventrolateral PAG stimulations are indicated by horizontal bars. C, a micrographshowing site ofKA injection confirmed by histochemical demonstration of HRP.


100 K. Inui, S. Murase and S. Nosaka J. Physiol. 477.1

the NRM. There are some results in the literature that arecompatible with ours. Projections from the ventrolateralPAG to the NRM in rats have been establishedanatomically (Gallager & Pert, 1978; Beitz et al. 1983;Fardin et al. 1984). Shah & Dostrovsky (1980) showed thatneurones in the PAG are excited antidromically by NRMstimulation, a result that indicates that there is a directprojection from the PAG to the NRM. A non-NMDA,excitatory amino acid mechanism involved in synaptictransmission from the ventrolateral PAG to the NRM hasalso been demonstrated (Wiklund et al. 1988). Further-more, the NRM has been reported to be involved inanalgesic effects elicited by the ventrolateral PAG (Prieto,Cannon & Liebeskind, 1983). Our finding that activation ofneurones in the NRM provoked baroreflex facilitation andvagal bradycardia is also in harmony with the hypothesisthat the NRM is a relay station of the ventrolateral PAGin terms of cardiovascular regulation.

Stimulation of the NRM produces hypotension,bradycardia (Henry & Calaresu, 1974; Yen et al. 1983),baroreflex facilitation (our observation), inhibition of therespiratory system (Lalley, 1986) and a decrease in muscletone (Lai & Siegel, 1990). All these effects are mimicked byventrolateral PAG stimulation, while being the opposite ofthe defence reaction elicited by the dorsolateal PAG.According to a very recent observation, activation ofneurones in the medullary raphe nucleus inhibits activitiesof neurones in the PAG in rats (Lovick, 1993b). Thesefindings may lead to the speculation that the ventrolateralPAG inhibits the defence reaction due to the dorsal PAG(Lovick, 1992), at least in part by way of a long-looppathway that involves the NRM. It may not be tooextreme to suggest that the ventrolateral PAG can exertits cardiovascular influence by suppressing a potentiallyinterfering action of the dorsal PAG via this circuitry.

Although the target site(s) of ventrolateral PAGfacilitation of the arterial baroreflex remains unclear fromour study, interesting suggestions regarding this issue havebeen made in the recent literature. In cats, injection ofserotonin receptor agonists into the nucleus ambiguus (NA)region has been demonstrated to elicit vagal bradycardia(Izzo, Jordan & Ramage, 1988). Furthermore, evidence wasrecently reported that cardiac vagal preganglionic neuronesreceive synaptic inputs from serotonin neurones in rats (Izzo,Deuchars & Spyer, 1993). With respect to sympatheticactivities, serotonin receptor agonists applied in the rostralventrolateral medulla (RVLM) produce hypotension(Mandal, Zhong, Kellar & Gillis, 1990). Since the NRM isknown to contain abundant serotonin neurones and both theRVLM and the NA region are involved in arterialbaroreflexes, it is possible that the ventrolateral PAG exertsits cardiovascular effects, including baroreflex facilitation, inthe RVLM and the NA region via the serotonin neurones inthe NRM. Although the nucleus tractus solitarii (NTS) isalso a potential candidate for mediation of the

ventrolateral PAG effects on the cardiovascular system,our results imply that, baroreflex facilitation occurs athigher-order structures in the baroreflex arc than the NTS,the site of primary terminations of baroreceptor afferents.If this baroreflex facilitation occurs at the NTS, it would bereasonable to expect that both the sympathetic andparasympathetic components of the baroreflex would beequally affected. In fact, however, our results showed thatthe ventrolateral PAG strongly facilitates the cardiaccomponent of the baroreflex, while the vascular componentis only slightly affected.

There are some conditions under which arterialbaroreflexes do not function to their full extent. Forexample, arterial baroreflexes are suppressed duringstimulation of the hypothalamus (Gebber & Snyder, 1970;Humphreys, Joels & McAllen, 1971; Coote, Hilton & Perez-Gonzalez, 1979; Nosaka, Nakase & Murata, 1989), duringnoxious somatosensory activation (Kumada, Nogami &Sagawa, 1975; Nosaka & Murata, 1989) and duringviscerosensory receptor activation (Nosaka et al. 1991). Bycontrast, however, facilitatory modulation of this reflexhas been less frequently described. The present studyprovides evidence that there is a centrally mediated,powerful facilitation of the arterial baroreflex. Moreimportantly, the results of the present study indicate thatthe PAG is an important structure that subservesbaroreflex modulation, exerting both inhibitory andfacilitatory effects on the arterial baroreflex, in particularon its cardiac component.

REFERENCESADAIR, J. R., HAMILTON, B. L., SCAPPATICCI, K. A., HELKE, C. J.& GILLIS, R. A. (1977). Cardiovascular response to electricalstimulation of the medullary raphe area of the cat. BrainResearch 128, 141-145.

BANDLER, R. & CARRIVE, P. (1988). Integrated defence reactionelicited by excitatory amino acid microinjection in the midbrainperiaqueductal grey region of the unrestrained cat. BrainResearch 439, 95-106.

BANDLER, R., CARRIVE, P. & ZHANG, S. P. (1991). Integration ofsomatic and autonomic reactions within the midbrainperiaqueductal grey: viscerotopic, somatotopic and functionalorganization. Progress in Brain Research 87, 269-305.

BEITZ, A. J., SHEPARD, R. D. & WELLS, W. E. (1983). Theperiaqueductal gray-raphe magnus projection containssomatostatin, neurotensin and serotonin but notcholecystokinin. Brain Research 261,132-137.

CARRIVE, P., BANDLER, R. & DAMPNEY, R. A. L. (1989). Somaticand autonomic integration in the midbrain of theunanesthetized decerebrate cat: a distinctive pattern evoked byexcitation of neurones in the subtentorial portion of themidbrain periaqueductal grey. Brain Research 483, 251-258.

COOTE, J. H., HILTON, S. M. & PEREZ-GONZALEZ, J. F. (1979).Inhibition of the baroreceptor reflex on stimulation in the brainstem defence centre. Journal of Physiology 228, 549-560.

FARDIN, V., OLIVERAS, J. L. & BESSON, J. M. (1984). Projectionsfrom the periaqueductal gray matter to the B3 cellular area(nucleus raphe magnus and nucleus reticularisparagigantocellularis) as revealed by the retrograde transportof horseradish peroxidase in the rat. Journal of ComparativeNeurology 223, 483-500.

J. Phlysiol.477.1 Baroreflex facilitation by the ventrolateral PAG 101

GALLAG(ER, D. WV. & PERT, A. (1978). Afferents to brain stem nuclei(brain stem raphe, nucleus reticularis pontis caudalis andnucleus gigantocellularis) in the rat as demonstrated bymicroiontophoretically applied horseradish peroxidase. BrainResearch 144, 257-275.

GEBBER, G. 1L. & SNYDER, D. W. (1970). Hypothalamic control ofbaroreceptor reflexes. A merican2 Journ(al of Physiology 218,124-131.

HENRY, J. L. & CALARESU, F. R. (1974). Excitatory and inhibitoryinputs from medullary nuclei projecting to spinalcardioacceleiatoirv neurons in the cat. Experiinental BrainResearch 20, 485-504.

HEYMANS, C. & NEIL, E. (1958). Baroreceptor reflexes other thancirculatory. In Refiexogenic Areas of the Cardiovascular System,pp. 95-100. Little, Brown & Co., Boston.

HILTON, 8. M. & REDFERN, WV. S. (1986). A search for brain stemcell groups integrating the defence reaction in the rat. Journalof Physiology 378, 213-228.

HILTON, S. MI. & SPYER, K. AM. (1971). Participation of the anteriorhypothalamus in the baroreceptor reflex. Journal of Physiology218, 271-293.

HOlISTEC E, C. (1991). Descending pathways fIom theperiaqueductal gray and adjacent areas. In The MidbrainPeriaqueductal Gray Mlatter: Functionial, A natomnical andNTeurochemical Organizatiotn, ed. DEPAULIS, A. & BANDLER, R.,pp. 239-265. Plenum Press, New York.

HUMPHREYS, P. WV., JOELS, N. & McALLEN, R. SM. (1971).AModification of the reflex response to stimulation of carotidsinus baroreceptors during and following stimulation of thehypothalamic defence area in the cat. Journal of Physiology216, 461-482.

Izzo, P. N., DEUCHARS, J. & SPYER, K. AM. (1993). Localization ofcardiac vagal preganglionic Inotoneurones in the rat:immunocytochemical evidence of synaptic inputs containing5-hydroxytryptamine. Journal of Comparative Neurology 327,572-583.

Izzo, P. N., JORDAN, D. & RAMAGE, A. G. (1988). Anatonmical andpharmacological evidence supporting the involvement ofserotonin in the central control of cardiac vagal motoneuronesin the anaesthetized cat. Journal of Physiology 406, 19P.

JONES, R. O., KIRKMAN, E. & LITTLE, R. A. (1990). Theinvolvement of the midbrain periaqueductal grey in thecardiovascular response to injury in the conscious andanaesthetized rat. Experimnental Physiology 75, 483-495.

JURGENS, U. (1991). Neurochemical study of PAG control of vocalbehavior. In The Mlidbrazin Periaqueductal Gray Mlatter:Functionial, A natomical and Neurochemical Organization, ed.DEPAULIS, A. & BANDLER, R., pp. 11-21. Plenum Pr ess, NewYork.

KUAMADA, M., NOGANI1, K. & SAGAWA, K. (1975). Modulation ofcarotid sinus baroreceptor reflex by sciatic nerve stimulation.A merican7 Journal of Physiology 228, 1535-1541.

LAI, Y. Y. & SIEGEL, J. AM. (1990). Cardiovascular and muscle tonechanges produced by microinjection of cholinergic andglutamatergic agonists in dorsolateiral pons ancl medialmedulla. Braint Research 514, 27-36.

LALLEY, P. AM. (1986). Responses of phrenic motoneurones of thecat to stimnulation of medullary raphe nuclei. Journal ofPhysiology 380, 349-371.

LoVICK, T. A. (1992). Inhibitory modulation of the cardiovasculardefence response by the ventrolateral periaqueductal grexmatter in rats. Experimental Brain Research 89, 133-139.

LoVICK, T. A. (1993a). Integrated activity of cardiovascular andpain regulatory systems: role in adaptive behaviouralresponses. Progress in Neurobiology 40, 631-644.

LOVICK, T. A. (1993b). Serotonergic influence fiom nucleus rapheobscurus on neurones in the periaquedcuctal grey matter in therat. Brain Research 606, 92-98.

MIANDAL, A. K., ZHONG, P., KELLAR, K. J. & GILLIS R. A. (1990).Ventrolateral medulla: an important site of action for thehypotensive effect of drugs that activate serotonin-1Areceptors. Journal of Cardiovascular Pharmacology 15, S49-60.

AIELLER, S. T. & DENNIS, B. J. (1991). Efferent projections of theperiaqueductal gray in the rabbit. Neuroscience 40,191-216.

NOSAKA, S., AIURASE, S. & MAlURATA, K. (1991). Arterial baroreflexinhibition by gastric distension in rats: mediation by splanchnicafferents. American Journal of Physiology 260, R985-994.

NOSAKA, S. & AIURATA, K. (1989). Somatosensory inhibition ofvagal baroreflex bradycardia: afferent nervous mechanisms.American Journal of Physiology 257, R829-838.

NOSAKA, S., AIURATA, K., INUI, K. & AIURASE, S. (1993). Arterialbaroreflex inhibition by midbrain periaqueductal grey inanaesthetized rats. Pfliigers Archiv 424, 266-275.

NOSAKA, S., NAKASE, N. & AIURATA, K. (1989). Somatosensory andhypothalamic inhibitions of baroreflex vagal bradyeardia inrats. Pfliigers Archiv 413, 656-666.

PRIETO, G. J., CANNON, J. T. & LIEBESKIND, J. C. (1983). N. raphemagnus lesions disrupt stimulation-produced analgesia firomventral but not dorsal midbrain areas in the rat. Brain Research261, 53-57.

SAKUMA, Y. & PFAFF, D. MW. (1979). Facilitation of femalereproductive behavior from mesencephalic central gray in therat. American Journal of Physiology 237, R278-284.

SAPRU, H. N. & KRIEGER, A. J. (1977). Carotid and aorticchemoreceptor function in the rat. Journal of AppliedPhysiology 42, 344-348.

SJAH, Y. & DoSTROVSKY, J. 0. (1980). Electrophysiologicalevidence for a projection of the periaqueductal gray matter tonucleus raphe magnus in cat and rat. Brain Research 193,534-538.

SPvER, K. Al. (1972). Baroreceptor sensitive neurones in theanterior hypothalamus of the cat. Journal of Physiology 224,245-257.

SWANSON, L. WV. (1976). An autoradiographic study of the efferentconnections of the preoptic region in the rat. Journal ofComparative Neurology 167, 227-256.

V'EENINCG, J., BUMIA, P., HORST, G. J. T., ROELING, T. A. P.,LUITEN, P. G. Al. & NIEUWENHUYS, R. (1991). Hypothalamicprojections to the PAG in the rat: topographical, immuno-electronmicroscopical and functional aspects. In The MIidbrainPeriaqueductal Gray Mlatter: Functional, Anatomical andNeurochemical Organizationi, ed. DEPAULIS, A. & BANDLER, R.,pp. 387-415. Plenum Press, New York.

WVIKLUND, L., BEHZADI, G. & KALEN, P., HEADLEY, P. Al.,NICOLOPOULOS, L. S., PARSONS, C. G. & WEST, D. C. (1988).Autoradiographic and electrophysiological evidence forexcitatory amino acid transmission in the periaqueductal grayprojection to nucleus raphe magnus in the rat. NeuroscienceLetters 93, 158-163.

YEN, C. T., BLUMI, P. S. & SPATH, J. A. (1983). Control ofcardiovascular function by electrical stimulation within themedullary raphe region of the cat. Experimental Neurology 79,666-679.

Received 22 July 1993; accepted 15 October 1994.

facilitation of the arterial baroreflex by the ventrolateral part of the

Documents