Brain Research 1059 (2005) 197 – 202

www.elsevier.com/locate/brainres

Research Report

Uptake blockade of endocannabinoids in the NTS modulates

baroreflex-evoked sympathoinhibition
Daniel T. Brozoski, Caron Dean, Francis A. Hopp, Jeanne L. Seagard*
Zablocki Department of Veterans Affairs Medical Center and Departments of Anesthesiology and Physiology, Medical College of Wisconsin,
Research Service 151, 5000 W. National Avenue, Milwaukee, WI 53295, USA

Accepted 19 August 2005

Available online 9 September 2005

Abstract

Previous studies supporting a possible physiological role for an endogenous cannabinoid, arachidonylethanolamide (AEA, anandamide),
showed a significant increase in AEA content in the nucleus tractus solitarius (NTS) after an increase in blood pressure (BP) and prolonged
baroreflex inhibition of renal sympathetic nerve activity (RSNA) after exogenous AEA microinjections into the NTS. These results, along
with other studies, support the hypothesis that endogenous AEA can modulate the baroreflex through cannabinoid CB1 receptor activation
within the NTS. This study was performed to characterize the physiological role of endogenously released cannabinoids (endocannabinoids)
in regulating baroreflex control of RSNA through actions in the NTS. Endocannabinoid effects were assessed by measuring the RSNA
baroreflex response to increased pressure after bilateral microinjections of AM404, an endocannabinoid transport inhibitor, into the NTS of
adult male Sprague Dawley rats. AM404 blocks uptake of endocannabinoids and enhances the effects of any endocannabinoids released [M.
Beltramo, et al., Functional role of high-affinity anandamide transport, as revealed by selective inhibition, Science 277 (5329) (1997) 1094 –
1097.] into the NTS. Therefore, it was hypothesized that microinjections of AM404 should exhibit effects similar to microinjections of
exogenous AEA. In this study, AM404 microinjections into the NTS were found to significantly prolong baroreflex inhibition of RSNA
compared to control, similar to effects of exogenous AEA. This effect is thought to result from an increased endocannabinoid presence in the
NTS, leading to prolonged CB1 receptor activation. These results indicate that endocannabinoids released in the NTS have the potential to
modulate baroreflex control at this site in the central baroreflex pathway.
Published by Elsevier B.V.

Theme: Endocrine and autonomic regulation

Topic: Cardiovascular regulation

Keywords: AEA; AM404; Baroreceptor; CB1 receptor; NTS; Sympathetic activity

1. Introduction a pivotal role by receiving afferent baroreceptor signals and

This study was performed to define the possible role that
endogenous cannabinoids (endocannabinoids) play in mod-
ulating barosensitive neurons in the nucleus tractus solitar-
ius (NTS) of rats. The arterial baroreceptor reflex helps
buffer transient variations in blood pressure (BP) via CNS
regulation of the parasympathetic and sympathetic branches
of the autonomic nervous system. Neurons in the NTS play
* Corresponding author. Fax: +1 414 645 6550.
E-mail address: jseagard@mcw.edu (J.L. Seagard).
0006-8993/$ - see front matter. Published by Elsevier B.V.
doi:10.1016/j.brainres.2005.08.030
processing these inputs to determine magnitude, pattern, and
duration of the signal that is then transmitted to distal central
sites. Within the NTS, inhibitory GABA inputs from
interneurons and descending inputs and excitatory gluta-
mate inputs from baroreceptor afferent fibers all contribute
to the net output of the baroreceptive neurons in the NTS.
Work from this laboratory and others suggests that this
neurotransmission can be modulated by cannabinoid (CB1)
receptor activation by endocannabinoids. Endocannabi-
noids such as arachidonylethanolamide (anandamide,
AEA), 2-arachidonyl glycerol (2-AG), and 2-arachidonyl
198 D.T. Brozoski et al. / Brain Research 1059 (2005) 197 – 202
glycerol ether (2-AGE) are endogenous products, synthe-
sized via a multi-enzymatic cascade from phospholipids.
Upon demand, neurons produce endocannabinoids intra-
cellularly, and when released, these endocannabinoids may
activate CB1 receptors. Endocannabinoid deactivation may
occur by degradation by the catabolic enzyme, fatty acid
amide hydrolase (FAAH), and/or removal via an uptake
transporter. N-(4-hydroxyphenyl)-5Z,8Z,11Z,14Z-eicosate-
traenamide (AM404) blocks this endocannabinoid uptake
transporter, yet exhibits little cannabimimetic activity in
vivo [1,6].
Studies by others have shown endocannabinoid modu-
lation of excitatory glutamatergic and inhibitory GABAer-
gic transmission in other regions of the CNS [2,5,7 – 9,
13 –18], primarily in brain slice studies. In this laboratory,
an acute increase in BP in rats was found to increase AEA
content in the NTS when compared to control [11],
indicating that endocannabinoid release could be evoked
by a physiological stimulus. Egertova and Elphick [4] and
preliminary evidence from this laboratory have also shown
immunohistochemical staining for CB1 receptors on fibers
within the NTS. These two findings indicate that the
components necessary for endocannabinoid modulation of
neurotransmission are present in the NTS. This possible role
of endocannabinoids was further supported by functional
studies which showed that microinjection of AEA into the
NTS prolonged baroreflex inhibition of renal sympathetic
nerve activity (RSNA), and this enhanced baroreflex control
was attenuated by SR141716, a CB1 receptor antagonist
[11]. In addition, the effects of CB1 receptor activation were
eliminated by prior microinjection of GABAA receptor
antagonists, suggesting that endocannabinoids alter barore-
flex function by modulating GABAergic transmission.
If endocannabinoids are released and physiologically
modulate baroreflex function, then microinjection of
AM404 into the NTS should prolong the presence of
extracellular AEA and mimic the effects of microinjection
of AEA or other CB1 receptor agonists. In the current study,
microinjection of AM404 into the NTS prolonged barore-
flex inhibition of RSNA in a manner similar to micro-
injection of AEA, suggesting that endogenously released
cannabinoids can contribute to baroreflex control at neurons
in the NTS.
2. Methods
2.1. General methods
The Animal Care and Use Committees at the Medical
College of Wisconsin and the Zablocki Department of
Veterans Affairs Center approved this protocol. The experi-
ments were performed in male Sprague Dawley rats (350 –
390 g) anesthetized with sodium pentobarbital (50 mg/kg,
i.p.). Throughout the experiments, a heating pad (Fine
Science Tools) maintained body temperature at 37 -C.
Catheters were inserted into femoral veins for supplemental
administration of anesthetic and administration of phenyl-
ephrine (PE). Arterial BP was monitored continuously from
a femoral arterial cannula connected via a pressure trans-
ducer (Statham) to a polygraph (Grass Model 7) and
recorded on tape (Vetter PCM recording adapter Model
3000A).
2.2. Microinjection studies
The rat was placed in a stereotaxic frame (Kopf), and
RSNA was recorded using flexible silver wire electrodes
(Medwire, AG5T) positioned on a renal nerve running
parallel to the renal artery, exposed via a retroperitoneal
approach. The electrodes were fixed in position with silastic
gel (Sil Gel) allowing body adjustments of the animal
without disturbing neural recordings. The electrophysiolog-
ical signal was directed to a high impedance differential
preamplifier (gain = 1000; 0.1– 10 kHz passband), followed
by a filter/amplifier (gain up to 400; high and low pass
filtering 10 Hz –3 kHz), and recorded on tape. The amplifier
output was directed to a precision full-wave rectifier and
averaged using a Bessel linear averaging filter (averaging
interval = 100 ms) to obtain an online moving time average.
The averaged RSNA, along with BP, was displayed on a
Grass Model 7 recorder to observe trends during the
experiment.
To allow microinjections into the NTS, the dorsal surface
of the medulla was exposed via an occipital craniotomy. A
four-barreled micropipette was used to microinject test
agents into the NTS using a system designed and
constructed in the laboratory. The volumes of injected
solutions were visually determined by measuring the change
in meniscus height in each barrel using a 50 monocular
microscope with a calibrated graticule (7 nl/div). The first
barrel was filled with d,l-homocysteic acid (DLH) (4 mM),
an excitatory amino acid receptor agonist. DLH (7 nl) was
microinjected to ensure that the micropipette was placed in a
site from which a depressor (decrease in BP of 30.2 T 15.2)
and sympathoinhibitory response could be evoked, indica-
tive of a baroreflex. The second barrel contained AM404
(50 AM), dissolved in aCSF and 0.01% Tocrisolve100, and
the third barrel contained AEA (50 AM), dissolved in aCSF
and 0.01% Tocrisolve100. Tocrisolve100 (Tocris Cookson,
Inc.) is an emulsion of a 1:4 soya oil/water, emulsified with
Pluronic F68. The fourth barrel contained sky blue dye in
aCSF and 0.01% Tocrisolve100 and was used to mark
injection sites. The vehicle of aCSF and 0.01% Tocri-
solve100 has been shown to have no effect on baseline
parameters or the baroreflex [11].
The micropipette tip (30 –50 Am) was initially inserted in
the medulla targeting the caudal region of the NTS at 0.5
mm rostral to calamus scriptorius, 0.5 mm lateral to the
midline and 0.5 mm beneath the dorsal surface, a region that
is known to receive primary afferent input from arterial
baroreceptors. To ensure that the micropipette was placed in
 D.T. Brozoski et al. / Brain Research 1059 (2005) 197 – 202
a site from which a depressor response could be evoked,
DLH (7 nl) was microinjected while BP and RSNA were
monitored. If no depressor and sympathoinhibitory
responses were observed, the micropipette was carefully
repositioned to an adjacent site until a depressor response
greater than 20 mm Hg and associated sympathoinhibitory
response could be evoked. Once established, corresponding
coordinates were utilized on the contralateral side for a
bilateral microinjection into the NTS.
With the micropipette tip positioned at a sympathoinhibi-
tory site, BP and RSNA were allowed to return to control
baseline values, after which a bolus of PE (0.02 mg%; 4 Al,
i.v.) was given to produce a transient increase in BP of
approximately 50 mm Hg (PE pressor test) to activate
barosensitive receptors. This PE dose was used for all
subsequent PE pressor tests to produce repeatable changes
in BP. Responses in BP and RSNA to this control PE pressor
test were recorded and monitored until all parameters
returned to baseline (preinjection) levels. In twelve trials in
ten rats, AM404 (70 Al) was then microinjected bilaterally
into the NTS (45 s), and the PE pressor test was repeated at 1,
3, and 10 min following AM404 microinjection.
After a 20-min recovery period, which was found to be
sufficient to allow effects of AM404 to be eliminated, the
PE pressor test was repeated to obtain a new control and
AEA (70 Al) was microinjected bilaterally into the NTS. The
PE pressor test was again repeated at 1, 3, and 10 min. After
another 20-min recovery, the trials were repeated, but the
order of AM404 and AEA was reversed.
At the conclusion of the protocol, 70 Al of dye/vehicle
was microinjected to mark the injection site. The animal was
then sacrificed, and the medulla was removed and frozen.
Transverse (25 Am) sections through the medulla were cut
and examined microscopically to determine the locations of
microinjections.
2.3. Data analysis and statistics
Data analysis was performed as described in an earlier
study [11]. Analog-to-digital conversion of recorded param-
eters was performed using a computer (model 310 Hewlett-
Packard Co., Palo Alto, CA, USA). Arterial BP and
averaged RSNA were sampled at 20 Hz for each control
and microinjection procedure and stored on disk files for
quantification and statistical analysis. Baseline values of BP
and averaged RSNA were obtained as 30-s averages of each
parameter sampled during the period immediately prior to
PE pressor tests for control and each microinjection and
compared using an analysis of variance. Peak blood
pressure for each PE baroreflex pressor test was compared
using an analysis of variance to verify the repeatability of
the stimuli. To examine the rate of recovery of RSNA
following each baroreflex stimulus, averaged RSNA was
plotted versus time, and nonlinear regression was used to fit
the data to a first-order exponential recovery equation using
methods described previously [10,11]. This analysis pro-
199
vided values for the rate of recovery time constant (TCR) for
RSNA. TCR is an index of the rate at which RSNA recovers
following the baroreflex-induced sympathoinhibition. For
each TCR, RSNA will recover 63.2% of the remaining
distance to the asymptotic value (baseline RSNA), such that
after four TCRs, RSNA recovers to 98.2% of the baseline
level. TCR for each pressor test following microinjection of
either AM404 or AEA was normalized as a percent of
control TCR, and responses for each treatment were
compared using an analysis of variance.
In all cases, normalized values for RSNA and BP for
each procedure were tested and confirmed for normal
distribution. Following one-way analyses of variance,
differences were identified using the Tukey test, with
significance set at P < 0.05. Data are presented as a percent
of control using mean T standard deviation (SD).
3. Results
Histological analysis of microinjections marked by
blue dye confirmed that the bilateral microinjection sites
were located within the commissural and medial sub-
nuclei of the NTS (Fig. 1). Baseline BP and RSNA
remained stable throughout each experiment and were not
significantly altered by microinjections of AM404 or
AEA. Importantly, the pressor response to bolus injec-
tions of PE was not significantly different for any trial,
and the blood pressure increase was reproducible,
averaging 49.5 T 14.3 mm Hg.
Microinjections of either AM404 or AEA resulted in a
significant prolongation of the reflex inhibition of RSNA as
reflected by an increase in TCR when compared to control.
As shown by the curve fits imposed on averaged RSNA
responses to the pressor test in Fig. 2, the length of time to
Fig. 1. Diagram of a transverse section through the dorsomedial medulla
representing levels 0.1 – 0.8 mm caudal to obex indicating the center of the
sites of microinjection of AM404 within the NTS. Nine of 10 micro-
injection sites were recovered. Microinjections were made bilaterally, but
only unilateral sites are shown to allow labeling of the section. AP, area
postrema; CC, central canal; G, nucleus gracilis; NTS, nucleus tractus
solitarius; TS, tractus solitarius; X, dorsal motor nucleus of the vagus nerve.
200 D.T. Brozoski et al. / Brain Research 1059 (2005) 197 – 202
Fig. 2. Effects of NTS microinjections of AM404 or AEA on baroreflex
inhibition and recovery of averaged renal sympathetic nerve activity
(average RSNA) in response to brief blood pressure (BP) increases induced
by bolus injections of phenylephrine (PE pressor test). Curve fits of RSNA
(light gray line) are shown for the control response and 3 min after either
AM404 or AEA microinjection into the NTS. In comparison to the control
response, the exponential fit of the changes in RSNA following micro-
injection of AM404 or AEA into the NTS shows a significant delay in
recovery of RSNA to baseline levels (dark gray line).
recovery to baseline was prolonged following either AM404
or AEA. Summed data presented in Fig. 3 show that
bilateral administration of AM404 significantly increased
the TCR at 1 min, 3 min, and 10 min post-AM404 injections
versus control (Fig. 3, upper panel), with the 3-min period
exhibiting the greatest increase for the time intervals tested.
Summed data for AEA also resulted in a significant
prolongation of the reflex inhibition of RSNA at 3 min,
reflected by an increase in TCR (Fig. 3, lower panel).
4. Discussion
Data from this study support a role for endogenously
released cannabinoids in the modulation of baroreflex
control of sympathetic activity. Prolongation of baroreflex
inhibition of RSNA produced by NTS microinjection of
AM404, an endocannabinoid transport inhibitor, mimicked
the effects of microinjection of the endocannabinoid, AEA.
The action of AM404 prolongs the presence of endocanna-
binoids in synapses, enhancing the action of endogenously
released cannabinoids. This suggests that endocannabinoids
are released in the NTS during periods of acute hypertension
and can lead to modulation of ensuing baroreflex control of
RSNA. In support of this finding, this laboratory previously
found that AEA content in the NTS was increased in rats
subjected to a short period of hypertension [11]. This data
were the first to link AEA production to a known autonomic
function and suggested that some neurons in the NTS were
capable of producing endocannabinoids. Additional support
for endocannabinoid modulation of central baroreflex
control arises from a previous study in which this laboratory
demonstrated that both AEA and AM404 increased the
discharge of baroreceptive neurons in the rat [12]. The
increase in neuronal discharge was prevented by prior
administration of the CB1 receptor antagonist SR141716,
which implicates a role for CB1 receptor activation in the
responses.
In the present study, the prolongation of TCR was
greatest at the 3-min pressor test following microinjection
of either AM404 or AEA. TCR was still significantly
prolonged 10 min following microinjection of AM404.
Since AM404 blocks uptake of endocannabinoids, the
prolonged effect reflects the protracted presence of endo-
cannabinoids in the synapse. In addition, AM404 slows the
elimination of both endocannabinoids (AEA and 2-AG) and
may implicate 2-AG as a second important contributor to
the modulation of the baroreflex in the NTS (Fig. 3).
What was not determined in the present study was the
duration or magnitude of acute hypertension needed to
evoke sufficient endocannabinoid release to modulate
baroreflex function. AM404 artificially allowed endocanna-
Fig. 3. Summed data from twelve trials in ten experiments showing the
effects of microinjection of AM404 (top panel) or AEA (lower panel) into
the NTS on the time constant for the rate of the recovery (TCR) of RSNA
obtained by nonlinear regression for 1, 3, and 10 min following
microinjections. Values are normalized as percents of control TCR for
each microinjection T standard deviation. For reference, the absolute value
of the TCR for control was 12.0 T 4.2 s. AM404 significantly increased
TCR, reflecting a prolonged recovery of RSNA, at 1, 3, and 10 min post-
AM404. AEA significantly increased TCR reflecting a prolonged recovery
of RSNA 3 min post-AEA microinjections. *Significantly different from
control at P < 0.05.
 D.T. Brozoski et al. / Brain Research 1059 (2005) 197 – 202
binoid accumulation in the NTS to an extent that affected
baroreflex control, but it is not clear if enough endocanna-
binoids are released in an amount sufficient to prolong
baroreflex control during a brief change in pressure, or if a
more prolonged change is required. A previous study
suggested that blockade of CB1 receptors with SR141716
could modulate baroreflex function, but this was not studied
in enough detail to draw any conclusions. This question
remains a future area of investigation.
Based on this laboratory’s hypothetical model from an
earlier study [11], CB1 receptor activation via endocanna-
binoids in the NTS may presynaptically attenuate GABA
release from interneurons or descending inputs in the NTS,
leading to disinhibition and enhanced baroreceptive neuro-
nal excitability, prolonging reflex inhibition of RSNA. The
finding that blockade of postsynaptic GABAA receptors by
bicuculline eliminates the endocannabinoid enhancement of
baroreflex inhibition of RSNA supports a role for GABAer-
gic transmission modulation by endocannabinoids. In
addition, activation of barosensitive neurons in the NTS
by AEA was attenuated by prior blockade of GABAA
receptors, again focusing the actions of endocannabinoids
on modulation of GABAergic transmission [12]. AEA has
been found to exert a presynaptic inhibition of GABA
release from neurons in other regions of the CNS
[2,5,7,15,17], and the presence of CB1 receptors on
GABAergic interneurons and their nerve terminals in the
human hippocampus supports the concept that they could
play a presynaptic modulatory role in GABA release [3].
Besides gaining a basic understanding of how the central
baroreflex pathway is controlled, more fully understanding
the baroreflex will allow development of novel therapies for
conditions where the central baroreflex pathway is mal-
functioning. For example, novel pharmacological therapies
for treatment of labile hypertension could include enhancing
the endogenous effects of endocannabinoids. Since short-
term hypertension results in an elevation of AEA content in
the NTS, a therapeutic approach could be the use of AEA
re-uptake or FAAH inhibitors to prolong the action of AEA
and enhance endogenous CB1 receptor activation. The short
effective half-life of exogenous AEA in NTS suggests that
the processes for inactivation would be good therapeutic
targets to enhance the action of the endogenous CB1 system.
This could result in localized CB1 receptor activation
without the psychoactive side effects that would result from
administration of a direct acting CB1 receptor agonist.
Studies examining the effects of chronic exposure to
AM404 could provide an insight into the feasibility of this
type of treatment for labile hypertension. Finally, future
studies should also be directed toward determining the
specific location of CB1 receptors within the subnuclei of
the NTS, including determining co-localization of CB1
receptors on GABAergic neurons, utilizing immunohisto-
chemical techniques. The anatomical relationship of CB1
receptors to GABAergic neurons and baroreceptive neurons
could provide support or direction for new studies.
201
In conclusion, AM404, an endocannabinoid transport
inhibitor, produced an increase in the TCR for baroreflex
inhibition of RSNA. This result indicates that endogenous
cannabinoids are released in the NTS by acute increases in
BP and have the potential to modulate transmission in the
central baroreflex pathway. Based on our hypothetical
model, CB1 receptor activation via endocannabinoids in
the NTS may have presynaptically attenuated GABA
release, leading to enhanced baroreceptive neuronal excit-
ability and prolonged reflex inhibition of RSNA.
Acknowledgments
VA Medical Research Funds, an AHA Grant-in-Aid
Award 0150518 and a MCW New Idea Research Grant
supported this study. The authors would like to acknowl-
edge the excellent technical and histological assistance of
Claudia A. Hermes and Victoria Woyach.
References
[1] M. Beltramo, et al., Functional role of high-affinity anandamide
transport, as revealed by selective inhibition, Science 277 (5329)
(1997) 1094 – 1097.
[2] A.V. Derbenev, T.C. Stuart, B.N. Smith, Cannabinoids suppress
synaptic input to neurones of the rat dorsal motor nucleus of the
vagus nerve, J. Physiol. 559 (Pt. 3) (2004) 923 – 938.
[3] T.P. Dinh, et al., Brain monoglyceride lipase participating in
endocannabinoid inactivation, Proc. Natl. Acad. Sci. U. S. A. 99
(16) (2002) 10819 – 10824.
[4] M. Egertova, M.R. Elphick, Localisation of cannabinoid receptors in
the rat brain using antibodies to the intracellular C-terminal tail of
CB1, J. Comp. Neurol. 422 (2000) 159 – 171.
[5] T.F. Freund, I. Katona, D. Piomelli, Role of endogenous cannabinoids
in synaptic signaling, Physiol. Rev. 83 (3) (2003) 1017 – 1066.
[6] A. Giuffrida, M. Beltramo, D. Piomelli, Mechanisms of endocanna-
binoid inactivation: biochemistry and pharmacology, J. Pharmacol.
Exp. Ther. 298 (1) (2001) 7 – 14.
[7] N. Hajos, et al., Cannabinoids inhibit hippocampal GABAergic
transmission and network oscillations, Eur. J. Neurosci. 12 (9)
(2000) 3239 – 3249.
[8] A.F. Hoffman, C.R. Lupica, Mechanisms of cannabinoid inhibition of
GABA(A) synaptic transmission in the hippocampus, J. Neurosci. 20
(7) (2000) 2470 – 2479.
[9] O.J. Manzoni, J. Bockaert, Cannabinoids inhibit GABAergic synaptic
transmission in mice nucleus accumbens, Eur. J. Pharmacol. 412 (2)
(2001) R3 – R5.
[10] D.J. Rademacher, et al., Microinjection of a cannabinoid receptor
antagonist into the NTS increases baroreflex duration in dogs, Am. J.
Physiol. 284 (2003) H1570 – H1576.
[11] J.L. Seagard, et al., Anandamide content and interaction of endocan-
nabinoid/GABA modulatory effects in the NTS on baroreflex-evoked
sympathoinhibition, Am. J. Physiol.: Heart Circ. Physiol. 286 (3)
(2004) H992 – H1000.
[12] J.L. Seagard, et al., Effects of endocannabinoids on discharge
of baroreceptive NTS neurons, Neurosci. Lett. 381 (3) (2005)
334 – 339.
[13] J.M. Sullivan, Mechanisms of cannabinoid-receptor-mediated inhi-
bition of synaptic transmission in cultured hippocampal pyramidal
neurons, J. Neurophysiol. 82 (3) (1999) 1286 – 1294.
[14] B. Szabo, et al., Inhibition of GABAergic inhibitory postsynaptic
202 D.T. Brozoski et al. / Brain Research 1059 (2005) 197 – 202

currents by cannabinoids in rat corpus striatum, Neuroscience 85 (2)
(1998) 395 – 403.
[15] B. Szabo, et al., Cannabinoids inhibit excitatory neurotransmission in
the substantia nigra pars reticulata, Neuroscience 97 (1) (2000) 89 – 97.
[16] K.A. Takahashi, D.J. Linden, Cannabinoid receptor modulation of
synapses received by cerebellar Purkinje cells, J. Neurophysiol. 83 (3)
(2000) 1167 – 1180.
[17] C.W. Vaughan, I.S. McGregor, M.J. Christie, Cannabinoid receptor
activation inhibits GABAergic neurotransmission in rostral ventro-
medial medulla neurons in vitro, Br. J. Pharmacol. 127 (4) (1999)
935 – 940.
[18] C.W. Vaughan, et al., Actions of cannabinoids on membrane proper-
ties and synaptic transmission in rat periaqueductal gray neurons in
vitro, Mol. Pharmacol. 57 (2) (2000) 288 – 295.