RESEARCH PAPER
Dexmedetomidine and ST-91 analgesia in the
formalin model is mediated by a2A-adrenoceptors:
a mechanism of action distinct from morphine
A Nazarian1,2, CA Christianson1, X-Y Hua1 and TL Yaksh1
1Department of Anesthesiology, University of California-San Diego, La Jolla, CA, USA and 2Department of Pharmaceutical Sciences,
Western University of Health Sciences, Pomona, CA, USA
Background and purpose: Intrathecal administration of a2-adrenoceptor agonists produces potent analgesia. This study
addressed the subtype of spinal a2-adrenoceptor responsible for the analgesic effects of i.t. dexmedetomidine and ST-91 in the
formalin behavioural model and their effects on primary afferent substance P (SP) release and spinal Fos activation.
Experimental approach: The analgesic effects of i.t. dexmedetomidine and ST-91 (a2 agonists) were tested on the formalin
behavioural model. To determine the subtype of a2-adrenoceptor involved in the analgesia, i.t. BRL44408 (a2A antagonist) or
ARC239 (a2B/C antagonist) were given before dexmedetomidine or ST-91. Moreover, the ability of dexmedetomidine and
ST-91 to inhibit formalin-induced release of SP from primary afferent terminals was measured by the internalization of neurokinin1
(NK1) receptors. Finally, the effects of dexmedetomidine on formalin-induced Fos expression were assessed in the dorsal horn.
Key results: Intrathecal administration of dexmedetomidine or ST-91 dose-dependently reduced the formalin-induced paw-
flinching behaviour in rats. BRL44408 dose-dependently blocked, whereas ARC239 had no effect on the analgesic actions of
dexmedetomidine and ST-91. Dexmedetomidine and ST-91 had no effect on the formalin-induced NK1 receptor
internalization, while morphine significantly reduced the NK1 receptor internalization. On the other hand, both
dexmedetomidine and morphine diminished the formalin-induced Fos activation. The effect of dexmedetomidine on
formalin-induced Fos activation was reversed by BRL44408, but not ARC239.
Conclusion and implications: These findings suggest that a2A-adrenoceptors mediate dexmedetomidine and ST-91 analgesia.
This effect could be through a mechanism postsynaptic to primary afferent terminals, distinct from that of morphine.
British Journal of Pharmacology (2008) 155, 1117–1126; doi:10.1038/bjp.2008.341; published online 1 September 2008
Keywords: Dexmedetomidine; ST-91; a2-adrenoceptors; substance P; neurokinin 1 receptor; primary afferent; nociception;
morphine; BRL44408; ARC239
Abbreviations: PBS, sodium phosphate buffer; SP, substance P
Introduction
Activation of spinal a2-adrenoceptors by noradrenaline or
synthetic agonists produces potent analgesia in animals and
humans (Eisenach et al., 1996; Pertovaara, 2006). There are
three different subtypes of a2-adrenoceptors (a2A, a2B and
a2C), which have similar homology and signal transduction
mechanisms (Bylund et al., 1994; Kable et al., 2000). Histo-
logical evidence indicates that the predominant subtypes of
a2-adrenoceptors in the spinal cord are the a2A- and
a2C-adrenoceptors. a2A-Adrenoceptors are largely located on
substance P (SP)-containing C-fibre primary afferent terminals;
although a smaller population also exists on other sites
within the superficial dorsal horn (Stone et al., 1998; Chen
et al., 2007). On the other hand, a2C-adrenoceptors are found
on inhibitory interneurons and axons of projections neurons
(Olave and Maxwell, 2002). The spinal cord contains
negligible levels of a2B-adrenoceptors (Zeng and Lynch,
1991; Nicholas et al., 1993). On the basis of the distribution
of binding and findings from transgenic animals, spinally
administered a2-adrenoceptor agonists are considered to
produce their analgesic actions through activation of
a2A-adrenoceptors. That is, disabling a2A-adrenoceptor func-
tion by a point mutation in the a2A gene resulted in the
inability of clonidine and dexmedetomidine to be analgesic
in acute thermal nociception and SP-elicited pain behaviour
in mice (Lakhlani et al., 1997; Stone et al., 1997).
The distribution of a2A-adrenoceptors raises the possibility
that spinal a2-adrenoceptor agonists produce their analgesic
effects by a presynaptic mechanism. Indeed, previous
Received 22 February 2008; revised 11 July 2008; accepted 22 July 2008;
published online 1 September 2008
Correspondence: Dr A Nazarian, Department of Pharmaceutical Sciences,
Western University of Health Sciences. 309 East Second Street, Pomona, CA
91766-1854, USA.
E-mail: anazarian@westernu.edu
British Journal of Pharmacology (2008) 155, 1117–1126
& 2008 Macmillan Publishers Limited All rights reserved 0007–1188/08 $32.00
www.brjpharmacol.org
findings have shown that dexmedetomidine, clonidine and
ST-91 reduced the evoked release of SP from spinal cord
preparations (Pang and Vasko, 1986; Ono et al., 1991; Takano
et al., 1993). Work from our laboratory and others (Honore
et al., 1999; Nazarian et al., 2008) demonstrated that
intraplantar formalin evokes the release of SP, as measured
by the internalization of neurokinin1 (NK1) receptors (NK1r).
Since a2A-adrenoceptors are involved in the analgesia
produced by a2-adrenoceptor agonists, it would be reason-
able to predict that a2A-adrenoceptors are also involved in
the presynaptic inhibition of SP release. Thus, in this study, it
was determined whether a2A-adrenoceptors mediate
the analgesic effects of dexmedetomidine and ST-91 in the
formalin-induced pain behaviour model, as well as regu-
lating formalin-induced primary afferent SP release, as
measured by NK1r internalization. Finally, the effects of
dexmedetomidine on formalin-evoked Fos expression in the
dorsal horn were also measured. The present findings
indicate that dexmedetomidine and ST-91 produce their
analgesic actions through a2A-adrenoceptors in the formalin
behaviour test; however, this effect appears to be mediated
postsynaptically in the spinal dorsal horn.
Materials and methods
Animals
Male Holtzman Sprague–Dawley rats (250–350 g; Harlan,
Indianapolis, IN, USA) were individually housed in standard
cages and maintained on a 12-h light/dark cycle (lights on at
07 h). Testing occurred during the light cycle. Animal care
was in accordance with the Guide for the Care and Use of
Laboratory Animals (National Institutes of Health publica-
tion 85–23, Bethesda, MD, USA) and approved by the
Institutional Animal Care and Use Committee of the
University of California, San Diego. Food and water were
available ad libitum.
Intrathecal catheter implantation
Rats were implanted with a single i.t. catheter for drug
delivery, as described previously (Malkmus and Yaksh, 2004).
In brief, rats were anaesthetized by induction with 4%
isoflurane in a room air/oxygen mixture (1:1), and the
anaesthesia was maintained with 2% isoflurane delivery by
mask. The animal was placed in a stereotaxic headholder
with the head flexed forward. A midline incision was made
on the back of the occipital bone and the neck to expose the
cisternal membrane. The membrane was carefully opened
with a stab blade, and a single lumen polyethylene-5 (outer
diameter 0.36 mm) catheter (8.5 cm) was inserted and passed
into the i.t. space surrounding L3–L4 spinal segments. The
other end of the catheter was jointed to a polyethylene-10,
which was tunnelled s.c. to exit through the top of the head.
Catheters were flushed with 10 ml of saline and then plugged.
The rats were given 5 mL of lactated ringer’s solution s.c. and
allowed to recover under a heat lamp; those showing motor
weakness or signs of paresis on recovery from anaesthesia
were killed immediately. The rats were allowed to recover for
5–7 days before the experiment.
a2-Adrenoceptor agonists and antagonists on formalin-induced
flinching
Formalin-induced flinching was measured using an auto-
mated detection system (Yaksh et al., 2001). A soft metal
band (10 mm wide and 27 mm long, shaped into a C, and
weighingB0.5 g) was placed around the left hindpaw of the
animal being tested. Animals were allowed to acclimate in
individual Plexiglas chambers for 30 min before experimen-
tal manipulations. For dexmedetomidine and ST-91 dose–
response studies, rats were administered with dexmedetomi-
dine (0.01, 0.1, 0.3, 1 or 2 mg), ST-91 (0.3, 0.9, 3, 10 or 30 mg)
or saline 10 min before a s.c. injection of formalin (5%,
50 mL) into the dorsal side of the banded paw. Immediately
after the formalin injection, rats were placed into the test
chamber and nociceptive behaviour was quantified by
automatic counting of spontaneous flinching and shaking
of the injected paw. Flinches were counted in 1-min intervals
for 60 min. The data are expressed as total number of flinches
observed during phase 1 (0–9 min) and phase 2 (10–60 min).
In studies with the antagonist drugs, the a2A-adrenoceptor
antagonist, BRL44408 (Young et al., 1989) or the a2B/C-
adrenoceptor antagonist, ARC239 (Bylund et al., 1992) were
administered 10 min before the administration of dexmede-
tomidine or ST-91. All drugs were injected i.t. in a volume of
10 mL followed by a 10 mL saline flush.
a2-Adrenoceptor agonists on formalin- and paw compression-
induced NK1 receptor internalization
After recovery from i.t. catheter implantation, rats were
administered i.t. with dexmedetomidine (0.3–10 mg), ST-91
(30 mg), morphine (20 mg) or saline. Five minutes after i.t.
drug administration, rats were anaesthetized with sodium
pentobarbital (50 mg kg�1, i.p.). For formalin-induced NK1r
internalization, intraplantar formalin injection (5%, 50 mL)
to the left hindpaw occurred 5 min after the anaesthesia. Rats
were killed and transcardially perfused with fixative 8 min
after the formalin injection. For paw compression-induced
NK1r internalization, dexmedetomidine (1 or 10 mg), ST-91
(30 mg), morphine (20 mg) or saline was administered i.t. Five
minutes after i.t. drug administration, rats were anaesthe-
tized with sodium pentobarbital (50 mg kg�1, i.p.) and 5 min
after anaesthesia, the hindpaw was positioned perpendicu-
larly across the jaws of a 6-inch mosquito forceps with non-
serrated jaws. The jaws were closed to the third click of the
hemostat ratchet. Compression was applied for 60 s. Applica-
tion of the forceps produces an evident compression of the
soft tissue resulting in approximately a 30% reduction in
paw thickness during the compression. This stimulus has
been previously used to evoke NK1r internalization (Ghilardi
et al., 2004; Kondo et al., 2005; Nazarian et al., 2008). Rats
were transcardially perfused 5 min after the paw compression
(see Immunocytochemistry).
Dexmedetomidine and a2-adrenoceptor antagonists on formalin-
induced Fos expression
After the rats had recovered from i.t. catheter implantation,
they were administered i.t. with BRL44408 (100mg), ARC239
(100mg) or saline. Ten minutes after antagonist administration,
a2-Adrenoceptor analgesia mechanism of action
A Nazarian et al1118
British Journal of Pharmacology (2008) 155 1117–1126
rats were administered with dexmedetomidine (1 mg), mor-
phine (20 mg) or saline. Intraplantar formalin (5%, 50 mL) was
injected 10 min after agonist administration. Rats were
anaesthetized with sodium pentobarbital (50 mg kg�1, i.p.),
1 h 55 min after formalin administration. Five minutes after
induction of anaesthesia rats were transcardially perfused.
Tissue preparation and immunocytochemistry
Anaesthetized rats were transcardially perfused with 0.9%
NaCl followed by 4% paraformaldehyde in 0.1 M sodium
phosphate buffer (PBS), pH 7.4. The lumbar spinal cord was
removed and postfixed overnight. After cryoprotection in
20% sucrose, coronal sections were made using a sliding
microtome (30 mm for NK1r staining and 50 mm for Fos
staining). Immunofluorescent staining was performed to
examine NK1r expression in the spinal dorsal horn. In brief,
sections were incubated in a rabbit anti-NK1r polyclonal
antibody overnight at room temperature. The antibody was
diluted to a concentration of 1:3000 in 0.01 M PBS contain-
ing 10% normal goat serum and 0.3% Triton X-100. After
rinses in PBS, sections were then incubated for 90 min at
room temperature in a goat anti-rabbit secondary antibody
conjugated with Alexa 488 diluted at 1:1000 in 0.01 M PBS
containing 10% normal goat serum and 0.3% Triton X-100.
All sections were finally rinsed and mounted on glass slides
and coverslipped with ProLong mounting medium.
Immunocytochemical staining for Fos consisted of
blocking endogenous peroxidate activity by incubating the
spinal cord sections in 0.3% H2O2 for 30 min. Non-specific
binding sites were blocked in 0.5% Triton X-100 and
5% normal goat serum in PBS. For primary antibody
incubation, tissues were incubated in rabbit anti-Fos
polyclonal antibody (1:10 000) for 72 h at 4 1C. The binding
sites were visualized by using an ABC kit. The signal was
developed in diaminobenzidine–tetrahydrochloride solution
containing H2O2 in PBS. Upon the presence of a light
background, the reaction was stopped by two washes in
distilled water. Sections were mounted on slides, air-dried,
dehydrated through graded ethanol solution followed by
citrisolve and then coverslipped with Permount mounting
medium.
Quantification of NK1 receptor internalization and Fos expression
Neurokinin1 receptor internalization was counted using an
Olympus fluorescence microscope at �60 magnification
and followed the standard of previous reports (Mantyh et al.,
1995; Abbadie et al., 1997). The total number of NK1r-
immunoreactive neurons in lamina I, with or without NK1r
internalization, was counted and taken as a ratio of cells
showing NK1r internalization vs all NK1r-positive cells and
then converted into a percentage of NK1r-immunoreactive
cells. Neuronal profiles that had 10 or more endosomes in
their soma and the contiguous proximal dendrites were
considered to have internalized NK1rs. NK1r-positive
neurons in both sides of the dorsal horn were counted. The
person counting the neurons was blinded to the experi-
mental treatments. Mean counts from three to five sections
per segment of the lumbar spinal cord were used as
representative counts for a given animal. Three to five
animals per drug treatment group were used for statistical
analysis (n¼3–5).
Fos immunoreactivity was quantified by counting Fos-
positive cells in lamina I of the lumbar L3–4 segments of the
spinal cord ipsilateral and contralateral to the formalin-treated
paw. The mean counts from five sections were used as
representative counts for a given animal. Four animals per drug
treatment group were used for the statistical analysis (n¼4).
Confocal microscopy and image processing
Confocal images of representative NK1r cells were acquired
by a Leica TCS SP2 confocal system equipped with AOBS
with a �63 objective (1.4 numerical aperture) and an argon
488 nm laser line with a pinhole of 1 airy unit. Images were
acquired at a digital size of 1024�1024 pixels. Five to ten
adjacent optical sections (B0.5-mm separation) along the
z-axis were projected together for demonstration. Images
were processed with Adobe Photoshop software (version 8.0)
by using the ‘curves’ option to adjust image contrast.
Drugs, antibodies and materials
Dexmedetomidine (Precedex, Hospira Inc., Lake Forest, IL,
USA) was purchased from the pharmacy of the University of
California, San Diego Medical Center. Dexmedetomidine in
powder form was kindly provided by Dr Donna
L Hammond from the University of Iowa. ST-91 (N-(2,6-diethyl-
phenyl)-4,5-dihydro-1H-imidazol-2-amine hydrochloride) was
provided by Boehringer Ingelheim (Ridgefield, CT, USA),
Formalin solution was purchased from Sigma Chemicals
(St Louise, MO, USA). BRL44408 (2-[(4,5-dihydro-1H-imida-
zol-2-yl)methyl]-2,3-dihydro-1-methyl-1H-isoindole maleate)
and ARC239 (2-[2-(4-(2-methoxyphenyl)piperazin-1-yl)ethyl]-
4,4-dimethyl-1,3-(2 H,4 H)-isoquinolindione dihydrochloride)
were purchased from Tocris Biosciences (Ellisville, MO, USA).
Dexmedetomidine, ST-91 and BRL44408 were dissolved in
saline, whereas ARC239 was dissolved in 10% 2-hydroxy-
propyl-b-cyclodextrin. All drugs were prepared on the day
of testing. The rabbit anti-NK1r polyclonal antibody was
purchased from the Advanced Targeting Systems (San Diego,
CA, USA) and the rabbit anti-Fos polyclonal antibody
was purchased from Calbiochem-EMD Biosciences (La Jolla,
CA, USA). Secondary Alexa 488-conjugated antibody was
purchased from Invitrogen (Eugene, OR, USA). ProLong
mounting medium was obtained from Invitrogen, ABC
kit from Vector Laboratories (Burlingame, CA, USA) and
Permount mounting medium was from Fisher Scientific
(Pittsburgh, PA, USA).
Nomenclature for drugs and receptors conform with the
guide to receptors and channels of the British Journal of
Pharmacology (Alexander et al., 2008).
Statistical analysis
Changes in formalin-induced paw-flinching behaviour were
analysed using separate one-way ANOVA for phases 1 and 2.
Upon detection of a significant ANOVA, Tukey’s post hoc tests
were performed for pair-wise comparisons of drug-treated
a2-Adrenoceptor analgesia mechanism of action
A Nazarian et al 1119
British Journal of Pharmacology (2008) 155 1117–1126
groups with their relative control within phase 1 or 2. The
analyses for NK1r internalization data consisted of one-way
or two-way ANOVAs. To detect the differences in the
presence of a significant one-way ANOVA, Tukey’s post hoc
analysis was conducted. In the presence of a significant two-
way ANOVA, Bonferroni post hoc tests were applied. The
analysis of the Fos data consisted of a paired t-test and a
one-way ANOVA. Tukey’s post hoc test was used for pair-wise
comparisons. In all analysis, probability to detect the
differences was set at the 5% level (Po0.05).
Results
Effects of dexmedetomidine and ST-91 on formalin-induced
paw-flinching behaviour
The effects of dexmedetomidine and ST-91 on formalin-
induced paw-flinching behaviour are shown in Figures 1a
and b, respectively. Dexmedetomidine dose-dependently
reduced the formalin-induced paw flinches in phase 1:
F(5,21)¼6.42, Po0.001; with doses of 1 and 2 mg producing
a significant reduction in flinching. In phase 2, dexmede-
tomidine reduced the formalin flinches at doses of 0.3, 1 and
2 mg: F(5,21)¼8.05, Po0.001. In this experiment, the 2mg
dose of dexmedetomidine produced mild sedation, but the
1 mg dose of dexmedetomidine did not appear to have
sedative effects; similarly, i.t. dexmedetomidine (1 mg)
was not found to alter the rotarod performance in rats
(Ida¨npa¨a¨n-Heikkila¨ et al., 1994). Accordingly, the 1mg dose of
dexmedetomidine was typically employed in subsequent
experiments, unless otherwise noted. ST-91 dose-depen-
dently reduced the formalin-induced paw flinches in phase
1: F(5,31)¼5.74, Po0.001, with 10 and 30 mg doses produ-
cing a significant reduction in flinching. In phase 2, ST-91 at
doses of 3, 10 and 30 mg reduced the formalin flinching:
F(5,31)¼11.36, Po0.001, with no effect upon arousal noted
at the highest dose.
a2-Adrenoceptor antagonists on dexmedetomidine and ST-91
analgesia
The effects of the a2A-adrenoceptor antagonist, BRL44408 on
dexmedetomidine and ST-91 analgesia are presented in
Figures 2a and b. In phase 1, dexmedetomidine reduced
the formalin-flinching behaviour: F(4,22)¼13.14, Po0.001,
Whereas the i.t. pretreatment with BRL44408 (10 and 100 mg)
blocked the dexmedetomidine-induced reduction in forma-
lin flinching. Equally, in phase 2, dexmedetomidine reduced
the formalin-flinching behaviour; whereas pretreatment
with BRL44408 (10 and 100 mg) blocked the dexmedetomi-
dine-induced reduction in formalin-flinching behaviour:
F(4,22)¼17.72, Po0.001. ST-91 did not significantly alter
phase 1 formalin-flinching behaviour and BRL44408 pre-
treatment had no effect on ST-91-treated rat