Department of Anesthesia and Critical Care, The University of
Chicago, Chicago, Illinois
 |
Introduction |
Persistent
nociceptive input to the spinal cord induces prolonged alterations in
the response properties, neurochemistry and phenotype of dorsal horn
neurons (Dubner and Ruda, 1992
) and primary afferent fibers (Neumann
et al., 1996; Woolf, 1996a). One such change is the
induction of central sensitization in dorsal horn neurons after
electrical stimulation of C-fiber afferents or the peripheral
application of noxious substances such as mustard oil, carrageenan or
formalin (Woolf and Wall, 1986; Neugebauer and Schaible, 1990; Woolf
et al., 1994; Xu et al., 1995). The enhanced excitability of dorsal horn neurons is thought to mediate the hyperalgesia and allodynia that develop after tissue injury. Studies of
the mechanisms responsible for the induction and maintenance of central
sensitization and the behavioral sequelae to tissue injury have
predominantly emphasized the role of excitatory neurotransmitters such
as glutamate and substance P, as well as intracellular messengers (Coderre et al., 1993). Despite substantial evidence that
GABA and its receptors are appropriately situated to modulate
nociceptive transmission in the dorsal horn of the spinal cord
(Hammond, 1997), comparatively little attention has been paid to the
role of GABA in central sensitization in the spinal cord (Sivilotti and
Woolf, 1994) or in the behavioral sequelae to tissue injury (Hao
et al., 1991; Yamamoto and Yaksh, 1991; Smith et
al., 1994; Dirig and Yaksh, 1995). Most of these studies examined
the effects of GABA receptor agonists, and only a very few used GABA
receptor antagonists to assess the role of endogenous GABA in central
sensitization and the behavioral sequelae to tissue injury (Yamamoto
and Yaksh, 1993). Yet, small-diameter primary afferent neurons are
known to make synaptic contacts on the dendrites of GABAergic neurons in the dorsal horn (Carlton and Hayes, 1990; Hayes and Carlton, 1992),
and stimulation of afferent inputs to slices of the spinal cord evokes
GABAA receptor-mediated inhibitory postsynaptic
potentials in dorsal horn neurons (Yoshimura and Nishi, 1995).
Activation of nociceptive afferents by injection of formalin,
carrageenan or the topical application of mustard oil is therefore
likely to evoke a release of GABA, as well as glutamate, substance P and calcitonin gene-related peptide, in the spinal cord. It is reasonable to expect that the behavioral sequelae to tissue injury reflect the summation of inhibitory processes mediated by
GABAA and GABAB receptors,
and excitatory processes mediated by NMDA and neurokinin receptors. The
recent report that i.t. administration of GABA receptor
antagonists does not alter formalin-evoked pain behaviors
(Dirig and Yaksh, 1995) is contrary to this expectation. However,
the use of a single, high concentration of formalin (5%) and the
existence of a "ceiling effect" for the number of flinches may have
precluded identification of an increase in formalin-induced pain
behaviors by GABA receptor antagonists. This observation led us to
reexamine the role of GABA and GABAA receptors in
the development and maintenance of persistent pain behaviors as modeled by the formalin test with the important distinction that the effects of
the GABAA receptor ligands were examined at
concentrations of formalin ranging from 0.25% to 2.5%. Subsequent
comparison of the concentration-effect curves for formalin in the
presence of increasing doses of the GABAA
receptor antagonist bicuculline or the GABAA
receptor agonists, muscimol and isoguvacine, permitted a quantitative
estimate of the extent to which antagonism or mimicry of the actions of
GABA at the GABAA receptor, respectively,
enhanced or suppressed nociception. It was hypothesized that i.t.
administration of low doses of bicuculline would enhance
formalin-induced pain behaviors. Moreover, a preferential enhancement
of pain behaviors in phase 2 of the formalin test, which are mediated
by activation of NMDA receptors (Coderre and Melzack, 1992; Yamamoto
and Yaksh, 1992), was expected as removal of an inhibitory GABAergic
input to dorsal horn neurons should facilitate the activation of NMDA receptors by glutamate. Intrathecal administration of isoguvacine or
muscimol was expected to suppress pain behaviors in both phase 1 and
phase 2. A preliminary report of some of these data has appeared
(Kaneko and Hammond, 1997).
| |
Materials and Methods |
Animal preparation.
This study was approved by the
Institutional Animal Care and Use Committee of the University of
Chicago. Male Sprague-Dawley rats (Sasco, Madison, WI; 300-350 g) were
housed in groups of three and maintained on a 12-hr light/dark cycle
with free access to food and water. The rats were anesthetized with
halothane, and a polyethylene catheter (PE-10) was inserted through an
incision in the atlanto-occipital membrane. One end of the catheter was advanced caudally to the rostral edge of the lumbar enlargement, and
the other end was tunneled subcutaneously and externalized on top of
the head (Yaksh and Rudy, 1976; Hammond, 1988). The rats were housed
individually after surgery and allowed at least 7 days to recover
before testing.
Formalin test.
Animals were placed individually in Plexiglas
testing chambers (30.5 × 30.5 × 30.5 cm) and allowed to
acclimate for at least 60 min. A mirror was situated behind the chamber
and another was situated at a 45° angle below the floor of the
chamber to allow an unobstructed view of the rat's paws. After
acclimation, 100 µl of formalin (0.25-2.5%) was injected s.c. into
the plantar surface of the left hind paw, and the rat was returned to
the testing chamber. Its behavior was observed for the next 60 min. The
time spent in each of four mutually exclusive categories of behavior
was determined by use of a BASIC computer program generously provided
by Dr. K.B.J. Franklin (Department of Psychology, McGill University,
Montreal, Canada). The behaviors were those originally described by
Dubuisson and Dennis (1977) and reiterated by Abbott et al.
(1995) as "0 = normal weight bearing on the injected paw, 1 = limping during locomotion or resting the paw lightly on the floor,
2 = elevation of the injected paw so that at most the nails touch
the floor, and 3 = licking, biting" or shaking the injected paw.
A weighted pain score was calculated by multiplying the amount of time
spent in each category by its assigned category weight, summing these
products and then dividing by the total time in each 5-min block of
time. In addition, the number of flinches that occurred was counted
(Wheeler-Aceto and Cowan, 1991). Proper placement of the i.t. catheter
was verified at the conclusion of the formalin test by the occurrence
of hindlimb paralysis after an i.t. injection of 10 µl of 2%
tetracaine hydrochloride or, in rats sacrificed by
CO2 inhalation, by direct visualization of the
catheter tip after laminectomy and injection of India ink.
Experimental design.
Animals were used only once in this
study and received only one dose of drug and one concentration of
formalin. The first series of experiments was designed to determine the
time course and dose dependence of the effect of the
GABAA receptor antagonist bicuculline methiodide
on nociceptive behaviors induced by injection of formalin in the
hindpaw. In the pretreatment study, rats received an i.t. injection of
either saline or 0.03, 0.1 or 0.3 µg of bicuculline 10 min before the
injection of a concentration of formalin ranging from 0.25% to 2.5%.
In the post-treatment study, either saline or 0.3 µg of bicuculline
was injected i.t. 7 to 8 min after the injection of a concentration of
formalin ranging from 0.25% to 1.25%. Doses of bicuculline greater
than 0.3 µg were not tested because 1) higher doses produce
spontaneous vocalization, allodynia and caudally directed biting and
scratching behavior (Yaksh, 1989; McGowan and Hammond, 1993) and 2) 0.3 µg of bicuculline effectively antagonizes the increase in tail-flick
latency produced by i.t. administration of the
GABAA receptor agonist isoguvacine (McGowan and
Hammond, 1993).
The second series of experiments examined the time course and dose
dependence of the effect of i.t. administration of a
GABAA receptor agonist on formalin-evoked
nociceptive behaviors. In these experiments, either 10 or 30 µg of
isoguvacine or 0.3 µg of muscimol was injected i.t. 10 min before the
injection of a concentration of formalin ranging from 0.25% to 2.5%.
Statistical analysis.
The number of flinches and weighted
pain scores were determined for each 5-min interval after the injection
of formalin, and the data were expressed as the mean ± S.E.M. for
that 5-min interval. Two approaches were used to assess the effect of
drug treatment on formalin-induced pain behaviors. The first approach
compared the effect of drug treatment with that of saline at each
concentration of formalin. This analysis was performed by a two-way
analysis of variance for repeated measures in which drug treatment was one factor and time was the second (repeated) factor. Post
hoc comparisons of individual mean values were made by the
Newman-Keuls test. The second approach examined the effect of drug
treatment on the stimulus-response functions of formalin. For this
analysis, concentration-effect curves for formalin were constructed in
drug- and saline-treated rats for phase 1, phase 2 and interphase
behaviors. For this purpose, phase 1 was defined as the first 5 min,
interphase was defined as the period 10 to 15 min after injection of
formalin and phase 2 was defined as the period 20 to 50 min after
injection of formalin. The analysis of phase 1 behaviors used the total number of flinches and the weighted pain score for that 5-min interval.
The analysis of interphase behaviors used the average number of
flinches and the average of the weighted pain scores determined 10 and
15 min after the injection of formalin. The analysis of phase 2 behaviors used the average number of flinches and the average of the
weighted pain scores determined between 20 and 50 min after the
injection of formalin, respectively. Least-squares linear regression of
the individual data was used to determine the concentration of formalin
(EC50) that produced one-half the maximal number
of flinches or increase in weighted pain score. These criteria
corresponded to a weighted pain score of 1.1 (maximum pain score was
2.2) and 50 flinches for phases 1 and 2 and to a weighted pain score of
0.6 (maximum pain score was 1.2) and 25 flinches for the interphase.
Fieller's theorem was used to determine CL (Finney, 1964). The
significance of differences in the EC50 values of
formalin in drug- and saline-treated rats was determined by analysis of
covariance (Zar, 1984). P 
.05 was considered significant.
Drugs and injections.
All drugs were injected i.t. in
a volume of 10 µl followed by 10 µl of saline to flush the
catheter. The drugs were freshly prepared, adjusted to pH 6.8 to 7.1 and filtered before administration. Bicuculline methiodide, muscimol
and tetracaine hydrochloride were purchased from Sigma Chemical Co.
(St. Louis, MO). Isoguvacine hydrochloride was purchased from Research
Biochemicals Inc. (Natick, MA).
| |
Results |
Stimulus-response function of formalin in saline-treated
animals.
Subcutaneous injection of formalin into the plantar
surface of one hindpaw of saline-pretreated rats evoked concentration- and time-dependent increases in the number of flinches (fig.
1) and in weighted pain score (fig.
2). These increases were biphasic, with
an initial increase occurring within the first 5 min (phase 1),
followed by a quiescent period characterized by fewer flinches and pain
behaviors (interphase; between 5 and 15 min), and a second increase in
flinching and nociceptive behaviors beginning about 20 min and
continuing for at least 50 min (phase 2) after the injection of each
concentration of formalin. Figure 3
illustrates the stimulus-response relationships of formalin for the
number of flinches and for weighted pain scores in phases 1 and 2, and in the interphase. The numbers of flinches and weighted pain scores in
phase 1, interphase and phase 2 were linearly related to the concentration of formalin in the range of 0.25% to 1.25%. In general, 1.25% formalin produced the maximal number of flinches or greatest weighted pain score, with the exception of the number of flinches in
the interphase in which 2.5% formalin elicited a significantly greater
number of flinches than did 1.25% formalin. Table
1 presents the EC50
values for formalin in saline-treated rats for phase 1, phase 2 and
interphase pain behaviors as determined by number of flinches and by
weighted pain score.
View larger version (37K):
[in this window]
[in a new window]
|
Fig. 1.
Effects of i.t. pretreatment with saline ( ) or
with 0.03 ( ), 0.1 ( ) or 0.3 ( ) µg of bicuculline on the
number of flinches evoked by either 0.25% (A), 0.5% (B), 1.25% (C)
or 2.5% (D) formalin. Saline or bicuculline was administered i.t. 10 min before s.c. injection of formalin into the plantar surface of one
hindpaw. Each symbol represents the mean ± S.E.M. of
determinations made in five to eight animals.
|
|
View larger version (38K):
[in this window]
[in a new window]
|
Fig. 2.
Effects of i.t. pretreatment with saline () or
with 0.03 (), 0.1 () or 0.3 () µg of bicuculline on weighted
pain score evoked by injection of either 0.25% (A), 0.5% (B), 1.25%
(C) or 2.5% (D) formalin. Saline or bicuculline was administered i.t. 10 min before s.c. injection of formalin into the plantar surface of
one hindpaw. Each symbol represents the mean ± S.E.M. of
determinations made in five to eight animals.
|
|
View larger version (27K):
[in this window]
[in a new window]
|
Fig. 3.
Concentration-effect relationship of formalin
during phase 1 (A, B), interphase (C, D) and phase 2 (E, F) in rats
pretreated i.t. with saline () or with 0.03 (), 0.1 () or 0.3 () µg of bicuculline. Responses during phase 1, interphase and
phase 2 are presented as the mean of the number of flinches (A, C, E) or weighted pain score (B, D, F) determined 0 to 5, 10 to 15 and 20 to
50 min after injection of formalin, respectively. The concentrations of
formalin are plotted on a log scale. Each symbol represents the
mean ± S.E.M. of determinations in five to eight animals.
|
|
View this table:
[in this window]
[in a new window]
|
TABLE 1
EC50 values and 95% CL of formalin for phase 1, interphase and
phase 2 responses in saline- or bicuculline-treated ratsa
|
|
Effects of i.t. pretreatment with the GABAA
receptor antagonist bicuculline.
Formalin-induced pain behaviors
in phase 1 were generally unaffected by i.t. pretreatment with
bicuculline. Although the number of flinches was slightly increased in
phase 1 by 0.03, 0.1 or 0.3 µg of bicuculline, the magnitude of the
effect was small and was not dose-dependent (figs. 1 and 3A).
Comparison of the EC50 values for formalin did
not reveal any differences among saline and bicuculline-treated rats
for number of flinches in phase 1 (table
1). Intrathecal pretreatment with these
doses of bicuculline did not increase the weighted pain scores during
phase 1 at any concentration of formalin (figs. 2 and 3B). The
EC50 of formalin for weighted pain scores in
phase 1 also did not differ among saline- or bicuculline-treated rats
(table 1).
Intrathecal pretreatment with 0.1 or 0.3 µg of bicuculline
significantly increased the number of flinches and weighted pain scores
determined during the interphase period as compared to saline-treated
rats at each concentration of formalin (figs. 1 and 2; table 1). Thus,
an enhancement of formalin-induced nociceptive behaviors began within
10 min of the injection of formalin. This enhancement was apparent as a
leftward shift in the concentration-effect curve of formalin in the
interphase (fig. 3). Both the number of flinches and the weighted pain
scores during interphase were significantly greater in rats treated
with 0.1 and 0.3 µg of bicuculline compared with saline-treated rats
(figs. 1 and 2); the greatest enhancement occurred at the lower
concentrations of formalin. The highest dose of bicuculline decreased
the EC50 of formalin for the number of flinches
to 0.33%, or nearly one-third the value of 1.1% determined in
saline-treated rats. This same dose of bicuculline decreased the
EC50 of formalin for weighted pain score to
0.43%, or nearly one-quarter the value of 1.67% determined in
saline-treated rats (table 1).
Intrathecal pretreatment with bicuculline also increased nociceptive
responses to formalin in phase 2. This increase was apparent as a
leftward shift in the concentration-effect curve for formalin (fig. 3,
E and F). Pretreatment with 0.1 or 0.3 µg of bicuculline markedly
increased the number of flinches in phase 2 evoked by 0.25, 0.5 or
1.25% formalin as compared with saline-treated rats (fig. 1).
Pretreatment with 0.03 µg of bicuculline also increased the number of
flinches; however, this enhancement was small and was most consistently
apparent at the two lowest concentrations of formalin. Although i.t.
pretreatment with 0.3 µg of bicuculline increased the number of
flinches evoked by 2.5% formalin, this effect was evident only during
the earliest aspect of phase 2 (i.e., 20-35 min). Table 1
illustrates that the EC50 values of formalin for
flinches in rats pretreated with bicuculline were decreased in a
dose-dependent manner with significant differences observed in rats
pretreated with either 0.1 or 0.3 µg as compared with saline-treated
rats. The 0.3-µg dose of bicuculline decreased the
EC50 of formalin for number of flinches to
0.48%, or nearly one-half the value of 0.85% in saline-treated rats.
Weighted pain scores during phase 2 were also enhanced by pretreatment
with bicuculline (fig. 2). In rats that received the two lowest
concentrations of formalin, doses as little as 0.03 µg of bicuculline
significantly increased weighted pain score as compared with
saline-treated rats. In rats that received 1.25% or 2.5% formalin,
the most consistent enhancement was produced by 0.3 µg of bicuculline
(fig. 2). The EC50 of formalin for weighted pain
score was decreased in a dose-dependent manner in the presence of
increasing doses of bicuculline, with a comparable effect produced by
either 0.1 or 0.3 µg of bicuculline (table 1). As observed for number
of flinches, 0.3 µg of bicuculline decreased the
EC50 of formalin for weighted pain score to
0.36%, or nearly one-half the value of 0.80% determined in
saline-treated rats.
In addition to the quantitative changes described above, i.t.
pretreatment with bicuculline produced qualitative changes in formalin-evoked nociceptive responses. For example, during the phase 2 response to injection of either 1.25% or 2.5% formalin, saline-treated rats exhibited two characteristic types of flinches. One
type of flinch was limited to the ipsilateral hindquarter and was
accompanied by small-amplitude, high-frequency shakes of the paw,
whereas the other was bilateral and involved the entire hindquarters.
These flinches were considered equivalent for purposes of quantitation.
These forms of flinching were not elicited by lower concentrations of
formalin in saline-treated rats. However, these two types of flinching
were frequently elicited by concentrations of formalin as low as 0.5%
in rats pretreated with bicuculline. Furthermore, bicuculline
pretreatment altered the amount of time spent in the three different
categories of pain behavior. One factor that contributed to the
increase in weighted pain scores in bicuculline-treated animals was the
increase in the amount of time spent in category 3 behaviors. For
example, 0.3 µg of bicuculline increased the total time spent in
category 3 during phase 2 to 167 ± 40 sec from 99 ± 13 sec
at 0.5% formalin. Moreover, licking of the contralateral paw (`mirror
pain' (Aloisi et al., 1993)), which was not observed in
saline-treated animals at 0.5% formalin, occurred in 8 of 11 rats
treated with either 0.1 or 0.3 µg bicuculline at this formalin
concentration.
The 0.3-µg dose of bicuculline was administered to four rats in the
absence of formalin to determine whether this dose by itself could
elicit formalin-like nociceptive behaviors. No formalin-like nociceptive behaviors were observed in three of the rats. In the fourth
rat, the injected hindpaw was favored for less than 4 min. A total of
two flinches occurred among the four rats during the 60-min observation
period.
Effect of i.t. post-treatment with bicuculline.
Post-treatment
with 0.3 µg i.t. bicuculline 7 to 8 min after injection of formalin
significantly enhanced both the number of flinches and the weighted
pain score in phase 2 as compared with saline-treated rats. This effect
was consistent at each concentration of formalin (0.25-1.25%) that
was tested (time course not shown). The magnitude of nociceptive
behaviors in phase 1, before the injection of drug, did not differ
between the saline- and bicuculline-treatment groups (table 1). In rats
that received 0.3 µg of bicuculline, the concentration-effect curves
for formalin were shifted to the left (fig.
4) and the EC50
values of formalin were significantly decreased as compared with values
in saline-treated rats for both number of flinches and for weighted
pain score (table 1). Importantly, the EC50
values of formalin for rats in which 0.3 µg of bicuculline was
administered 7 to 8 min after formalin did not differ from the
EC50 values determined in rats in which this same
dose was administered 10 min before formalin (table 1). This result
indicates that bicuculline was equally effective when administered
either as a pre- or post-treatment. The effects of post-treatment with bicuculline on nociceptive behaviors during the interphase could not be
determined as the injection interfered with assessment of these
behaviors.
View larger version (17K):
[in this window]
[in a new window]
|
Fig. 4.
Concentration-effect relationship of formalin for
(A) number of flinches and (B) weighted pain score in phase 2 in rats
that received an i.t. injection of either saline () or 0.3 µg of
bicuculline () 7 to 8 min after injection of formalin in one
hindpaw. The concentrations of formalin are plotted on a log scale.
Each symbol represents the mean ± S.E.M. of determinations in six
to seven animals.
|
|
Effect of i.t. pre-treatment with a GABAA
receptor agonist.
The effect of GABAA
receptor agonists was somewhat dependent on the measure of nociception
(figs 5 and
6). The
numbers of flinches in both phase 1 and phase 2 were significantly
decreased in a dose-dependent manner by i.t. pretreatment with 10 or 30 µg of isoguvacine at each concentration of formalin (fig. 5). Both
doses of isoguvacine also significantly decreased the number of
flinches in the interphase, albeit not in a dose-dependent manner (fig.
5). However, isoguvacine appeared to be less effective with respect to
its ability to decrease weighted pain score in either phase 1 or phase
2. Only pretreatment with 30 µg of isoguvacine decreased weighted
pain scores in phase 1 and phase 2 at each concentration of formalin,
and this suppression was small. (fig. 6). In rats pretreated with 10 µg of isoguvacine, a significant decrease in weighted pain scores
occurred only at the 0.25% concentration of formalin. The reduction in
weighted pain scores by isoguvacine during phase 2 was predominantly
the result of a decrease in the time spent in category 3 behaviors. The
time spent in category 3 behaviors in rats pretreated with 30 µg of
isoguvacine at 0.25%, 0.5% and 1.25% formalin was 0 ± 0, 24 ± 19.9 and 97.1 ± 26.8 sec, respectively. By comparison,
the time spent in this category in saline-treated rats was 22.6 ± 12.1, 106.2 ± 17.5 and 169.8 ± 25.6 sec, respectively
(P < .05, each concentration). Pretreatment with 30 µg of
isoguvacine shifted the concentration-effect curve of formalin for
weighted pain scores to the right (fig. 7) and significantly increased
the EC50 value (CL) of formalin in phase 1 and
phase 2 to 0.55 (0.41-0.69)% and 1.26 (1.07-1.52)%, respectively (P < .05 compared with saline, table 1). The
EC50 values of formalin could not be calculated
for the number of flinches as this response did not exceed the
criterion value of 50 even in those isoguvacine-treated rats that
received 2.5% formalin. However, fig. 7 illustrates that the
concentration-effect curve of formalin for the number of flinches was
shifted to the right in a nonparallel manner in rats pretreated with
either 10 or 30 µg of isoguvacine. It was estimated that the
EC50 of formalin for number of flinches in rats
pretreated with 30 µg of isoguvacine was increased by at least 4-fold
compared with saline-treated rats. As qualitative observations
indicated that 30 µg of isoguvacine caused mild muscle weakness of
the hindlimbs in 3 of 21 rats, higher doses were not administered in
this study. These three rats were not included in the data analysis.
View larger version (37K):
[in this window]
[in a new window]
|
Fig. 5.
Effects of i.t. pretreatment with saline (), 10 ( ) or 30 () µg of isoguvacine or 0.3 µg of muscimol () on
the number of flinches evoked by injection of either 0.25% (A), 0.5%
(B), 1.25% (C) or 2.5% (D) formalin. Saline, isoguvacine or muscimol
was administered i.t. 10 min before s.c. injection of formalin into the
plantar surface of one hindpaw. Each symbol represents the mean ± S.E.M. of determinations made in five to eight animals.
|
|
View larger version (36K):
[in this window]
[in a new window]
|
Fig. 6.
Effects of i.t. pretreatment with saline (), 10 () or 30 () µg of isoguvacine, or 0.3 µg of muscimol () on
weighted pain score after injection of either 0.25% (A), 0.5% (B),
1.25% (C), or 2.5% (D) formalin. Saline, isoguvacine or muscimol was
administered i.t. 10 min before s.c. injection of formalin into the
plantar surface of one hindpaw. Each symbol represents the mean ± S.E.M. of determinations made in five to eight animals.
|
|
View larger version (25K):
[in this window]
[in a new window]
|
Fig. 7.
Concentration-effect relationship of formalin
during phase 1 (A, B) and phase 2 (C, D) in rats pretreated i.t. with
saline (), 10 () or 30 () µg of isoguvacine, or 0.3 µg of
muscimol (). Responses during phase 1 and phase 2 are presented as
the mean of the number of flinches (A, C) or weighted pain score (B, D)
determined 0 to 5 min and 20 to 50 min after injection of formalin, respectively. The concentrations of formalin are plotted on a log
scale. Each symbol represents the mean ± S.E.M. of determinations in five to eight animals.
|
|
The disparate effects of isoguvacine on the number of flinches and
weighted pain scores elicited by formalin prompted examination of the
effects of another GABAA receptor agonist,
muscimol. Muscimol produced effects similar to those of isoguvacine.
Intrathecal pretreatment with 0.3 µg of muscimol significantly
decreased the number of flinches in phase 1 and phase 2 at each
concentration of formalin (fig. 5). It also shifted the
concentration-effect curve of formalin for the number of flinches to
the right in a nonparallel manner, increasing the
EC50 of formalin by at least four-fold (fig. 7).
An EC50 value for formalin in muscimol-treated rats could not be calculated as the number of flinches did not exceed
the criterion value of 50 even in rats that received 2.5% formalin.
Like isoguvacine, muscimol appeared to be less effective in decreasing
the weighted pain score (fig. 6). A small, but significant increase in
the EC50 (CL) of formalin to 1.07 (0.95-1.22)%
occurred in rats pretreated with 0.3 µg of muscimol. Higher doses of
this GABAA receptor agonist could not be tested
as these caused mild muscle weakness of the hindlimbs (D. L. Hammond and M. K. McGowan, unpublished observations).
| |
Discussion |
A different approach to the formalin test.
The formalin test
is generally used to measure persistent nociception. However, most
studies with the formalin test have restricted their examination to an
analysis of drug effects at a single concentration of formalin and
often to a single measure, either the number of flinches or weighted
pain score. This approach, although efficient and cost-effective, has
limitations. Use of a high concentration of formalin can hinder
identification of weak analgesics and, because of the existence of a
ceiling effect, preclude identification of agents that increase
nociceptive behaviors. The inability of Dirig and Yaksh (1995) to
detect an enhancement of formalin-induced pain behaviors by i.t.
administered bicuculline most likely arises from these factors.
Conversely, low concentrations of formalin, which are suitable for
studies of hyperalgesia, may not produce an effect of sufficient
magnitude to permit reliable detection of antinociception. The present
study, in which the effects of different doses of bicuculline, muscimol
or isoguvacine were examined over a wide range of concentrations of
formalin, represents an important advance in the use of the formalin
test as a measure of nociception. This systematic approach generates a
"matrix" of data that can be used for several different types of
analyses. First, comparison of the EC50 values of
formalin in saline- and drug-treated animals enables a quantitative
analysis of the extent to which a drug treatment alters the perceived
intensity of the noxious stimulus. For example, i.t. pretreatment with
0.3 µg of bicuculline halved the EC50 values of
formalin for both weighted pain score and number of flinches in phase
2, which suggests that formalin is nearly twice as noxious in the
presence of bicuculline. Conversely, i.t. pretreatment with either 0.3 µg of muscimol or 30 µg of isoguvacine shifted the
concentration-effect relationship of formalin for the number of
flinches in phase 2 at least 4-fold to the right, which suggests that
formalin was perceived to be one-quarter as noxious in the presence of
these GABAA receptor agonists. Generation of
concentration-effect curves for formalin therefore enabled detection of
both the enhancement, as well as the suppression of nociceptive
behaviors. Second, it is possible that different concentrations of
formalin may induce different pharmacologic mechanisms in the spinal
cord and in the periphery, or alter the balance of central and
peripheral mechanisms that contribute to the nociception. The
systematic analysis of drug effects over a wide range of concentrations
of formalin can identify such important occurrences. Finally, this
approach remains suitable for the standard determination of the
ED50 of a drug at any specified concentration of
formalin. Moreover, comparison of the ED50 values of the drug across increasing concentrations of formalin, which are
presumably increasingly more noxious, can be used to determine the
relative efficacy and fractional receptor occupancy requirements of
antinociceptive agents in a model of persistent chemically induced
nociception. This approach is analogous to previous studies in which
different intensities of noxious thermal stimuli were used to determine
the relative efficacy and fractional receptor occupancy requirements of
i.t. administered opioid or alpha-2 adrenoceptor agonists
(Saeki and Yaksh, 1992; Saeki and Yaksh, 1993).
Intrathecal bicuculline increases formalin-induced pain
behaviors.
A principal finding of this study was that i.t.
pretreatment with bicuculline, a GABAA receptor
antagonist, increased formalin-induced pain behaviors in a
dose-dependent manner in the interphase and phase 2, but did not affect
pain behaviors in phase 1. Bicuculline was equally effective when
administered 7 to 8 min after formalin. Antagonists, by definition,
bind to receptors but have no efficacy. For bicuculline to enhance
formalin-induced pain behaviors, there must be an inhibitory action of
GABA mediated by GABAA receptors. These data
therefore suggest that within minutes of the injection of formalin
there is a release of GABA and an activation of
GABAA receptors in the spinal cord. These doses
of bicuculline by themselves did not produce allodynia or hyperalgesia,
nor did they increase formalin-induced pain behaviors in phase 1. The
lack of effect of bicuculline in phase 1 does not reflect an
insensitivity of phase 1 to modulation by GABAA
receptor ligands as GABAA receptor agonists were
able to suppress pain behaviors in phase 1. Rather, these data suggest
that there is little tonic activation of spinal GABAA receptors before or during the first few
minutes after injection of formalin.
The GABA that is released in the spinal cord in response to s.c.
injection of formalin may originate from two sources. One source is
interneurons, which are the principal source of GABA in the dorsal horn
(Miyata and Otsuka, 1975; Todd and McKenzie, 1989). Primary afferents,
including small-diameter afferents containing calcitonin gene-related
peptide, make synaptic contacts with the dendrites of GABAergic
interneurons in the dorsal horn (Carlton and Hayes, 1990; Hayes and
Carlton, 1992). Also, activation of A
primary afferent fibers in
slices of rat spinal cord evokes a polysynaptic inhibitory postsynaptic
potential in substantia gelatinosa neurons that is mediated by
GABAA receptors and that is abolished in the
presence of 6-cyano-7-nitroquionoxaline-2,3-dione, a non-NMDA receptor
antagonist (Yoshimura and Nishi, 1995). Thus, GABA is likely to be
released from interneurons in the dorsal horn as a result of the direct
activation by formalin of small-diameter glutamatergic and peptidergic
primary afferents. A second source of GABA in the dorsal horn is the
spinal projections of GABAergic neurons in the ventromedial medulla
(Blessing, 1990; Reichling and Basbaum, 1990; Jones et al.,
1991; Antal et al., 1996). In addition, there is evidence
that serotonergic bulbospinal neurons synapse on GABAergic interneurons
in the dorsal horn (Alhaider et al., 1991). Activation of
medullary neurons produces an antinociception that is antagonized by
i.t. administration of bicuculline, which suggests that bulbospinal
pathways can modulate acute nociception by a
GABAA receptor-mediated mechanism (McGowan and
Hammond, 1993). There is also evidence that the activity of bulbospinal
pain modulatory pathways is increased during the development of acute
inflammation (Schaible et al., 1991). Therefore, tissue
injury induced by injection of formalin may also elicit a release of
GABA from the terminals of GABAergic bulbospinal neurons or from
GABAergic interneurons via a spino-bulbospinal loop.
The ability of bicuculline to increase pain behaviors in phase 2 suggests that the responses in this phase are normally diminished by a
coincident inhibitory process mediated by GABAA
receptors. Removal of this inhibitory influence by antagonism of
GABAA receptors permits full expression of the
behavioral sequelae to formalin-induced tissue injury. The increase in
pain behaviors in phase 2 is also consistent with the original
interpretation that this phase of the formalin test reflects the
occurrence of central sensitization in the spinal cord (Coderre
et al., 1990; Yamamoto and Yaksh, 1992; Coderre et
al., 1993). Although more recent studies have questioned this
proposal and concluded that pain behaviors in phase 2 depend on
continued activity in primary afferent fibers (Dallel et
al., 1995; Puig and Sorkin, 1995; Taylor et al., 1995; McCall et al., 1996), these two mechanisms are not mutually
exclusive. In fact, continued low-frequency input by C fibers could
induce "wind-up" in the spinal cord during this phase (McCall
et al., 1996; Woolf, 1996b). If pain behaviors in phase 2 were solely dependent on activation of primary afferent fibers, then
bicuculline would be expected not to enhance
formalin-induced pain behaviors in phase 2 just as it failed to
increase these behaviors in phase 1, which more clearly depends on
primary afferent input. The ability of bicuculline to preferentially
increase pain behaviors in phase 2 is not compatible with the proposal
that phase 2 is entirely dependent on primary afferent activity and can
therefore be viewed as additional evidence for the occurrence of
central sensitization in the formalin test.
There are interesting parallels between central sensitization in the
spinal cord and LTP in the hippocampus. For example, LTP (Bliss and
Collingridge, 1993; Nicoll and Malenka, 1995) and central sensitization
(Woolf, 1983; Woolf and Wall, 1986; Woolf et al., 1994) are
each induced by brief, repetitive, high-threshold afferent stimulation.
Activation of NMDA receptors mediates the LTP in certain regions of the
hippocampus (Collingridge and Davies, 1989; Bliss and Collingridge,
1993; Nicoll and Malenka, 1995), as well as the development and
maintenance of central sensitization in the spinal cord (Haley et
al., 1990; Woolf and Thompson, 1991; Neugebauer et al.,
1994). Finally, LTP and central sensitization share a similar
dependence on increases in intracellular Ca++
(Bliss and Collingridge, 1993; Coderre et al., 1993; Woolf,
1996b). Numerous studies now indicate that GABA and
GABAA receptors also play an important role in
the development and induction of LTP. LTP is enhanced in the presence
of GABAA receptor antagonists (Wigström and
Gustafsson, 1983; Hirai et al., 1993; Tomasulo et
al., 1993; Yasui et al., 1993). Indeed, in certain
regions of the central nervous system, antagonism of
GABAA receptors is a prerequisite for the
occurrence of LTP (Bear et al., 1992; Kanter and Haberly,
1993; Watanabe et al., 1995). A similar situation may exist
with respect to the induction of central sensitization in the spinal
cord after the injection of formalin. It was recently reported that
injection of formalin did not elicit central sensitization in the
spinal cord (Xu et al., 1995). This finding was surprising because a wide variety of stimuli induce central sensitization in the
spinal cord, including repetitive electrical stimulation of C-fibers,
topical application of capsaicin or mustard oil and injection of other
inflammatory agents such as kaolin and carrageenan. The inability to
detect central sensitization after the injection of a single, high
concentration of formalin (5%) may actually reflect the large amounts
of GABA released in the spinal cord and the strength of
GABAA receptor-mediated inhibition. It is likely
that under certain conditions of afferent stimulation the amounts of
GABA that are released are sufficient to cause hyperpolarization of
dorsal horn neurons, thereby hindering or preventing the activation of
NMDA receptors by glutamate. In this respect, it is noteworthy that
GABAergic and glutamatergic synapses often exist in close apposition on
neurons in the dorsal horn of the spinal cord (Maxwell et
al., 1995) and so are situated to effectively modulate local membrane potential (Staley and Mody, 1992; Tomasulo et al.,
1993), which is likely to be of greater import for NMDA than non-NMDA receptors (Staley and Mody, 1992). Exclusion of central sensitization as a mechanism that contributes to pain behaviors in phase 2 of the
formalin test is therefore premature until the effects of lower
concentrations of formalin are examined alone and in the presence of
bicuculline to remove opposing inhibitory influences. In this regard,
it has been demonstrated that antagonism of GABAA receptors in the spinal cord enhances central sensitization (Sivilotti and Woolf, 1994).
Relatively little is known about the mechanisms responsible for the
suppression of pain behaviors that occurs 10 to 20 min after the
injection of formalin. The mean firing rate of C-fiber afferents
declines during interphase, which suggests that the quiescent period
results from a diminished afferent input. However, some C-fibers
continue to discharge during this period (Puig and Sorkin, 1995; McCall
et al., 1996). There is also evidence for central modulation
because the interphase period is absent in decerebrate rats (Matthies
and Franklin, 1992). The present finding that i.t. administered
bicuculline increases pain behaviors in interphase provides additional
evidence for central modulation. Pretreatment with 0.3 µg of
bicuculline shifted the concentration-effect curve of formalin in the
interphase to the left to a greater extent (3- to 4-fold) than it did
for phase 2 (approximately 2-fold). This difference suggests that the
inhibitory effects of GABA are largely unopposed by excitatory
mechanisms during the interphase period.
Effects of GABAA receptor agonists in the
formalin test.
In agreement with a previous report (Dirig and
Yaksh, 1995), pretreatment with a GABAA receptor
agonist suppressed formalin-induced pain behaviors in phase 1, as well
as phase 2. The suppression of pain behaviors in phase 1, which are
attributed to direct acute activation of nociceptors by the injection
of formalin and therefore analogous to acute nociception, is consistent
with the antinociceptive effects of isoguvacine and muscimol in other
measures of acute nociception such as the tail-flick and hot-plate
tests (Hammond and Drower, 1984; McGowan and Hammond, 1993). The
effects of muscimol and isoguvacine depended to a certain extent on the
dependent measure used in the formalin test. Isoguvacine produced large decreases in the number of flinches in phase 1 and phase 2, but its
reduction of weighted pain score was less remarkable. Because muscimol
had similar effects, it is likely that this differential effect is
characteristic of GABAA receptor agonists. At
first glance, the large decrease in number of flinches is at odds with the modest decrease in weighted pain score. However, closer examination of weighted pain scores revealed that proportionately greater decreases
occurred in the time spent in category 3 (to nearly 30% of values in
saline-treated rats), than in category 2 (to only 50 to 60% of values
in saline-treated rats). The greater effect observed with use of number
of flinches as the dependent measure may therefore reflect a
specificity of action of GABAA receptor agonists
for inhibition of reflexive measures. Alternatively, because the
therapeutic index between the antinociceptive effects and the adverse
motor effects of i.t. administered GABAA receptor agonists is small (Hammond and Drower, 1984; Dirig and Yaksh, 1995), it
is possible that the robust decrease in number of flinches may be
confounded by an additional effect on motor function that was not
behaviorally apparent. Indeed, Dirig and Yaksh (1995) noted that the
upper CL for muscimol extended into the dose range at which motor
depression occurs.
Summary.
These findings provide strong evidence for a
physiological role of GABA in modulating the behavioral responses to
tissue injury produced by formalin. Specifically, injection of formalin
evokes a coincident "compensatory" release of GABA and activation
of GABAA receptors in the spinal cord. This
release of GABA may contribute to the decrease in pain behaviors that
occurs 10 to 20 min after the injection of formalin (i.e.,
the interphase period). More importantly, the resulting coincident
increase in inhibition may limit the development of central
sensitization by NMDA receptor-dependent mechanisms in the spinal cord
and thereby diminish the magnitude of pain behaviors in phase 2.
![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
-Aminobutyric AcidA
Receptors in Formalin-Induced Nociception in the Rat1
Top
Abstract
Introduction
-aminobutyric acid (GABA) and
GABAA receptors in the spinal cord in the
expression of pain behaviors evoked by injection of formalin in
concentrations ranging from 0.25 to 2.5% in the hindpaw of the rat.
Two approaches were used. The first approach compared the effect of
drug treatment to saline at each concentration of formalin. The second
approach examined the effect of drug treatment on the
concentration-response functions of formalin, i.e., its
EC50. Intrathecal (i.t.) pretreatment with 0.03 to 0.3 µg of bicuculline, a GABAA receptor
antagonist, dose-dependently increased the number of flinches and
weighted pain scores in the interphase and phase 2, but did not alter
responses in phase 1. In the interphase, the EC50
values of formalin for number of flinches or weighted pain score in
bicuculline-pretreated rats were decreased to one-third or one-fourth,
respectively, of their values in saline-pretreated rats. In phase 2, the EC50 values of formalin for number of
flinches or weighted pain score in bicuculline-pretreated rats were
similarly decreased to one-half of their value in saline-pretreated
rats. These results suggest that formalin was a significantly more
noxious stimulus in the presence of bicuculline. Pretreatment with the
GABAA receptor agonists, muscimol (0.3 µg) or
isoguvacine (10 or 30 µg i.t.), significantly decreased the number of
flinches in phase 1 and phase 2, but produced only a marginal decrease
in the weighted pain score at the highest doses. These findings suggest
that there is little tonic activation of GABAA
receptors by GABA in the spinal cord before or immediately after the
injection of formalin. However, approximately 10 min after the
induction of injury by formalin, there is a release of GABA and
activation of GABAA receptors in the spinal cord
that 1) contributes to the period of quiescence between phase 1 and
phase 2 and 2) coincidentally diminishes the magnitude of pain
behaviors in phase 2, possibly by limiting the development of central
sensitization in the spinal cord.
