Antagonism of the Antinocifensive Action of Halothane by Intrathecal Administration of GABA-A Receptor Antagonists

Male Sprague-Dawley rats weighing 250-375 g were instrumented with intrathecal catheters that terminated at the thoracolumbar junction of the spinal cord as described by Yaksh and Rudy. [23]Briefly, rats were anesthetized with 2-3% halothane and, under aseptic conditions, one end of a PE-10 catheter was introduced through the atlantooccipital membrane and threaded 8.6 cm caudally in the spinal cord subarachnoid space. The other end was externalized at the dorsum of the head. Rats were allowed to recover for 1 week after surgery, during which neurologic function and weight were monitored daily. Rats exhibiting motor dysfunction as a consequence of catheter implantation were excluded from the study.

The acute experiments were performed 1 week after intrathecal catheter implantation. Rats were initially anesthetized with halothane in an airtight container. After induction, 1.1-1.5% halothane was administered by a nose cone. The halothane concentration was monitored throughout the experiment by a Datex Capnomac Anesthetic Monitor (Helsinki, Finland) and measured as the average of concentrations sampled at 3 Hz over a 1-min period.

Core body temperature was monitored via a rectal probe and maintained at 37.0-39.0 degrees C for the duration of the experiment using a warm water blanket, a plastic cover, and a heat lamp. Although core body temperature varied among rats, each individual rat was maintained within 0.5 degree C of its core body temperature determined 40-70 min after anesthetic induction. After a 40-min equilibration period at 1.0-1.4% halothane, the rat was tested for a motor response to the application of a 1-cm serrated alligator clip to the proximal third of the tail. Application of the clip evoked a vigorous escape response in awake rats and was judged to be very painful when applied to a fold of the investigator's skin. The clip was oscillated for 1 min or until the rat responded by purposeful movement of one or more of its limbs. In addition to recording the occurrence of nocifensive movement, the latency to movement was measured to the nearest second. Rats in which application of the clip did not elicit nocifensive movement were assigned the cutoff latency of 60 s. If application of the clip evoked a movement, then the halothane concentration was increased by 0.2%. If no response occurred, the concentration of halothane was decreased by 0.2%. Forty minutes were allowed for equilibration at each new concentration of halothane, after which the response to application of the clip was redetermined. During MAC determination, the clip was not applied at intervals less than 40 min. The above process was repeated until two independent move-no move crossovers were obtained. The average of the midpoints obtained in each crossover was defined as the minimum anesthetic concentration (MAC). [4].

After determination of MAC, the halothane concentration was then either maintained at or adjusted to and then equilibrated at the level at which the rat had most recently responded to stimulation. The rat was then tested to confirm its responsiveness to application of the clip at this concentration of halothane. Within 15 min of the confirmatory response, either saline, 0.3 micro gram bicuculline methiodide, 0.3 micro gram or 1.0 micro gram picrotoxin, 1.0 micro gram strychnine hemisulfate, or 30 micro gram CGP 35348 was injected intrathecally. Five minutes later, the rat's response to application of the alligator clip was redetermined. The concentration of halothane was then increased by 0.2% and responsiveness was redetermined at 7-min intervals for the subsequent 35-min period. In 14 rats, the concentration of halothane was not increased, but remained at the sub-MAC concentration, and responsiveness was redetermined as described earlier to evaluate the effects of saline or antagonist on baseline responsiveness. Each rat received only one dose of a drug and the identity of the drug was unknown to the person testing the rat. At the conclusion of the experiment, rats were killed with an overdose of chloral hydrate or pentobarbital, exsanguinated, and the location and patency of the catheter was verified by visual inspection and injection of india ink.

Although the initial determination of MAC was made by conventional methods, pharmacodynamic factors precluded a conventional redetermination of MAC after intrachecal drug administration. MAC determination requires that two independent move-no move crossovers be obtained, a process that can take 3-4 h in the case of halothane with its 40-min equilibration time. This requirement was at odds with the comparatively short duration of action of intrathecally administered drugs. In the case of the agents used in this study, onset to effect occurs within 5 min with peak effect observed by 15 min after administration. [15,21]The effect of these drugs is substantially diminished 40 min after administration. Although continuous infusion of the antagonists would enable a conventional determination of MAC, this approach was not attempted. It was considered unlikely that the doses (both molar amount and volume) of the antagonists in the spinal cord subarachnoid space could be titrated to levels that (1) were consistently just sufficient for receptor antagonism for a period of 3-4 h, (2) did not eventually redistribute beyond the lumbar segments of the spinal cord, and (3) did not eventually accumulate to concentrations that produce allodynia or seizures, or alter baseline response latency. Thus, it was necessary for the purposes of this study to forgo a determination of the change in MAC and rather examine the drug effects on the decrease in motor responses produced by a small increase in halothane concentration from just below MAC to just above MAC.

All drugs were obtained from Sigma Chemical (St. Louis, MO), with the exception of CGP 35348, which was obtained from Ciba-Geigy (Basel, Switzerland). The solutions were freshly prepared, passed through a 0.2-micro meter filter and their pH was adjusted to a range of 6.9-7.2. Drugs were administered in a volume of 10 micro liter followed by 10 micro liter saline to clear the catheter. The doses of antagonists were based on their ability to antagonize the effects of their respective receptor agonists in the spinal cord of the rat without altering responsiveness to noxious or non-noxious stimuli. For example, 0.3 micro gram bicuculline produces a 3.4-fold rightward shift in the dose-response curve of intrathecally administered isoguvacine but does not alter response latency to noxious thermal stimuli or produce touch-evoked allodynia in awake rats [15](unpublished observations, DL Hammond). Intrathecal administration of 30 micro gram CGP 35348 shifts the dose-response curve of intrathecally administered baclofen tenfold to the right but does not alter response latency to noxious thermal stimuli or produce touch-evoked allodynia [21,22](unpublished observations, DL Hammond). Analogous information was not available for either strychnine or picrotoxin. Therefore, the doses of these antagonists were based on literature reports, and confirmed in preliminary dose-ranging experiments, of doses that were submaximal for the production of touch-evoked allodynia and overt "pain" or motor behaviors such as scratching, vocalization, or myoclonic twitches after intrathecal administration in the rat. In the case of picrotoxin, this dose corresponded to 1.0 micro gram [24](unpublished observations, DL Hammond). In the case of strychnine, intrathecal administration of 2.8 micro gram was reported not to produce significant biting, scratching, twitching, or vocalization in the rat. [24]However, in our preliminary studies, 1.5 or 3.0 micro gram strychnine produced myoclonic seizures, biting of the flanks, and distress vocalization; the intensity was dose-dependent in nature. The dose of 1.0 micro gram was thus chosen because it was just subthreshold for these effects and for touch-evoked allodynia (unpublished observations, DL Hammond).

The effects of the GABA or glycine receptor antagonists on the latency to response were compared to that of saline by two-way analysis of variance for repeated measures. Post hoc comparisons of individual mean values were made by Newman-Keuls test. Fisher's exact test was used to compare the percentage of rats in the saline- and drug-treated groups that moved in response to application of the alligator clip. A P value of less than or equal to 0.05 was considered significant.

During early experiments, it became evident that core temperature influenced the probability of movement in response to application of the alligator clip. Examination of the relationship between core temperature and response to application of the alligator clip during MAC determination indicated that rats were more likely to move at warmer temperatures than at colder temperatures. [25]As shown in Figure 1, core temperature had the greatest effect on the probability of nocifensive movement at intermediate concentrations of halothane. At 1.20-1.39% halothane, rats with temperatures of 37.0-37.5 degrees C moved in only 23% of trials whereas rats with a core temperature of 38.5-39.0 degrees C moved in 75% of trials. Because of this trend, core body temperature was recorded at the time of each stimulus trial and was maintained at a steady level throughout the experiment. The mean temperature deviation over time within each treatment group was less than 0.1 degree C. Two-way analysis of variance for repeated measures showed no significant difference in core body temperature among the different treatment groups and no significant difference within each treatment group over time. Under conditions in which core body temperature was controlled within a range of 37.0-39.0 degrees C, the mean MAC of halothane was 1.10+/-0.02 (n = 49) and did not differ among drug treatment groups (P > 0.2). This value is in good agreement with previous reports. [6,26,27].

Responsiveness to application of the alligator clip was redetermined 5 min after the intrathecal administration of saline or antagonist and before the concentration of halothane was increased, i.e., while the rats were still at a sub-MAC concentration of anesthetic. Within treatment group comparisons revealed that intrathecal administration of saline did not alter either the latency to respond or the percentage of rats that moved in response to the alligator clip as compared to its baseline value (Figure 2(A and B)). Similarly, intrathecal injection of 0.3 micro gram picrotoxin (Figure 2(A and B)) or 1.0 micro gram strychnine (Figure 3(A and B)) did not alter the latency to response or the percentage of rats that moved in response to application of the alligator clip as compared to their respective baseline values. In rats that received 30 micro gram CGP 35348, a modest increase (P < 0.01) in response latency, but no decrease in the percentage of rats that moved occurred (Figure 3(A and B)). This increase could be attributed to one rat whose latency increased to 55 s at this one time. However, in rats receiving intrathecal injections of either 0.3 micro gram bicuculline or 1.0 micro gram picrotoxin a significant decrease in response latency occurred 5 min later as compared to their respective baseline values (Figure 2(A and B)). Importantly, between-group comparisons indicated that none of the antagonist treatment groups differed from the saline control group with respect to the latency to movement or the percentage of rats that moved at either the baseline timepoint or 5 min after intrathecal injection.

Nonetheless, these results prompted an ancillary study to further examine whether bicuculline or picrotoxin decreased baseline response latency when administered by themselves to a rat maintained at a sub-MAC concentration of halothane and whether repetitive application of the alligator clip induced sensitization in saline-treated rats. For this experiment, either saline, 0.3 micro gram bicuculline, or 1.0 micro gram picrotoxin was administered intrathecally and responsiveness to application of the alligator clip was redetermined for the subsequent 40 min in the absence of an increase in halothane concentration, i.e., while the rats remained at a sub-MAC concentration of halothane. Although within-treatment group comparisons to baseline latency indicated that latency to movement decreased 19-33 min after intrathecal administration of picrotoxin (P < 0.05), latency to movement was similarly decreased 26-40 min after the intrathecal administration of saline (P < 0.05). No significant change in latency to movement occurred after intrathecal administration of picrotoxin. Importantly, between-group comparisons indicated that response latencies in picrotoxin- or bicuculline-treated rats did not differ from those of saline-treated rats at any time (Figure 4).

After redetermination of responsiveness 5 min after intrathecal drug administration, the concentration of halothane was then increased by approximately 0.2%. The mean increase in halothane concentration for all animals was 0.22+/-0.01% and did not differ among the treatment groups (P > 0.5). In saline-treated rats, the latency to movement increased and the percentage of rats responding to the alligator clip decreased at 12 min after the increase in halothane concentration (Figure 2(A and B)). These effects stabilized by 19 min at which time none of the saline-treated rats moved in response to oscillation of the alligator clip for 1 min. The 12-and 19-min postinjection time points corresponded to 7 and 14 min, respectively, after the increase in halothane concentration.

The antinocifensive effects of the small increase in the concentration of halothane were attenuated in rats pretreated with either bicuculline or picrotoxin (Figure 2). In rats pretreated with 0.3 micro gram of bicuculline, the response latency determined at 12 and 19 min postinjection (corresponding to 7 and 14 min after the increase in halothane) was significantly less than in saline-treated rats (Figure 2(A)). In rats pretreated with 0.3 micro gram picrotoxin, the response latency determined after the increase in halothane concentration was significantly less than in saline-treated rats at 12 and 19 min (Figure 2(A)). Increasing the dose of picrotoxin to 1.0 micro gram resulted in a more prolonged attenuation of the effects of halothane. Response latency was significantly less than that in saline-treated rats at 19, 26, and 33 min (Figure 2(A)). By comparison, intrathecal pretreatment with either 1 micro gram strychnine, a glycine receptor antagonist, or 30 micro gram CGP 35348, a GABA^Breceptor antagonist, did not attenuate the antinocifensive effects of halothane (Figure 3(A and B)).

The antinocifensive effect of the GABA^Areceptor antagonists also was evident from comparisons of the percentage of rats that moved at each time point (Figure 2(B) and Figure 3(B)). At 19 min postinjection, a significantly greater percentage (P < 0.05) of rats treated with 0.3 micro gram picrotoxin moved in response to application of the alligator clip when compared to saline-treated rats (Figure 2(B)). Increasing the dose of picrotoxin to 1.0 micro gram increased the duration of antagonism. Thus, between 19 and 33 min postinjection, 80-100% of rats treated with 1.0 micro gram picrotoxin moved in response to the alligator clip, whereas none of the saline-treated rats moved in response to this stimulus at these times (Figure 2(B)). The effect of bicuculline on the percentage of rats that moved in response to application of the alligator clip was not statistically different from that of saline (P < 0.1). However, the inability to detect a significant effect of bicuculline is likely attributable to the poor power of the analysis (0.46) and therefore represents a type II statistical error (failure to detect a difference that exists). Neither strychnine nor CGP 35348 significantly increased the percentage of rats that moved in response to the alligator clip (Figure 3(B)).

The current finding that intrathecally administered GABA^Areceptor antagonists attenuate the antinocifensive effects of halothane provides direct evidence for a spinal site of action of inhalational anesthetics, as well as insight into a pharmacologic basis for halothane's actions within the spinal cord.

Several studies now support the contention that the spinal cord is an important site for the antinocifensive effect of inhalational anesthetics. A variety of anesthetic agents, including isoflurane, barbiturates, 2,6-diisopropyl phenol, and ether, are reported to depress monosynaptic and polysynaptic reflex activity in the isolated or intact spinal cord. [28-31]Furthermore, acute spinal transection does not alter the antinocifensive potency of isoflurane in the rat. [8]The current finding that intrathecally administered GABA^Areceptor antagonists attenuate the antinocifensive effect of halothane provides additional, direct support for a spinal site of action of inhalational anesthetics. This conclusion is based on the premise that the actions of the receptor antagonists used in this study were limited to the spinal cord. Substantial evidence supports a spinal site of action for intrathecally administered drugs during the 40-min period after their administration. This evidence, which includes analyses of the spinal and supraspinal distribution of radiolabeled compounds after intrathecal administration and studies of the dependence of drug effect on the segmental level of administration, has been reviewed at length. [32]For example, less than 0.5% of [sup 3 Hydrogen]-naloxone or [sup 14 Carbon]-urea, and less than 0.1% of [sup 3 Hydrogen]-morphine is found at supraspinal levels after intrathecal injection. [23,33,34]It therefore is unlikely that more than 1.5 ng of the 0.3 micro gram bicuculline or 5 ng of the 1.0 micro gram picrotoxin administered in the current study would be expected to reach supraspinal sites. These amounts are insufficient to alter nociceptive responsiveness when administered directly to supraspinal sites. [35-37]Additionally, local application of GABA^Areceptor antagonists at several supraspinal sites produces antinociception, rather than nociception. [35-39]Thus, in the unlikely event that pharmacologically relevant amounts (40-100 ng) of bicuculline or picrotoxin were to reach supraspinal sites, these drugs would be expected to enhance, rather than attenuate the antinocifensive effects of halothane.

Although anesthetic actions within the spinal cord appear to be sufficient to block nocifensive movement, [40]an additional action of inhalational anesthetics at supraspinal structures cannot be excluded. For example, activation of GABA^Areceptors in the periaqueductal gray or raphe magnus results in an enhanced responsiveness to noxious stimuli, [35-37,41]presumably by inhibiting the activity of neurons in these nuclei that project to the spinal cord and inhibit nociceptive transmission. [42-44]As halothane and other inhalational anesthetics potentiate the actions of GABA at the GABA^Areceptor, [10]it is likely that halothane can act supraspinally at sites in the periaqueductal gray or raphe magnus to inhibit the activity of these neurons and so enhance spinal nociceptive transmission by decreasing tonic descending inhibition. Indeed, nociceptive inhibitory neurons in the raphe magnus are inhibited by supra-MAC concentrations of isoflurane. [45]By comparison, in the spinal cord, enhancement of GABAergic transmission results in antinociception. Selective antagonism of the actions of halothane at spinal GABA^Areceptors may concomitantly permit the expression of or unmask the pronociceptive actions of halothane at supraspinal sites. Thus, the ability of intrathecally administered GABA^Areceptor antagonists to attenuate the antinocifensive effects of inhalational anesthetics may reflect both a direct antagonism of the actions of halothane at GABA^Areceptors in the spinal cord and an indirect unmasking of a pronociceptive action of halothane at supraspinal GABA^Areceptors. Clarification of the contribution of supraspinal sites of action must await complimentary studies of the effects of supraspinally administered GABA^Aand GABA^Breceptor ligands on the antinocifensive potency of halothane.

This study used two GABA^Areceptor antagonists of different physicochemical structure and different mechanisms of action to assess the contribution of GABA^Areceptors to the antinocifensive effects of halothane. Bicuculline is a competitive antagonist of GABA at the GABA^Areceptor, whereas picrotoxin is a non-competitive antagonist of the chloride channel associated with the GABA^Areceptor. [46]Both bicuculline and picrotoxin effectively antagonized the antinocifensive effect of a small increase in halothane concentration. In contrast, antagonists for either the closely related glycine receptor or the GABA sub B receptor were ineffective. These findings suggest that the antinocifensive effects of threshold concentrations of inhalational anesthetics result from an enhancement of the actions of GABA at GABA^Areceptors within the spinal cord.

Intrathecal administration of GABA^Areceptor antagonists similarly antagonized the suppression of nociceptive reflexes by systemic barbiturates. [47]However, the large doses of bicuculline (25 micro gram) or picrotoxin (12 micro gram) used in that study induce seizures when administered alone. [47]Thus, the antagonism of the antinocifensive effects of the barbiturate may simply have been due to gross alterations of excitability in the spinal cord. This confounder was not an issue in the current study because the doses of GABA^Areceptor antagonists were carefully chosen to be just sufficient for antagonism of their receptor, yet too low to enhance sensitivity to innocuous mechanical stimuli, augment baseline nociceptive motor responses, or produce myoclonic twitches in awake rats. Although intrathecal administration of GABA^Areceptor antagonists can produce allodynia and hyperalgesia, the minimum doses that produce such effects are threefold- to tenfold higher [15](unpublished observations, DL Hammond) than those used to antagonize halothane in the current study. The allodynia, hyperalgesia, and spontaneous myoclonic twitches that are commonly reported after intrathecal administration of GABA^Areceptor antagonists in unanesthetized rats, [19,20]occur after administration of doses of 30-60 micro gram, 100- to 200-fold higher than the doses used in the current study. Administration of these high doses also causes an increase in the spontaneous activity of dorsal horn cells. [48].

The lack of effect of the low doses of GABA^Areceptor antagonists used in this study was further verified in rats anesthetized with sub-MAC concentrations of halothane, a condition in which the actions of GABA at the GABA^Areceptor may be enhanced [49]and so might be more sensitive to antagonism by bicuculline or picrotoxin. Although initial experiments indicated that the latency to nocifensive movement in rats anesthetized with sub-MAC concentrations of halothane was significantly decreased 5 min after intrathecal administration of 0.3 micro gram bicuculline or 1.0 micro gram picrotoxin, this observation was not replicated in an ancillary study. In this study, rats treated with 0.3 micro gram bicuculline or 1.0 micro gram picrotoxin did not differ from saline-treated rats at any time before or after drug administration. The most parsimonious explanation for this finding is that there is little or no potentiation of GABA^Areceptor-mediated transmission in the lightly anesthetized condition and that an increase in the tonic activation of GABA^Areceptor-mediated transmission occurs only at anesthetic levels equal to or greater than MAC. Taken together, these findings indicate that the antagonism of the antinocifensive effects of halothane by intrathecal administration of low doses of bicuculline or picrotoxin cannot be ascribed to a nonspecific excitation or a general reduction in inhibition in the spinal cord, but rather to selective antagonism of the GABA^Areceptor.

Although Barany [50]suggested that general anesthetics suppress central nervous system activity in proportion to the number of synapses in the pathway, general anesthetics preferentially suppress monosynaptic reflexes. [51]Thus, the effect of general anesthetics on somatomotor transmission is more likely due to the differential vulnerability of specific synapses (i.e. the Ia afferent to alpha-motoneuron synapse vs. the polymodal nociceptor to dorsal horn cell synapse) to modulation by general anesthetics than to the cumulative number of synapses used. [51,52]The current results suggest that synapses with GABA^Areceptors, located either on the presynaptic terminal or on the postsynaptic membrane, may be selectively targeted for enhancement by halothane at concentrations near the threshold for blocking nocifensive movements. The depression of somatomotor activity by general anesthetics in the spinal cord may be mediated by inhibition of excitatory neurotransmitter release via GABA^Areceptors situated presynaptically on the terminals of Ia afferents, [53-56]low threshold primary afferents, [57]and myelinated nociceptors. [58,59]In addition, GABA^Areceptors located postsynaptically may mediate inhibition of dorsal horn cells, [60,61]Ia interneurons, [62]and motoneurons. [63-65]Thus, activation or potentiation of GABA^Asynaptic transmission, at numerous sites within the spinal cord, is likely to contribute to the suppression of somatomotor transmission by inhalational general anesthetics.

In conclusion, the current results provide direct evidence that the spinal cord is an important site of action for the antinocifensive action of inhalational anesthetics. These results further indicate that enhancement of the action of GABA at spinal GABA^Areceptors is one mechanism by which halothane exerts its antinocifensive actions.