Tooth Pulp– and Facial Hair Mechanoreceptor–evoked Responses of Trigeminal Sensory Neurons Are Attenuated during Ketamine Anesthesia 

Experiments were conducted in four intact, unanesthetized, chronically instrumented cats, which were implanted under deep gaseous anesthesia (45–60% N²0 in a 1.5–2.5% halothane/oxygen mixture) with a head-restraining device and electrodes for monitoring behavioral states. 11,19Electrodes were also implanted into the frontal sinus (electroencephalogram [EEG]), lateral geniculate nucleus of the thalamus (pontogeniculooccipital [PGO] wave activity), the orbital plate (electrooculogram), and neck muscles (electromyogram [EMG]) and were used for monitoring behavioral state. Through the use of these electrodes, the behavioral state of quiet wakefulness could be clearly differentiated from sleep or anesthesia by the presence of a desynchronized, low-voltage cortical EEG waveform pattern that was also accompanied by a tonic EMG activity, and a paucity of eye movements or PGO wave activity.

In one cat, tripolar strut electrodes were chronically implanted in the (contralateral) ventrobasal thalamus (Horsley-Clarke coordinates [HC]: A 6–8, L 4–6, H -1 to 2) to permit antidromic activation of trigeminothalamic (TGT) neurons. 12An indwelling jugular catheter was chronically implanted and used for the intravenous administration of ketamine. Silver–silver chloride dentinal pit electrodes were implanted in the mandibular and maxillary canines for electrical stimulation of tooth-pulp afferents. 11,15All surgical and chronic restraint procedures reported complied with international 20,21and institutional (University of British Columbia Animal Care Committee) guidelines.

During recording sessions, cats were loosely wrapped in a soft canvas bag and their heads painlessly maintained in a stereotaxic position. Behavioral state was constantly monitored by recording EEG, electrooculogram, PGO, and nuchal EMG activities. Behavioral state was scored using previously established criteria. 11,12,15Cats normally cycled between quiet drowsy wakefulness and quiet sleep; in addition, they readily entered episodes of active sleep. Using this preparation, we previously reported that the tooth pulp–evoked responses of feline TSNC neurons are suppressed during active sleep. 11,22Accordingly, the behavioral state of quiet wakefulness served as a control state for testing the effects of ketamine.

The extracellular spike activity of neurons located within the rostral TSNC was recorded via  tungsten electrodes (2 MΩ) using an AC-coupled amplifier (A-M Systems, Carlsborg, WA, model 1800, bandpass, 0.3–10.0 KHz, 1,000×). The boundaries of the rostral TSNC were demarcated using extracellularly recorded orthodromic mass action potentials evoked by electrical stimulation of the inferior alveolar nerve and tooth pulps as well as antidromic mass action potentials evoked via  stimulation of the digastric and masseter muscles and facial nerve (fig. 1). 11Poststimulus time histograms depicting cumulative responses of TSNC neurons were constructed from consecutive canine tooth-pulp stimuli at threshold (0.2 ms, 5–100 μA) and suprathreshold (approximately 125% of threshold;fig. 1) intensities. Recorded TSNC neurons were categorized as being of the stimulus intensity–independent or stimulus intensity–dependent type by their responses to graded electrical tooth-pulp stimuli (figs. 1A and 1B; see Cairns et al.  11). The response characteristics of rostral TSNC neurons may be associated with their axonal projection; for example, trigeminothalamic tract (TGT) neurons exhibit stimulus intensity–independent responses to electrical stimulation of the tooth pulps. 11,12,23,24 

FHM-evoked responses of TSNC neurons were also investigated by delivering innocuous air-puff stimuli (10–50-ms puff, 0.5 Hz) to facial receptive fields. In one cat, FHM-evoked TSNC neurons were further identified as TGT neurons based on the following criteria in response to low intensity (0.2 ms, 500 μA) thalamic stimuli:(1) constant latency (variability less than 0.2 ms), (2) high frequency following (≥ 333 Hz), and (3) collision with ongoing or peripherally evoked action potentials. 15,19,25 

Cats rested quietly in the stereotaxic setup without exhibiting any signs of distress or discomfort during stimulation of the teeth, face, or thalamus. In our experiments, we observed a gradual recovery from ketamine anesthesia (2.2 mg/kg), which was uneventful and allowed trigeminal sensory neuron activity to be monitored for periods exceeding 2 h.

Prior to the administration of ketamine, a control response to tooth pulp or FHM stimuli was obtained during the state of quiet wakefulness. Then, normal saline (0.5 ml) was slowly injected intravenously over 1 min and a second control response collected. Finally, a relatively low dose of ketamine HCl (Ketalean, MTC Pharmaceuticals, Cambridge, Ontario, Canada; 2.2 mg/kg diluted to 0.5 ml in normal saline) was slowly injected over 1 min and neuronal activity monitored at regular intervals for 1 h after drug administration. A dose of 2.2 mg/kg ketamine was chosen because it is the lowest dose employed in previous studies of lumbar dorsal horn neurons that reliably provides 5–10 min of surgical anesthesia followed by a 30–60-min recovery period. 3,26A maximum of two experiments was performed per week. The minimum time period between successive drug administrations during each week was 48 h.

All data were videotaped, and analyses were performed “off-line” using computerized data acquisition software (A/DVANCE Fine Science Tools, Vancouver, Canada). Poststimulus time histograms were constructed from 50 consecutive responses and the mean activity in spikes/stimulus calculated. Evoked activity was sampled every 5 min over the first 30 min after drug infusion, and then every 10 min thereafter for a total of 60 min. The magnitude of the control response at time zero was equated to 100% and subsequent values were calculated based on this value. Ketamine-related changes in neuronal excitability were assessed for each neuron by comparing the mean activity during wakefulness with that obtained during and upon recovery from ketamine anesthesia. 11A repeated-measures analysis of variance (ANOVA) and post hoc  Dunnett's method were applied to determine whether ketamine exerted significant changes in the magnitude of evoked responses over time. A paired Student t  test was used to investigate differences in latency to onset and the slope of the stimulus–response relationship between control and ketamine. A Mann–Whitney test was used to compare the relative magnitude of ketamine-related suppression between air puff– and tooth pulp–evoked neuronal responses. The α value for all tests was set at 0.05. All values are reported as means ± SD.

Indices of behavioral state were monitored throughout each experiment. Control neuronal activity was collected during the state of quiet wakefulness, which was characterized by desynchronized, low- voltage cortical EEG activity, tonic nuchal EMG activity, and a paucity of eye movement or PGO wave activity. 11,12,19In contrast, the anesthetic state induced by ketamine was clearly differentiated from quiet wakefulness primarily by the presence of high-voltage EEG activity with characteristic intermittent “spike-like” slow waves, and by a tonic nuchal EMG of decreased amplitude. 8,9In each animal, ketamine's action on EEG and EMG reached a peak effect 5 min after administration that recovered within 20 min. Upon recovery from ketamine, the behavioral state resembled predrug wakefulness. A representative trace of EEG, electrooculogram, PGO, and EMG activity during wakefulness, ketamine anesthesia, and recovery is shown in figure 2.

The effect of ketamine anesthesia was examined on a total of 14 tooth pulp–evoked TSNC neurons (cat 1: n = 2; cat 2: n = 6; cat 3: n = 4; cat 4: n = 2). Six of these cells were characterized as stimulus intensity–dependent neurons, and the remaining eight neurons were of the stimulus intensity–independent type. The response magnitude and latency to onset for these tooth pulp–evoked TSNC neurons during the control state of quiet wakefulness did not differ from that reported previously. 11 

During the state of quiet wakefulness (W), baseline responses to threshold stimuli applied to tooth pulps were not affected by control injections of saline. However, the mean response magnitude was maximally suppressed by 50%, 5 min after the intravenous administration of ketamine (mean activity^W: 1.6 ± 0.4 spikes/stimulus; mean activity^KETAMINE: 0.8 ± 0.7 spikes/stimulus; F = 7.924, P < 0.001 on repeated-measures ANOVA). The ketamine-related suppression of tooth pulp–evoked activity was not associated with a change in the mean latency to onset of the first action potential (mean latency^W: 5.9 ± 0.7 ms; mean latency^KETAMINE: 5.9 ± 0.6 ms;P > 0.05, paired Student t  test). Spike activity recovered to predrug baseline levels approximately 30 min after intravenous administration of ketamine.

The graded responses of stimulus intensity–dependent TSNC neurons provided the opportunity to examine the effect of ketamine on trigeminal neuronal responsiveness to both threshold and suprathreshold stimuli. The response characteristics of a stimulus intensity–dependent TSNC neuron before, during, and after recovery from ketamine anesthesia are illustrated in figure 2. Overall, tooth pulp–evoked responses of stimulus intensity–dependent neurons were maximally suppressed 5 min after the administration of ketamine at threshold (mean activity^W: 1.9 ± 0.5 spikes/stimulus; mean activity^KETAMINE: 0.7 ± 0.7 spikes/stimulus). Compared with baseline, the mean response magnitude was significantly suppressed at 5, 10, and 15 min after ketamine (F = 2.183, P = 0.034 on repeated-measures ANOVA, P < 0.05 by Dunnett's method).

The suprathreshold responses of stimulus intensity–dependent neurons were also maximally suppressed 5 min after ketamine (mean activity^W: 4.1 ± 1.0 spikes/stimulus; mean activity^KETAMINE: 1.9 ± 0.2 spikes/stimulus). However, ketamine's suppressive action was longer for suprathreshold versus  threshold responses. Indeed, the mean response magnitude remained significantly suppressed, relative to baseline responses for 30 min, or twice as long as the mean response magnitude at threshold (F = 7.276, P < 0.001 on repeated-measures ANOVA, P < 0.05 by Dunnett's method;fig. 3).

We previously reported 11that the stimulus–response curve of stimulus intensity–dependent TSNC neurons is linear (fig. 1). To further investigate the effect of ketamine on the response characteristics of this type of TSNC neuron, the slope of the stimulus–response curve for each cell was calculated as the difference between suprathreshold and threshold responses divided by the difference in stimulus intensity. Collectively, the slope of the stimulus–response relationship for all six neurons was significantly decreased (mean slope^W: 2.0 ± 0.5 spikes/μA; mean slope^KETAMINE: 1.0 ± 0.3 spikes/μA, P < 0.05 by paired Student t  test). Ketamine suppressed both threshold and suprathreshold responses. However, the prolonged recovery of suprathreshold responses and the decrease in the slope of the stimulus–response curve suggests that ketamine may exert a preferential action on suprathreshold tooth pulp inputs. 6,7 

Although ketamine anesthesia was accompanied by a marked reduction in response magnitude, it did not change the mean latency to onset of the first action potential at threshold (mean latency^W: 6.5 ± 2.4 ms; mean latency^KETAMINE: 6.8 ± 2.4 ms;P > 0.05, paired Student t  test) or suprathreshold (mean latency^W: 5.0 ± 1.2 ms; mean latency^KETAMINE: 5.4 ± 1.5 ms;P > 0.05, paired Student t  test) stimulus intensities.

The action of ketamine on a stimulus intensity–independent neuron is presented in figure 4. Ketamine reversibly suppressed the tooth pulp–evoked response of this neuron. Overall, the threshold tooth pulp–evoked responses of eight stimulus intensity–independent neurons were maximally suppressed by 33% 5 min after ketamine administration (TSNC neurons; mean activity^W: 1.2 ± 0.6 spikes/stimulus; mean activity^KETAMINE: 0.8 ± 0.6 spikes/stimulus;fig. 5). Similar to stimulus intensity–dependent neurons, the mean response magnitude was significantly suppressed, relative to baseline responses, at 5, 10, and 15 min after ketamine administration (F = 4.336, P < 0.001 on repeated-measures ANOVA, P < 0.05 by Dunnett's method). The relative magnitude of ketamine's peak suppression of stimulus intensity–independent neurons was not different from that of stimulus intensity–dependent neurons at threshold (P > 0.05, Mann–Whitney test;figs. 3 and 5). Administration of ketamine did not change the mean latency to onset of the first action potential of these stimulus intensity–independent neurons (mean latency^W: 5.7 ± 2.8 ms; mean latency^KETAMINE: 5.9 ± 3.1 ms;P > 0.05, paired Student t  test).

Experiments were performed on eight TSNC neurons (cat 3: n = 3; cat 4: n = 5) that responded to air-puff stimuli directed at facial receptive fields. None of these neurons responded to electrical stimulation of the tooth pulps; however, two of these neurons were identified as TGT neurons. An example of ketamine's action on an air puff–driven TSNC neuron is shown in figure 6. Note that intravenous administration of ketamine, but not saline, reversibly suppressed the air puff–evoked response of this neuron by 65%. This particular neuron was subsequently identified as a TGT neuron (fig. 1B).

The air puff–driven responses of all eight TSNC neurons (n = 6 TSNC, and n = 2 TGT neurons) were maximally suppressed (range, 25–92%) 5–10 min after the administration of ketamine. Overall, air puff–driven responses of all eight TSNC neurons were suppressed by 45%, 5 min after the administration of ketamine (fig. 7; mean activity^W: 2.5 ± 1.7 spikes/stimulus; mean activity^KETAMINE: 1.4 ± 1.4 spikes/stimulus). Ketamine exerted a suppressive action that was similar in time course to tooth pulp–evoked neurons (fig. 8). Compared with baseline responses, the mean response magnitude was significantly suppressed at 5, 10, and 15 min after ketamine administration (F = 4.336, P < 0.001 on repeated-measures ANOVA, P < 0.05 by Dunnett's method).

In agreement with previous studies, 8,9the systemic administration of a relatively low dose of ketamine in chronically instrumented cats rapidly and reversibly induces a behavioral state of dissociative anesthesia that is characterized by distinct intermittent spike-like slow waves in the cortical EEG and reduced nuchal EMG activity compared with preceding episodes of quiet wakefulness. In addition, the state of ketamine anesthesia is associated with a marked reversible suppression of both tooth pulp– and FHM-evoked responses of rostral TSNC neurons.

A major limitation of the present study is that ketamine was administered systemically, and thus the site of ketamine action in exerting the suppressive action on TSNC neurons is not clear. The analgesic properties of ketamine are thought to be caused, in part, by a noncompetitive channel blockade of the postsynaptic N -methyl-D-aspartate (NMDA) receptor. 4,26–28Systemically administered ketamine (up to 5 mg/kg) has been reported to antagonize the NMDA but not non-NMDA excitatory amino acid–evoked responses in spinal cord dorsal horn neurons. 4,26However, other actions of ketamine are also noteworthy with respect to its putative postsynaptic actions in vivo . These same relatively low doses of ketamine are known to antagonize excitation of Renshaw cells by acetylcholine, 4and there is other in vitro  evidence to indicate that ketamine is also a nicotinic cholinergic receptor antagonist. 29In addition, it is important to note that ketamine may affect γ-aminobutyric acid, 30serotonergic, 31noradrenergic, 32and muscarinic cholinergic 33systems.

The identity of the excitatory neurotransmitters responsible for mediating synaptic transmission between orofacial afferents and trigeminal sensory neurons is not known. However, tooth pulp primary afferent terminals display glutamate-like immunoreactivity, suggesting that an excitatory amino acid neurotransmitter such as glutamate or aspartate may be released in response to tooth-pulp stimulation. 34Neurons comprising the TSNC and, in particular, the TGT are reactive with antibodies raised against both NMDA and non-NMDA receptor subtypes. 35Microiontophoretic recording experiments performed in vivo  have shown that local application of both NMDA and non-NMDA receptor agonists excites TSNC and TGT neurons. 36,37Thus, the ketamine-related suppression of orofacial-evoked activity observed herein might be explained by a blockade of postsynaptic NMDA receptor channels located on the soma of TSNC neurons.

We have suggested that stimulus intensity–dependent neurons may act as local interneurons in the rostral TSNC. 11The work of Katakura and Chandler indicates that digastric motoneuronal responses driven by tooth pulp stimuli are blocked by NMDA antagonists and antidromically identified premotor interneurons are excited by microiontophoretically applied NMDA. 38The findings of Boucher et al.  39that microinjection of ketamine (0.02 mg/kg) as well as the competitive NMDA receptor antagonist 2-amino-5-phosphonovalerate into the TSNC of awake rats suppresses the tooth pulp–evoked jaw-opening reflex corroborate the aforementioned findings. Collectively, these studies support the suggestion that antagonism of NMDA receptors by ketamine could contribute to a suppression of the synaptic transmission between tooth pulp afferents and TSNC neurons. Consistent with previous findings, 6,7, ketamine also reduced the slope of the stimulus–response curve of all stimulus intensity–dependent neurons tested. Moreover, the recovery of suprathreshold responses of these cells from ketamine anesthesia was also prolonged, lasting up to 30 min after intravenous bolus administration. These data indicate that the drug attenuates to a greater extent high threshold rather than low threshold inputs to these cells. These data may be relevant to partially explaining the analgesic properties of low doses of ketamine in conscious patients suffering from various pain syndromes. 1,2 

Ketamine also decreased the tooth pulp–evoked activity of stimulus intensity-independent TSNC neurons. Perhaps as many as one half of all rostral TSNC neurons recorded in vivo  that exhibit stimulus intensity–independent response characteristics similar to ours 11may project to the ventrobasal thalamus. 12,23,24Thus, ketamine's action on stimulus–independent TSNC neurons indicates that transmission of tooth-pulp information to the thalamus and, ultimately, to the cerebral cortex is attenuated during the state of ketamine anesthesia at the level of the TSNC. This suggestion is supported by the finding that the amplitude of tooth pulp–evoked somatosensory potentials recorded in the brainstem, thalamus, and cerebral cortex of awake-behaving monkeys are suppressed after ketamine (5–15 mg/kg) administration. 40 

Under our experimental conditions, the state of ketamine anesthesia was also found to be associated with decreased FHM-evoked activity in both rostral TSNC and TGT neurons. This contrasts with results reported for experiments performed in acute decerebrate cat preparations, in which local or systemic administration (2–10 mg/kg) of ketamine did not affect LTM-evoked responses of unidentified spinal cord dorsal horn neurons. 3,4Furthermore, in the single study performed on intact, behaving cats, Collins 41concluded that ketamine (10–20 mg/kg intravenously) had no effect on non-noxious mechanoreceptor-evoked responses of unidentified lumbar dorsal horn neurons. However, Collins reported that ketamine exerted either suppressive or facilatory actions on the non-noxious responses in 25% of the cells examined. Therefore, it is conceivable that the effect of ketamine and other NMDA antagonists on low-threshold mechanoreceptor–evoked responses of spinal cord dorsal horn neurons might depend on the type of neuron tested (i.e.,  tract cells vs.  segmental interneurons) or on the level of the neuraxis (i.e. , trigeminal vs.  lumbar sensory neurons).

With regard to the possibility that ketamine may exert different effects on local interneurons versus  tract neurons, TGT neurons are known to display a higher density of postsynaptic NMDA receptors than other rostral TSNC neurons. 35This suggests that TGT neurons could exhibit a greater sensitivity to ketamine than local TSNC interneurons. In the current study we did find that FHM-evoked responses of two TGT neurons were markedly suppressed by ketamine, but it is not clear whether this suppression was greater than for FHM-evoked responses in non-TGT TSNC neurons.

It is also conceivable that ketamine-related suppression of FHM-evoked rostral TSNC neurons may result from a distinct trigeminal site of drug action, such as on acetylcholine receptors, or an effect of ketamine on the caudal  TSNC. Certain caudal TSNC neurons that respond to low-threshold mechanoreceptor stimuli are excited by local application of acetylcholine. 42,43In this mechanistic scenario, ketamine could act to decrease FHM-evoked responses in TSNC and TGT neurons via  the suppression of a “tonic” excitatory cholinergic influence impinging on these neurons. In addition, it is thought that the caudal TSNC may exert a facilitatory influence on rostral TSNC neurons, and thus part of the suppressive effect of ketamine on FHM-evoked rostral TSNC neurons may be secondary to a ketamine-mediated decrease in the excitability of caudal TSNC neurons. 44Nevertheless, it is also important to recognize that ketamine-related suppression of FHM-evoked responses in rostral TSNC neurons may arise by ketamine's actions on areas of the central nervous system remote from the TSNC.

Finally, in view of the fact that ketamine was administered systemically, it is important to consider recent evidence pointing to the existence of peripheral NMDA and non-NMDA receptors located in nonglabrous skin, the knee, and the temporomandibular joint, as well as the tooth pulp. 45–47Activation of peripheral NMDA receptors appears to play a role in modulating the sensitivity of cutaneous primary afferents to mechanical stimuli. 45It remains to be determined whether or not such receptors play a functional role in resetting the excitability of FHMs by natural stimuli.

Irrespective of the site and mechanism of action, the results of the present study indicate that responses of trigeminal sensory neurons to innocuous stimuli are suppressed during the dissociative state of surgical anesthesia induced by intravenous ketamine.