| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
COMPLEX FUNCTIONS OF THE CENTRAL NERVOUS SYSTEM, SLEEP AND LOCOMOTION
Section of Oral Biology, Ohio State University, Columbus, Ohio, 43218-2357
Submitted 28 January 2003 ; accepted in final form 21 March 2003
 |
ABSTRACT |
|---|
 TOP ABSTRACT METHODSRESULTSDISCUSSIONREFERENCES |
|---|
central pattern generator; feeding; gasping; multifunctional substrate
The neural mechanism by which the lingual-masticatory apparatus switches from licking to gaping in response to an aversive taste stimulus is unknown. In Aplysia, switching between rhythmic ingestion and rejection responses is mediated by the action of neuropeptides in a multifunctional substrate (38, 39). Although the neurochemistry of the medullary RF in rodents is exceedingly complex, the oromotor nuclei receive glutamatergic, GABAergic, and glycinergic input from the medullary RF (48, 49, 61, 85) and receptors for these neurotransmitters are found within this substrate as well (2, 8, 23, 33, 64, 65, 77). In addition, physiological studies suggest important roles for inhibitory and excitatory amino acids in the generation of fictive masticatory-like jaw movements (11, 20, 36, 41, 47). Thus the purpose of the present study was to determine whether blockade of ionotropic glutamate receptors, GABAA receptors, or glycine receptors in the rostrolateral medullary RF altered the expression of ingestion and rejection elicited by natural stimuli in an awake preparation.
|
METHODS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
After implantation of IO cannulas and EMG electrodes, the head of the rat was fixed in a conventional stereotaxic instrument with blunt ear bars and the skull leveled with respect to bregma and lambda. After removal of a small portion of bone 4 mm posterior to lambda, the dura was removed and two
stainless guide cannulas (26 gauge, 24 mm) were positioned symmetrically in
the lateral medullary RF at the level of the rostral (r) nucleus of the
solitary tract (NST). Before the guide cannula was implanted, the rNST was
identified by recording unit responses to a taste mixture applied to the tip
of the tongue. The rNST was 
4.5 mm caudal and 1.8 mm lateral to lambda at
a depth of 6.5 mm. During subsequent placement of the guide cannula, a
fine wire extending through the guide cannula could record neural activity to
reidentify landmarks such as the IV ventricle, the surface of the brain stem,
and the rNST to assure accurate placement into the RF. All locations were
subsequently verified histologically. The guide cannulas, as well as the IO
cannulas and Amphenol strip connector, were secured to the skull with dental
acrylic. Thirty-three-gauge stainless steel tubing was used as a stylet. The
incision was closed with suture, and rats were given penicillin G procaine
(30,000 U im daily) for 3–4 postoperative days. During the recovery
period, rats were fed a mixture of powdered rat chow and Crisco to enhance
weight gain. All procedures were approved by the Institutional Animal Care and
Use Committee and conformed with the "Guiding Principles for Research
Involving Animals and Human Brings" of the American Physiological
Society.
Adaptation. Training, adaptation, and testing took place in either a Plexiglas chamber (24.5-cm diameter x 26.5 cm) that allowed the rat to move freely or a smaller restraint-chamber (8.5-cm diameter x 21.5 cm; Kent Scientific). Fourteen rats were tested under free-moving conditions in
the Plexiglas chamber, and twenty-two rats were tested in the smaller chamber.
After 4 days of recovery from surgery, rats were readapted to the observation
chamber for 2–3 days and subsequently adapted to the stimulation
protocol. For adaptation to the stimulation protocol, rats received six blocks
of IO taste stimuli. Each block consisted of a 50-µl infusion of sucrose
(0.1 M), quinine monohydrochloride (QHCl; 0.003 M) and NaCl (0.1 M). Stimuli
were delivered via a pressurized stimulus delivery system and
computer-controlled solenoids. Each taste stimulus was followed by three (or 6
after QHCl) infusions of distilled water. After the last water rinse, the rats
were given access to a water bottle containing 0.3 M sucrose for 10 s. They
were then given three IO water rinses to end the stimulus block. Each block
lasted 15 min. The interval between consecutive blocks was 10–15
min.
Experimental design and EMG recording. The highly selective drug 6-cyano-7-nitroquinoxaline-2,3-dione (disodium salt CNQX) was used as a non-N-methyl-D-aspartate (NMDA) receptor antagonist and D-[(3)-2-carboxypiperazin-4-yl]-propyl-1-phosphonic acid (D-CPP) as an NMDA receptor antagonist (17, 34). Bicuculline was used as a GABAA antagonist (6, 7) and strychnine as a glycine antagonist (16). All drugs were purchased from Sigma and dissolved in saline. After the adaptation trials, rats were first tested with saline infusions (100 nl; vehicle control), followed 48 h later by a test with the GABAA agonist muscimol. Muscimol (0.06 nmol/100 nl) was infused bilaterally to functionally verify that subsequent drug infusions would be made into sites previously established as essential for the generation of oral consummatory responses (14). Infusions with CNQX, DCPP, bicuculline, or strychnine commenced 48 h after the muscimol test. Drugs, doses, and the numbers of animals in each group are shown in Table 1. These doses were chosen on the basis of published behavioral studies (see, e.g., Refs. 3, 59, 75, 89) and pilot data. In the case of strychnine, an effective dose was at the saturation level and more could only be given by increasing the volume. Because each animal received saline infusions in addition to at least one drug treatment, drug effects for a given dose were determined by comparing postinfusion effects to pre-drug infusion effects with a repeated-measures design.
|
The procedure of EMG recording was similar to those described previously (14, 19, 80). On the test day, the rat was placed in the observation chamber for 1 h before testing and the Amphenol connector on the head was mated to a cable that connected the EMG electrodes to conventional AC amplifiers. The EMG signals were monitored and recorded online through a CED interface (Power 1401, Cambridge Electronic Design) and stored in a microcomputer. After EMG responses to two blocks of stimuli were obtained, the stylets were removed and infusion cannulas containing drug or saline were inserted into the brain via the guide cannulas. Infusers constructed from 33-gauge stainless tubing extended 0.5–1.0 mm beyond the guide cannula, and the other end was fitted to PE-10 tubing. The PE-10 tubing was attached to a 10-µl syringe (Hamilton) that was driven by a syringe pump (KD Scientific). After infusion of the drug or saline into the brain, the infusers were left in place for 30 s and then removed. During the test session, the animals' behaviors were monitored and videotaped.
At the termination of the experiment, Fluorogold (2%) was injected to mark the infusion sites. Under deep anesthesia with Nembutal (150 mg/kg), the rat was perfused transcardially with 0.9% saline followed by 10% formalin. The brain was removed, sectioned (50 µm) into two series, and mounted. Injection sites were identified with fluorescent microscopy.
Data analysis. EMG activity was analyzed with custom software written for the CED Spike2 system. EMG activity was rectified and filtered at 80 Hz, and baseline activity from 1 s before oral responses was calculated from a number of trials and used to set a threshold (Fig. 1). Crossing the threshold defined a contraction onset, and recrossing the threshold defined a contraction offset. Very short intervals between contractions were amalgamated with longer contractions, and very short contractions were eliminated (see Fig. 1). From the individual burst onsets and offsets, a number of parameters were determined for each response to stimulus-induced activity including mean burst duration, mean total integrated activity between onset and offset (amplitude), and mean peak activity. In addition, the rate of activity was determined by dividing the number of contractions by the total response time (bout duration). Response parameters were normalized to the predrug values obtained in block 1 for each test session. Drug effects were compared against saline with a repeated-measures ANOVA for blocks 2–6 and the four stimulus conditions. A separate ANOVA was conducted for each muscle and response parameter. EMG electrodes failed in some muscles over the course of the experiment, and thus the n for a given ANOVA was sometimes less than the maximum n.
|
|
RESULTS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
2 h after injection.
|
|
The prevalence of complete suppression of oromotor activity by D-CPP made it difficult to determine whether the different response parameters of burst duration, amplitude, and rate were differentially affected by the drug. To gain insight into this, we determined recovery functions for oromotor activity by comparing the blocks immediately after the end of complete suppression to the predrug blocks 1 and 2. Because suppression lasted for varying times for different animals, the first postsuppression block was determined individually for each animal and the results were pooled. Figure 4 compares recovery functions for four response parameters collapsed across AD and Gen for the four stimulus conditions. For the two blocks after complete suppression, response rate and peak response were significantly depressed compared with predrug blocks (P < 0.0003 and P < 0.013, respectively) but burst duration approached a nearly significant longer value (P = 0.086) Thus the total (integrated) activity of contractions reflected the decreased peak values offset by longer contractions and was essentially flat (P = not significant).
|
Infusion of the GABAA agonist muscimol (0.06 nmol/100 nl) that preceded the D-CPP infusions produced either complete or partial suppression of the AD and Gen EMG response during licking and gaping. Overall, the amplitude of the response was reduced by 50% for both muscles across all
four stimulus conditions. Within-subjects ANOVAs for AD (n = 9) and
Gen (n = 8) amplitude showed significant differences between muscimol
and saline (AD: P < 0.05; Gen: P < 0.05). The
suppression effect appeared immediately after muscimol infusion and returned
to baseline by block 7, 2.5 h after infusion. These results were
consistent with our previous report
(14), indicating that we were
in the same region of the RF. Histological verification confirmed that the
infusion sites were distributed in the parvocellular (PCRt) and the
intermediate (IRt) subdivisions of the reticular formation ventral to the rNST
(Fig. 5).
|
CNQX. Of the six cases receiving a high (3.62 nmol/100 nl) dose of CNQX in the lateral medullary RF, three produced a long-lasting (>1 h) complete suppression of oromotor activity elicited by gustatory stimulation. Two other cases showed complete suppression that lasted 0.5 h, and one
case showed only a transient suppression of bottle licking. In a second group
of five cases that received a lower dose of CNQX (1.81 nmol/100 nl), only one
case showed a delayed and transient suppression of oromotor activity; the
other cases showed no effect.
In Fig. 6, the normalized mean amplitudes of EMG activity for AD and Gen before and after the high dose of CNQX and saline are plotted against stimulus block. The reduction in blocks 3 and 4 began to recover by block 5. Repeated-measures ANOVAs (within-subjects design; n = 6) revealed only a marginal main effect for normalized integrated amplitude after CNQX and saline treatments (AD: P = 0.067; Gen: P = 0.06), but there was a significant drug x block interaction for both AD (P = 0.015) and Gen (P = 0.027). The drug x stimulus interaction was not significant (AD: P = 0.424; Gen: P = 0.205). In the group receiving the low dose of CNQX, repeated-measures ANOVAs for amplitude revealed no main effect (AD: P = 0.696; Gen: P = 0.594) and no drug x block interaction (AD: P = 0.345; Gen: P = 0.957). During the suppression of oromotor response after the drug infusion, the rats showed an appetitive response to the sucrose bottle and the ability to find and orient to it. However, they were unable to lick. Recovery functions after the effective dose of CNQX were determined as they were for D-CPP (Fig. 7). Similar to the recovery after D-CPP, there was a significant decrease in the contraction rate in the two blocks after suppression (ANOVA: P < 0.018). In contrast to D-CPP, however, this was accompanied by a significant increase in peak EMG activity (P < 0.046) and burst duration (P < 0.022). The increase in peak activity, combined with the increased burst duration, made for a large increase in the total (integrated) amplitude associated with each contraction (P < 0.002). The enhanced peak and amplitude of EMG activity in association with a slower rate during recovery to CNQX infusions can also be seen in the rectified and integrated activity of a single case illustrated in Fig. 8.
|
|
|
Although licking and gaping were severely suppressed by CNQX, an unexpected result was spontaneous gasping that appeared in the course of the oromotor suppression (Fig. 8). Gasps were characterized by slow rate, long duration, large amplitude, and synchronized activity across AD, Gen, and Sty (Fig. 8) that was clearly distinct from the EMG pattern of licking and gaping. Although gasps were evident in all six cases receiving the high dose of CNQX, they occurred to varying degrees. In some animals there were long periods of time when animals did not gasp but still showed suppression of licking, and other cases had long periods of gasping. During recovery, gasps were sometimes seen interspersed with the licking response. As evident from Fig. 8, gasping and licking responses could be distinguished by multiple criteria including amplitude, rate, and the phase relationship between muscles.
Figure 9 shows the mean normalized amplitude and the rate of oromotor activity from three muscles as a function of block for four cases receiving the high dose of CNQX that displayed a high incidence of gasps. Although the amplitude of EMG activity
from AD, Gen, and Sty was greater during gasping than licking, the increased
activity from the tongue muscles during gasping was significantly greater than
activity from the AD. Collapsing across postdrug blocks 3–6 and
comparison to predrug block 2 indicated that the change in amplitude
for both Gen (2-sample t-test; P < 0.009) and Sty
(P < 0.012) was significantly greater than the change for AD. In
addition, compared with licking responses, the rate was much slower, at
1.5 Hz (paired t-test; P < 0.001).
|
In 8 of 11 cases, infusion of the GABAA agonist muscimol (0.06 nmol/100 nl) in the lateral medullary RF completely eliminated oromotor EMG activity in response to taste stimuli. The other three cases showed partial suppression. Overall, the mean amplitude of EMG response for AD and Gen was reduced by 60% across blocks 3 and 4. In nine cases with complete data, repeated-measures ANOVA (within-subject design) revealed a significant difference in the amplitude reduction for AD and Gen between muscimol and saline injection (AD: P = 0.013; Gen: P = 0.015). These results suggested that the injection sites in this experiment were located in the muscimol effective zone, and histological reconstruction verified that all injection sites were located in the PCRt and IRt, ventral to the rNST (Fig. 5).
Bicuculline. Infusions of the GABAA antagonist bicuculline augmented the amplitude of oral responses induced by IO taste stimulation, an increase accompanied by a reduction in rate. The drug effect was reversible and lasted 2 h. An example of AD activity before and after
bicuculline is shown in Fig.
10. It is evident that responses to IO sucrose and QHCl
stimulation were affected by the bicuculline infusion but that the response to
licking sucrose from a bottle was not. Higher doses of bicuculline were
generally less effective than doses in the low range. A scatterplot of the
integrated EMG responses of AD in block 4 after IO stimulations as a
function of drug dose is shown in Fig.
11A. When drug doses were divided into a
"high" and a "low" range and plotted against stimulus
block, the relative effectiveness of the two dose ranges was even more
pronounced. Figure
11B compares the normalized mean amplitudes for
IO-elicited responses collapsed across AD and Gen for the high and low ranges
of bicuculline and saline. Infusion of the high dose of bicuculline produced
only a moderate increase in the amplitude of oral responses compared with
infusions of the low dose. Similar changes were also observed for other
response parameters, i.e., there were small increases in the contraction
duration and decreases in the rate after infusion of the high-dose
bicuculline. An ANOVA comparing the high range of bicuculline to saline
indicated a significant reduction in rate (P < 0.001) and a
borderline increase in amplitude (P = 0.057).
|
|
The effects of the low dose of bicuculline on other EMG response were further analyzed. A within-group ANOVA comparing the effects of saline to those of the low range of bicuculline indicated significant drug effects for amplitude (P < 0.002) and peak response (P < 0.012). Significant drug x stimulus interactions were also evident for amplitude (P < 0.005) and peak (P < 0.05), and inspection of the data suggested that bottle licking was unaffected (see, e.g., Fig. 10, G–I). Because of the significant drug x stimulus interaction and the apparent lack of changes to bottle licking, the results of IO stimulation (collapsed across sucrose, NaCl, and QHCl) were pooled and compared with bottle licking for four EMG parameters (Fig. 12). When the data were collapsed in this fashion, amplitude (P < 0.001), peak (P < 0.012), and burst duration (P < 0.007) all increased and rate decreased (drug x block interaction: P < 0.01). In contrast, none of the EMG parameters for bottle licking even approached significance (P > 0.4).
|
The responses of Gen to bicuculline were similar to those of AD but somewhat weaker (Fig. 12C). Thus an ANOVA showed drug x stimulus interactions for amplitude (P < 0.005) and peak (P < 0.007), but rate and burst duration were unchanged. Collapsing across IO-elicited responses, however, yielded significant drug effects for rate (P < 0.02). The oral responses induced by sucrose presented in a bottle were not altered after bicuculline infusion (Fig. 12D). Repeated ANOVAs for Gen (within-subjects design; n = 8) revealed no significant difference for peak (P = 0.876), contraction duration (P = 0.292), amplitude (P = 0.675), or rate (P = 0.718).
Intracranial injection sites were plotted in coronal sections (Fig. 13). Targeted infusion sites were distributed in the lateral medullary RF including subdivisions of PCRt and IRt ventral to and at the level of the rNST. Infusion of the
GABAA agonist muscimol (0.06 nmol/100 nl) in the lateral medullary
RF either completely eliminated or reduced the AD and Gen EMG responses during
licking and gaping. Overall, the amplitude of the response was reduced by
50% for both AD and Gen across all four stimulus conditions.
Repeated-measures (within-subjects design) ANOVAs for amplitude showed a
significant difference between muscimol and saline infusions (AD, n =
7; Gen, n = 6; P < 0.05). These results confirm that the
sites of injection were in the muscimol effective zone
(14).
|
Strychnine. Infusion of the glycine receptor antagonist strychnine (6.74 nmol/100 nl) produced a moderate enhancement in the amplitude and contraction duration of oral responses to taste stimulation without altering the frequency. An ANOVA comparing the effects of strychnine against saline for AD activity indicated significant increases for amplitude (P < 0.007) and peak (P < 0.03). Neither burst duration nor rate was altered, and there were no drug x stimulus interactions. Similar results were obtained for Gen. Figure 14 shows the effects of strychnine on four response parameters collapsed across all four stimulus conditions as a function of stimulus block for AD and Gen. Collapsing the data in this fashion produced significant increases in burst duration (AD: P < 0.013; Gen: P < 0.005) together with integrated amplitude (AD: P < 0.008; Gen: P < 0.002) and peak (AD: P < 0.027; Gen: P < 0.024). There was no significant change in rate.
|
Infusion of the GABAA agonist muscimol (0.06 nmol/100 nl; n = 6) in the lateral medullary RF either completely abolished or reduced the AD and Gen EMG response of ingestion and rejection. Overall, the amplitude of the response was reduced by 50% for AD and Gen across all
four stimulus conditions. Repeated-measures (within-subjects design) ANOVAs
for the amplitude revealed a significant difference between muscimol and
saline infusions (AD: P < 0.05; Gen: P < 0.05). These
results demonstrate that the sites of injection were in the muscimol effective
zone. All intracerebral injection sites were plotted in coronal sections
(Fig. 13) and were distributed
in the lateral medullary RF overlapping with the PCRt and IRt ventral to and
at the level of the rNST.
|
DISCUSSION |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
-amino-3-hydroxy-5-methylisoxazole-propionic acid
(AMPA)/kainate or NMDA receptors profoundly suppressed oromotor activity in
response to IO gustatory stimuli or sucrose presented in a bottle. The drug
effects were reversible and dose dependent in suppressing motor activity of
both the tongue and jaw. Although the differential recovery of specific
parameters of the EMG responses after the drug infusions is indicative of
different roles for these receptors in orchestrating oroconsummatory activity,
neither DCPP nor CNQX differentially affected the responses of ingestion or
rejection, suggesting that the switch from one behavior to the other does not
rely exclusively on these specific glutamate receptor subtypes. The present study provides further evidence that this same substrate is under tonic inhibition. Blocking GABAA receptors significantly increased the amplitude and contraction duration of consummatory responses of ingestion and rejection following IO stimulation but had no effect on oromotor responses following stimulation with sucrose presented in a bottle. Thus the inactivation of GABAA receptors produced large gapelike jaw openings in response to IO sucrose stimulation, a stimulus that normally induces licking with only small jaw openings. Disinhibition of these GABAA receptors could provide one mechanism by which aversive gustatory stimuli produce gape (rejection) responses. Table 2 summarizes some of the major effects of the amino acid antagonists on various EMG parameters.
|
Site of action. Determining the effective site of a drug infusion in the brain is an empirical question. A volume of 100 nl will fill a cavity with a radius of 288 µm (57), but diffusion through extracellular space will enlarge the area. In a previous study, we infused 100 nl of muscimol (0.06 nmol) into the medullary RF and estimated the effective zone in suppressing licking to be <1 mm (14). Thus infusions of muscimol centered dorsally in the vestibular nucleus, medial in nucleus gigantocellularis, or rostral to the NST were ineffective in suppressing licking compared with infusions centered in the lateral RF.
Although all of the drug infusions in the present study were made into these same "muscimol-effective" sites, there is no guarantee that these drugs were effective over the same distance. The effectiveness of a particular drug depends on numerous factors, including diffusibility and degradation rate to name but two (54). Thus, as evident from Figs. 5 and 13, infusions into the lateral medullary RF may have spread to adjacent structures including the rNST (dorsal), the ventral RF, the spinal trigeminal complex (lateral), and the medial RF. In addition, the spherical volume equation does not account for the likelihood that infused drugs "run up" the infusion cannula (35). Thus we cannot rule out the possibility that some of the drug effects were mediated, at least in part, by neurons in the overlying NST. Likewise, infusion of the non-NMDA receptor antagonist 1,2,3,4-tetrahydro-6-nitro-2,3-dioxobenzo[f]quinoxaline-7-sulfonamide (NBQX) into the IV ventricle suppressed sucrose intake, and the investigators concluded that one mechanism might involve diffusion of the antagonist into the underlying rostral (taste) NST (90). However, lesions of the rNST have only a minimal impact on oromotor activity (5, 21, 67), and in the present study the volume used was very small (100 nl), suggesting that spread to other substrates should be relatively minor. Moreover, differential effects on oromotor activity induced by different tastants were not detected, consistent with a minimal impact on neurons in the NST. Spread into the lateral spinal trigeminal complex is more problematic, as deafferentation of oral trigeminal structures reduced mouth opening during eating (37, 88). However, on the basis of the location of the infusion sites for D-CPP and CNQX, and the estimates of spread, it is more likely that the suppressed oral activity resulted from targets of second-order projections from the spinal trigeminal complex in the RF rather than the first-order sensory neurons themselves. Nevertheless, it will be interesting to test whether activation of GABAA receptors or blockade of glutamate receptors in the spinal trigeminal complex can suppress oral ingestive behavior.
Role of glutamatergic receptors in ororhythmic activity. A role for both NMDA and non-NMDA receptors in the medullary control of oroconsummatory behavior is supported by immunohistochemical studies localizing receptors to neurons in this region (60, 63). In addition, the iontophoretic application of NMDA and non-NMDA antagonists onto motor trigeminal-projecting RF neurons in the rostral dorsal medulla inhibited rhythmic activity produced by cortical electrical stimulation (36). However, in apparent contrast to the present study, differential effects of NMDA and non-NMDA receptor antagonists on jaw opener activity were observed. In the Inoue et al. study (36), application of CNQX inhibited the discharge of both jaw closer and jaw opener premotor neurons in the pontine-medullary RF during fictive mastication whereas application of the NMDA receptor antagonist CPP only suppressed the activity of jaw closer interneurons. These results suggest that non-NMDA receptors exist on both jaw opener and jaw closer premotor neurons but that NMDA receptors exist only on jaw closer premotor neurons in the area examined. In the present study we did not monitor jaw closing activity because it is relatively quiescent during licking (82). However, the NMDA receptor antagonist D-CPP did suppress jaw opening activity during licking. The apparent inconsistency between the electrical brain stimulation model and our chronic model may reflect activation of multiple neurons in a local circuit with the larger infusions. For example, the suppression of jaw opening by D-CPP in the present study could reflect suppression of motor trigeminal (jaw opener)-projecting RF neurons that require local interneuron activation via NMDA receptors.
Involvement of NMDA and non-NMDA receptors in fictive mastication was also observed in in vitro studies (45, 74). In the presence of the GABAA receptor antagonist bicuculline, bath application of the NMDA receptor agonist N-methyl-aspartate (NMA) or the non-NMDA agonist kainate induced rhythmic discharge in the motor branches of the motor trigeminal nerve in a brain stem slice preparation. Blockade of NMDA receptors by DL-2-amino-5-phosphonobutyric acid (AP5) eliminated the discharge induced by either NMA or kainate. In contrast, blockade of AMPA/kainate receptors with CNQX abolished the discharge elicited by kainate but not by the NMDA receptor agonist NMA. These findings suggest that chemically induced rhythmic trigeminal discharge requires activation of either NMDA or non-NMDA receptors; however, the use of a bath application approach makes it difficult to determine the exact location of the receptors, i.e., it remains unclear in this study whether glutamate receptor antagonists acted on premotor and/or motor neurons. It is also significant that these in vitro studies were in brain stem slices rostral to the RF sites examined in the present studies. Although the RF region adjacent to motor V may play an important role in the control of fictive mastication, it is unclear whether the rhythmic discharge of an efferent nerve is functionally equivalent to natural mastication or licking that involves complex jaw-tongue coordination. In a previous study (14), we demonstrated that rhythmic jaw-tongue coordination during licking was suppressed by the GABAA agonist muscimol infused into the lateral medullary RF caudal to the exiting facial nerve but not when infusions were made immediately rostral to it in an area more closely approximating the area studied in the in vitro studies.
Although not identical to the distinctions described above, the differential recovery functions after complete suppression by D-CPP and CNQX also argue for different roles for these receptors in producing consummatory oromotor responses. Thus, immediately after complete suppression, the magnitude of the EMG burst (peak activity) was significantly suppressed with D-CPP but enhanced with CNQX. After complete suppression with either drug, however, the rate of licking was slower and each lick had longer burst duration. Thus, within this substrate, contraction rate and contraction duration did not appear as independent mechanisms. Other studies, however, suggest different control mechanisms for amplitude and rate. In the anesthetized rabbit, an increase in the frequency of cortically induced jaw movement (fictive mastication) was elicited by increasing the stimulus intensity whereas stimulation of different areas of the cortex could change the amplitude without altering the frequency (53). Likewise, in the anesthetized guinea pig, hemitransection of the medulla between the obex and the rostral third of the inferior olivary nucleus attenuated the amplitude of the AD EMG without changing the cycle or burst duration (12). These two studies imply separate neural mechanisms controlling amplitude and frequency. Results from the present study indicate that one mechanism for these differential effects might involve different ionotropic glutamate receptors. On the other hand, differences in glutamate receptor subtype do not appear to contribute to a differential control of burst duration and rate, two parameters for which independent mechanisms were previously suggested (9, 10).
Inhibition. Several lines of evidence suggest that both GABAA and glycine receptors in the pontomedullary RF have inhibitory roles in oromotor activity. Receptors for both GABAA and glycine are located in this region (2, 8, 23, 33, 64, 65, 77). In the anesthetized guinea pig, blocking GABAA or glycine receptors by iontophoretically applying bicuculline or strychnine onto trigeminal-projecting RF neurons in the rostral medulla increased neuronal discharge (36). Likewise, eliciting rhythmic discharge in the motor branches of the trigeminal nerve in a brain stem slice preparation required bath application of an excitatory amino acid, together with bicuculline, presumably to remove tonic inhibitory influences (45, 46, 74). Differential effects of blockade of GABAA and glycine receptors were also reported (36, 46).
In anesthetized guinea pig, blocking GABAA receptors with bicuculline increased the discharge of jaw closer trigeminal-projecting premotor neurons only during the nonopening phase of fictive mastication (36). In contrast, application of the glycine receptor antagonist strychnine increased the discharge of the same premotor neurons only during the jaw-opening phase (36). These findings suggest that GABAA receptors mediate tonic inhibition of jaw closer interneurons, whereas glycine receptors mediate phasic inhibition of jaw closer premotor neurons during jaw opening. This study, however, did not test the effects of either antagonist on jaw opening premotor neurons. In the present study, it appeared that both GABAA and glycine receptors provide tonic inhibition to a jaw opening substrate: blocking either GABAA or glycine receptors increased the amplitude of jaw opening activity, although the effect of glycine blockade was less robust. In addition, increases in Gen (tongue protruder) amplitude were associated with either bicuculline or strychnine infusion. Thus GABAA and glycine receptors may provide tonic inhibition to both motor trigeminal- and hypoglossal-projecting interneurons. This is supported by anatomic evidence that premotor neurons in the lateral RF project to multiple oromotor nuclei (1, 50, 76).
In addition to the disruption of phasic inhibition of jaw-closing interneurons during jaw opening after the iontophoretic application of strychnine (36), bath application of the excitatory amino acid agonist NMA combined with strychnine in a brain stem block connected to jaw muscles elicited noncoordinated rhythmic EMG bursts between jaw opener and jaw closer muscles (46). A role for glycine receptors in the phasic inhibition of motor patterns is also found in the locomotor system (15, 42). The present study did not provide much evidence for the disruption of phasic inhibition during natural licking after strychnine or bicuculline infusions. Uncoordinated lingual or lingual-digastric activity after either strychnine or bicuculline was not observed. However, with the use of cross-correlation (19), a phase shift between AD (jaw opener) and Sty (tongue retractor) contractions from their normal out-phase relation to an abnormal in-phase relation was observed in one case in which we used a higher dose of strychnine (13.68 nmol/200 nl).
Although blocking GABAA receptors in the lateral medullary RF affected licking induced by IO stimulation, it had no effect on appetitive licking in response to sucrose presented in a bottle. One explanation for this differential effect is that appetitive licking, which requires descending input from the forebrain (86), may interact with RF circuitry to strengthen the impact of local GABAergic inhibition, thus making it more refractory to bicuculline. There is certainly evidence that local circuitry can function independently of forebrain influences, and both ingestion and rejection to IO stimulation can occur without an intact forebrain (29). A role for GABAA receptors in this local circuitry can be postulated. Specifically, the large mouth openings observed to sucrose after bicuculline infusion appeared very similar to the normal gape response after QHCl stimulation. The striking similarity of the response to sucrose after the bicuculline infusion and the normal response to QHCl is shown in the video clips (supplemental materials video clips A–C; available online at http://ajpregu.physiology.org/cgi/content/full/00054.2003/DC1). Video clip A shows the normal response to IO stimulation with sucrose, which consists of small mouth openings and several lateral tongue protrusions. Video clip B shows the response to sucrose after the bicuculline infusions, and much larger mouth openings are evident to the extent that the lower incisors can be seen. These very large mouth openings appear very similar to the classic gape responses to QHCl seen in video clip C.
Thus we speculate that gapes may be formed, in part, by the inhibition of normally present tonic (GABAergic) inhibition, i.e., disinhibition and that this disinhibition originates from QHCl-sensitive neurons in the nearby rNST. A specific population of neurons in the central subnucleus of the rNST expresses fos-like immunoreactivity after stimulation with QHCl (31, 43, 44, 83). These neurons do not project extensively to the medullary RF (84) but may project to neurons in the ventral subdivision of the rNST that do (30). Thus we can speculate that one source of tonic inhibition may be from output neurons in the ventral subdivision of the rNST that are, in turn, inhibited after stimulation with QHCl.
Multifunctional substrate. The lateral medullary RF is implicated in diverse physiological functions including respiration, autonomic responses, and ingestion (62, 68, 76, 78). The role of the lateral medullary RF in orchestrating orolingual movements in multiple functions is supported by anatomic evidence that premotor neurons in the lateral RF project to multiple oromotor nuclei (1, 50, 76) and receive input from both brain stem orosensory structures as well as descending inputs from forebrain areas associated with ingestive function (4, 32, 66).
Because many neurons in the lateral medullary RF are active during both licking and gaping (79), and inactivation of neurons in this area suppresses both patterns of oromotor activity in awake rats (13, 14), network reconfiguration of a multifunctional substrate is one possible mechanism underlying the switch between licking and gaping in response to different gustatory stimuli. Licking and gaping share the same muscles but differ in terms of contraction duration, magnitude, and the phase relation between the tongue and jaws (82). Recently, Lieske and colleagues (51) proposed that different patterns of masticatory-like activity of ororhythmic neurons elicited by electrical stimulation implied network reconfiguration in the brain stem RF. Behavioral switching by multifunctional systems has been extensively studied in invertebrate preparations. For example, the switch between ingestion and rejection (ejection) in Aplysia may be mediated by the action of a neuropeptide in a multifunctional CPG (38, 39). It is unknown, however, whether similar mechanisms operate in the taste-induced switch between licking and gaping. Indeed, we observed that blocking GABAA receptors in the lateral medullary RF produced large gaping-like jaw opening in response to IO sucrose stimulation. This suggests a role for GABAA receptors, at least in part, in mediating the taste-induced switch between licking and gaping.
Because the suppressive effect of DCPP and CNQX was equivalent for licking as well as for gaping (oral rejection response) evoked by QHCl, it is unlikely that glutamatergic inputs alone produce the switch from ingestion to rejection. Instead, it appears that glutamatergic input to this substrate is necessary for both behaviors. The specific source of the glutamatergic inputs may include projections from any of the forebrain and brain stem orosensory nuclei known to project to this region as well as RF interneurons (4, 40, 66, 81, 87).
Although a switch from licks to gapes appeared independent of glutamate receptor subtype, a glutamate receptor-mediated switch controlling the jaw and tongue was evident. After CNQX infusions, motor patterns of the jaw and tongue switched from a characteristic licking pattern to one associated with gasping. This could indicate a specific role for non-NMDA receptors in a multifunctional substrate, one capable of generating both licking and gasping. In such a substrate, stimulation of AMPA/kainate receptors supports licking but appears to have the opposite role in gasping, i.e., they tonically inhibit this motor pattern. Alternatively, the expression of gasping and concomitant suppression of licking could reflect an indirect interaction between spatially separate substrates, one controlling respiration and the other controlling oroconsummatory function. In other words, it is possible that gasping occurred because CNQX diffused ventrally.
Several studies have established that neurons in the ventral medulla caudal to the facial nucleus in the pre-Bötzinger complex form a necessary respiratory "kernel" (62, 68). When normal respiration is impeded, for example by hypoxia or the removal of descending inputs from the pons, eupnea is replaced by gasps (70, 71). Increased respiratory drive, as likely occurs during gasps, is associated with coactivation of lingual protrudor and retractor muscles similar to that observed in the present study (22, 24, 25). Because the respiratory kernel is more sensitive to non-NMDA antagonists (including CNQX) than NMDA antagonists (27), it is conceivable that gasps observed in the present study resulted from the spread of CNQX to the ventral medulla.
Several lines of evidence, however, argue against this interpretation. The site of our infusions appeared too far dorsal to influence the ventral medulla, and infusions of muscimol into the same sites as CNQX did not overtly influence the respiratory rhythm. Infusions of muscimol will, however, suppress eupnea when made more ventral (14, 56, 69). This does not rule out the possibility that CNQX can diffuse more widely than an effective dose of muscimol. However, further arguing against the diffusion of CNQX as a mechanism of action are studies implicating our sites of infusion in the genesis of gasping (70, 71). Specifically, lesions in sites that overlapped our infusion sites prevented gasping responses (26, 72). Thus it could be argued that gasps were released by CNQX, perhaps by blocking non-NMDA (excitatory) receptors on neurons that function to tonically suppress gasping.
Although licking responses are suppressed during gasping, both jaw and tongue muscles participate in the response, demonstrating that the lack of licking/gaping was not due to total paralysis of orolingual muscles. The present study cannot differentiate between the likelihood that different populations of neurons in the lateral tegmental field participate in consummatory and gasping responses or that the same network is reconfigured to produce the gasp.
|
ACKNOWLEDGMENTS |
|---|
This work was supported by National Institute of Deafness and Other Communications Disorders Grant DC-00417.
|
FOOTNOTES |
|---|
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
|
REFERENCES |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
1 mRNA
subunit in the brainstem nuclei controlling esophageal peristalsis.
Brain Res Mol Brain Res 40:
143–147, 1996.
subunit mRNA in the rat brain.
Brain Res 560:
23–37, 1991.
1 subunit mRNA-containing neurons in the
lower brainstem of the rat. Brain Res Mol Brain Res
7: 335–345,
1990.
, Infusion
of D-CPP; 
, CNQX.
