|Home | About | Journals | Submit | Contact Us | Français|
The appetitive component of feeding is controlled by forebrain substrates but the consummatory behaviors of licking, mastication and swallowing are organized in the brainstem. The target of forebrain appetitive signals is unclear but likely includes regions of the medullary reticular formation (RF). The present study was undertaken to determine the necessity of different RF regions for mastication induced by a descending appetitive signal. We measured solid food intake in response to third ventricular (3V) infusions of the orexigenic peptide, neuropeptide Y 3-36 in awake, freely-moving rats and determined whether focal RF infusions of the GABAA agonist muscimol suppressed eating. Reticular formation infusions were centered in either the lateral tegmental field, comprised of the intermediate (IRt) and parvocellular (PCRt) RF, or in the nucleus gigantocellularis (Gi). Infusions of NPY 3-36 (5 ug/5ul) into 3V significantly increased feeding of solid food over a 90 minute period compared to the non-infused condition (4.3 ± 0.56 versus 0.57 ± 0.57g, p < .001). NPY 3-36 induced food intake was suppressed (1.7g ± 0.48) by simultaneous infusions of muscimol (0.6 mM/100 nl) into the IRt/PCRt (p < .01). Coincident with the decrease in feeding was a decrease in the amplitude of anterior digastric muscle contractions in response to intra-oral sucrose infusions. In contrast, infusions of muscimol into Gi had no discernible effect on food intake or EMG amplitude. These data suggest that the IRt/PCRt is essential for forebrain-initiated mastication but that the Gi is not a necessary link in this pathway.
Feeding is controlled by structures located in widespread regions of the brain (Broberger, 2005) (Berthoud, 2002; Morton, Cummings, Baskin, Barsh, & Schwartz, 2006) (Saper, Chou, & Elmquist, 2002). Within this distributed system are certain critical nodes that subserve specific functions. One node, the hypothalamus, is sensitive to humoral and neural signals reflecting homeostatic state and receives numerous other inputs; e.g., from the nucleus accumbens, a forebrain structure implicated in reward (Kelley, Baldo, & Pratt, 2005). Many of the integrative capacities of the hypothalamus are shared by the caudal nucleus of the solitary tract in the medulla (Grill, 2006), but an animal reliant solely on the brainstem does not feed spontaneously when deprived (Grill & Norgren, 1978a). For this reason, the appetitive component of feeding is thought to be organized in the forebrain. Critically, however, the forebrain structures that monitor energy balance, generate craving, and sense the rewarding properties of food must at some point interface with substrates that actually produce the behavior of eating. Compared to the remarkable progress in unraveling the intricate circuitry of the hypothalamus, relatively little is known of these pathways. It is generally accepted that the lower brainstem contains the neural circuitry responsible for the generation of consummatory behavior but the precise locations of the substrate underlying the various components of this complex act are still debated.
Although there is general agreement that a central pattern generator (CPG) for swallowing is localized to the caudal nucleus of the solitary tract and the reticular formation near nucleus ambiguus (Jean, 2001), there is less consensus on the location(s) for the CPGs for licking and mastication. Based on studies of fictive mastication elicited by cortical (orbital cortex) stimulation, a pathway from nucleus gigantocellularis (Gi) to the lateral tegmental field, composed of the intermediate (IRt) and parvocellular (PCRt) reticular formation of the medulla, is hypothesized to be essential for mastication (Nakamura & Katakura, 1995). Other studies, however, have observed rhythmic motor trigeminal activity in an en bloc preparation devoid of the medulla, suggesting a pontine locus for a masticatory CPG (Kogo, Funk, & Chandler, 1996; Kogo, Tanaka, Chandler, & Matsuya, 1998; Lund, Kolta, Westberg, & Scott, 1998; Tanaka, Kogo, Chandler, & Matsuya, 1999). Indeed, these latter studies suggest that the medulla may be inhibitory to the expression of masticatory movements; i.e. blocs that included successively more medulla showed smaller-amplitude rhythmic activity (Tanaka et al., 1999). Thus, it is actually unclear whether the medulla is excitatory or inhibitory to mastication.
To more directly assess the role of the medulla in mastication, we measured food intake in an awake, freely-moving preparation following infusions of the GABA agonist muscimol into selected regions of the medullary RF. Previous studies in awake animals suggested that laterally placed muscimol infusions in the IRt/PCRt suppressed licking induced by intraoral stimulation, but that inactivation of the Gi had no effect (Chen & Travers, 2003; Chen, Travers, & Travers, 2001). These studies are consistent with anatomical studies indicating that neurons in the IRt constitute a major source of projections to the hypoglossal nucleus (Holstege, Kuypers, & Dekker, 1977; Travers & Norgren, 1983). However, the functional inactivation studies with muscimol only measured licking of fluids, limiting generalizations regarding the mastication of solid food and because the neuroanatomical studies suggest that the IRt/PCRt is a major source of projections to the motor trigeminal nucleus as well, we wanted to clarify its role in mastication. In addition, because intraoral stimulation can induce consummatory behavior in decerebrate preparations, its use in awake, intact rats leaves open the question of whether the IRt/PCRt or GI are essential for consummatory behavior initiated by a forebrain “appetitive” stimulus. Thus, the present study combined infusions of muscimol into selected RF regions with 3rd ventricle (3V) infusions of NPY 3-36, a highly orexigenic compound (Flynn, Plata-Salaman, & Ffrench-Mullen, 1999; O’Shea et al., 1997).
Sixteen adult Sprague-Dawley rats (252 – 404 g) were anesthetized with pentobarbital sodium (Nembutal; 50 mg/kg ip) and fit with intraoral (IO) cannulae for delivery of fluid solutions into the oral cavity. The cannulae were guided into oral mucosa and exited via an incision on the skull (Grill & Norgren, 1978b). With a ventral approach, bipolar electromyogram (EMG) electrodes made of twisted fine wires (67 μm, NiCr) insulated except for 0.5 mm at the tip were implanted in the anterior digastric (AD; jaw opener) (Travers & Norgren, 1986). In several cases, we adapted the technique of Pearson and colleagues (Pearson, Acharya, & Fouad, 2005) for sewing the electrodes into the muscle, which we have now found to be a superior technique. With either method, leads from the EMG electrodes were guided through a subcutaneous path to the top of the head and attached to an Amphenol connector. Hindlimb areflexia was used as an index of the surgical level of anesthesia, and supplemental Nembutal was administered when needed. A DC electric body warmer maintained body temperature at 37°C throughout the surgery. After implantation of intra-oral cannulae 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 either symmetrically in the lateral medullary RF at the level of the rostral nucleus of the solitary tract (rNST) or a single cannula was placed along the midline at the same depth. To aid in accurate cannula placement, the rNST was first identified by recording unit responses to a taste mixture applied to the tip of the tongue.
During subsequent placement of the guide cannula, a fine tungsten electrode extending through the cannula could record neural activity to re-identify landmarks such as the fourth ventricle, the surface of the brainstem, and the rNST to assure accurate placement into the RF. For injections into the 3rd ventricle (3V), a 3rd stainless steel guide cannula (26 gauge, 24 mm) was placed stereotaxically along the midline 2.3 mm caudal to bregma at a depth of 8 mm ventral to the brain surface. All locations were subsequently verified histologically. The guide cannulae, as well as the IO cannulae 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 wound clips, and rats were given penicillin G procaine (30,000 U im daily) for 3–4 postoperative days. During the recovery and subsequent testing period, rats were fed with standard rat chow and also had a mixture of powdered rat chow and Crisco (“mash”, see recipe below) available to enhance weight gain after surgery and to accustom them to the food that would be presented during test sessions. All procedures were approved by the Institutional Animal Care and Use Committee and conformed to the “Guiding Principles for Research Involving Animals and Human Beings” of the American Physiological Society.
Testing occurred during the light phase of the light-dark cycle in a Plexiglas chamber (24.5-cm diameter 26.5 cm) that allowed the rat to move freely. After 4 days of recovery from surgery, rats were adapted to the observation chamber for 2–3 days prior to testing.
There were 4 test sessions that took place on consecutive days except for a 48 – 72 hr break between sessions three and four. A test session consisted of 8 fifteen minute blocks (fig. 1). Each block began with access to palatable wet mash (27 g crumbled rat chow plus 11 g Crisco) presented in a small dish. At the end of the 15 minute period, the dish was removed and weighed to determine the amount of food consumed and the rat was presented with intraoral (IO) taste stimulation consisting of a 50-μl infusion of sucrose (0.1 M) and 3 water rinses. After the water rinses the food dish was placed back in the test chamber for the next 15 minute block. The first test session was a baseline session with no drug infusions. Test sessions 2–4 included intracranial drug infusions given between blocks 2 and 3. The second test session consisted of infusing NPY 3-36 (5 μg/5μl) into 3V. NPY 3-36 was selected among several potential NPY agonists because of its high potency in increasing feeding when infused into the 3V (Flynn, Turrin, Plata-Salaman, & Ffrench-Mullen, 1999; O’Shea et al., 1997). The third test session consisted of simultaneous infusions of NPY 3-36 into 3V and muscimol into the medullary RF. Infusions were made bilaterally into the IRt/PCRt (0.06nmol/100nl – .09 nmol/150 nl) or a single infusion into Gi (0.06nmol/100 nl – 0.12 nmol/300 nl) or, in one instance a bilateral midline infusion (0.09 nmol/150 nl). The fourth test session was again NPY 3-36 into 3V with a saline infusion into the medullary RF. With this protocol we could measure food intake and EMG responses both within a session, i.e. compare pre- and post-drug, and across sessions, i.e. drug, no drug, saline.
The procedures for EMG recording were similar to those described previously (Chen et al., 2001; Travers & Norgren, 1986). 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 and food consumption were obtained for the first two blocks, the stylets were removed and infusion cannulae containing drug or saline were inserted into the brain via the guide cannulae. 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). NPY 3-36 was delivered at a rate of 1 μl/min; muscimol at a rate of 100 nl/min. 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. One series was stained with cresyl violet for identification of neural structures, the other series was merely dehydrated and cleared to preserve the fluorescence. Injection sites were identified with fluorescent microscopy.
Using 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 (Chen & Travers, 2003; Chen et al., 2001). 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. From the individual burst onsets and offsets we measured mean peak activity normalized to the pre-drug values obtained in block 1 for each test session. Drug effects were compared against saline using a repeated-measures ANOVA that included blocks 2–8.
Infusions of NPY 3-36 had a profound effect on food intake. Under ad lib conditions, rats consumed relatively little while in the test chamber in the no drug condition (Fig. 2). In contrast, infusions of NPY 3-36 alone or in combination with saline into the IRt/PCRt resulted in a significant increase in food consumption. When NPY 3-36 infusions were combined with muscimol infusions in the IRt/PCRt, however, there was a 69% reduction in food intake compared to the NPY/saline condition. A repeated measures ANOVA across the four conditions showed a significant drug effect (p < .001), and post hoc Bonferroni-adjusted comparisons indicated that both the baseline (no drug) and the NPY 3-36/muscimol conditions resulted in significantly less intake compared to either NPY 3-36 alone or NPY 3-36 combined with saline. Examples of EMG activity recorded under these different conditions are presented in figure 3. Although animals typically did not ingest the wet mash in the absence of NPY 3-36 stimulation, they were responsive to intraoral stimulation with sucrose (Fig. 3A). In the presence of NPY 3-36, rats consumed more and rhythmic AD activity captured these ingestive movements (Fig. 3B). Infusions of muscimol into the IRt/PCRt at the same time as 3V NPY 3-36 infusions suppressed both licking and masticatory movements. Figure 3C shows a trial in which licking was totally suppressed; however, during the 15 minutes preceding the IO stimulation, the animal ingested a small amount of food associated with masticatory movements that were reduced in amplitude.
The total amount of mash consumed when NPY 3-36 was combined with IRt/PCRt muscimol was only reduced by 69% compared to NPY 3-36 combined with saline. However, an analysis of the time course of ingestion revealed a more pronounced effect in the blocks immediately after drug infusion (Fig. 4A). Infusions of NPY 3-36/saline resulted in an immediate increase in food consumption (block 3) that diminished over the subsequent blocks. In comparison, NPY 3-36/muscimol did not produce a spike in mash consumption but rather, consumption increased slightly over the course of the next 90 minutes. An ANOVA indicated a significant main effect of drug (p < .001) as well as a significant block X drug interaction (p <.005).
In parallel with the suppression of solid food intake following IRt/PCRt infusions of muscimol, the amplitude of the anterior digastric EMG diminished following intra-oral infusions of sucrose (Fig. 4B). The amplitude returned to baseline by block 7 in parallel with the increased food consumption evident in figure 4A. A repeated measures ANOVA indicated a significant drug (p <.005) and block effect (p <.014) but no block X drug interaction.
Compared to IRt/PCRt infusions, muscimol into Gi had little effect on food consumption evoked by NPY 3-36 (Fig. 5). An ANOVA revealed a significant drug effect (p < .001) but post-hoc Bonferoni-adjusted t-tests showed that only the baseline condition was significantly different from the other blocks; i.e., in the absence of any central drug infusions, these non-deprived rats ingested only a minimal amount of food but NPY 3-36 3V injections potently increased intake, regardless of whether or not the GI was inactivated by muscimol. This conclusion held up to the greater scrutiny permitted by analyzing the time course of ingestion following NPY 3-36 infusions after saline or muscimol infusions into Gi. The NPY 3-36 infusion combined with saline into the GI showed a peak in ingestion immediately after drug infusion (Fig. 6A), similar to the group receiving lateral tegmental field infusions. However, this condition did not differ from NPY 3-36 paired with the muscimol infusions into Gi (ANOVA, p>.1). Likewise, an analysis of the EMG contractions associated with IO stimulation indicated no difference between NPY 3-36 combined with saline and NPY 3-36 paired with Gi muscimol infusions (Fig. 6B).
A summary of the injection sites is shown in figure 7. Lateral infusions were centered subjacent to the rostral nucleus of the solitary tract at a level approximately 0.5 mm caudal to the rostral pole, in the PCRt/IRt lateral to Gi and medial to the trigeminal sensory complex Midline infusions were centered at the same rostral-caudal level. Inspection of the pattern of spread of 100nl of Fluorogold suggested that the effective injection sites were mostly limited to the PCRt/IRt for the lateral infusions and to the Gi for the medial infusions.
The present experiments demonstrate that the consummatory response of mastication induced by a forebrain appetitive signal is profoundly suppressed when neurons in the IRt/PCRt are inactivated with muscimol. These results extend previous observations showing that neurons in the IRt/PCRt are necessary for the consummatory response of licking induced with intra-oral stimulation (Chen & Travers, 2003; Chen et al., 2001). It is unlikely that the suppression of consummatory responses was as a result of direct suppression of either the motor trigeminal or hypoglossal nucleus via the spread of the muscimol to these structures. Examination of the Fluorogold injections made at the conclusion of each experiment did not reveal any encroachment on either motor nucleus.
Together, these observations suggest the general importance of the IRt/PCRt in constituting a final common pathway for ingestive behavior. The essential role of the IRt/PCRt in the ingestive act is in conflict with models postulating that the medulla is mostly inhibitory to consummatory rhythm generation. These models are based on in vitro experiments demonstrating that glutamate-evoked rhythmic responses of motor trigeminal neurons become less pronounced when the medulla is included in the preparation (Tanaka et al., 1999). However, the functional import of these in vitro observations is difficult to reconcile with the present in vivo data that imply a predominant excitatory drive from IRt/PCRt to these same cells in behaving animals. An excitatory role for the medullary IRt/PCRt is also more in line with in situ hybridization and immunohistochemical studies demonstrating that, despite a significant GABAergic minority, the majority of IRt/PCRt neurons with direct projections to the mV are glutamatergic (Travers, Yoo, Chandran, Herman, & Travers, 2005; Turman & Chandler, 1994).
As with previous reports, NPY 3-36 infused into 3V initiated a robust feeding response (Flynn, Plata-Salaman et al., 1999; O’Shea et al., 1997). Rats in the present study ate 4.85 g in a 90 min period, comparing favorably with a previous study demonstrating an intake ~ 4.5 g over 2 hrs with a comparable dose (Flynn, Plata-Salaman et al., 1999). The precise pathway through which this response is effected is unknown, but likely involves the NPY Y5 receptor (MacNeil, 2007). Although NPY 3-36 has an affinity for the Y2 receptor that is primarily anorexogenic (Abbott et al., 2005; Batterham et al., 2002), it also has an affinity for Y5 receptors that are orexigenic (Gerald et al., 1996). Highly selective Y5 agonists induce food intake (Cabrele et al., 2000;E. M. Parker et al., 2000), an effect blocked by Y5 antagonists (Criscione et al., 1998; E. M. Parker et al., 2000). Because Y5 receptors are found within the hypothalamus, including the arcuate nucleus, paraventricular nucleus and lateral hypothalamus (R. M. Parker & Herzog, 1999), it is likely that NPY 3-36 infused into 3V activates these structures to promote food intake.
There are multiple descending pathways through which a forebrain appetitive signal could engage brainstem areas that mediate consummatory responses (Berthoud, 2002; Kelley & Berridge, 2002; Morton et al., 2006; Saper et al., 2002; Travers, DiNardo, & Karimnamazi, 1997), including direct projections to the IRt/PCRt from the lateral hypothalamus (Luiten, ter Horst, & Steffens, 1987; Notsu, Tsumori, Yokota, Sekine, & Yasui, 2008; Shammah-Lagnado, Costa, & Ricardo, 1992). In addition to integrating signals related to energy balance, the hypothalamic projection may be particularly critical in serving as an interface with ventral forebrain regions implicated in motivation and reward (Kelley et al., 2005). The present results indicate that the IRt/PCRt represents an essential link in transforming hypothalamic signals into feeding-related motor patterns. Other ventral forebrain structures that could impact the IRt/PCRt directly to trigger or modulate feeding include the central nucleus of the amygdala (Cassell, Gedney, & Agassandian, 2003; Jongen-Relo & Amaral, 1998) (which also contains some Y5 receptors) and the bed nucleus of the stria terminalis (Holstege, Meiners, & Tan, 1985).
Other data emphasize the role of the orbitofrontal cortex and its projections as a circuit for initiating mastication (Nakamura & Katakura, 1995) (Lund et al., 1998). Stimulation of the cortical masticatory area in the orbitofrontal cortex produces masticatory movements in anesthetized preparations and these cortical outputs have been proposed to influence the brainstem via the pyramidal tract. According to one model, Gi neurons transform a “tonic” cortical signal via the pyramidal tract to the fundamental masticatory rhythm which is then relayed to pre-oromotor neurons in IRt/PCRt (Nakamura & Katakura, 1995). Although the present experiment did not directly address cortically-evoked chewing, we observed no effect of Gi inactivation on appetitive chewing elicited by 3V NPY 3-36. Thus, it remains unclear what role Gi plays in ingestive behavior under awake conditions. However, there are direct projections to IRt/PCRt from the cortical masticatory area (Shammah-Lagnado et al., 1992; Zhang & Sasamoto, 1990) and these projections could mediate the fine motor control of the masticatory apparatus that is compromised following cortical lesions (Hiraba et al., 2007).
Although the present experiment highlights the critical role of the IRt/PCRt for consummatory behavior, it does not negate the likelihood of a more distributed brainstem consummatory substrate, one that includes the pons. Studies using cortical stimulation in anesthetized animals and en bloc preparations suggest that there is a masticatory rhythm generator in the pons (Lund & Kolta, 2006; Lund et al., 1998). In this model, descending cortical outputs to the pontine nucleus caudalis influence motor trigeminal (mV) neurons via short axon pontine pathways to the RF immediately surrounding mV. Interestingly, the present study replicated a previous observation suggesting that the amplitude and rate of anterior digastric contractions are independent (Chen et al., 2001). Thus, in cases where IRt/PCRt muscimol produced only a partial reduction in anterior digastric amplitude, allowing the residual mastication to be observed, contraction rate was essentially invariant (Fig. 8). Thus, it is possible that a pontine rhythm generator could influence interneurons and pre-oromotor neurons in the IRt/PCRt to produce the complete consummatory pattern. On the other hand, recent observations in our laboratory suggest that the rhythm-generating substrate itself might extend to the IRt/PCRt since infusions of a mu opiate agonist into this region profoundly slow lick rate (Kinzeler & Travers, 2008).
A number of experimental approaches lead to the conclusion that the IRt/PCRt represents an important node in the interface between brainstem and forebrain structures regulating homeostatic control of ingestion and the execution of the consummatory response (Fig. 9). Neurons in the IRt/PCRt that project to the oromotor nuclei are rhythmically active during both cortically induced (electrical) stimulation (Inoue, Chandler, & Goldberg, 1994; Sahara, Hashimoto, & Nakamura, 1996) as well as intraoral stimulation in the awake, freely moving preparation (Travers, DiNardo, & Karimnamazi, 2000). In addition to the forebrain projections discussed above, neurons in the IRt/PCRt receive input from the (gustatory) rNST (Nasse et al., 2008) and parabrachial nucleus (Karimnamazi & Travers, 1998), as well as trigeminal somatosensory input (Dauvergne, Pinganaud, Buisseret, Buisseret-Delmas, & Zerari-Mailly, 2001; Zerari-Mailly, Pinganaud, Dauvergne, Buisseret, & Buisseret-Delmas, 2001). Thus, the IRt/PCRt may be an important integrative node for consummatory ingestive behavior that receives input from multiple descending forebrain as well as local orosensory sources.
This work was supported by NIH DC-00416 and NIH DC-00417.
Publisher's Disclaimer: The following manuscript is the final accepted manuscript. It has not been subjected to the final copyediting, fact-checking, and proofreading required for formal publication. It is not the definitive, publisher-authenticated version. The American Psychological Association and its Council of Editors disclaim any responsibility or liabilities for errors or omissions of this manuscript version, any version derived from this manuscript by NIH, or other third parties. The published version is available at www.apa.org/pubs/journals/bne