CCK-Induced Reduction of Food Intake and Hindbrain MAPK Signaling Are Mediated by NMDA Receptor Activation

Abstract

The dorsal vagal complex of the hindbrain, including the nucleus of the solitary tract (NTS), receives neural and humoral afferents that contribute to the process of satiation. The gut peptide, cholecystokinin (CCK), promotes satiation by activating gastrointestinal vagal afferents that synapse in the NTS. Previously, we demonstrated that hindbrain administration of N-methyl-d-aspartate (NMDA)-type glutamate receptor antagonists attenuate reduction of food intake after ip CCK-8 injection, indicating that these receptors play a necessary role in control of food intake by CCK. However, the signaling pathways through which hindbrain NMDA receptors contribute to CCK-induced reduction of food intake have not been investigated. Here we report CCK increases phospho-ERK1/2 in NTS neurons and in identified vagal afferent endings in the NTS. CCK-evoked phospho-ERK1/2 in the NTS was attenuated in rats pretreated with capsaicin and was abolished by systemic injection of a CCK1 receptor antagonist, indicating that phosphorylation of ERK1/2 occurs in and is mediated by gastrointestinal vagal afferents. Fourth ventricle injection of a competitive NMDA receptor antagonist, prevented CCK-induced phosphorylation of ERK1/2 in hindbrain neurons and in vagal afferent endings, as did direct inhibition of MAPK kinase. Finally, fourth ventricle administration of either a MAPK kinase inhibitor or NMDA receptor antagonist prevented the reduction of food intake by CCK. We conclude that activation of NMDA receptors in the hindbrain is necessary for CCK-induced ERK1/2 phosphorylation in the NTS and consequent reduction of food intake.

Increases in meal size are associated with the development of obesity in both rodents (see for example Refs. ^1–3) and humans (see for example Refs. ⁴ and ⁵), suggesting that alterations in the process of satiation may contribute to excess caloric intake and weight gain. The size of individual meals is controlled by neural and humoral gastrointestinal signals that promote satiation and result in meal termination. The dorsal vagal complex (DVC) in the caudal hindbrain is the site where vagal afferents terminate and from which signals that participate in the control of meal size are relayed to other brain regions. Cholecystokinin (CCK), the first gut peptide shown to reduce meal size (6, 7), activates gastrointestinal vagal afferent fibers that synapse in the nucleus of the solitary tract (NTS). Although vagal mediation of reduction of food intake by CCK is well established, the processing of CCK-evoked signals in the NTS is not fully understood.

Glutamate is the primary neurotransmitter released from vagal afferents (8) and is released by some NTS neurons and glial cells as well (9, 10). Electrophysiological and immunohistochemical experiments reveal that vagal afferent stimulation activates NTS neurons via NMDA- and α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA)-type glutamate receptors (11–13). Hence, glutamatergic transmission in the NTS must play an integral role in the processing of gastrointestinal satiation signals like CCK. Indeed, previous reports from our group and others indicate that antagonism of hindbrain NMDA-type glutamate receptors (NMDAR) delays satiation and increases meal size (14–16). Recently, we reported that fourth ventricle (4V) or NTS injections of NMDAR antagonists prevent CCK-induced reduction of food intake (17). This observation suggests that activation of hindbrain NMDAR is required for CCK-induced reduction of food intake.

Several reports suggest that activation of the MAPK signaling pathway in the hindbrain plays a role in the reduction of food intake by CCK (18, 19). Specifically, ip injection of CCK triggers phosphorylation of extracellular signal-related kinase 1/2 (ERK1/2) in the NTS and inhibition of ERK1/2 phosphorylation [phosho-ERK1/2 (pERK1/2)] attenuates CCK-induced reduction of food intake (18). These observations are intriguing because MAPK signaling can be activated by increased intracellular calcium generated by NMDAR activation (20–22).

We hypothesized that CCK-induced phosphorylation of ERK1/2 in the DVC and consequent reduction of food intake depends on activation of NMDAR. To test this hypothesis, we evaluated whether CCK-induced ERK1/2 phosphorylation occurs in NTS neurons and identified vagal afferent endings, both of which are known to express NMDAR (23–28). We also examined the effects of 4V NMDAR and AMPAR antagonists on pERK1/2 immunoreactivity after ip injection of CCK. Finally, we examined the CCK-induced reduction of food intake after antagonism of hindbrain NMDAR or MAPK kinase (MEK) inhibition. Results of our experiments indicate that activation of NMDAR in the NTS is required for CCK-evoked ERK1/2 phosphorylation in NTS neurons and central terminals of gastrointestinal vagal afferents and that NMDAR-mediated ERK1/2 phosphorylation is required for reduction of food intake by CCK.

Materials and Methods

Animals and housing

Male adult Sprague Dawley rats (Simonsen Laboratories, Gilroy, CA) were individually housed in suspended wire mesh cages in a vivarium with 12-h light, 12-h dark cycle (lights out 1800 h) and controlled temperature (22 C) and humidity (70%). The rats weighed 280–300 g at the beginning of experiments and were handled daily and habituated to laboratory conditions before surgery or testing began. They had ad libitum access to water and pelleted rodent diet (Teklad, Kent, WA) except during overnight fasts, as described below. All animal housing and experiments reported here were conducted in compliance with the National Institutes of Health Guide for Care and Use of Laboratory Animals under a protocol approved by the Washington State University Institutional Animal Care and Use Committee.

Drugs and antibodies

2,3-Dioxo-6-nitro-1,2,3,4,tetrahydrobenzo[f]quinoxaline-7-sulfonamide (NBQX), d-3-(2-carboxypiperazine-4-yl)-1-propenyl-1-phosphonic acid (d-CPP-ene), and MEK inhibitor U0126 were purchased from Tocris (Ellisville, MO). Sulfated CCK octapeptide (CCK-8) was from Peptides International (Louisville, KY). Cell Signaling Technology (Beverly, MA) supplied the primary antibody for phospho-p44/42 MAPK (Thr202/Tyr 204 pERK1/2). The antibody for the NMDA receptor NR1 subunit was obtained from Santa Cruz Biotechnology (Santa Cruz, CA). Isolectin B4 conjugated with biotin (IB4) was purchased from Vector Laboratories (Burlingame, CA). Biotinylated dextran amine (BDA) was obtained from Invitrogen (Carlsbad, CA). Capsaicin and devazepide were obtained from Sigma-Aldrich (St. Louis, MO).

Immunohistochemical detection of pERK1/2 and double-labeling procedures

Rats were adapted to experimental conditions by being handled and given ip injections of sterile 0.9% NaCl (1.0 ml/kg body weight) daily for 1 wk preceding the experiment and were fasted overnight (16 h) before the experiment. On the test day, the rats were injected with 0.9% sterile NaCl or CCK (10 μg/ml; 1 ml/kg in 0.9% NaCl) and rapidly anesthetized 13 min later with isoflurane inhalation anesthetic (Vedco, St. Joseph, MO). Two minutes later, the animals were perfused intracardially with 0.1 m phosphate buffer NaCl followed by 4% paraformaldehyde (Electron Microscopy Sciences, Hatfield, PA) in 0.1 m phosphate buffer (pH 7.4). Immediately after perfusion, brains were collected, postfixed in the same fixative for 2 h, and cryoprotected in 0.1 m phosphate buffer containing 25% sucrose overnight at 4 C. Thirty-micrometer coronal cryostat sections of the hindbrain were collected for immunostaining. For detection of pERK1/2 and NR1, sections were incubated for 36 h at room temperature in mouse anti-pERK1/2 (1:250) and goat anti-NR1 (1:100) antisera. Then, sections were washed and incubated in Alexa 555-conjugated donkey antimouse (1:600; Invitrogen) and Alexa 488-conjugated donkey antigoat (1:600; Invitrogen) antisera for 3 h. For pERK1/2 and IB4 double labeling, hindbrain sections were incubated in mouse anti-pERK1/2 (1:250) antisera and biotinylated IB4 (1:400) for 36 h at room temperature. Subsequently, the sections were incubated with Alexa 488-conjugated goat antimouse (1:600; Invitrogen) antiserum and Cy3 conjugated with ExtrAvidin (1:200; Sigma, St. Louis, MO). The stained sections were then mounted on slides and coverslipped with ProLong Gold (Invitrogen) before microscopic examination.

BDA anterograde injections

Rats (n = 3) were fasted overnight and anesthetized with a mixture consisting of ketamine 50 mg/kg (Pfizer, New York, NY), xylazine 25 mg/kg (Vedco Inc., St. Joseph, MO), and acepromazine 2 mg/kg (Boehringer Ingelheim Vetmedica, St. Joseph, MO). The left cervical vagus nerve was exposed via a ventral midline neck incision, and the left nodose ganglion was exposed. BDA (10% solution of 10-kDa BDA in 0.1 m PBS) was injected into the nodose ganglion using a 36-gauge stainless steel needle (World Precision Instruments, Sarasota, FL) attached to a micro-syringe pump (World Precision Instruments). The needle was inserted under the perineurium just distal to the ganglion, and a total volume of 2 μl of the BDA solution was delivered at a rate of 25 nl/sec with microscopic observation. Fast Green (1%) was added to the BDA solution to visualize diffusion of the tracer throughout the nodose ganglion.

Ten days after unilateral BDA injection, overnight-fasted rats were given an ip injection of CCK (10 μg/kg) and perfused 15 min later, as described above. Brain tissue was harvested and prepared for immunohistochemical detection of pERK1/2 and BDA, as described above.

Capsaicin treatment

Systemic treatment with capsaicin was used to destroy small unmyelinated primary afferent neurons, including those in the vagi, as previously described (29–31). The total capsaicin dose (125 mg/kg) was administered ip as a series of three injections (25, 50, and 50 mg/kg) at an injection volume of 1 ml/kg (n = 8). All three injections were made within a 24-h period (0, 6, and 24 h, respectively). An additional group of rats were injected with the vehicle solution (n = 8; 10% ethanol and 10% Tween 80 in 0.9% NaCl) at the same volumes and schedule as above. Rats were maintained at a surgical plane of anesthesia using isoflurane for all capsaicin and vehicle injections, and artificial respiration was provided as needed. Rats were allowed to recover for 2 wk during which both capsaicin- and vehicle-treated animals gained weight. The eye wipe response to mild corneal stimulation, a response mediated by the capsaicin-sensitive trigeminal innervation of the cornea, was tested to screen the effectiveness of capsaicin treatment (29, 30). The test involved application of a drop of 1% NH₄OH to one eye. All of our vehicle-treated rats immediately wiped the stimulated eye. However, none of the capsaicin-treated rats exhibited any eye-wipe response to the test, confirming that our capsaicin treatment successfully destroyed afferent C fibers.

After the recovery period, capsaicin- and vehicle-treated groups were fasted overnight and administered an ip injection of CCK (2 μg/kg) or 0.9% NaCl to achieve the following treatment conditions: vehicle/NaCl (n = 4), vehicle/CCK (n = 4), capsaicin/NaCl (n = 4), and capsaicin/CCK (n = 4). Fifteen minutes after ip injection, animals were anesthetized and perfused, and brain tissue was prepared for immunohistochemical detection of pERK1/2 and IB4. Subsequently, the sections were incubated with Alexa 555-conjugated goat antimouse (1:600; Invitrogen) and Alexa 488 conjugated with streptavidin (1:600; Invitrogen) secondary antisera.

The number of pERK1/2-immunoreactive cell bodies were counted manually in anatomical regions of interest (ROI) from each rat: one section at each of four different brain levels, corresponding to 14.1, 13.8, 13.6, and 13.3 mm caudal to bregma, according to the stereotaxic atlas of Paxinos and Watson (32). ROI were determined using landmarks established from the atlas of Paxinos and Watson (32): the medial NTS (mNTS), commissural NTS (cNTS), interstitial NTS (isNTS), and area postrema (AP). The pERK1/2-immunoreactive cell bodies were counted on both sides of the hindbrain in each section containing a given ROI for each rat. The numbers of pERK1/2-immunoreactive cells are presented as averages for each brain area across all of the rostrocaudal levels at which the particular structure appeared. pERK1/2 immunoreactivity in the neuropil was quantified by digitally excluding cell bodies and measuring the mean fluorescence intensity of the nonsomal area of mNTS. These values were normalized to the background intensity, which was sampled from the noncellular area of the hypoglossal nucleus. The data are presented as the average relative neuropil fluorescence intensity for each brain across the four rostrocaudal levels listed above. The relative neuropil intensity was calculated by applying the following formula: relative intensity = (intensity_neuropil/intensity_background × 100) − 100.

Effects of systemic administration of devazepide on CCK-induced stimulation of pERK1/2

Overnight-fasted rats (total n = 16) were divided into the following treatment conditions: vehicle/NaCl (n = 4), vehicle/CCK (n = 4), devazepide/NaCl (n = 4), and devazepide/CCK (n = 4). Devazepide (300 μg/kg) or vehicle [50% dimethylsulfoxide (DMSO) in 0.9% NaCl] was administered ip 10 min before ip injection of CCK (2 μg/kg) or 0.9% NaCl. Fifteen minutes after the final ip injection, animals were anesthetized, perfused, and prepared for immunohistochemical detection of pERK1/2. After incubation in the primary antiserum, hindbrain sections were incubated with Alexa 488-conjugated goat antimouse (1:600) secondary antiserum. Quantification of hindbrain pERK1/2-immunoreactive neurons and neuropil intensity was done as previously described.

Fourth ventricular cannulations

Rats were anesthetized as previously described with a ketamine/xylazine/acepromazine mixture, placed in a stereotaxic instrument with head leveled between λ and bregma, and implanted with 26-gauge stainless steel (McMaster-Carr, Santa Fe Springs, CA) guide cannulas aimed for the 4V (2.0 mm anterior to the occipital suture, on the midline, 6.6 mm ventral from surface of the dura). Cannulas were secured to the skull using three stainless steel screws and methacrylate (Orthojet; Patterson Dental Supply, Spokane, WA). Immediately after surgery, analgesic, flunixin meglumine 2.5 mg/kg (MWI, Meridian, ID) and antibiotic, procaine penicillin G 300,000 U/kg (Norbrook, Lenexa, KS) was administered sc. Rats were allowed 2 wk postsurgical recovery, during which time they all exceeded their presurgical body weights.

Effects of hindbrain administration of d-CPP-ene or NBQX on CCK-induced pERK1/2

A group of rats (total n = 24) were divided into six subgroups of four rats each. After an overnight (16 h) fast, each subgroup received one of the following treatments: NaCl/NaCl (n = 4), NaCl/CCK (n = 4), d-CPP-ene/CCK (n = 4), d-CPP-ene/NaCl (n = 4), NBQX/CCK (n = 4), and NBQX/NaCl (n = 4). Fourth ventricle injections of d-CPP-ene (40 ng), NBQX (250 nmol), or 0.9% NaCl in volumes of 3 μl were made via a 26-gauge guide cannula using a 33-gauge injector attached to a 10-μl Hamilton syringe over a 2-min period. Fourth ventricle injections were followed immediately by an ip injection of either CCK (10 μg/kg) or 0.9% NaCl. A 10-μg/kg dose of CCK was used to get a maximal pERK1/2 response and because we previously reported that 4V injection of d-CPP-ene (40 ng) effectively attenuates CCK-induced (10 μg/kg) c-Fos in the hindbrain (17). Fifteen minutes after ip injection, animals were anesthetized, perfused, and prepared for immunohistochemical detection of pERK1/2 as previously described. After incubation in the primary antiserum, sections were incubated in Alexa 555-conjugated goat antimouse secondary antiserum (1:600). Quantification of hindbrain pERK1/2-immunoreactive neurons and neuropil intensity was done as previously described.

Effects of MEK inhibitor on CCK-induced pERK1/2 immunoreactivity in hindbrain neurons and neuropil

A group of fasted (16 h) rats (total n = 15) were divided into the following treatment groups: vehicle/NaCl (n = 4), vehicle/CCK (n = 4), U0126/CCK (n = 4), and U0126/NaCl (n = 3). Fourth ventricle injections of vehicle (50% DMSO dissolved in 0.9% NaCl) or U0126 (2 μg dissolved in vehicle solution) in volumes of 3 μl were injected over a 2-min period. Forty-five minutes after 4V injections, rats received an ip injection of either CCK (2 μg/kg) or 0.9% NaCl. Fifteen minutes after ip injection, animals were anesthetized, perfused intracardially, and prepared for immunohistochemical detection of pERK1/2, as described above. Quantification of hindbrain pERK1/2-immunoreactive neurons and neuropil intensity was done as previously described.

Effects of 4V administration of d-CPP-ene and U0126 on CCK-induced reduction of food intake

Rats (n = 13) implanted with 26-gauge cannulas aimed for the 4V were used to assess the effect of 4V U0126 injection on the reduction of food intake by ip CCK. Rats were adapted to eat a meal of pelleted rodent diet over 30 min after an overnight 16-h fast. Briefly, food, but not water, was removed 1 h before lights out, and a weighed amount was returned between 0900 and 1000 h the next day. Food intake and spillage were measured over the ensuing 30 min, after which the rats were allowed ad libitum access to both food and water for at least the next 48 h before another overnight fast was imposed. On experimental days, each rat received a 3-μl injection of either vehicle (50% DMSO/0.9% NaCl) or U0126 (2 μg in 50% DMSO/0.9% NaCl) into the 4V. Forty-five minutes after the 4V injection, animals received an ip injection of either CCK (2 μg/kg in 0.9% NaCl) or 0.9% NaCl. Immediately after the ip injection, rats were given immediate access to a preweighed amount of pelleted rodent diet and intake less spillage was recorded for 30 min, as described above. Each rat received the following testing injection combinations in the following order: 4V vehicle/ip NaCl, 4V vehicle/ip CCK, 4V U0126/ip CCK, 4V U0126/ip NaCl, 4V vehicle/ip CCK, 4V U0126/ip CCK, and 4V vehicle/ip NaCl. For data analysis and presentation, intakes for repeated experimental conditions were averaged.

In a replication of a past experiment, we used 11 rats from the same group of animals to assess the effect of d-CPP-ene on the reduction of food intake. Rats were fasted overnight, and a weighed amount of food was returned between 0900 and 1000 h the next day. Each rat received a 3-μl injection of either 0.9% NaCl or d-CPP-ene (40 ng in 0.9% NaCl). Five minutes after the 4V injection each received an ip injection of either 0.9% NaCl or CCK (2 μg/kg in 0.9% NaCl). Immediately after the ip injection, the rats were given access to a preweighed amount of pelleted rodent diet, and intake less spillage was recorded for 30 min. Each rat received the following testing injection combinations in the following order: 4V NaCl/ip NaCl, 4V d-CPP-ene/ip CCK, 4V NaCl/ip NaCl, and 4V d-CPP-ene/ip NaCl. Intakes for 4V NaCl/ip NaCl were averaged for data analysis and presentation.

Statistics

Data were analyzed using appropriate repeated-measures ANOVA, followed by Tukey's post hoc analysis. In behavioral experiments, the repeated factor was treatment condition, whereas the repeated factor for counts of pERK-positive neurons and fluorescence intensity was the brain area analyzed. The confidence limit for statistical significance was set at P < 0.05. However, wherever actual confidence limits were substantially less than 0.05, those P values are provided. Results are presented as means ± sem.

Results

CCK-induced ERK1/2 phosphorylation in hindbrain neurons

NR1-immunoreactive neurons were abundant in the DVC of the hindbrain, and ip administration of CCK (10 μg/kg) triggered phosphorylation of ERK1/2 in NR1-positive neurons. Virtually all of the neurons that exhibited pERK1/2 after CCK injection were NR1 immunoreactive. We observed a high density of NR1 and pERK1/2 colocalization in the mNTS (Fig. 1A) and AP (Fig. 1C). NR1 and pERK1/2 colabeling was observed in neurons of various sizes including neurons of smaller phenotype in the AP. Intraperitoneal injection of NaCl produced minimal ERK1/2 phosphorylation; thus, negligible levels of NR1 and pERK1/2 colocalization was observed (data not shown).

Fig. 1.

Dual-label immunofluorescent images showing CCK-induced pERK1/2 immunoreactivity colocalized with NMDAR NR1 immunoreactivity in mNTS and AP neurons. A–C, pERK1/2-immunoreactive cells (red); A′–C′, NR1 immunoreactivity (green); A″–C″, merged images illustrating colocalization of pERK1/2 and NR1 immunoreactivity (white arrows); A–A″, ×20 images of the mNTS; B–B″, ×100 images of a double-labeled neuron in the mNTS; C–C″, ×20 images of the SP. Scale bars, 100 μm (A and C) and 25 μm (B).

CCK triggers ERK1/2 phosphorylation in hindbrain neuropil and identified vagal afferent endings

In addition to triggering ERK1/2 phosphorylation in hindbrain neuronal cell bodies, peripheral administration of CCK (10 μg/kg) increased pERK1/2 immunoreactivity in the NTS neuropil in a pattern consistent with the distribution of gastrointestinal vagal afferent innervation. This overlap was evident in immunohistochemically prepared hindbrain tissue that was stained for pERK1/2 and IB4 binding, which is limited to primary afferent C fibers in the NTS (Fig. 2, A–C). At a lower magnification (×20), there appeared to be an overlap of pERK1/2 and IB4 fiber staining in the cNTS, mNTS, isNTS, and AP. Colocalization of pERK1/2 and IB4 was confirmed by examining high magnification (×100) z-stacks of optical sections (Fig. 2D).

Fig. 2.

Dual-label immunofluorescent images showing CCK-induced pERK1/2 immunoreactivity localized in vagal afferent endings in the NTS. A, Stitched images (×20) illustrating pERK1/2 immunoreactivity (green) in the NTS and AP; B, IB4 binding in the same section shown in A; C, merged image of pERK1/2 immunoreactivity and IB4 binding; D, images (×100) illustrating colocalization of pERK1/2 (green) with IB4 (red) in vagal afferent endings identified by IB4 binding; E, images (×100) illustrating colocalization of pERK1/2 (green) and BOA anterogradely labeled vagal afferents (red). Scale bars, 200 μm (A–C) and 10 μm (D and E).

To further establish that CCK triggers pERK1/2 in vagal afferents, BDA was injected into the nodose ganglion of the vagus nerve. The anterograde tracer was definitively colocalized with pERK1/2-immunoreactive fibers and terminals in the NTS visualized using high-magnification (×100) z-stacks of optical sections (Fig. 2E).

CCK-induced ERK1/2 phosphorylation in hindbrain neuropil is attenuated after chemical lesion of vagal afferent C fibers

Surgical vagotomy (33, 34) or destruction of vagal C fibers with capsaicin (29) attenuates reduction of food intake by CCK. However, the role of vagal afferents in CCK-induced increase of hindbrain pERK1/2 has not previously been examined. In our rats, capsaicin treatment successfully destroyed vagal afferent C fibers, indicated by the lack of an eye-wipe response and the absence of IB4 binding in the mNTS of capsaicin-treated rats (Fig. 3A). Intraperitoneal CCK induced a significant increase in pERK1/2 immunoreactivity in mNTS neuropil of the vehicle-treated rats (P < 0.001) but not in the capsaicin-treated rats (Fig. 3B). CCK also induced a significant increase in the number of pERK1/2-positive neuron cell bodies in the mNTS in the vehicle group (P < 0.001), cNTS (P < 0.001), isNTS (P < 0.001), and AP (P < 0.01) compared with ip NaCl-injected rats. Compared with ip NaCl, ip CCK in the capsaicin group triggered an increase in pERK1/2-positive neurons in the mNTS (P < 0.001) and cNTS (P < 0.05) but not in the isNTS or AP. Compared with the vehicle-treated group, capsaicin treatment did not attenuate the number of CCK-triggered pERK1/2-immunoreactive cell bodies in the mNTS.

Fig. 3.

Capsaicin lesion of vagal afferent C fibers eliminates CCK-induced pERK1/2 immunoreactivity in NTS neuropil. A, Representative images of dorsal hindbrain sections stained to reveal pERK1/2 immunoreactivity (red) and IB4 binding (green) from capsaicin- or vehicle-treated rats. Row labels and column labels indicate experimental conditions (2 μg/kg CCK or 0.9% NaCl administered ip). Note strong IB4 binding and pERK1/2 immunoreactivity in vehicle-treated rat NTS (top row) and near absence of both IB4 binding and CCK-induced pERK1/2-immunoreactive neuropil in capsaicin-treated rat (bottom row). B, Average relative fluorescence intensity of neuropil sampled from the mNTS across four rostrocaudal levels of the hindbrain and normalized to background intensity. C, Numbers of pERK1/2-immunoreactive neurons counted from coronal sections at four rostrocaudal levels of the hindbrain, as established using the stereotaxic atlas of Paxinos and Watson (32). Counts were made for the following areas of interest at each rostrocaudal level where the given area occurs: mNTS, cNTS, isNTS, and AP. The counts for areas of interest at all levels were summed for each animal and averages and se calculated for each area under each treatment condition. Scale bar, 200 μm. Lowercase “b” above bars indicates significant difference from bars marked with “a” (see Results for confidence limits). Bars marked with the same letter “b” or “a” do not differ significantly.

Blockade of CCK1 receptors prevents CCK-induced ERK1/2 phosphorylation in hindbrain neurons and neuropil

Reduction of food intake by endogenously released CCK (31, 35, 36) or ip injection of exogenous CCK (29, 33, 34) is mediated by gastrointestinal vagal afferents. Electrophysiological results indicate that vagal afferent CCK1 receptor expression is concentrated in afferents that innervate the stomach and upper small intestine (37). Therefore, we examined whether a CCK1 receptor antagonist could prevent CCK-induced increase of pERK1/2 in the DVC. We found that CCK-induced increase in pERK1/2-labeled neurons in the mNTS, cNTS, isNTS, AP, and mNTS neuropil after ip CCK (Fig. 4, E–G) was eliminated by previous ip injection of the CCK1 receptor antagonist devazepide: mNTS (P < 0.001), cNTS (P < 0.001), isNTS (P < 0.001), AP (P < 0.001), and mNTS neuropil (P < 0.001).

Fig. 4.

Intraperitoneal injection of CCK1 receptor antagonist devazepide (Dev) (300 μg/kg) attenuates CCK-induced (2 μg/kg) pERK1/2 immunoreactivity in hindbrain neurons and neuropil. A–D, Representative images of dorsal hindbrain sections stained to reveal pERK1/2 for the following treatments: A, vehicle/NaCl; B, devazepide/NaCl; C, vehicle/CCK; D, devazepide/CCK. E, Average relative fluorescence intensity of neuropil sampled from the mNTS across four rostrocaudal levels of the hindbrain and normalized to background intensity. F and G, Numbers of pERK1/2-immunoreactive neurons counted in coronal sections at four rostrocaudal levels of the hindbrain from the following areas of interest at each rostrocaudal level where the given area occurs: mNTS, cNTS, isNTS, and AP. Scale bar, 200 μm. Lowercase “b” above bars indicates significant difference from bars marked with “a” or “c” (P < 0.001). In panel E, “c” indicates significant difference from bars topped with “a” (P < 0.05). Bars marked with the same letter, “b” or “a” do not differ significantly.

CCK-induced ERK1/2 phosphorylation in the hindbrain is NMDA receptor dependent

Before ip injection of CCK (10 μg/kg), d-CPP-ene (40 ng) was administered via the 4V to determine whether hindbrain NMDAR activation was necessary for CCK-induced ERK1/2 phosphorylation (Fig. 5). Compared with ip NaCl, ip CCK triggered a significant increase in pERK1/2-immunoreactive neurons in the mNTS (P < 0.001), cNTS (P < 0.001), isNTS (P < 0.001), AP (P < 0.01), and mNTS neuropil (P < 0.001) (Fig. 5, B and C). Pretreatment with d-CPP-ene significantly reduced CCK-induced pERK1/2-positive neurons in the mNTS (P < 0.001), cNTS (P < 0.001), isNTS (P < 0.001), AP (P < 0.05), and mNTS neuropil (P < 0.001).

Fig. 5.

The 4V injection of NMDA receptor antagonist d-CPP-ene attenuates CCK-induced pERK1/2 immunoreactivity in the hindbrain DVC. A, Representative images of dorsal hindbrain sections stained to reveal pERK1/2 immunoreactivity after 4V d-CPP-ene (40 ng) or vehicle and ip CCK (10 μg/kg) or vehicle. Column labels indicate treatment conditions, and row labels indicate the distance caudal to Bregma for sections in that row. B, Average relative fluorescence intensity of neuropil sampled from the mNTS across four rostrocaudal levels of the hindbrain and normalized to background intensity. C, Quantification of pERK1/2-immunoreactive neurons counted in coronal sections at four rostrocaudal levels of the hindbrain from the following areas of interest at each rostrocaudal level where the given area occurs: mNTS, cNTS, isNTS, and AP. Lowercase “b” above bars indicates significant difference from bars marked with “a” or “c” (P < 0.001). “c” indicates significant difference from bars topped with “a” (P < 0.05). Bars marked with the same letter do not differ significantly.

Blockade of hindbrain AMPA receptors does not abolish CCK-induced pERK1/2 in the hindbrain

As shown in Fig. 6, pretreatment with NBQX (250 nmol) did not abolish CCK-induced (10 μg/kg) pERK1/2 in the hindbrain. Compared with ip CCK, 4V administration of NBQX before ip CCK did not result in a significant reduction in pERK1/2 immunoreactivity except for in the isNTS (P < 0.001) (Fig. 6, E–G).

Fig. 6.

The 4V administration of an AMPA receptor antagonist (NBQX; 250 nmol) failed to attenuate CCK-induced (10 μ/kg) pERK1/2 immunoreactivity in the hindbrain. A–D, Representative images of dorsal hindbrain sections stained to reveal pERK1/2 for the following treatments: A, 4V NaCl/ip NaCl; B, 4V NBQX/ip NaCl; C, 4V NaCl/ip CCK; D, 4V NBQX/ip CCK. E, Average relative fluorescence intensity of neuropil sampled from the mNTS across four rostrocaudal levels of the hindbrain and normalized to background intensity. F and G, Numbers of pERK1/2-immunoreactive neurons counted in coronal sections at four rostrocaudal levels of the hindbrain from the following areas of interest at each rostrocaudal level where the given area occurs: mNTS, cNTS, isNTS, and AP. Scale bar, 200 μm. Lowercase “b” above bars indicates significant difference from bars marked with “a” or “c” (P < 0.001). In panel G, “c” indicates significant difference from bars topped with “a” (P < 0.05). Bars marked with the same letter do not differ significantly.

Central injection of MEK inhibitor attenuates CCK-induced pERK1/2 immunoreactivity in hindbrain neurons and neuropil

Increases in pERK1/2-labeled neurons in the mNTS, cNTS, isNTS, AP, and NTS neuropil after ip CCK (Fig. 7, E–G) were significantly reduced by previous 4V injection of U0126: mNTS (P < 0.001), cNTS (P < 0.01), AP, and mNTS neuropil (P < 0.05).

Fig. 7.

MEK inhibitor U0126 (2 μg) attenuates CCK-induced (2 μg/kg) pERK1/2 immunoreactivity in hindbrain neurons and neuropil. A–D, Representative images of dorsal hindbrain sections stained to reveal pERK1/2 immunoreactivity after the following treatments: A, 4V vehicle/ip NaCl; B, 4V U0126/ip NaCl; C, 4V vehicle/ip CCK; D, 4V U0126/ip CCK. E, Average relative fluorescence intensity of neuropil sampled from the mNTS across four rostrocaudal levels of the hindbrain and normalized to background intensity. F and G, Quantification of pERK1/2-immunoreactive neurons counted in coronal sections at four rostrocaudal levels of the hindbrain from the following areas of interest at each rostrocaudal level where the given area occurs: mNTS, cNTS, isNTS, and AP. Scale bar, 200 μm. Within individual areas of interest bars topped with different lowercase letters differ significantly (P < 0.05). Bars marked with the same letter do not differ (see Results text for confidence limits).

Fourth ventricle administration of U0126 or d-CPP-ene inhibits CCK-induced reduction of food intake

Repeated-measures ANOVA revealed significant treatment effects on food intake when rats were injected with CCK after 4V NaCl, NMDAR antagonist, or MEK inhibitor [F_(5,50) = 21; P < 0.0001]. As previously reported (17), 4V injection of d-CPP-ene (40 ng) completely (P < 0.05) blocked the reduction of food intake by CCK (Fig. 8A). Sutton et al. (18) previously reported that 4V administration of U0126 attenuates CCK-mediated reduction of 1 h chow intake. We report here that U0126 (2 μg) significantly attenuates (P < 0.01) the reduction of chow intake by CCK (2 μg/kg) as early as 30 min after CCK injection (Fig. 8B).

Fig. 8.

The 4V injection of d-CPP-ene or U0126 attenuated CCK-induced reduction of 30 min food intake. A, 4V injection of d-CPP-ene (40 ng) attenuated CCK-induced reduction of food intake; B, 4V injection of U0126 (2 μg) attenuated CCK-induced (2 μg/kg) reduction of food intake. Bars topped with different lowercase letters differ significantly (P < 0.05). Bars marked with the same letter do not differ.

Discussion

CCK triggers ERK1/2 phosphorylation in the dorsal hindbrain, and hindbrain ERK1/2 phosphorylation is necessary for reduction of food intake by CCK (18). Our results indicate that CCK-induced hindbrain ERK1/2 phosphorylation and reduction of food intake requires activation of NMDAR in the hindbrain NTS. Specifically, CCK triggered ERK1/2 phosphorylation not only in hindbrain neurons but also in identified primary vagal afferent endings in the NTS. Like CCK-induced reduction of food intake, CCK-induced ERK1/2 phosphorylation in vagal afferents was abolished after destruction of vagal C fibers. Systemic injection of a CCK1 receptor antagonist or hindbrain NMDAR antagonist dramatically attenuated CCK-induced ERK1/2 phosphorylation in NTS neurons and vagal afferent fibers. Finally, inhibition of ERK1/2 phosphorylation by 4V injection of a MEK inhibitor attenuated CCK-induced reduction of food intake in a manner identical to that observed after NMDAR antagonist administration. Collectively, our results indicate that hindbrain NMDAR-mediated ERK1/2 phosphorylation in NTS neurons and/or vagal afferent endings is essential for CCK-induced reduction of food intake.

NMDAR coupling to the MAPK signaling cascade leading to ERK1/2 phosphorylation is well established in central and peripheral neurons (21, 38–40). However, ours is the first report of NMDAR-dependent ERK1/2 phosphorylation in the NTS. Expression of NMDAR subunit immunoreactivity and mRNA in hindbrain NTS neurons is well documented (23–27). Moreover, electrophysiological studies using whole-cell patch-clamp recordings in hindbrain slices indicate that functional NMDAR are expressed by NTS neurons that receive primary vagal afferent input (13, 41, 42). Therefore, it is plausible that NMDAR antagonists attenuate CCK-induced pERK1/2 and reduction of food intake through an action on intrinsic neurons. Our immunohistochemical observations revealed that virtually all NTS neurons that exhibited pERK1/2 after CCK injection are immunoreactive for the NMDAR NR1 subunit, an obligatory component of all NMDAR (43). Therefore, the presence of NR1 is indicative of NMDAR expression. The high correlation between NMDAR NR1 immunoreactivity and CCK-evoked pERK1/2 expression, taken together with our observation that NMDAR antagonist prevents CCK-evoked ERK1/2 phosphorylation, is consistent with coupling of ERK1/2 phosphorylation to NMDAR activation in the NTS.

We found that CCK-evoked pERK1/2 immunoreactivity in NTS neuropil was localized largely to fibers that bound IB4. IB4 binding does not occur on intrinsic NTS neurons or fibers but is highly expressed by unmyelinated vagal afferent endings in the NTS (44, 45). Hence, colocalization of pERK1/2 immunoreactivity with IB4 in the NTS indicates that CCK triggers MAPK activation in central vagal afferent endings. Consistent with IB4 localization to vagal afferent C fibers, we found that capsaicin treatment abolished IB4 binding and CCK-induced increase of pERK1/2 in NTS neuropil. We also found CCK-evoked pERK1/2 immunoreactivity in NTS fibers and terminals anterogradely labeled after nodose ganglion injection of an anterograde tracer, BDA. Hence, it is clear that CCK triggers ERK1/2 phosphorylation in vagal afferents, the majority of which are C fibers.

Reduction of food intake by endogenously released CCK (36) or by exogenous CCK is attenuated by abdominal vagotomy (33) or by treatment with capsaicin (29). Likewise, administration of a CCK1 receptor antagonist abolishes reduction of food intake by both endogenous (46–48) and exogenous CCK (49, 50). Although capsaicin-sensitive vagal afferents innervate both gastrointestinal and nongastrointestinal viscera, electrophysiological results indicate that vagal afferent CCK1 receptor sensitivity is localized primarily, if not exclusively, in afferents that innervate the stomach and upper small intestine (37). We found that capsaicin lesion of vagal afferents or injection of a CCK1 receptor antagonist blocked CCK-induced ERK1/2 phosphorylation in hindbrain neuropil. Collectively, our observations indicate that CCK triggers ERK1/2 phosphorylation in vagal afferents that innervate the gastrointestinal tract.

Although capsaicin treatment abolished CCK-induced ERK1/2 phosphorylation in the mNTS and cNTS neuropil, it did not eliminate pERK1/2 in NTS neuron cell bodies. This result suggests that CCK triggers ERK1/2 phosphorylation in some NTS neuron via activation of capsaicin-insensitive A-type afferents, an interpretation consistent with reports that CCK activates A-type vagal afferents as well as C-type afferents (51, 52). In addition, the fact that capsaicin treatment attenuates both reduction of food intake and pERK1/2 phosphorylation in NTS neuropil but does not reduce numbers of CCK-induced pERK1/2-immunoreactive neurons in the mNTS and cNTS suggests that ERK1/2 phosphorylation in presynaptic vagal C-type endings is a critical site for the participation of pERK1/2 in the reduction of food intake.

CCK-evoked pERK1/2 expression in NTS neurons and neuropil, including vagal afferent endings, was dramatically attenuated by 4V NMDAR antagonist administration, indicating that MAPK signaling in vagal afferents depends on NMDAR activation. NMDAR NR1 subunit was expressed by most NTS neurons that exhibited pERK1/2 after ip CCK. Although we were not able to resolve NMDAR immunoreactivity on fibers in the NTS, vagal afferent expression of NMDAR is well documented (53–55). Indeed, Aicher et al. (28) have detected NMDAR NR1 immunoreactivity on vagal afferent endings in the NTS, using ultrastructural analysis. The expression of NMDAR by vagal afferents therefore is consistent with our observations that CCK-induced ERK1/2 phosphorylation in vagal afferent endings is prevented by NMDAR antagonist administration.

The mechanisms by which MAPK activation leads to CCK-induced reduction of food intake remain speculative. Although pERK1/2 alters gene transcription via CREB phosphorylation (56–59), the fact that both CCK-induced NTS ERK1/2 phosphorylation and reduction of feeding occur over a 5- to 15-min time course (18) makes a transcriptional mechanism unlikely. In addition to transcriptional actions, pERK1/2 phosphorylates a number of membrane-bound or cytosolic target proteins (60) that may alter synaptic transmission or neuronal excitability. For example, in the hippocampus, ERK-dependent phosphorylation of the Kv4.2 potassium channel subunit reduces its membrane-surface expression resulting in increased neuronal excitability during long-term potentiation (61). Sutton et al. (18) reported that ip CCK injection results in pERK1/2 phosphorylation of Kv4.2 subunit in the hindbrain and suggested that this phosphorylation might mediate increased NTS neuronal response to CCK. pERK1/2 also phosphorylates the NR2B NMDAR subunit (62) and reportedly modulates postsynaptic AMPA receptor trafficking and insertion as well (63, 64).

Mechanisms by which ERK1/2 phosphorylation can alter neuronal function include effects at presynaptic axon terminals. For example, pERK1/2 has been shown to modulate neurotransmitter release by phosphorylating synapsin I (65, 66). Synapsin I is a synaptic vesicle protein that tethers synaptic vesicles to the actin cytoskeleton, thereby maintaining transmitter in a distal reserve pool (67). The affinity of synapsin I for both actin and synaptic vesicles is controlled in part by pERK1/2-mediated phosphorylation of specific serine residues (68–70). Synapsin I phosphorylation by pERK1/2 reportedly increases synaptic strength by allowing more vesicles to enter the readily releasable pool (71, 72). Conceivably, a process such as this in NTS vagal afferent endings could enable an increase in synaptic strength necessary to sustain the behavioral effect of CCK on food intake.

Activation of the NMDAR requires both glutamate binding and membrane depolarization sufficient to displace Mg²⁺ from the NMDAR ion channel (73, 74). Most electrophysiological experiments indicate that activation of AMPA-type glutamate receptors mediates most fast excitatory vagal afferent transmission and depolarization of NTS neurons (41). Consistent with these electrophysiological observations, 60% of NTS neurons that express Fos immunoreactivity after gastrointestinal stimulation express AMPA receptor GluR2/3 subunit immunoreactivity (12). Nevertheless, we found that 4V administration of NBQX, a potent AMPA/kainite receptor antagonist, did not significantly reduce CCK-induced pERK1/2 in the hindbrain. It is noteworthy that our observations are consistent with published results indicating that hindbrain injection of an AMPA receptor antagonist also is ineffective for attenuating CCK-induced reduction of food intake (75). It is unlikely that the 250-nmol dose of NBQX used in our experiments was insufficient to block hindbrain AMPA receptors because our NBQX dose produced prominent and sustained sedative effects, which are associated with AMPA receptor antagonism (76, 77). Additionally, we used a dose of NBQX that is larger than those that have been reported to effectively block AMPA receptor activation after ventricular injection of the antagonist (78, 79). Therefore, future experiments also must consider the possibility that neuronal responses to non-AMPA receptor signaling contribute to NMDAR activation and consequent phosphorylation of ERK1/2 in the NTS. In this regard, a potential contribution of hindbrain melanocortin 4 (MC4) receptors is especially intriguing. MC4 receptor transcript has been detected in hindbrain neurons (80, 81) and in vagal afferents themselves (82). Moreover, hindbrain injection of MC4 agonists triggers hindbrain ERK1/2 phosphorylation, detected by Western blotting, and MC4 receptor antagonist attenuates CCK-induced reduction of food intake and MAPK signaling in the hindbrain (19). Therefore, it is possible that MC4 receptor interaction with proopiomelanocortin peptides plays a permissive role in NMDAR-mediated ERK1/2 phosphorylation.

In summary, our results indicate that activation of NMDAR in the NTS is essential for both MAPK signaling in the NTS and reduction of food intake after ip CCK injection. In addition, our immunohistochemical analysis indicates that CCK-evoked phosphorylation of ERK1/2 depends on NMDAR activation in vagal afferent terminals as well as in NTS neuronal cell bodies. Our results also suggest that ERK1/2 phosphorylation in gastrointestinal vagal afferent C fibers may be critical for reduction of food intake by CCK. Considering the importance of CCK as a direct controller of meal size, further examination of NMDAR participation in modulation of these signals could prove interesting and potentially important as an avenue of intervention in control of food intake.