Sensory suppression during feeding

Abstract

Feeding is essential for survival, whereas withdrawal and escape reactions are fundamentally protective. These critical behaviors can compete for an animal's resources when an acutely painful stimulus affects the animal during feeding. One solution to the feeding-withdrawal conflict is to optimize feeding by suppressing pain. We examined whether rats continue to feed when challenged with a painful stimulus. During feeding, motor withdrawal responses to noxious paw heat either did not occur or were greatly delayed. To investigate the neural basis of sensory suppression accompanying feeding, we recorded from brainstem pain-modulatory neurons involved in the descending control of pain transmission. During feeding, pain-facilitatory ON cells were inhibited and pain-inhibitory OFF cells were excited. When a nonpainful somatosensory stimulus preactivated ON cells and preinhibited OFF cells, rats interrupted eating to react to painful stimuli. Inactivation of the brainstem region containing ON and OFF cells also blocked pain suppression during eating, demonstrating that brainstem pain-modulatory neurons suppress motor reactions to external stimulation during homeostatic behaviors.

Behaviors such as eating, drinking, micturition, and defecation are essential for an organism's survival and are affected by exposure to aversive stimuli. Stress-induced eating and defecation occur across species and can be triggered by exposure to a painful stimulus (1-4). Yet, very little is known about the effects of feeding (eating and drinking), micturition, and defecation on pain sensitivity. In food-deprived animals, eating takes precedence over pain-motivated behaviors. During eating, food-deprived cats were less likely to withdraw from acute noxious cutaneous heat (5), and food-deprived chickens showed fewer pain-motivated behaviors in response to chronic pain induced by sodium urate (6). However, responses to a painful stimulus are affected by the hunger/satiation state (7), and it remains to be determined whether these analgesic effects can be generalized to animals that have been fed freely. Consequently, the aim of Experiment 1 was to examine whether the drive to satisfy hunger still overrides the need to avoid pain in non-food-deprived animals. We found that feeding suppressed pain in rats that were fed ad libitum.

The suppression of pain during feeding indicates the activation of endogenous pain modulatory pathways. The brainstem ventromedial medulla (VMM) is a critical area in the descending control of pain and is the final common pathway from the brain to the spinal cord (8-11). The VMM modulates pain bidirectionally: its activation can produce either pain facilitation or pain inhibition (10-16). The pain-facilitatory and pain-inhibitory effects are thought to be mediated by two populations of neurons with opposing responses to noxious stimulation and morphine (17-20). Cells activated by noxious stimulation are inhibited by opioids. These cells are classified as ON cells and are thought to facilitate nociception. Cells inhibited by noxious stimulation are excited by opioids and are termed pain-inhibitory OFF cells. These putative pain-facilitatory and pain-inhibitory neurons have spontaneous activity that varies across the sleep-wake cycle; most ON cells discharge in bursts during waking and are inactive during slow-wave sleep, and most OFF cells are continuously active during slow-wave sleep and sporadically active during waking (21, 22). Experiments 2-4 were designed to study the contributions of VMM ON and OFF cells to sensory suppression during feeding.

Materials and Methods

Subjects. All procedures were reviewed and approved by the University of Chicago Institutional Animal Care and Use Committee. Subjects were male Sprague-Dawley rats (Charles River Laboratories) weighing 350-450 g at the start of the experiment. They were housed in pairs (experiments 1 and 3) or singly (experiments 2 and 4) on a 12-h light-dark schedule with food and water available ad libitum. All rats were familiarized with the handling procedures and apparatus.

Test Apparatus. Rats were tested in a Plexiglas box (25 × 45 × 35 cm) with a wire-mesh floor. A modified Hargreave's apparatus (23) located underneath the cage was used to deliver radiant heat to the hind paw. The heat intensity was set to elicit baseline withdrawal latencies of 3-4 s and terminated at a maximum duration of 8 s.

Experiment 1: Noxious Heat Stimulation During Quiet Waking, Grooming, and Eating. Five rats were tested on each of 2-3 days for paw withdrawal from noxious heat during quiet waking, grooming, and eating. Noxious heat stimulation was given at intervals >5 min. The total number of trials each rat received ranged from 41 to 60. Grooming of the face and head with the forepaws was used as a control for active movements. The rats were given laboratory chow (global 18% protein rodent diet, Harlan no. 2918; Teklad, Madison, WI), chocolate chips, and yogurt drops to eat.

Experiment 2: Cellular Discharge During Feeding. Surgery. A small craniotomy was performed in the area overlying VMM. For extracellular recording, a threaded microdrive base with a guide tube was implanted (anterior-posterior, -11.3 mm from bregma, at midline) for the introduction of tungsten recording electrodes (5 MΩ, A-M Systems, Carlsborg, WA). Stainless steel screws were placed in the frontal and parietal bones for differential electroencephalographic (EEG) recording, and stranded stainless steel microwires were sutured into the nuchal muscles and the femoris muscles of the right hind paw for electromyographic (EMG) recording. The EEG and EMG leads were attached to a microconnector (Omnetics, Minneapolis) and affixed to the skull. The rats were allowed at least 1 week to recover from surgery.

Data acquisition. A preamplifier within the commutator cable amplified the EEG and EMG signals 5-fold. These signals were then further amplified 10,000-fold by a differential ac amplifier (A-M Systems) and digitized at 1 kHz by a Power1401 interface (Cambridge Electronic Design, Cambridge, UK). The preamplifier buffered the unit signal at unity gain. The unit signal was then amplified 10,000-fold by an ac differential amplifier (Warner Instruments, Hamden, CT) and digitized at 20 kHz by the Power1401 interface. Extracellular recordings were taken from all units that provided stable recordings. For each isolated unit, an amplitude threshold was set by using spike2 acquisition software (Cambridge Electronic Design). When the signal crossed this threshold, the time of that crossing was stored. In addition, 46 digitized points of the waveform were saved: 21 points before and 25 points after threshold crossing. Individual waveforms were reviewed offline, sorted according to amplitude and shape, and assigned to a particular unit by using a template-matching algorithm provided by spike2.

Behavioral classification. Behavioral observations and/or EEG and EMG measures were used to determine wake and sleep states. State assignments were made continuously with no minimum bout length. Four states were recognized: active wake, quiet wake, slow-wave sleep, and paradoxical sleep. The rat was considered to be in active wake if he was engaged in active behaviors such as grooming, eating, drinking, urinating, defecating, or exploring. During such behaviors, the EEG was desynchronized and low in amplitude, and the EMGs showed both tonic and phasic activity. The rat was judged to be in quiet wake when he was sitting quietly with occasional head movements and postural adjustments. The EMG showed low tonic activity except during the brief, isolated moves. The EEG was desynchronized with short periods of synchronized, high-amplitude activity. Brief arousals from sleep during which the rat made slight postural adjustments before returning to sleep were also classified as quiet wake. During these microarousals, the EEG was desynchronized, and the nuchal and/or femoral EMG showed a short burst of phasic activity. The rat was considered to be in slow-wave sleep when he adopted a sleep posture and displayed primarily respiratory-related movements. In this sleep state, the EEG was synchronized and high in amplitude, and all EMG measures showed low tonic activity. During paradoxical sleep, the EEG was desynchronized and low in amplitude. Behaviorally, we observed and recorded on video tape “twitching” of whiskers, ears, and/or paws. Such activity in distal musculature was typically not evident in the nuchal EMG that was low to very low in amplitude during paradoxical sleep.

Electrophysiological and behavioral recordings. A microdrive-electrode assembly was connected to the threaded base on the rat's head. The rat was placed into the test cage and was able to move freely with free access to food and water. The recording electrodes were advanced manually into the VMM region until a unit was isolated. The rat's behavior was continuously monitored and videotaped. Some animals were tested with noxious thermal stimulation of the hind paw at intervals of >5 min. The recording depth for every unit was noted. At the end of each session, the cable and the microdrive-electrode assembly were removed. Rats were tested once a week. The final recording site from each animal was marked by applying a 20-μA anodal current for 4 min.

Histology. All rats were overdosed with sodium pentobarbital (i.p.) and perfused with a fixative containing 10% formalin in 0.1 M PBS (pH 7.4). The brainstem was removed and placed in 30% sucrose. Serial coronal sections (50 μm) were cut on a freezing microtome, mounted on slides, and stained with cresyl violet. The marked recording sites were examined microscopically and plotted onto standard sections. All other recording sites were calculated from the marked site. All recorded sites were within VMM.

Experiment 3: Effects of Air-Puff Stimulation on Eating-Induced Analgesia. Rats (n = 4) were tested for withdrawal responses to noxious radiant heat during quiet waking (no air, no eat), after air-puff stimulation (air, no eat), during eating (no air, eat), and during eating after air-puff stimulation (air, eat). On each of three days, rats were tested under all four conditions in random order. Air-puff stimulation averaged 7.4 ± 0.2 s in length at a rate of 5 liters/min directed at the rat's head. The rats were given chocolate chips to eat.

Experiment 4: Effects of Intra-VMM Muscimol on Eating-Induced Analgesia. Surgery, drugs, and histology. A 26-gauge stainless steel guide cannula (Plastics One, Roanoke, VA) was implanted above VMM (anterior-posterior, -11.3 mm; lateral, 0.0 mm; ventral, -7.7 mm relative to bregma) and cemented to the skull with dental acrylic. The internal microinjection cannula (33 gauge) extended 3 mm beyond the tip of the guide cannula and was connected to a swivel and a Hamilton 1-μl microsyringe via PE20 tubing. Muscimol (50 ng in 500 μl) or an equivalent volume of saline was delivered at a rate of ≈0.5 μl/min. All injection sites were histologically verified to be within VMM (see Fig. 5B, which is published as supporting information on the PNAS web site).

Procedure. Rats were microinjected with muscimol (n = 3) or saline (n = 4) and tested for hind paw withdrawal from noxious radiant heat during eating and not-eating, with an interstimulus interval of ≈5 min. The testing order for eating and not-eating was counterbalanced across animals. The rats were given laboratory chow or chocolate chips to eat. On the second test day, rats received the reverse drug and were tested in the same eating/not-eating order as used on day 1.

Data Analyses

Experiment 1. The percent maximal effects of withdrawal latencies during grooming, eating chow, and eating chocolate/yogurt relative to withdrawal latencies during quiet waking were calculated for each animal on each day to control for individual differences across the test days. An ANOVA was used to test for significant differences in percent maximal effects, with α set at a 0.05 level.

Experiment 2. The discharge rate during each eating bout was compared with the discharge rates during periods of equivalent duration before and after each eating bout. An eating bout was considered to be the period during which a rat ate a single ingestant without interruptions of more than 4 s. This analysis was conservative because (i) an eating bout was frequently punctuated with brief pauses in eating that were accompanied by transient changes in cellular activity, and (ii) rats were generally asleep before eating, and most of the cells discharged infrequently during sleep.

Experiments 3 and 4. Repeated-measures ANOVAs were used to test for significant differences in withdrawal latencies and were followed by a post hoc Tukey's test for all pairwise comparisons. The α level was set at 0.05.

Results

Experiment 1. The purpose of this experiment was to examine whether pain responses to an acute noxious heat stimulus were suppressed in non-food-deprived rats. Motor withdrawals from a noxious heat stimulus were measured in rats during quiet waking, grooming, and eating. When the heat stimulus was presented during quiet waking (see Movie 1, which is published as supporting information on the PNAS web site) or grooming (see Movie 2, which is published as supporting information on the PNAS web site), rats attended to their stimulated paw. In contrast, when the heat stimulus was presented during eating (see Movie 3, which is published as supporting information on the PNAS web site), motor withdrawal latencies were significantly longer than when rats were in quiet waking or grooming (P < 0.05, ANOVA). The suppression of withdrawals was evident across different ingestants (laboratory chow, chocolate chips, and yogurt drops; Fig. 1).

Fig. 1.

Motor withdrawals from a painful stimulus are suppressed during eating. The percent maximal effect of withdrawal latency (+SE) during eating was significantly greater than that during quiet waking (^*, P < 0.05, ANOVA). Sensory suppression during eating occurred regardless of whether the ingestant contained sugar (chocolate or yogurt) or not (chow). The percent maximal effect during grooming was not significantly different from that during quiet waking (P > 0.05, ANOVA).

Experiment 2. The results from experiment 1 suggested that endogenous pain disfacilitatory and/or inhibitory mechanisms were engaged during feeding. Experiment 2 was conducted to investigate the activity of pain-modulatory neurons in VMM during feeding. Sixty-seven VMM cells were recorded across 201 eating bouts. All cells discharged either irregularly and/or at high rates (see Fig. 6, which is published as supporting information on the PNAS web site) and were therefore likely to be nonserotonergic. By using an algorithm described in ref. 24, all cells were verified to be nonserotonergic.

Most of the recorded cells (≈70%) were more active during waking than sleeping (49/67). Of those tested, most cells (≈70%) were excited by noxious heat stimulation (18/26). Fifty cells had one or both of these characteristics and were, therefore, classified as ON cells. More than half of the ON cells (n = 27) decreased their discharge rate during eating (Figs. 2 A and B and 3A). Two neurons displayed the physiological characteristics of pain-inhibitory OFF cells; one cell had sleep-active discharge, and the other cell was inhibited by noxious stimulation. These cells increased their discharge during eating and drinking (Fig. 3B). The locations of these ON and OFF cells are shown in Fig. 5A.

Fig. 2.

Cellular discharge of VMM cells during eating, drinking, and micturition. Cellular activity of an ON cell decreased during ingestion of chocolate chip (A1), wafer (A2), and cookie (A3). Cellular activity of another ON cell during ingestion of laboratory chow (B1), yogurt drop (B2), or nibbling of paper (B3) is shown. (C) Decrease in cellular activity of an ON cell during drinking. (D) Decrease in cellular activity of an ON cell during micturition.

Fig. 3.

Discharge rates of ON and OFF cells during feeding. (A) Discharge rates of all ON cells (n = 27) that decreased their firing during eating are shown for the periods before, during, and after eating. The upper and lower edges of each box represent the 25th and 75th percentiles, respectively, with the midline representing the median. The upper and lower whiskers show the 10th and 90th percentiles. The filled circles show the means. (B) The discharge rates of two OFF cells increased during eating and drinking.

The decreases in ON cell discharge during eating occurred regardless of the sugar or sucrose content of the ingestant (Fig. 2). Such decreases were observed during feeding on highly palatable, sweet foods (chocolate chips, yogurt drops, wafers, and butter cookies) as well as during ingestion of laboratory chow or even feces. However, the degree of cellular inhibition depended on whether the substance was ingested. For example, ingestion of laboratory chow or yogurt drop was associated with an almost complete cessation in cellular activity, whereas nibbling (biting with little or no ingestion) of paper was associated with a smaller decrease in cellular activity (Fig. 2B).

The rats were given middle- to large-sized pieces of food and displayed typical foraging and eating behaviors (25, 26). The general sequence consisted of the following: (i) foraging and location of food by sniffing, (ii) retrieval of food by mouth, (iii) transfer of food from mouth to forepaws, (iv) sitting back on haunches to eat, (v) manipulating the food with digits while eating, (vi) consumption of the last morsel of food with grounded forepaws, (vii) postprandial foraging by head scanning of the vicinity for more food, and (viii) facial and head grooming. The changes in cellular discharge occurred during selected portions of the eating sequence (iii-v), but not during the food handling movements before consumption (i-ii) or behaviors after the central eating period (vi-viii). Cellular discharge increased back to baseline levels when the forepaws were grounded and while the rat chewed and swallowed the final morsels and during postprandial behaviors (see Movie 4, which is available as supporting information on the PNAS web site). Moreover, cellular activity did not decrease during oro-facial or forepaw movements that occurred during grooming.

Some of the VMM cells that were inhibited during eating were also recorded during drinking and micturition. Most ON cells (9/10) inhibited during eating were also inhibited when the rats drank water (Fig. 2C); the remaining cell was excited. Further, 13/14 ON cells inhibited during eating were also inhibited during micturition (Fig. 2D), with the remaining cell excited.

Experiment 3. The inhibition of pain-facilitatory ON cells and excitation of pain-inhibitory OFF cells during eating predicted the suppression of sensory inputs during eating. To test whether these changes in VMM discharge are sufficient to suppress sensory inputs during eating, we examined the effects of antagonizing ON cell inhibition and OFF cell excitation on sensory suppression during eating by applying an innocuous air puff, a stimulus that excites ON cells and inhibits OFF cells (21). An innocuous air puff delivered before eating reversed the long latencies to withdraw from noxious heat during eating that were otherwise observed (Fig. 4A). Air-puff stimulation alone did not alter withdrawal latencies (Fig. 4A).

Fig. 4.

Sensory suppression during feeding was eliminated by prior stimulation with an innocuous air puff or by inactivation of VMM. Plotting conventions are as in Fig. 3. (A) Air-puff stimulation: Withdrawal latencies were significantly longer during eating (No air/Eat vs. No air/No-eat; ^*, P < 0.05, repeated-measures ANOVA). Air-puff stimulation before eating blocked the increases in withdrawal latencies that were otherwise observed during eating (Air/Eat). A preceding stimulation with an air puff did not affect withdrawal latencies during quiet waking (No air/No eat). (B) Inactivation of VMM with muscimol: Sensory suppression during feeding was evident when saline was microinjected into VMM (Sal/Eat) but was blocked by pretreatment with muscimol (Mus/Eat) (^*, P < 0.05, repeated-measures ANOVA).

Experiment 4. The results from experiment 3 are consistent with the idea that ON cell inactivation and OFF cell activation produce sensory suppression during eating. However, they do not directly implicate VMM cells. To test whether VMM neurons are necessary for sensory suppression during eating, VMM was inactivated by microinjection of the GABA_A receptor agonist muscimol. Inactivation of VMM eliminated the suppression of motor withdrawals during feeding (Fig. 4B), demonstrating that VMM cell activity is critical to sensory suppression during eating.

Discussion

The current results show that the precedence of eating over protective noxious-evoked motor withdrawals observed in food-deprived animals (5, 6) also occurs in animals fed ad libitum. The suppression of motor withdrawal from noxious heat was observed when rats ate either highly palatable, sweet substances or laboratory chow (Fig. 1). This eating-induced analgesia is distinct from the analgesia elicited by an acute intraoral infusion of sucrose (27-29) for two reasons. First, sucrose-induced analgesia is present only in infant rats (28). Second, eating-induced analgesia was evident when rats ingested laboratory chow, a food that contains no sugar or sucrose.

Consistent with the behavioral data, ON cell activity decreased during eating of highly palatable sweet foods or sugar-free laboratory chow, but not during grooming. The changes in ON cell activity during eating appear to be behaviorally specific, occurring when specific movements, such as chewing and bringing the forepaws to the mouth, are centered within a feeding bout, but not when the same or similar movements are made during head and oro-facial grooming. We have also noticed no apparent change in ON cell activity during bruxism or chattering. Although chewing and bruxism produce different mandibular trajectories and activity patterns in the anterior temporalis muscles (30), both oral behaviors require activation of masticatory muscles. These observations suggest that decreases in ON cell activity during feeding are not due to the activation of masticatory muscles per se.

In further support for the idea that decreases in ON cell activity during feeding are not due to the activation of specific muscles or muscle groups, ON cell activity decreased during other homeostatic behaviors. Most of the ON cells tested decreased their discharge across multiple homeostatic behaviors: eating, drinking, and micturition. In anesthetized rats, decreases in ON cell activity and increase in OFF cell activation also occur during micturition (31). Because sensory suppression during feeding was reversed by VMM inactivation, VMM activity is necessary for eating-induced analgesia. Sensory suppression was also prevented by a stimulus that inhibits OFF cells and excites ON cells. Taken together with the electrophysiological data, this finding shows that the mechanisms underlying sensory suppression during eating involves decreases in VMM ON cell activity and increases in VMM OFF cell activity.

It is interesting that after preactivation of ON cells and preinhibition of OFF cells with an innocuous air puff, sensory suppression did not occur and rats interrupted feeding to respond to noxious heat. It is possible that our air-puff stimulation mimicked a situation frequently encountered in the natural environment where animals are extremely vulnerable to predation while feeding. Chipmunks that have heard a playback of conspecific alarm calls continue to feed but do so more slowly as they spend more time on vigilant behaviors (32). We speculate that predation-related stimuli activate ON cells and thereby evoke a state of vigilant eating without concurrent sensory suppression.

In sum, the present results provide direct evidence that VMM ON and OFF cells are critical to the completion of a homeostatic behavior. Because VMM stimulation modulates the cellular responses and behavioral reactions evoked by both noxious and innocuous inputs (33-35), a generalized decrease in reactivity to somatosensory stimulation may accompany feeding, allowing an animal to nourish itself without being distracted by every falling leaf or gust of wind. Taken together, our work has shown that decreases in ON cell firing and increases in OFF cell firing occur during sleep, eating, drinking, and micturition (21, 22, 31). All of these behaviors serve critical roles in maintaining homeostasis and survival. Thus, the inhibition of VMM ON cells and excitation of VMM OFF cells during homeostatic behaviors suppress sensory processing and thereby protect critical behaviors from disruption.

The activation of descending pain inhibition during eating and drinking provides a critical function for VMM cells in modulating sensory receptivity on a daily basis. VMM is situated in a phylogenetically ancient and conserved region of the hindbrain. Its anatomical location and evolutionary conservation is consistent with VMM having a fundamental and critical role in everyday, basic physiology. The suppression of incoming sensory input during homeostatic behaviors is a daily life-preserving function befitting VMM's medullary location.