Hippocampus oxytocin signaling promotes prosocial eating in rats

Abstract

Background –

Oxytocin (OT) is a hypothalamic neuropeptide involved in diverse physiological and behavioral functions, including social-based behavior and food intake control. The extent that OT’s role in regulating these two fundamental behaviors is interconnected is unknown and is a critical gap given that social factors have a strong influence on eating behavior in mammals. Here we focus on OT signaling in the dorsal hippocampus (HPCd), a brain region recently linked with eating and social memory, as a candidate system where these functions overlap.

Methods –

HPCd OT signaling gain- and loss-of-function strategies were employed in male Sprague-Dawley rats that were trained in a novel social eating procedure to consume their first nocturnal meal under conditions that vary with regards to conspecific presence and familiarity. The endogenous role of HPCd OT signaling was also evaluated for olfactory-based social transmission of food preference learning, sociality, and social recognition memory.

Results –

HPCd OT administration had no effect on food intake under isolated conditions, yet significantly increased consumption in the presence of a familiar, but not an unfamiliar conspecific. Supporting these results, chronic knockdown of HPCd OT receptor expression eliminated the food intake-promoting effects of a familiar conspecific. HPCd OT receptor knockdown also blocked social transmission of food preference learning and impaired social recognition memory without affecting sociality.

Conclusion –

Collective results identify endogenous HPCd OT signaling as a novel substrate where OT synergistically influences eating and social behaviors, including the social facilitation of eating and the social transmission of food preference.

INTRODUCTION

Oxytocin (OT) is an evolutionarily conserved neuropeptide produced in the paraventricular and supraoptic nuclei of the hypothalamus. Its release either into the periphery or within the brain plays a role in a diverse set of behavioral functions, including food intake regulation (1, 2). Acting as an anorexigenic signal, administration of OT reduces food intake in experimental rodent models, as well as in nonhuman primates and humans (1, 3). In rodents, either central or peripheral administration reduces food intake, especially for highly palatable foods (4, 5) and in diet-induced obese models (6). These findings have identified OT as a potential therapeutic for the treatment of obesity (7–10).

OT has also been extensively studied for its effects on modulating social behaviors, including maternal bonding, pair bonding, sociability, and social-based memory (11, 12). While OT’s influence on social behaviors is generally considered to be prosocial (e.g., cooperation, caregiving), recent findings suggest that its effects are complex and highly-dependent on the social context, with OT administration increasing outgroup discrimination and reducing affiliative behaviors under some conditions (12–14). Given that OT plays a role in mediating both eating and social behaviors, it is highly plausible, yet unknown whether critical overlap in these functions exists.

Eating behavior is strongly influenced by social factors in humans. For example, the “social facilitation of eating” (SFE) effect is a phenomenon in which both the number of people at a meal, as well as the familiarity with and gender of those present during the meal can powerfully modulate caloric consumption (15–20). Rats, like humans, are also highly social eaters, as food choice and preference, the amount consumed, and foraging strategies in rats are influenced by the behavior of conspecifics (21, 22). The SFE effect has also been documented in rodent models (23, 24), thus supporting that results from mechanistic rodent models on interactions between eating and social factors have translational relevance. Despite evidence that social factors influence eating behavior in both humans and rodents, the overwhelming majority of rodent model research on food intake control has evaluated consumption with subjects in isolated conditions. Because social-based eating is a more ecologically valid model of food intake in both humans and rats compared to isolated eating conditions, it is critical to understand the extent that OT’s effects on food intake are modulated by the social environment.

We hypothesize that the dorsal subregion of the hippocampus (HPCd) is a substrate where OT synergistically influences eating and social behaviors. The HPCd has recently been linked with food intake control in rodents (25, 26) and OT receptors (OXTR) are expressed in the HPCd and have been established in mediating social behaviors (26–35). To investigate the mechanisms via which HPCd OT influences food intake, social factors, memory, and potential interactions between these behaviors, we developed a novel eating paradigm in rats allowing for evaluations of consumption under conditions that vary with regards to social presence, conspecific familiarity, and context familiarity. By combining this paradigm with established rodent social behavioral procedures (e.g., social transmission of food preference, social recognition memory) and neuropharmacological and virogenetic manipulations, our results identify neurobiological mechanisms connecting the OT system with food intake and socially-relevant behaviors.

METHODS and MATERIALS

Animals

Male Sprague-Dawley rats (Envigo, Indianapolis, IN; postnatal day [PND] 60–70; 250–275g on arrival) were individually housed in a temperature-controlled vivarium with ad libitum access (except where noted) to water and food (LabDiet 5001, LabDiet, St. Louis, MO) on a 12h:12h reverse light/dark cycle. All procedures were approved by the Institute of Animal Care and Use Committee at the University of Southern California.

Experiment 1: The effects of hippocampal oxytocin and social presence on eating

To determine the effects of hippocampal OT administration on eating, rats implanted with bilateral HPCd cannulae underwent the social eating procedure depicted in Figure 1A (surgical and behavioral procedures in Supplemental Methods, experimental design summary in Supplemental Table 1). Animals in the Isolated Home Cage (n=8), Isolated Neutral Cage (n=8), and Familiar Conspecific (n=8) groups were injected with either artificial cerebral spinal fluid (aCSF) or 0.05μg OT (within-subjects drug design, counterbalanced treatments) in the HPCd approximately 15 min prior to the start of the test (OT dose selection described in Supplemental Methods). Food was returned to the animals at the start of the testing period, coinciding with the onset of the dark cycle, and food intake and meal pattern parameters were recorded for one hour.

Figure 1. Overview of social eating procedures.

(A) An experimental timeline for surgeries and training in social eating procedures. (B) In a two-chamber food intake monitoring system, spontaneous meal patterns are evaluated during the 1^st hour of the nocturnal feeding period under isolated conditions that vary with regards to context familiarity, or under social conditions that vary with regards to conspecific familiarity. A divider separates the chambers and the animals physically to allow for precise consumption measures for individual animals, absent competition for the food source. The divider is transparent with various holes to still allow for transmission of visual, olfactory, and auditory social cues between chambers.

Experiment 2: The effects of social presence and conspecific familiarity on eating

Animals with bilateral HPCd cannulae (n=12) underwent social eating procedure training. Animals were then divided into two groups differing by conspecific familiarity. Those in the Group Familiar (n=6) received counterbalanced aCSF and 0.05 μg OT drug treatments in the presence of the familiar conspecific from training. Group Unfamiliar (n=6) received the treatments in the presence of a novel, unfamiliar conspecific that had also underwent social eating procedure training.

Experiment 3: The effects of hippocampal OXTR reduction on social eating

To examine the physiological role of hippocampal oxytocin signaling on social eating, we utilized a viral-mediated approach (described in Supplemental Methods) to reduce OXTR expression within the HPCd (DG subregion).Following surgery, body weight changes were measured for 36d, and home cage food intake (including meal patterns) was tracked for 5d (~3wks after surgery) (Supplemental Methods) before animals completed the social eating procedure. During training all animals were paired with the same conspecific for 12 training days. On test days Control (n=6) and OXTR KD (n=8) animals were placed into social eating cage with either the same “Familiar” conspecific from training or a novel “Unfamiliar” conspecific (within-subject design, counterbalanced tests) and one hour food intake was measured (experimental conditions summarized in Supplemental Table 1).

Experiment 4: The effects of hippocampal OXTR reduction on Social Transmission of Flavor Preference (STFP)

To examine the effects of HPCd OXTR knockdown on food preference memory based on olfactory and social cues, we utilized the STFP task (described in Supplemental Methods) (36–39). Non-experimental animals with intact OT signaling were used as Demonstrator animals. Observer animals had either control AAV (n=6) or OXTR KD AAV (n=9) injected into the HPCd at least 3 weeks prior to completing the STFP protocol.

Experiment 5: The effects of hippocampal OXTR reduction on sociability, social recognition memory (SRM), and object recognition memory

In this study we utilized a social discrimination (SD) task (described in Supplemental Methods) to assess both sociability (Figure 6A) and SRM (Figure 6D) (40–43). SD tests were completed in control AAV (n=7) and OXTR KD (n=7) animals that had undergone viral-mediated HPCd OXTR KD surgery at least 3 weeks prior to testing.

Figure 6. Reduction in HPCd OXTR expression impairs social recognition memory but not sociability.

(A) Sociability is assessed by placing experimental animals into an arena for 5 min with an empty enclosure and another containing a stimulus animal. The time spent investigating each enclosure is measured. (B-C) Both control and HPCd OXTR KD animals spent significantly more time investigating the stimulus animal over the empty enclosure, indicating normal sociability in both groups (CON: t(6)=8.352, p=0.00032 empty vs. stimulus, OXTR KD: t(6)=7.317, p=0.00033 empty vs. stimulus). (D) Following an interval of 30 min, animals are placed back into the arena to assess social recognition memory, now with the previously experienced “familiar” stimulus animal or a new “novel” animal. (E-F) Control animals spent more time investigating the novel stimulus animal while the HPCd OXTR KD animals did not (E: CON t(6)=3.358, p=0.03 familiar vs. novel; F: t(12)=2.432, p=.032 CON vs. OXTR KD, one-sample t-test comparing to 0.50 CON: t(6)=4.046, p=0.0068). (G) Animals were also tested in novel object recognition task to assess non-social recognition memory. (H-I) Control and OXTR KD groups performed similarly in the novel object recognition test (CON: t(6)=4.719, p=0.00326 familiar vs. novel, OXTR KD: t(6)=6.457, p=0.00039 familiar vs. novel). (Between-subjects for group; control n=7; OXTR KD n=7; Data are means ± SEM; *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001).

After SD testing, the rats underwent novel object recognition memory testing (NOR) (described in the Supplemental Methods) to evaluate novelty-based learning, absent social factors (44). Control AAV (n=7) and OXTR KD (n=7) animals both underwent NOR procedure at least 3 weeks after surgery (Figure 6G).

RESULTS

HPCd oxytocin effects on food intake are dependent on the social context.

Under “Isolated-Home Cage” conditions, OT administration to the HPCd had no effect on cumulative 1-hr food intake (Figure 2A), 1^st meal size (Figure 2B), or 1-hr meal frequency (Figure 2C). HPCd OT administration also did not influence 1-hr cumulative food intake (Figure 2D), 1^st meal size (Figure 2E), or meal frequency (Figure 2F) for animals in the “Isolated-Neutral Cage” condition.

Figure 2. HPCd oxytocin effects on food intake differ by social context.

Under isolated home cage (A-C) and isolated neutral cage (D-F) conditions, hippocampal administration of 0.05 μg oxytocin did not affect 1-hr cumulative caloric intake, 1^st meal size, or meal frequency (G) Oxytocin administration to the HPCd increased 1-hr cumulative chow intake when consumed in the presence of a familiar conspecific (t(8)=3.937, p=0.004). (H, I) This effect was mediated by a significant increase in the 1^st nocturnal meal size (t(8)=4.319, p=0.003), without affecting meal frequency (home cage n=9; isolated neutral n=8; familiar conspecific n=9; all between-subjects design for eating condition and within-subjects design for drug treatments; Data are means ± SEM; *p<0.05, **p<0.01).

However, for animals in group “Familiar Conspecific-Neutral Cage”, HPCd OT administration in the presence of a familiar conspecific significantly increased 1-hr cumulative food intake (Figure 2G), driven by a significant increase in 1^st meal size (Figure 2H) as there was no significant difference in meal frequency (Figure 2I), meal duration, number of eating bouts, average eating bout size, or average eating bout duration (Supplemental Figure 2).

Hippocampus oxytocin signaling augments the social facilitation of eating (SFE) by a familiar, but not an unfamiliar conspecific.

Here we sought to assess whether conspecific familiarity influences food intake under social eating conditions, and whether this effect is influenced by HPCd oxytocin signaling. All animals were trained to eat their first nocturnal meal with a familiar conspecific in the social eating procedure. When we compared intake at the start of training to the end of training and to home cage non-social conditions, there was no significant difference in 1hr cumulative intake. However, end of social eating training cumulative intake was significantly increased in comparison to both non-social and start of training conditions (Figure 3A). This familiarity-dependent SFE effect was mediated by an increase in 1^st meal size, as end of training 1^st meal size was significantly larger compared to the non-social condition and trending higher compared to the start of training condition (Figure 3B). There were also no significant differences in meal frequency between the three conditions (Figure 3C). There was a significant increase in 1^st meal duration between non-social and both training conditions and an increase in average bout size between non-social and end of training conditions, with no differences in bout frequency or duration (Supplemental Figure 3A–D). Collectively, these results establish a critical component of the SFE effect in rats: eating more in the presence of familiar (end of training) compared to unfamiliar (start of training) individuals or eating alone.

Figure 3. Social facilitation of eating in the presence of a familiar conspecific is augmented by hippocampal oxytocin.

(A) When compared to average isolated “Non-social” home cage intake, there is no significant difference in 60-min cumulative chow intake at the start of the social eating training (“Start Training”), however by the end of training (“End Training”) rats consume significantly more chow than under isolated conditions as well as compared to the start of the social eating training (n=12, RM one-way ANOVA, Days: F_{(1.590,17.49)}=9.587, p=0.0026, Sidak’s adjusted p=0.0048 non-social vs. end training and p=0.0092 start training vs. end training). (B-C) These outcomes were based on an increase in 1^st meal size by the end of training in comparison to the non-social and start of training (n=12, RM one-way ANOVA, Days: F_{(1.512,16.63)}=5881, p=0.0171, Sidak’s adjusted p=0.0281 non-social vs. end training) as there were no significant differences in meal frequency. (D) During the pharmacological testing phase, rats assigned to Group Familiar that received both drug treatments in the presence of a familiar conspecific, consumed significantly more food than rats assigned to Group Unfamiliar with an unfamiliar conspecific’s presence under both the vehicle aCSF and OT conditions. Group Familiar also saw a significant increase in food intake with the administration of OT in comparison to vehicle aCSF conditions while Group Unfamiliar saw no effect of OT administration (RM two-way ANOVA, conspecific familiarity: F_(1,10)=12.63, p=0.0052, treatment: F_(1,10)=7.083, p=0.0238 Fisher’s LSD planned comparisons aCSF: t(20)=2.2.381 p=0.0273 unfamiliar vs. familiar and OT: t(20)=3.569, p=0.0019 unfamiliar vs. familiar, and Familiar: t(10)=2.649, p=0.0244 aCSF vs. OT). (E-F) These outcomes were based on the presence of a familiar, but not an unfamiliar conspecific increasing 1^st nocturnal meal size (RM two-way ANOVA, conspecific familiarity: F_(1,10)=15.17, p=0.003, Fisher’s LSD planned comparisons aCSF: t(20)=2.247 p=0.0361 unfamiliar vs. familiar and OT: t(20)=4.086, p=0.0006 unfamiliar vs. familiar, and Familiar: t(5)=2.325, p=0.0424 aCSF vs. OT) without influencing meal frequency and hippocampal OT treatment enhancing these effects. (Within-subject n=12 design for training comparing intake in contexts differing in social presence; Mixed design for testing with conspecific familiarity as between-subjects and drug treatment as within-subjects; Group Unfamiliar n=6; Group Familiar n=6; Data are means ± SEM; *p<0.05, **p<0.01, ***p<0.001).

To determine whether elevated SFE by HPCd OT administration (Figure 2G) was driven simply based on social presence, or whether familiarity is an important component of this effect, animals were divided into two groups after training, differing with regards to conspecific familiarity for drug tests (Unfamiliar vs. Familiar group). Results revealed that animals in the Familiar group consumed more food than the Unfamiliar group following both aCSF and OT treatments (Figure 3D). Animals in the Unfamiliar group showed no difference in intake between drug treatments, whereas OT significantly increased 1hr cumulative intake vs. aCSF treatment in the Familiar group. Similar results were seen for 1^st meal size, where the Familiar group consumed a larger 1^st meal vs. the Unfamiliar group following aCSF, and OT increased intake in the Familiar group but not the Unfamiliar group (Figure 3E). There was a significant main effect of group but not drug treatment for 1^st meal size. The Familiar group had a significantly larger 1^st meal size under aCSF and OT conditions, with OT treatment significantly increasing 1^st meal size in comparison to aCSF in the Familiar, but not the Unfamiliar group. There were no significant differences in meal frequency between groups or treatments (Figure 3F). OT administration significantly increased the bout frequency in the Familiar but not the Unfamiliar group with no effect on average bout size or duration (Supplemental Figure 3F–H). Collective results reveal that rats exhibit hyperphagia when eating in the presence of a familiar conspecific relative to either eating alone or in the presence of an unfamiliar conspecific, and that this familiarity-dependent SFE effect is augmented by HPCd OT administration.

SFE by a familiar conspecific requires endogenous hippocampal OXTR signaling.

The viral-mediated approach to knockdown HPCd OXTR (Figure 4A) involved OXTR and control (scrambled) shRNA conjugated to GFP to confirm viral transfection in the HPCd DG (Figure 4B,C). The OXTR shRNA (KD) significantly reduced OXTR mRNA expression in the HPCd relative to controls (CON) by approximately 80% (Figure 4D). This targeted viral knockdown approach had no effect on OXTR expression in adjacent cortex (CON mean=1.00, SEM=0.3794; OXTR KD mean=0.878, SEM=0.2209; t(5)=0.2966, p=0.779), or in the ventral hippocampus subregion (CON mean=1.00, SEM=0.2741; OXTR KD mean=4.119, SEM=1.323; t(11)=2.136, p=0.056). Thus, this vector-mediated approach was effective in significantly reducing OXTR expression in the HPCd without affecting OXTR expression in adjacent brain regions.

Figure 4. Conspecific familiarity-based social facilitation of eating requires endogenous hippocampal oxytocin receptor signaling.

(A) Diagram depicting viral vector-mediated OXTR shRNA for chronic knockdown of OXTR expression in the HPCd. (B-C) Following the infusion of scrambled sequence control (scrmb) or OXTR shRNA AAVs, dentate gyrus target sites and viral-induced GFP expression were confirmed using immunohistochemistry (representative photomicrographs from each group depicted). (D) Quantification of relative OXTR mRNA expression showed a significant reduction in knockdown animals of approximately 80% relative to controls (t(13)=2.496, p=.027; control n=7; KD n=7). (E-G) During the training phase of the social eating procedure, OXTR KD animals consumed significantly less food within the hour in the social arena compared to controls (t(12)=2.553, p=0.025), with a trend towards a reduction in first meal size (t(12)= 2.10, p=0.057) but no effect on meal frequency. (H-J) Control, but not OXTR KD animals consumed more food within the hour-long test in the presence of a familiar vs. an unfamiliar conspecific (Fisher’s LSD planned comparisons CON: t(12)=2.217, p=0.0467 unfamiliar vs. familiar), an effect driven by an increased 1^st meal size (Fisher’s LSD planned comparisons CON: t(12)=2.655, p=0.021 unfamiliar vs. familiar) with no change in meal frequency. (K-M) Knockdown of dorsal hippocampal oxytocin receptors did not yield long-term changes in daily caloric intake, meal frequency, or body weight under isolated conditions in the home cage. (Between-subjects design for group; control n=6; OXTR KD n=9; Data are means ± SEM; *p<0.05; Abbreviations, DGmo = dentate gyrus molecular layer, DGsg = dentate gyrus granule layer, DGpo = dentate gyrus polymorph layer).

The OXTR KD group consumed significantly less food during the social eating training sessions versus the CON group (Figure 4E). There was a trend towards a reduction in average 1^st meal size during training (Figure 4F) with no group effect on meal frequency (Figure 4G). These results suggest that endogenous HPCd OXTR signaling mediates SFE by a familiar conspecific, and consistent with our pharmacological data, this effect is driven by meal size and not eating frequency.

On test days CON animals consumed significantly more food within the 1hr social eating test in the presence of a “Familiar” conspecific in comparison to an “Unfamiliar” conspecific, an effect driven by increased 1^st meal size, whereas OXTR KD animals showed no effect of conspecific familiarity on 1hr food intake or meal size (Figure 4H–I). There were no significant between- or within-group differences based on conspecific familiarity in meal frequency (Figure 4J). Overall results show that CON animals consumed more with a familiar conspecific whereas OXTR KD animals did not, thus further supporting that endogenous hippocampal OXTR signaling mediates familiarity-based SFE.

Reduced expression of HPCd OXTR did not significantly influence long-term daily caloric intake, average meal size, or change in body weight under isolated home cage conditions (Figure 4K–M). There were also no significant differences in other meal pattern parameters (1^st meal size, average 1^st meal duration; Supplemental Figure 4). These results, consistent with the pharmacological data, highlight the selectivity of HPCd OT signaling in influencing food intake under conditions involving social interactions.

Hippocampal oxytocin receptor knockdown impairs STFP learning.

In addition to influencing the amount of food consumed, social factors can influence food preference and choice (20, 45, 46). An STFP protocol (depicted in Figure 5A) was used to assess the role of HPCd oxytocin receptor signaling in social-based learning about food-associated olfactory cues. Data from the social interactions following Demonstrator consumption revealed no difference in time spent investigating the Demonstrators between CON and OXTR KD groups (Figure 5B), indicative of comparable sociability between groups. For food preference testing, consistent with previous STFP studies (36, 37, 47), CON animals showed a strong preference for the demonstrator-paired flavor during the 30-min consumption test. In contrast, OXTR KD animals showed no significant preference between the two food choices, as indicated by no significant difference from chance (Figure 5C). When comparing groups, the CON group had a significantly higher paired flavor preference than the OXTR KD group. There was no significant group difference in total consumption during the two-choice preference test (Figure 5D). Taken together, these data indicate that hippocampal OXTR KD impairs social-based learning about food preference.

Figure 5. Endogenous HPCd oxytocin receptor signaling is required for social transmission of food preference (STFP) learning.

(A) Diagram depicting STFP procedure in which an experimental or observer rat is exposed to a novel food flavor from the breath of a demonstrator that has recently consumed the flavored chow in a separate room. 24 hours after exposure to demonstrator rats, observer rats are tested in the two-choice preference consumption test. (B) There was no difference between Control and OXTR KD rats in the time spent investigating the demonstrator during social interaction. (C) Control animals successfully demonstrated a significant preference for the flavor their demonstrator had consumed, whereas OXTR KD rats showed no flavor preference (Welch’s t-test CON vs. KD t(9.532)= 2.429, p=0.037; one-sample t-test comparing to 0.50 CON: t(5)=12.58, p<0.0001). (D) The total amount of food consumed during the preference test did not differ by group. (Between-subjects design for group; control n=6; OXTR KD n=9; Data are means ± SEM; *p<0.05, ****p<0.0001).

Hippocampal oxytocin administration has no effect on risk-associated appetitive and consummatory behavior in a nonsocial context

To determine whether effects of HPCd OT administration on SFE and STFP were due to reduced anxiety in a food-related environment (i.e., functioning as a “safety signal” to eat), animals completed an approach-avoidance consumption task (detailed in Supplemental Methods) where animals enter a brightly-lit environment, leaving the safety of a darkened familiar space, to consume a highly palatable peanut butter food reward. Results showed there was no effect of HPCd OT administration on latency to enter the light side (Supplemental Figure 5C), time spent in the light side (Supplemental Figure 5D), or on peanut butter consumption (Supplemental Figure 5E). Overall, these results show that HPCd OT does not appear to promote SFE and STFP via reducing anxiety in a food-related context.

Hippocampal oxytocin receptor signaling mediates social recognition memory (SRM) but not sociability or object recognition.

In the SD task, both CON and OXTR KD animals spent significantly more time investigating the novel stimulus animal than the empty enclosure (Figure 6B). Both groups were significantly above chance (0.5) in their investigation of the stimulus animal and there were no significant group differences in exploration ratio (Figure 6C). Thus, HPCd OXTR KD does not impact sociability.

Results of the subsequent SRM test (Figure 6D) revealed that while the CON group spent significantly more time investigating the novel vs. the familiar animal, the OXTR KD group did not (Figure 6E). Likewise, the CON group had a novel stimulus exploration ratio that was significantly above chance, while the OXTR KD group did not, consistent with the significant group difference in exploration ratio (Figure 6F). These results indicate that HPCd OXTR KD impairs SRM without affecting measures of sociability.

The Novel Object Recognition (NOR) procedure was used to evaluate novelty-based memory, outside of a social context (Figure 6G). Results showed that both CON and OXTR KD animals spent significantly more time investigating the novel vs. the familiar object (Figure 6H), that both CON and OXTR KD were significantly above chance (0.5) for exploration ratio, and that there were no group differences for exploration ratio (Figure 6I). Thus, reduction of HPCd OXTR expression had no effect on object recognition memory, suggesting that OT’s action in the hippocampus preferentially encodes social cue-related memory processes.

DISCUSSION

Oxytocin (OT) administration reduces food intake in rodents and in humans (3). However, the overwhelming majority of previous studies evaluating oxytocin’s effects on food intake involve eating under isolated conditions. Given that social factors potently influence eating behavior in both rodents and humans, and that both species regularly consume food in the presence of conspecifics (16, 20–24), social-based eating is a more ecologically valid model to assess the impact of OT on food intake control, particularly given that OT has been extensively studied for its role in mediating social behaviors (2, 12, 48), including those relating to food intake (14, 49, 50). Here we examined how OT signaling in the DG subregion of the HPCd, an OXTR-expressing brain area linked with social-based memory and more recently with food intake control (33–35, 51–54), influences consumption in rats in a novel behavioral paradigm that varies eating testing conditions with regards to social presence and familiarity. Results from neuropharmacological studies reveal that while HPCd OT administration had no effect on food intake under isolated conditions in either the home cage or in a familiar neutral cage, HPCd OT significantly increased consumption in the presence of a familiar conspecific in a social eating arena, an effect based on increasing the size of the first nocturnal meal. These findings differ from previous studies conducted under isolated eating conditions that reveal a potent anorexigenic role for central OT driven by a reduction in meal size (3, 55–58). However, previous studies did not examine OT’s effects on food intake specifically within the hippocampus, but rather its action in traditional feeding centers in the brain, including the caudal brainstem and the hypothalamus (1, 59, 60). Our findings taken together with previous work reveal that central OT signaling bidirectionally influences meal size, dependent on the targeted brain region and the social context.

The SFE effect is a robust phenomenon studied in humans that involves two primary components: eating more in the presence of a group vs. in isolation, and eating more in the presence of familiar vs. unfamiliar individuals (15, 17, 18). While the former component has been demonstrated in rats (23, 24), to our knowledge the latter has not. Here we demonstrate that rats consume more in the presence of a familiar conspecific vs. either isolated conditions or in the presence of an unfamiliar conspecific. Further, using complementary gain- and loss-of-function approaches, we reveal that familiarity-based SFE is mediated by OT signaling in the HPCd. These findings are in line with previous work showing that OT, which has typically been thought to elicit predominantly prosocial effects, has a more complicated role in social interactions based on social familiarity and perceived in- vs. out-group dynamics (12, 13). Related, previous findings in mice revealed increased sucrose consumption following peripheral OXTR antagonist injections in both non-social and social contexts in dominant mice, but only in a non-social context in subordinate mice (50). Results from the present study expand these findings by revealing that HPCd OXTR signaling promotes prosocial effects on eating that are dependent on conspecific familiarity. Additional research is needed to determine whether either the number of conspecifics present or the dominant vs. submissive status of the animals influence HPCd OXTR-mediated effects on eating.

The hippocampus is important in mediating social behaviors, including SRM (61, 62), and knockout of OXTR in the CA2/CA3 region in mice impairs SRM (30). Here we focused on the DG subregion in the HPCd, a region established as an important center for adult neurogenesis and memory function, particularly for context recognition and memory tasks involving pattern separation (63–66). The ability to distinguish individuals based on auditory, visual, and olfactory cues is essential for SRM, and can be considered an analogous process to DG-mediated pattern separation. Indeed, adult-born DG neurons are associated with social memory maintenance in mice (53, 67). Here we extend these previous studies by identifying a novel role for DG OXTR signaling in mediating SRM. Using a viral vector-mediated approach to knockdown HPCd DG oxytocin receptors by ~80%, our results revealed normal sociability, yet an absence of SRM in the knockdown group. Given these findings taken together with our additional results showing that DG OXTR knockdown eliminates the familiarity-based SFE effect, it may be that HPCd OT signaling promotes prosocial eating by enhancing the mnemonic familiarity of a conspecific. Additional work is required to determine whether these effects are mediated by OXTR expressed on adult-born DG neurons.

Social-based learning about eating behavior is extremely beneficial from an evolutionary perspective as it allows an animal to mitigate food-related risks by learning from another individual’s experience. This is especially the case for rodents such as rats that lack the ability to dispel harmful chemicals after consumption through emesis. As such, rats exhibit robust food neophobia (68–71). In the STFP procedure, animals learn to prefer food with a flavor that was previously experienced through an interaction with a “demonstrator” animal that recently consumed food with that flavor (36, 38, 39, 72). OT facilitates STFP learning in rats (47), however, this previous work did not provide insight into the neural site(s) of action. Given our results indicating a role for HPCd OT signaling in SFE, and previous findings that the HPCd is critical for STFP (51), here we hypothesized that HPCd OT signaling mediates STFP. Our results confirmed this, as HPCd OXTR knockdown prevented STFP learning without influencing sociability during the social interaction or the total amount consumed during the preference test.

It is possible that HPCd OT signaling functions as a general “safety signal” to eat (SFE), or to eat specific foods (STFP). However, this is unlikely as HPCd OT administration had no effect on either appetitive or consummatory behaviors in a nonsocial risk-associated approach-avoidance task. Moreover, effects of HPCd OT signaling on familiarity-dependent SFE, or STFP are unlikely to be based on a general role in novelty recognition memory, as HPCd OT knockdown did not impact novel object recognition memory. Rather, given that both SFE and STFP involve learning about social experiences relevant to eating, we hypothesize that endogenous HPCd OT signaling facilitates the acquisition and/or retention of social-based food-relevant memories to guide future eating behavior.

A limitation of the present study is the exclusion of females. We have recently identified sex differences in the effect of ICV OT administration on food intake, mediated by gonadal hormones and the estrus cycle (73). Moreover, there are known sex differences in social behavior, feeding behavior, OT production, and central OXTR expression (74, 75). Investigating whether HPCd OT signaling mediates SFE in females is an important follow-up direction. Given that ICV OT preferentially reduces consumption of highly palatable foods under isolated conditions (76–78), it would also be interesting for future work to examine how macronutrient composition alters the effect of OT administration on food intake in social contexts.

Our results demonstrate a novel SFE effect in a rat model that is dependent on conspecific familiarity, similar to the phenomenon observed in humans (15, 17, 20). Both gain- and loss-of-function approaches identify oxytocin signaling in the HPCd DG as a neural substrate mediating SFE, as well as STFP, potentially via SRM-mediated processes. Collective findings thus reveal novel mechanisms through which OT intersects the control of food intake and social behaviors. Given that OT is currently being investigated as a potential therapy for the treatment of obesity (8–10), present results should be taken into consideration regarding clinical obesity relevance, as OT’s role as an anorexigenic system may be more complex than previously considered.

KEY RESOURCES TABLE.

Resource Type	Specific Reagent or Resource	Source or Reference	Identifiers	Additional Information
Add additional rows as needed for each resource type	Include species and sex when applicable.	Include name of manufacturer, company, repository, individual, or research lab. Include PMID or DOI for references; use “this paper” if new.	Include catalog numbers, stock numbers, database IDs or accession numbers, and/or RRIDs. RRIDs are highly encouraged; search for RRIDs at https://scicrunch.org/resources.	Include any additional information or notes if necessary.
Bacterial or Viral Strain	AAV1-GFP-U6-r-rOXTR-shRNA(4)	Vector Biolabs	RefSeq#NM_012871, rat OXTR-shRNA transgene, eGFP reporter
Bacterial or Viral Strain	AAV1-GFP-U6-ShRNA	Vector Biolabs	Cat#7040, scrambled shRNA transgene, eGFP reporter
Chemical Compound or Drug	Oxytocin (OT)	Bachem	Cat# 4016373
Chemical Compound or Drug	5-Thio-D-Glucose (5-TG)	Sigma-Aldrich	Ca#88635-1G
Chemical Compound or Drug	Lithium Chloride (LiCl)	Sigma-Aldrich	Cat# L4408-100G
Chemical Compound or Drug	Prolong Gold AntiFade Reagent	Cell Signaling Technology	Cat#9071s
Commercial Assay Or Kit	RNAscope Fluorescent Multiplex Detection Reagent Kit	Advanced Cell Diagnostics	Cat#323100
Commercial Assay Or Kit	RNeasy Lipid Tissue Mini Kit (50)	Qiagen	Cat#74804
Commercial Assay Or Kit	QuantiTect Rev. Transcription Kit (50)	Qiagen	Cat#205311
Commercial Assay Or Kit	TaqMan Fast Advanced Master Mix for qPCR	AppliedBioSystems by Thermo Fisher Scientific	Cat#4444557
Organism/Strain	Rat, Sprague Dawley, male	Envigo	RRID:RGD_737903
Sequence-Based Reagent	OXTR GE Assay (primers for ct-qPCR)	Thermo Fisher Scientific	ID: Rn00563503_m1
Sequence-Based Reagent	Gapdh GE Assay (primers for ct-qPCR)	Thermo Fisher Scientific	ID: Rn01775763_g1
Sequence-Based Reagent	RNAscope Probe-Rn-Oxtr-C3	Advanced Cell Diagnostics	Cat#483671
Software; Algorithm	BioDAQ Data Viewer	Research Diets Inc.	N/A
Software; Algorithm	Prism v10.2.3	GraphPad	N/A
Software; Algorithm	NIS-Elements BR Software	Nikon Instruments Inc.	N/A