Characterizing Brainstem GLP-1 Control of Sensory-Specific Satiety in Male and Female Rats Across the Estrous Cycle

Abstract

BACKGROUND:

Meal variety promotes overconsumption by delaying sensory-specific satiety (SSS), the transient reduction in reward value of a recently consumed food. Despite its role in meal cessation, the neuroendocrine mechanisms that underlie SSS are largely unknown.

METHODS:

Here, we developed a preclinical model of SSS wherein rats consume more of a different food compared with the same food presented again, leading to greater caloric intake. Using pharmacological and molecular approaches targeting the brainstem, we investigated the involvement of the satiation signal glucagon-like peptide-1 (GLP-1) in mediating SSS in male rats (n = 96) and in female rats (n = 85) across their estrous cycle. We also evaluated the sufficiency of the hormone estradiol to modulate GLP-1 and SSS.

RESULTS:

In males, brainstem GLP-1 receptors (GLP-1Rs) were necessary for the SSS-induced decrease in same food intake, while agonizing brainstem GLP-1Rs was sufficient to attenuate overconsumption of the different food. Female rats showed SSS in an estrous cycle–dependent manner and did not consume more of the different food in diestrus-to-proestrus and proestrus-to-estrus. However, blockade of brainstem GLP-1Rs restored different food overconsumption. Furthermore, the brainstem’s nucleus tractus solitarius and area postrema showed increased expression of the GLP-1 precursor glucagon (Gcg), during diestrus-to-proestrus and proestrus-to-estrus and greater Glp1r expression in proestrus-to-estrus. Similarly, 17β-estradiol injections in males not only increased Glp1r and Gcg expression but also reduced SSS.

CONCLUSIONS:

We identified a bidirectional role for brainstem GLP-1R signaling in modulating SSS, effects that are estrous cycle dependent. Moreover, our data indicate that estradiol regulates Glp1r and Gcg expression and likely influences SSS.

One reason for the continuous rise in obesity prevalence is the wide variety of available foods in the modern environment (1,2). Meal variety promotes overconsumption by delaying sensory-specific satiety (SSS) (3–5), the transient decline in pleasantness and motivation for a recently consumed food (6,7). While SSS suppresses intake of recently eaten foods, it comparatively increases the perceived relative value of different, un-eaten foods (3). Thus, a meal with multiple different foods can reduce overall SSS, thereby extending a meal and potentially promoting caloric overconsumption (3–5,8). While SSS is an important factor in meal termination, preclinical studies have primarily focused on reward-associated mechanisms in SSS, including opioid signaling in the nucleus accumbens (9) or the encoding of reward value in the prefrontal cortex (10–12). It is unclear which neural substrates that control satiation are involved in SSS.

Within the nucleus tractus solitarius (NTS) of the caudal brainstem, the neuropeptide glucagon-like peptide-1 (GLP-1) plays a critical role in meal termination (13). The NTS contains both preproglucagon (PPG) neurons that synthesize endogenous brain GLP-1, as well as GLP-1 receptor (GLP-1R)–expressing cells (14). PPG neurons integrate satiation signaling from the gastrointestinal tract and are necessary for limiting large meal size by promoting meal termination (15,16). Moreover, NTS GLP-1R–expressing cells are essential for restricting meal size and food-motivated behaviors (17,18). Given its role in meal termination, we sought to understand the contribution of brainstem GLP-1–GLP-1R signaling in mediating SSS in both male and female rats.

Although there is an abundance of literature disentangling the mechanisms of brainstem GLP-1 in attenuating food intake and food-motivated behaviors, most studies have included only male preclinical animal models (18,19). This is generally done to avoid confounding anorexigenic effects associated with the cycle-dependent sex hormone estradiol (20,21). Among reports that do include females (22,23), none to date have examined the relationship between brainstem GLP-1–GLP-1R signaling and the rodent estrous cycle. However, the results of several studies suggest that phases associated with greater estrogen signaling may enhance the effects of hypothalamic/forebrain GLP-1R action (24–26). To address this knowledge gap, we endeavored to develop an SSS model to investigate estrous cycle–dependent changes in brainstem Glp1r and Gcg expression. To our knowledge, the data described here are the first to show the involvement of a satiation signal, GLP-1, in mediating SSS, as well as the first to show that the central GLP-1 system is highly responsive to estrogen signaling. Finally, we investigated the bed nucleus of the stria terminalis (BNST) as a neural integrator of reward-associated mechanisms underlying SSS to modulate homeostatic brainstem GLP-1–GLP-1R signaling. Our collective findings may have broader implications for understanding the potential mechanisms underlying the discrepancies between sexes in the magnitude of weight lost while on GLP-1–based pharmacotherapies (27,28).

METHODS AND MATERIALS

For additional details, refer to Supplemental Methods.

Animals

For all behavioral studies, adult male and female Sprague Dawley rats (Charles River) were housed individually in hanging metal wire cages at the University of Pennsylvania. Female Fischer 344 (wild-type or transgenic animal model of Alzheimer’s disease, TgF344-AD) were group housed in shoebox cages at Georgia State University and used only to investigate gene expression changes following sham or ovariectomy. All animals were under a 12-hour light/dark cycle in a temperature- and humidity-controlled vivarium and provided ad libitum access to water and lab chow diet (Purina LabDiet) except when noted. All behavioral testing occurred in the home cage. The University of Pennsylvania and Georgia State University Institutional Animal Care and Use Committee approved all protocols.

Sensory-Specific-Satiety Studies

General SSS Protocol: Days 1 to 4.

We aimed to develop a novel rat SSS procedure that truly recapitulates the human behavioral phenomenon. To ensure that our procedure evaluated behavioral changes attributed to recent food experiences, we sought to identify 2 equally preferred snack foods. Briefly, after repeated attempts, we decided on a behavioral assay using Pringles BBQ Crisps (5.35 kcal/g) and Original Ritz crackers (5.0 kcal/g) (nutrition facts and ingredients are described in Tables S1 and S2, respectively). As described in Figure 1A, we habituated rats to 1.5 g each of these snacks for 2 days. On day 3, we conducted a 30-minute preference test, providing rats access to 10 g of each food in the absence of lab chow. On day 4, we provided 1.5 g of each snack.

Figure 1.

Male rats show SSS in a novel SSS procedure. (A) SSS procedure timeline. (B) Male rats equally preferred BBQ Pringles (22.34 ± 2.34 kcal) and Ritz crackers (18.75 ± 2.88 kcal) in a 30-minute preference test. (C) Rats demonstrated SSS, consuming more of the different food (30.04 ± 2.61 kcal) than the same food (17.75 ± 2.18 kcal) during the second food-access period. (D) Food intake measured over 24 hours, including SSS snacks (teal) and lab chow (gray), was increased in rats given the different food (same 87.40 ± 3.44 kcal vs. different 104.0 ± 3.32 kcal). These rats did not compensate for the additional kilocalories, consuming a similar amount of lab chow as animals that received the same food (same 69.65 ± 3.38 kcal vs. different 73.93 ± 3.96 kcal) (n = 14 males, within-subjects design). Data were analyzed with paired t tests and are expressed as mean ± SEM. **p < .01. SSS, sensory-specific satiety.

General SSS Protocol: Day-5 Onward SSS Testing.

Our SSS procedure is based on methods derived from Ahn and Phillips (12). Immediately before testing, we removed lab chow from the home cage. We provided rats with two 30-minute food-access periods separated by a 45-minute recess without food. During the first food-access period, we allowed rats access to 10 g of Pringles BBQ Crisps or Original Ritz crackers, which is enough to permit satiation/voluntary meal termination. For the second food-access period, we supplied rats with 12 g of either the same or different food compared to their first food-access period food. When relevant, we administered drug into the fourth ventricle during the recess. We organized testing in a counterbalanced, within-subjects design, with each test separated by at least 48 hours. On nontest days, we gave rats 1.5 g of each snack to avoid a binge-like feeding response (29). We counterbalanced each test by individual snack preference and body weight. We also counterbalanced the identity (Pringles BBQ Crisps or Original Ritz crackers) of each rat’s different food. However, we kept the identity of each individual rat’s same or different food consistent across tests to allow for within-subject comparisons. We conducted all testing during the light cycle, approximately 1 to 3.5 hours before the onset of the dark cycle.

Estrous Cycle Phase Assignment

We monitored the estrous cycle using daily vaginal lavage, collecting samples 3.5 to 4 hours before the onset of the dark cycle, approximately 30 to 60 minutes before SSS testing or sacrifice. Based on (30,31), we assigned cycle phase labels to the 24-hour period that ended at the time of sampling. Because our SSS tests and brain extractions occurred near transitions between phases, such as at the start of metestrus directly after estrus, we note this transition in phase assignments (i.e., estrus to metestrus).

Food Intake Measurements

For SSS tests, we determined intake by calculating the cumulative change in weight in grams of the snack food before and after the 30-minute food-access period. Similarly, we measured lab chow intake or body weight change by determining the cumulative change in the weight of the chow hopper or the animal starting before SSS tests and 24 hours after SSS tests. To account for spillage of snacks or chow, we placed a brown paper under each hanging wire cage to collect any crumbs and subtracted crumbs from the weight change of the snack or chow hopper. For each snack, we then converted intake to kilocalories.

Study Design

We designed this study to investigate the role of brainstem GLP-1–GLP-1R signaling in mediating SSS. To start, we used male rats to determine the necessity and sufficiency of brainstem GLP-1Rs by infusing an established (13,32) subthreshold dose (10 μg/2 μL) of the GLP-1R antagonist exendin-9–39 (Ex-9) (Cayman Chemical) or a subthreshold dose (0.025 μg/1 μL) of the GLP-1R agonist exendin-4 (Ex-4) (Bachem) into the fourth ventricle and testing SSS. We selected the fourth ventricle to target the hindbrain, including the NTS, while avoiding any forebrain drug effects (because of the rostral-to-caudal flow of cerebrospinal fluid) (33). When we expanded our interrogations to female rats, we uncovered estrous cycle–dependent SSS and tested the necessity of brainstem GLP-1R activation in estrous cycle–dependent SSS. To complement our behavioral results, we extracted tissue from the NTS and nearby area postrema (AP) and used quantitative polymerase chain reaction (qPCR) to examine expression changes in Glp1r and precursor messenger RNA for GLP-1, Gcg, across the estrous cycle and in ovariectomized Fischer 344 females.

Using qPCR, we also determined the sufficiency of the sex hormone estradiol to regulate Glp1r and Gcg expression in the NTS/AP. We injected male rats (habituated to subcutaneous injections) with endogenous estrogen receptor agonist 17β-estradiol (Bio-Techne). We selected subcutaneous 10 μg/kg 17β-estradiol based on studies that have shown that this dose and administration route produced serum levels similar to physiological estrogen in females (34,35). Finally, we also investigated the ability of 17β-estradiol to modulate SSS in male rats.

To understand how estradiol might affect brainstem Glp1r and Gcg expression, we determined coexpression with estrogen receptors. First, we analyzed our previously published Sprague Dawley rat AP/NTS single-nucleus RNA sequencing data from male rats (36). We used the gene Gcg to identify PPG neurons. Previously identified neuronal clusters were subclustered at a resolution of 2, and gene expression was visualized using Seurat 5.0.1 (37). This dataset is available under GEO accession nos. GSE167981, GSE167991, and GSE216247. Second, we conducted single-molecule fluorescent in situ hybridization (smFISH) for Gcg and Esr1 in the NTS.

Lastly, we began to consider the role of subcortical regions in modulating the control of brainstem GLP-1–GLP-1R signaling in SSS. To investigate a putative BNST-to-NTS connection, we used a mono-transsynaptic tracing approach (38) paired with smFISH for Gcg and Glp1r.

Statistical Analysis

For preference tests, we compared the number of kilocalories consumed by snack type using 2-tailed paired t tests. For SSS tests, we did not include the between-subjects factor of snack type (BBQ Pringles, Ritz Crackers) in analyses because our preliminary data showed no effect of snack type on intake during the second food-access period (Figure S1). We analyzed the data using either 2-tailed paired t tests, 1-way analysis of variance (ANOVA), or mixed- or repeated-measures ANOVA, as appropriate. Behavioral statistical analyses, including main effects and interactions, are presented in Table S3. Where relevant, we followed significant main effects with Fisher’s least significant difference (LSD) or Tukey’s post hoc tests (specified in the figure legends). We considered p < .05 significant. All data are expressed as mean ± SEM. To analyze data, we used GraphPad Prism version 10.2.3 software.

RESULTS

Establishing the SSS Procedure

We developed a preclinical model of SSS to investigate the brainstem action of GLP-1 in mediating SSS. We used 2 snack foods, BBQ Pringles and Ritz Crackers, that rats (n = 14, male) can discriminate between and equally preferred (Figure 1B). In our SSS tests, we found that rats that were provided a different food consumed significantly more calories than rats that were given the same food a second time (p = .0024) (Figure 1C). Interestingly, these rats did not compensate for the additional kilocalories, consuming a similar amount of lab chow after SSS regardless of whether they received the same or different food. As a result, animals consumed more total kilocalories within a 24-hour period when provided with a different snack (p = .0099) (Figure 1D). These experiments support the rigor of our novel SSS procedure and suggest that it is translationally relevant for studying excess caloric intake.

Role of GLP-1R in Mediating SSS

To begin to investigate the necessity of hindbrain GLP-1R signaling in mediating SSS, we administered the GLP-1R antagonist Ex-9 (10 μg/2 μL; fourth intracerebroventricular) in male rats. We selected a subthreshold dose of Ex-9 based on dose-response literature indicating this dose does not increase intake on its own (13,32). While vehicle-treated rats showed SSS and consumed more of the different food (p = .0237), Ex-9–treated animals consumed similar amounts of the same food and the different food (Figure 2A). Relative to vehicle-treated animals, Ex-9 increased intake of the same food (p = .0197) without affecting intake of the different food. We interpret these data to suggest that brainstem GLP-1Rs are necessary for the SSS-induced reduction of same food intake.

Figure 2.

GLP-1Rs in the caudal brainstem are necessary for the SSS-induced decline in same food intake and sufficient to prevent overconsumption of different food intake. (A) GLP-1R antagonist exendin-9–39 (10 μg/2 μL) injected into the fourth ventricle increased intake of the same food (vehicle 17.79 ± 3.04 vs. exendin-9 29.56 ± 4.53 kcal), without affecting different food intake (n = 9 males, within-subjects design). (B) In a different cohort, GLP-1R agonist exendin-4 (0.025 μg/1 μL) injected into the fourth ventricle selectively attenuated intake of the different food (vehicle 25.66 ± 1.45 vs. exendin-4 17.30 ± 2.15 kcal) (n = 20 males, within-subjects design). Post hoc tests used Fisher’s least significant difference, and we report adjusted p values. Data are expressed as mean ± SEM. *p < .05, **p < .01, ***p < .001. Dif, different food; GLP-1R, GLP-1 receptor; S, same food; SSS, sensory-specific satiety.

Next, we evaluated the sufficiency of hindbrain GLP-1R to control SSS. In a new cohort of rats (n = 20 males), we administered a subthreshold dose (39) of the GLP-1R agonist Ex-4 (0.025 μg/1 μL; fourth intracerebroventricular). Agonizing GLP-1Rs attenuated overconsumption of the different food (p = .0008) without affecting intake of the same food (Figure 2B). Interestingly, Ex-4 did not entirely prevent SSS because animals consumed significantly more of the different food (p = .0062). There was also no significant effect or interaction of order of exposure to Ex-4, suggesting that data were not impacted by the aversive effects of Ex-4 (40) (Figure S2). Together, these data suggest that activation of brainstem GLP-1R is sufficient to selectively reduce different food intake.

Role of GLP-1 in Mediating Estrous Cycle– Dependent SSS

To extend the translational relevance of our SSS model, we examined SSS in female rats. Surprisingly, while females (n = 14) equally preferred BBQ Pringles and Ritz crackers (Figure 3A), they did not consume significantly more of the different food (p = .0871) (Figure 3B). Given that energy intake varies across the female estrous cycle (41), we hypothesized that SSS might be masked by hormonally induced variation in food intake. To investigate, we determined each phase of the estrous cycle using vaginal cytology (42). Consistent with previous literature (30), lab chow intake and body weight declined during the diestrus-to-proestrus and proestrus-to-estrus phases (Figure S3A, B). Then, we conducted SSS tests during each of the 4 estrous phases: estrus to metestrus, metestrus to diestrus, diestrus to proestrus, and proestrus to estrus. Female rats consumed more of the different food (p < .0001) during estrus-to-metestrus and metestrus-to-diestrus, but not during phases associated with greater estradiol signaling, diestrus to proestrus (p = .0870) or proestrus to estrus (p = .2309) (Figure 3C). Specifically, rats suppressed intake of the different food during diestrus-to-proestrus compared with metestrus-to-diestrus (p = .0214) and during proestrus-to-estrus compared with metestrus-to-diestrus (p < .0001) and estrus-to-metestrus (p = .0078). In contrast, intake of the same food was similar across all phases of the estrous cycle. These data show that the increase in different food intake is an estrous cycle phase-dependent behavior in female rats.

Figure 3.

Different food overconsumption is dependent on the estrous cycle. (A) Female rats equally preferred BBQ Pringles (14.27 ± 2.47 kcal) and Ritz crackers (15.93 ± 1.02 kcal) in a 30-minute preference test. (B) Female rats did not consume significantly more of the different food (25.12 ± 2.14 kcal) than the same food (19.39 ± 1.84 kcal). (C) Female rats demonstrated SSS during e-M and m-D but did not overconsume the different food during d-P (same: 16.56 ± 2.17 vs. different: 20.19 ± 1.09 kcal) or p-E (same: 13.18 ± 1.84 vs. different: 15.80 ± 2.30 kcal) (n = 14 females, within-subjects design). Data are expressed as mean ± SEM. Data in (A) and (B) were analyzed with paired t tests. Significant comparisons were otherwise followed up with Tukey’s post hoc tests. **p < .01, ***p < .001. d, diestrus; e, estrus; m, metestrus; p, proestrus; SSS, sensory-specific satiety.

Having demonstrated the necessity and sufficiency of brainstem GLP-1R signaling to regulate SSS in male rats, we hypothesized that the estrous cycle phase-dependent alterations in female rat SSS behavior may be due to phase-dependent alterations in hindbrain GLP-1–GLP-1R signaling. We reasoned that the brainstem GLP-1–GLP-1R system might be engaged in diestrus-to-proestrus and proestrus-to-estrus, consequently limiting intake of the different food. We hypothesized that if this were true, antagonizing brainstem GLP-1Rs would restore overconsumption of the different food. Therefore, we tested SSS in a new cohort of female rats (n = 22) and administered a subthreshold dose of Ex-9 (10 μg/2 μL; fourth intracerebroventricular) during combined diestrus-to-proestrus/proestrus-to-estrus or estrus-to-metestrus/metestrus-to-diestrus. These phases are frequently pooled (25,43) due to overlapping hormonal and behavioral profiles (34), and we observed no differences between diestrus-to-proestrus and proestrus-to-estrus or estrus-to-metestrus and metestrus-to-diestrus in the previous experiment. Our results corroborated our estrous cycle–dependent SSS data; rats showed SSS during estrus-to-metestrus/metestrus-to-diestrus (vehicle treated, p = .0273; Ex-9 treated, p = .0312), while vehicle-treated rats in diestrus-to-proestrus/proestrus-to-estrus did not (Figure 4A). However, vehicle-treated animals in diestrus-to-proestrus/proestrus-to-estrus showed decreased intake of both the same food (p = .0312) and the different food (p < .0001) relative to animals in estrus-to-metestrus/metestrus-to-diestrus. Moreover, Ex-9–treated animals showed decreased intake of the same food relative to animals in estrus-to-metestrus/metestrus-to-diestrus (p = .0109).

Figure 4.

Fluctuations in estradiol contribute to increased brainstem Gcg and Glp1r expression. (A) GLP-1R antagonist exendin-9 (fourth ventricle injection; 10 μg/2 μL) selectively rescued overconsumption of the different food during d-P/p-E food (vehicle 14.50 ± 1.24 vs. exendin-9 24.92 ± 1.54 kcal). (B) In female rats (n = 9–10/group, between-subjects design), transcript expression of Glp1r was increased during p-E (4.03 ± 0.20) compared with e-M (1.97 ± 0.20), m-D (1.65 ± 0.20), and d-P (2.51 ± 0.30). In male rats (n = 12–13/group, between-subjects design), 17β-estradiol injected subcutaneously at a dose of 10 μg/kg increased Glp1r expression (vehicle 1.39 ± 0.01 vs. 17β-estradiol 3.60 ± 0.11). (C) In female rats, transcript expression of Gcg was increased during d-P (2.74 ± 0.30) and estrus (3.23 ± 0.44) compared with e-M (1.17 ± 0.09) and m-D (1.28 ± 0.12). In male rats, 17β-estradiol increased Gcg expression (vehicle 1.33 ± 0.19 vs. 17β-estradiol 2.75 ± 0.52). (D) Transcript expression of Glp1r was similar in sham surgicated females (1.00 ± 0.10) and ovariectomized females (0.97 ± 0.38) (n = 11–15/group, between-subjects design). (E) Transcript expression of Gcg was increased in sham surgicated females (1.092 ± 0.09) compared with ovariectomized females (0.84 ± 0.05). Data in (D) and (E) were analyzed with a between-subjects, 2-way analysis of variance, and because we observed no main effect of genotype (wild-type, TgF344-AD) and no interaction, we grouped data here by surgical intervention. (F) In male rats, 17β-estradiol (10 μg/kg, subcutaneous) prevented SSS (n = 18). Significant comparisons were otherwise followed up with Fisher’s least significant difference test, and we report adjusted p values or Tukey’s multiple comparisons post hoc tests. *p < .05, **p < .01, ***p < .001. 17β-Est, 17β-estradiol. d, diestrus; Dif, different food; e, estrus; GLP-1R, GLP-1 receptor; m, metestrus; Ovx, ovariectomized; p, proestrus; S, same food; SSS, sensory-specific satiety; Veh, vehicle.

Importantly, Ex-9 selectively restored SSS during diestrus-to-proestrus/proestrus-to-estrus, with animals consuming more of the different food than the same food (p < .0001) and compared with different food intake of vehicle-treated animals (p < .0001). Furthermore, Ex-9 did not affect food intake during estrus-to-metestrus/metestrus-to-diestrus or same food intake in diestrus-to-proestrus/proestrus-to-estrus. These results reveal that brainstem GLP-1Rs are necessary for the suppression of different food intake that occurs selectively in diestrus-to-proestrus/proestrus-to-estrus.

Changes in Brainstem Glp1r and Gcg Transcription in Male and Female Rats

Next, we sought to understand the mechanism that underlies the diestrus-to-proestrus/proestrus-to-estrus phase-specific control of GLP-1R in suppressing intake of the different food. Notably, the sex hormone estradiol is associated with anorexigenic behaviors (20,21). Although estradiol concentrations rise during diestrus and peak during proestrus (34), estradiol’s anorexigenic effects are delayed (44), occurring during proestrus and estrus. Given the lack of SSS during those 2 phases, we hypothesized that Glp1r and Gcg expression might be increased in diestrus-to-proestrus and proestrus-to-estrus. In nearly all cases, the endogenous ligand for brain GLP-1Rs originates from caudal brainstem PPG neurons (45,46) located within the NTS (14,47). Moreover, brainstem GLP-1R–expressing cells are located within the NTS and AP (48). Therefore, we used qPCR to examine Gcg and Glp1r expression in microdissected NTS/AP tissue from female rats across each stage of the cycle, using tissue from the 2 previous experiments (see Supplemental Results).

Gcg Neurons Express the Estrogen Receptor 1

Next, we sought to understand how estradiol injections in males increased NTS/AP Gcg and Glp1r expression (Figure 4B, C). We reanalyzed our previously described single-nucleus transcriptomic dataset of combined NTS/AP tissue from male rats (36) to determine the expression pattern of estrogen receptors. Given 17β-estradiol’s delayed action in males, we hypothesized that one of the nuclear receptors, ERα or Erβ (49), would be involved. We identified the previously detailed major cell types (36) (Figure S6) and subclustered the neurons to generate a detailed profile of cholinergic, excitatory, and inhibitory neurons in the rat NTS/AP (Figure 5A). Gcg-expressing PPG neurons also expressed the gene for ERα, Esr1 (Figure 5A, red box).

Figure 5.

(A) A dot plot depicting the expression of the GLP-1 and estrogen signaling–related genes in cholinergic excitatory and inhibitory neuronal subtypes of the rat NTS/area postrema. Gcg-expressing preproglucagon neurons express Esr1 (red box). (B) Representative caudal NTS section stained for messenger RNA of the Glp-1 precursor Gcg (red) and estrogen receptor 1 Esr1 (green) and cover slipped with DAPI-containing antifade mounting media (blue). (C, D) Enhancement of the white outlines in (A). (E, F) Quantification of NTS Gcg neurons that coexpress Esr1 at the level of the NTS posterior to the obex (approximately −15.4 to −14.7 posterior to bregma) and at the level of the obex (approximately −14.7 to −14.1 posterior to bregma in male rats). Expressed as percentage of Gcg cells that express Esr1 (E) and (F) total number of Gcg cells that do (blue) or do not (orange) express Esr1 (n = 5 males). CC, central canal; Chol., cholinergic; NTS, nucleus tractus solitarius; Post., posterior.

To independently validate these data, we performed smFISH in male rat (n = 5) brainstem slices and identified cells that coexpress Gcg and Esr1 (Figure 5B–D). Like previous studies (14), we observed Gcg-expressing cells in the caudal NTS and detected a prominent number of Esr1+ Gcg-expressing cells (Figure 5E). About 71.73% of Gcg-labeled NTS cells caudal to the obex and 68.44% of Gcg-labeled cells at the level of the obex were Esr1 positive (Figure 5F).

BNST-to-NTS–Projecting Cells Express Gcg and Glp1r Messenger RNA

We interpret our data to suggest that hindbrain GLP-1–GLP-1R signaling is not engaged in the presence of the different food. Given this interpretation, we speculate that higher-order brain regions known to ascribe value in SSS (10) may play a role in limiting GLP-1–GLP-1R signaling to enable overconsumption. In the presence of the different food, a brain region involved in integrating cortical reward information might send inhibitory signals to the brainstem. Notably, the BNST is thought to encode changes in motivational properties, including food (50,51). Given this function, the BNST is an attractive substrate for the top-down modulation of brainstem GLP-1–GLP-1R signaling. To investigate the presence of top-down projections, we determined whether the BNST projects to Gcg or Glp1r-expressing cells in the NTS. We used a mono-transsynaptic tracing approach (38) (n = 4) in males because we hypothesized that the projections would not vary as a function of sex or estrous cycle (Figure 6A–F). The BNST projected to nearly 50% of NTS Gcg cells and 20% of Glp1r-expressing cells (Figure 6G, H). Given these percentages, future studies should investigate the role of projections from the BNST-to-NTS PPG or GLP-1R–expressing cells using transgenic animals.

Figure 6.

Mono-transsynaptic anterograde labeling of NTS cells receiving BNST input. (A) Representative image of the BNST (white dashed outline) co-injected with an anterograde cre-expressing AAV1 and CTB-555 (red) and cover slipped with DAPI-containing antifade mounting media (blue). (B) Representative image of the NTS injected with AAV1 expressing cre-dependent EGFP (green). (C, E) Representative caudal NTS sections stained for messenger RNA of the EGFP (white), GLP-1 precursor, Gcg (pink) and Glp1r (green) near the level of the obex (C) and AP (E). (D) Enhancement of the white outline in (C). (F) Enhancement of the white outline in (E). (G) Quantification of Glp1r and Gcg-expressing cells that coexpress Cre-dependent EGFP in the caudal NTS at the level of the obex (approximately −14.3 to −14.7 posterior to bregma) and (H) at the level of the AP (approximately −13.4 to −14.2 posterior to bregma). Quantification is expressed as percentage of Glp1r or Gcg cells that express EGFP. AP, area postrema; BNST, bed nucleus of the stria terminalis; CC, central canal; CTB-555, cholera toxin B; EGFP, enhanced green fluorescence protein; LV, lateral ventricle; NTS, nucleus tractus solitarius.

DISCUSSION

Evolutionarily, SSS serves to promote intake of nutritionally diverse foods (52). However, in the modern food environment, varied, palatable foods can stimulate excess caloric consumption. Surprisingly, given the prevailing obesity epidemic, few preclinical studies have examined the neural mechanisms underlying SSS that lead to changes in energy intake. SSS is frequently used as a method of reward devaluation, predominantly measuring instrumental responding or cue seeking without food intake or providing a different food (53,54). Importantly, the current report represents the first preclinical evidence that establishes a role for the satiation peptide GLP-1 in mediating SSS.

The current pharmacological data targeting the brainstem GLP-1R system in males and females, paired with the endogenous changes in Glp1r and Gcg expression across the estrous cycle, identified a bidirectional role for brainstem GLP-1–GLP-1R control of SSS. Antagonizing brainstem GLP-1Rs in males selectively affected same food intake such that animals no longer consumed significantly less of the same food than the different food. In contrast, agonism of brainstem GLP-1Rs using a subthreshold dose attenuated intake of the different food. These data begin to point to greater levels of brainstem GLP-1R engagement during exposure to the same food than during exposure to the different food. However, it is also plausible our results are due to ceiling/floor effects: GLP-1R antagonism in males did not impact different food intake because intake was already high, and GLP-1R agonism did not decrease same food intake because it was already low. However, our female SSS data also suggest the necessity for brainstem GLP-1–GLP-1R signaling in mediating SSS. Female rats did not demonstrate SSS when in diestrus-to-proestrus and proestrus-to-estrus, and subthreshold GLP-1R antagonism restored increased consumption of the different food without affecting same food intake only during proestrus/estrus. It is compelling that this phase-dependent restoration occurred during phases associated with greater Glp1r expression (proestrus-to-estrus) and Gcg expression (diestrus-to-proestrus and proestrus-to-estrus). We speculate that GLP-1–GLP-1R signaling is onboard in the presence of the same food, and increased Glp1r and Gcg expression during diestrus-to-proestrus or proestrus-to-estrus lowers the threshold for triggering brainstem GLP-1–GLP-1R signaling. Altogether, we interpret our data to suggest that brainstem GLP-1–GLP-1R signaling is involved in SSS to limit same food intake and to limit different food intake in female rats in diestrus-to-proestrus and proestrus-to-estrus.

The role of brainstem GLP-1 in controlling ingestive behavior, particularly within the NTS (55), is widely supported. GLP-1Rs in the NTS play a critical role in satiation (13) and are involved in suppressing food reward or motivation (18,56). Notably, NTS PPG neurons show activation in response to a large meal (57), and in mice, their activation promotes termination of a second meal after a preload (15). While the current study is the first to identify a role for GLP-1 in SSS, it is likely not the only neural signal involved in SSS control.

There is little preclinical research that has examined estrous cycle–dependent changes in brain GLP-1. However, there is precedent for the estrous cycle to impact behaviors regulated by GLP-1R. During estrus, but not metestrus/diestrus, GLP-1R activation in the lateral hypothalamus attenuates food reinforcement (24). Furthermore, Glp1r expression in the hypothalamus is increased during proestrus/estrus (26), a pattern consistent with our NTS/AP expression data. These findings are particularly intriguing when contextualized with our results showing that females have increased NTS/AP Glp1r expression in proestrus-to-estrus and Gcg expression in diestrus-to-proestrus and proestrus-to-estrus compared with baseline expression in males. Collectively, these data highlight a possible mechanism for why female rats demonstrate more sensitivity to GLP-1R agonists (23,58). Our results have translational relevance when considering the sex-divergent effects of GLP-1R pharmacotherapies on weight loss. In most clinical trials, compared with men, women show a greater percentage of weight loss in response to GLP-1R agonists, from exenatide (59,60), liraglutide (61), semaglutide (62,63), dulaglutide (61,64), and triple receptor agonist retatrutide (65). Future clinical research is needed to examine the ability of menstrual cycle phase to moderate the effects of GLP-1R pharmacotherapies on weight loss. Furthermore, our results in ovariectomized rats point to the requirement of the estrous cycle for increased Gcg expression, underscoring a need to compare the efficacy of GLP-1 agonists in both pre- and postmenopausal women.

It is also notable that estradiol itself enhances the fluid- and food-satiating properties of GLP-1 agonists (23,58,66,67). In the current study, 17β-estradiol increased Glp1r and Gcg transcripts in male rats, revealing a possible mechanism for this enhancement. Although our experiments in females do not directly identify fluctuations in estradiol as the cause of estrous cycle–dependent SSS, our collective data in males indicate that estradiol is one of the sex hormones that contributes to cycle-dependent SSS.

Our reanalysis of rat NTS/AP single-nucleus transcriptomics data clearly defines a Gcg+ neuronal population that coexpresses Esr1, the messenger RNA for Erα. Hindbrain ERα is both important for and sufficient to mediate the anorexigenic effects of estradiol (68,69), and ERα signaling is specifically necessary for the actions of GLP-1 on food-reward behavior in both sexes (58). Although ERα-expressing cells and PPG neurons in the caudal NTS are activated in response to ingestion of a meal (57,70), these data are the first to show that NTS Gcg neurons coexpress Esr1. It is intriguing that Gcg but not Glp1r clusters coexpressed Esr1 considering that NTS/AP Glp1r expression was increased during proestrus-to-estrus, while Gcg transcripts were increased during diestrus-to-proestrus and proestrus-to-estrus. It is possible that Glp1r expression changes may occur via increased receptor translation/trafficking in the face of increased endogenous GLP-1 input or indirectly through another mechanism.

Finally, our data revealing that SSS in female rats is estrous cycle dependent were initially surprising because female rats (71–73) and humans demonstrate SSS (74). However, clinical studies have frequently detected changes in motivation, food cue reactivity, and liking across the menstrual cycle (75–78). In short, we believe that our initial SSS test in female rats was not significantly different because of fortunate happenstance; in our hands, nearly 30% of cycling Sprague Dawley rats had an extended proestrus-to-estrus. This overrepresentation likely drove our initial null results. Studies directly comparing male and female SSS behaviors were likely underpowered to detect estrous cycle–dependent differences in females. Moreover, the SSS literature using solely female rodents is surprisingly sparse, and those studies used different dependent measures such as taste reactivity (71), preference change (72), or did not use 2 equally preferred foods (72,73), which may decrease the chance of detecting small differences in SSS in females. Similarly, without explicitly investigating SSS across the menstrual cycle, studies with humans are unlikely to detect phase-specific differences. Ultimately, our study highlights a need to examine SSS across the menstrual cycle.

Conclusions

Our data contribute to a growing body of research that has identified a variety of behaviors under the control of brainstem GLP-1–GLP-1R signaling and demonstrate the ability of the estrous cycle to modulate GLP-1–controlled SSS. Considering that many patients receiving GLP-1R–based pharmacotherapies are females who are experiencing a regular menstrual cycle, our findings emphasize the significant translational value in investigating the impact of the estrous cycle on the therapeutic effects of GLP-1 analogs.

Supplementary Material

Supplementary material cited in this article is available online at https://doi.org/10.1016/j.biopsych.2025.01.012.

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.
Chemical Compound or Drug	Sterile Saline Solution 0.9% 1000 ml Bottle	Midwest Veterinary Supply	193.74504.3
Chemical Compound or Drug	Crystal violet powder	Sigma-Adrich	C0775
Chemical Compound or Drug	Sesame oil	Sigma-Adrich	S3547
Chemical Compound or Drug	17β-Estradiol	Bio-Techne	2824
Chemical Compound or Drug	artificial cerebral spinal fluid (aCSF)	Bio-Techne	3525
Chemical Compound or Drug	Exendin-4	Bachem	4019602
Chemical Compound or Drug	Exendin-9-39	Cayman Chemical	19890
Chemical Compound or Drug	Ketamine	Covetrus North America	80524
Chemical Compound or Drug	Xylazine	Butler Animal Health	61035
Chemical Compound or Drug	Acepromazine	Midwest Veterinary Supply	090.00500.3
Chemical Compound or Drug	Meloxicam	Midwest Veterinary Supply	515.50000.3
Chemical Compound or Drug	Crystal Violet Solution	Fisher Scientific	S25275B
Commercial Assay Or Kit	RNAscope Multiplex Fluorescent V2	Advanced Cell Diagnostics, Inc.	323110
Commercial Assay Or Kit	Rn-Gcg FISH probe	Advanced Cell Diagnostics, Inc.	315471-C2
Commercial Assay Or Kit	Rn-Esr1 FISH probe	Advanced Cell Diagnostics, Inc.	317151-C3
Commercial Assay Or Kit	Rn-Glp1r FISH probe	Advanced Cell Diagnostics, Inc.	315221-C1
Commercial Assay Or Kit	EGFP FISH probe	Advanced Cell Diagnostics, Inc.	400281-C3
Commercial Assay Or Kit	Opal 520 Reagent Pack	Akoya Biosystems	FP1487001KT
Commercial Assay Or Kit	Opal 570 Reagent Pack	Akoya Biosystems	FP1488001KT
Commercial Assay Or Kit	Opal 690 Reagent Pack	Akoya Biosystems	FP1497001KT
Commercial Assay Or Kit	RNeasy Mini Kit	QIAGEN	74104
Commercial Assay Or Kit	iScript™ Reverse Transcription Supermix	Bio-Rad	1708841
Commercial Assay Or Kit	SsoAdvanced Universal SYBR® Green Supermix	Bio-Rad	1725272
Genetic Reagent	AAV1.hSyn.Cre.WPRE.hGH	Addgene	105553-AAV1
Genetic Reagent	Invitrogen™ Cholera Toxin Subunit B (Recombinant), Alexa Fluor™ 555 Conjugate	Invitrogen	C34776
Genetic Reagent	AAV1 CAG-FLEX-EGFP-WPRE	Addgene	51502-AAV1
Organism/Strain	Rat: Sprague-Dawley	Charles River
Other	Microinjection Syringe Pump	World Precision Instruments	UMP3T-2
Other	NanoFil 10 μL Syringe	World Precision Instruments	NANOFIL
Other	bilateral guide cannula: 26-gauge	protech International Inc
Other	33G beveled NanoFil	World Precision Instruments	NF33BV-2
Other	Antifade Mounting Medium with DAPI	Vector Laboratories, Inc
Other	Microscope slides	Superfrost Plus, Fisher Scientific	12–550-15
Other	mircopump depressed: PHD 2000	Harvard Apparatus
Other	Infinite F Plex	Tecan Life Sciences
Other	TempAssure 0.2 mL PCR Pull-Apart 8-Tube Strips, Attached Individual Dome Caps	USA Scientific	#1402–2900
Other	Fluorescence Microscope	Keyence	Model: BZ-X810
Other	Chow	LabDiet	5001
Other	Pringles BBQ Crisps	Food Service Direct	21407534
Other	RITZ Original Crackers	Amazon		Amazon.com: RITZ Fresh Stacks Original Crackers, Party Size, 23.7 oz (16 Stacks)
Software; Algorithm	GraphPad Prism 10.2.3	GraphPad Software