Insights into the Neurobiology of Weight Loss after Bariatric Surgery and GLP-1R Agonists

Abstract

Obesity and its related complications are growing in prevalence worldwide, with increasing impact to individuals and healthcare systems alike. Currently, the leading treatment approaches for effective and sustained weight loss are bariatric surgery and gut peptide therapeutics. At a high level, both treatment strategies work by hijacking gut-brain axis signaling to reduce food intake. However, we predict that each modality has distinct neuronal mechanisms that are responsible for their success and complications. This review compares the neurobiology of feeding behavior between these two weight loss strategies via a discussion of both clinical and pre-clinical data. The most compelling evidence points to signaling within the hindbrain, hypothalamus, and reward circuits contributing to weight loss. Considerations for treatment, including differing complications between the two treatment approaches, will also be discussed. Based on the data, we pose the hypothesis that these two interventions are acting via distinct mechanisms to induce weight loss. Both interventions have variable degrees of weight loss across the patient population, thus, understanding these distinct mechanisms could help drive individualized medicine to optimize weight loss.

Introduction

By the year 2030, approximately half of adults and over one third of children in America will have obesity, emphasizing the need for more diverse and affordable anti-obesity treatments. Obesity results from a positive energy balance favoring caloric intake over energy expenditure. Many factors contribute to this shifted energy balance, but the most effective weight loss strategies work by reducing caloric intake rather than increasing energy expenditure (Hall et al., 2017). Feeding behavior is governed by a complex interaction of neuronal circuits in the hindbrain, hypothalamus, and higher order brain areas. The coordinated sensing of dietary caloric and macronutrient intake by different brain areas orchestrates hunger, satiety, reward, and aversion to alter food intake and food choices.

Up until recently, bariatric surgery was the most effective strategy for significant and sustained weight loss. However, the new era of anti-obesity medications includes gut peptide analogs that either near or match the weight loss seen with bariatric surgeries. Still, individual responses to these interventions remain quite variable. Further, cost and availability limit access to bariatric surgery and anti-obesity medications. Additionally, concerns over loss of lean mass and counterregulatory responses to weight loss that contribute to weight regain limit widespread use of both weight loss strategies. Lastly, as opposed to surgery, weight loss medications require consistent use throughout the patient’s life to prevent weight regain. All this highlights the need for continued research and development of anti-obesity therapies. Since both these interventions target the gut-brain axis, it might be predicted that bariatric surgery and gut peptide therapeutics share similar mechanisms to mediate weight loss. Although there are certainly similarities in the impact on some aspects of feeding behavior, there are clearly differences in the neurobiology mediating the weight loss and complications of each therapy. This review will compare the behavioral and neuronal mechanisms that drive weight loss after bariatric surgery and gut peptide pharmaceuticals with a focus on preclinical work, and we will discuss future directions for these therapies.

History of bariatric surgery and GLP-1R agonists

During the mid-1900’s, physicians observed significant weight loss after gastrectomy to treat cancer (Baker, 2011). This sparked the development of variations of intestinal bypass surgeries to produce weight loss, but early attempts often resulted in severe malabsorption and other digestive complications (Faria, 2017). By the 1990’s, significant advancements were made to improve the safety and efficacy of bariatric surgery for weight loss, including laparoscopic strategies. Bariatric surgery is now the most effective strategy to induce significant and sustained weight loss for both adults and adolescents (Hampl et al., 2023). Although there are several types, the two most utilized forms of bariatric surgery are vertical sleeve gastrectomy (VSG) and Roux-en-Y gastric bypass (RYGB), with VSG currently being performed at over a three-fold higher frequency than RYGB (Alsuhibani et al., 2023). VSG consists of an 80% reduction in gastric volume by removing the greater curvature of the stomach, while RYGB includes the formation of a small stomach pouch that is attached to the jejunum such that ingested nutrients bypass the majority of the stomach and the upper GI tract. While variable complications accompany VSG vs. RYGB surgeries, both procedures produce similar changes in feeding behavior and long-term degrees of weight loss with RYGB having slightly greater efficacy. Accordingly, we will discuss potential mechanisms of weight loss shared between the two forms of bariatric surgery.

Glucagon like peptide-1 receptor agonists (GLP-1RA) mimic the physiology of the predominantly gut-secreted peptide, GLP-1. Gut-produced GLP-1 is secreted by a subset of enteroendocrine cells called L-cells. The dogma is that GLP-1 is secreted by the intestine and signals onto many different peripheral tissues and the central nervous system (CNS) to suppress feeding and improve glucose homeostasis. Several key studies demonstrating that the GLP-1 receptor (GLP-1R) is crucial for glucose homeostasis paved the way for development of GLP-1 based therapies for type 2 diabetes mellitus (T2DM)(Drucker, 2018; Holst, 2007). Rapid clearance from the plasma, poor stability, and enzymatic degradation by the protease dipeptidyl-peptidase-4 (DPP-4) limits the bioavailability of endogenous GLP-1, which led to the development synthetic compounds that inhibit DPP-4 or that are a longer-acting agonist for the GLP-1R. DPP-4 inhibitors have been effective at treatment of T2DM but have limited weight loss efficacy. Therefore, for obesity treatment, GLP-1RAs have been the primary therapeutic focus.

The first GLP-1RA, exendin-4, was approved in 2012 and was indeed effective for management of T2DM and produced an additional benefit of inducing weight loss. Additional GLP-1RAs have since been developed, such as liraglutide and the longer-acting semaglutide, which are now prescribed for both glucose management and weight loss. Interestingly, the longer-acting formulation, semaglutide has a greater efficacy at inducing weight loss compared to either liraglutide or exendin-4 (Rubino et al., 2022), and may be more affordable based on a cost-needed analysis (Azuri et al., 2023). Currently, clinical trials demonstrate that continued use of semaglutide over a year induces nearly 20% weight loss (Wilding et al., 2021). A newer combined therapy of GLP-1RA with a gastric inhibitory polypeptide (GIP) receptor agonist (tirzepatide) has also been approved for obesity and T2DM treatment with an even greater efficacy than semaglutide (Chakhtoura et al., 2023). This has spurred a pipeline of combination therapies for obesity treatment including a triple agonist for GLP-1R, GIPR, and the glucagon receptor.

The question we explore in this review is whether bariatric surgery and GLP-1RAs, which both take advantage of gut-brain-axis signaling, work via similar mechanisms. We contend that the different impacts on physiology, and differing complications associated with each of these interventions, suggest that they work via non-overlapping mechanisms.

Bariatric surgery: the argument against GLP-1 as a mechanism for metabolic success

By far, the most hypothesized mechanism for the success of bariatric surgery is increased GLP-1. Following a meal, plasma GLP-1 increases ten-fold after both VSG and RYGB in both clinical and preclinical models (Hutch and Sandoval, 2017). Intestinal L-cells are the primary source of GLP-1, and in mice after VSG, intestinal stem cell differentiation towards enteroendocrine cells has been found to contribute to increased GLP-1 (Kim et al., 2022). However, native GLP-1 has a half-life of less than two minutes leading many to question its ability to reach the brain via the circulation to regulate feeding. DPP-4 inhibitors also increase plasma GLP-1 two-fold without altering body weight (Reid, 2012), further demonstrating that GLP-1 is not a likely mechanism of weight loss. In contrast, surgery-induced increases in GLP-1 and peptide YY (PYY), which is also secreted from intestinal L-cells, were associated with marked decreases in fMRI-measured activation of brain regions in response to seeing high-calorie foods (Scholtz et al., 2014). Whether these changes in gut peptides and fMRI-assessed activation are a cause or consequence of weight loss remains to be determined. Certainly, both clinical and preclinical data demonstrate that administration of exendin 9–39, a potent and specific GLP-1R antagonist, impaired glucose in controls and after bariatric surgery (Chambers et al., 2011; Jørgensen et al., 2013; Salehi et al., 2011). Another study demonstrated that exendin 9–39 similarly impaired glucose responses to a mixed meal in healthy controls, and in patients that had VSG with or without T2DM remission. The authors concluded that GLP-1 was neither necessary nor sufficient to cause T2DM remission in response to VSG (Shah et al., 2022). Multiple studies using genetic mouse models that knockout (KO) GLP-1 or the GLP-1R from birth have demonstrated that bariatric surgery produces weight loss independent of GLP-1 secretion (Kim et al., 2019; Mokadem et al., 2014) or GLP-1R signaling (Wilson-Pérez et al., 2013a) indicating that GLP-1, in and of itself, is not necessary for the success of surgery. Thus, one possibility is that GLP-1 regulates acute meal processes (postprandial glucose excursions and meal patterning) but is not necessary for the long-term weight loss or remission of T2DM.

Alternative mechanisms for bariatric surgery

If GLP-1 does not mediate the long-term weight loss or improvements in glucose homeostasis after surgery, then what gut factors might be driving these changes? Some possibilities include circulating nutrients, other gut peptides, neuronal signals from the gut to the brain, or a combination of these signals. GLP-1 is not the only gut peptide that is altered by bariatric surgery. Despite similar weight loss, VSG and RYGB induce different temporal and overall changes in several gastrointestinal (GI) hormones that control feeding such as ghrelin, cholecystokinin (CCK), and GIP, which we previously reviewed (Hutch and Sandoval, 2017). Briefly, the body of literature suggest that ghrelin and GIP are increased after VSG but not RYGB. CCK is increased in humans after either surgery, but not in rodent models of VSG. It is difficult to draw conclusions based on the latter data as the assays for CCK are notoriously problematic.

In contrast, PYY (like GLP-1) is consistently increased after both VSG and RYGB (Scholtz et al., 2014). Some groups have argued that PYY is necessary for sustained weight loss and long-term improvements in glucose homeostasis, but this remains to be clearly demonstrated (Brzozowska et al., 2023; Guida et al., 2019; Guijarro et al., 2007). In contrast to this hypothesis, mice with genetic deletion of receptors for both GLP-1 and PYY gained weight normally, and successfully lost weight after RYGB (Boland et al., 2019). Thus, evidence for any single hormone as a primary driver of weight loss or the long-term improvements in glucose homeostasis after bariatric surgery is lacking, but surgery-induced increases in GLP-1 and PYY acutely impacts postprandial glucose fluctuations (Larraufie et al., 2019) and satiety. Given that many redundant mechanisms control feeding, it may be that a combination of gut-brain signals altered after bariatric surgery collectively contribute to weight loss and prevent compensatory increases in appetite to promote weight regain.

Impact of surgery and gut peptide therapeutics on feeding behavior

A clear similarity between surgical and GLP-1R-targeted therapeutics is that they primarily induce weight loss via reductions in food intake. Although it is important to note that some data demonstrate that semaglutide prevents a drop in energy expenditure seen with comparable weight loss and that RYGB specifically increases energy expenditure in mice (Gabery et al., 2020; Stylopoulos et al., 2009). With bariatric surgery there is an initial reduction in food intake that produces a negative energy balance and up to 30% weight loss. After weight loss has stabilized, changes in feeding behavior persist, such as reduced appetite, reduced cravings, increased fullness/satiety, and healthier dietary choices (reduced fat preference and increased protein intake) (Quercia et al., 2014; Søndergaard Nielsen et al., 2018). In support of this being a physiological rather than behavioral impact on feeding, animal models also demonstrate a clear impact of bariatric surgery procedures on aspects of feeding such as an alteration in meal patterns (with smaller but increased frequency of meals), and there is a shift in macronutrient preference towards lower caloric density foods (Chambers et al., 2012; Stefater et al., 2010; Wilson-Pérez et al., 2013b). Additionally, bariatric surgery decreased liking of, and fMRI signals in response to, high calorie foods (Faulconbridge et al., 2016). However, these differences in macronutrient or caloric preference may be dependent upon experimental protocol as more recent data suggest that patients do not alter what they eat but reduce how much they eat instead (Livingstone et al., 2022). A common misconception of bariatric surgery is that food intake decreases purely due to a restricted stomach size. However, if physiologically necessary, individuals and rodents can dramatically increase food intake after surgery. For instance, lactating female rats that previously had VSG have an appropriate increase in caloric consumption (Grayson et al., 2013), refeeding after prolonged fasting is similar in rats after sham vs. VSG surgeries (Stefater et al., 2010), and mice do not prefer more energy dense foods as would be expected if stomach size were the primary constraint (Chambers et al., 2012). Finally, accelerated emptying of stomach contents into the small intestine, known as gastric emptying, likely offsets the diminished gastric volume after surgery (Chambers et al., 2014). Thus, surgery-induced weight loss is likely explained by more complex mechanisms within the brain that govern feeding behavior and energy balance.

Similar to bariatric surgery, GLP-1RAs lower body weight by reducing caloric intake through changes in multiple components of feeding behavior. Across rodents, non-human primates, and humans, exogenous GLP-1 and GLP-1RAs potently reduce caloric intake by increasing satiety (Flint et al., 1998; Scott and Moran, 2007), decreasing hunger (Van Can et al., 2014), and decreasing food palatability (Eren-Yazicioglu et al., 2021). Importantly this latter point is not supported by all data and although differences in the acute vs. chronic responses to the drug and/or experimental protocol may contribute to these differences, it may be that GLP-1RA-induced regulation of feeding is more focused on increasing satiety and decreasing hunger (Cawthon et al., 2023). Semaglutide, liraglutide, and dual agonists like tirzepatide also change food preference towards lower caloric density foods (Geisler et al., 2023). Thus, both interventions potently impact multiple aspects of feeding behavior in similar ways, but via different mechanisms.

Neuronal mechanisms of weight loss

While both interventions reduce food intake and alter feeding behaviors, a key question is whether the neurobiology that drives these changes are similar between the two modalities. While the GLP-1R is the obvious target of GLP-1 mimetics, the location of the receptors targeted, and the downstream neurobiology of their activation continues to be studied. Even less understood are the neural circuits altered by bariatric surgery, where there are many hormones and physiological processes that are likely driving CNS-induced changes in feeding behavior. Comparing neural changes induced by these two interventions provides an opportunity to explore shared and distinct mechanisms and identify new treatment strategies.

The various gut factors increased after bariatric surgery and GLP-1RAs themselves have the possibility to directly regulate CNS control of feeding and/or to indirectly activate the CNS via peripheral nerve activation. A key component in understanding the neurobiology of gut-brain signaling after bariatric surgery or GLP-1RAs is the degree to which these factors or drugs penetrate the blood brain barrier (BBB). Because of this, circumventricular organs and peripheral neurons innervating the gut are thought to be key players in mediating the impact of both interventions (Adams et al., 2018; Gabery et al., 2020). CNS GLP-1R have been found to be necessary for the ability of liraglutide to reduce body weight (Sisley et al., 2014), and the caudal brainstem and hypothalamus are key initial sites of action for gut hormones and GLP-1RAs given their lack of protection or proximity to the BBB (Yamamoto et al., 2003). While some data do suggest a deeper penetrance of GLP-1RAs beyond the BBB (Adams et al., 2018; Gabery et al., 2020; Salinas et al., 2018), the data also indicate that non-GLP-1R neurons are activated presumably downstream of GLP-1R-expressing neurons. Bypassing the BBB, one paper chronically administered exendin 9–39 directly into the lateral ventricle of the brain and this stimulated feeding and increased body weight similarly in RYGB and sham animals (Ye et al., 2014). Because the effects were not greater in the RYGB animals, the authors concluded that central GLP-1R signaling is not the critical mechanism for body weight loss with surgery. It is also possible that diabetes, obesity, or even gut peptides themselves may alter BBB permeability (Kaiyala et al., 2000; Samms et al., 2021), which could influence endogenous and pharmacological gut peptide penetrance. Regardless of whether it is through direct or indirect action on the CNS, we contend that the brain is critical in driving weight loss in response to both surgical and pharmaceutical interventions. In this section we will compare the brain regions activated by bariatric surgery and GLP-1RA.

Caudal hindbrain

The dorsal vagal complex within the caudal hindbrain is a key relay center connecting the gut and brain in energy balance control. The dorsal vagal complex is comprised of the area postrema (AP), the nucleus of the solitary tract (NTS), and the dorsal motor vagus (DMV). The AP resides outside of the BBB and is thus exposed to systemic circulation making it conceivable that endogenous gut peptides and analogs act on this region. The NTS receives inputs from spinal and vagal sensory neurons, as well as other hindbrain regions, rapidly integrating meal-related cues to reflexively regulate short-term feeding behavior. Both GLP-1 and GLP-1RAs have been found to robustly activate neurons within the AP leading to increased satiety, aversion, nausea, and GI symptoms (Zhang et al., 2021). Further, rodent studies that peripherally administered a fluorescently labeled liraglutide or semaglutide found that the labels accumulated in the caudal hindbrain. The AP and NTS were key initial sites of GLP-1 action and this led to the engagement of a neuronal circuit that includes non-GLP-1R expressing neurons within the parabrachial nucleus, central amygdala, and bed nuclei of the stria terminalis (Gabery et al., 2020). However, brain penetrance and activation patterns differed between liraglutide and semaglutide, (Gabery et al., 2020; Salinas et al., 2018), which could explain the higher weight loss efficacy with semaglutide. Glutamatergic GLP-1R+ neurons were necessary for liraglutide to reduce body weight in mice (Adams et al., 2018), however, inhibition of GLP-1R+ GABAergic neurons specifically within the NTS of rats attenuated liraglutide-induced weight loss (Fortin et al., 2020). It is unclear whether these discrepancies are due to differences in technical strategies (genetic mouse models with developmental manipulation of the GLP-1R vs. chemogenetic inhibition of GLP-1R in adult rodents) or the species studied (mouse vs. rat respectively). Interestingly, GIPR agonists also targetted neurons with an inhibitory neurotransmitter, GABA, within the AP to reduce food intake (Liskiewicz et al., 2023). Thus, the greater potency of tirzepetide could be driven by targeting both glutamatergic GLP-1R expressing neurons and GABAergic GIPR expressing neurons.

Bariatric surgery also impacts caudal hindbrain signaling. After bariatric surgery, rodents have increased activation of neurons within the AP and NTS in response to nutrients compared to obese sham surgery animals (Chambers et al., 2012). We have recently found that silencing neuronal activity of calcitonin receptor, leptin receptor (LepR), and CCK neurons within the NTS induced fat-mass gain in mice, regardless of whether the animals were on chow or high fat diet (Qiu et al., 2023). Despite their large role in regulating body weight, these neurons comprise only 30% of those nutrient-activated neurons within the NTS after VSG, and silencing these neurons did not prevent weight loss from VSG (Qiu et al., 2023). These data suggest that additional populations of NTS neurons are nutrient-sensing after VSG (via direct or indirect mechanisms), but it remains unknown whether these activated neurons are critical for the success of surgery. Catecholaminergic and prolactin-releasing hormone NTS neurons have been found to mediate hunger (Aklan et al., 2020; Sayar-Atasoy et al., 2023) and mechanosensation of ingested food (Ly et al., 2023), whereas preproglucagon NTS neurons have been found to encode alimentary nutrient signals that regulate feeding (Holt et al., 2019). Given the overlap of preproglucagon and LepR populations (Cheng et al., 2020), future efforts may be best focused on determining if catecholaminergic NTS neurons contribute to weight loss mediated by bariatric surgery. Although rodent data demonstrate that GLP-1R signaling was not necessary for the success of surgery, neurons that express the GLP-1R could still be activated by VSG via other stimulatory mechanisms. This remains to be tested, but there is some overlap in GLP-1R-expressing and the calcitonin receptor and CCK-expressing neurons within the NTS (Ludwig et al., 2021), which we found did not contribute to weight loss (Qiu et al., 2023). Understanding how various neuronal populations within the hindbrain respond to gut signals postbariatric surgery could provide mechanistic insights into the drivers of weight loss and potentially uncover new therapeutic targets.

Vagus nerve

Vagal afferent neurons innervating the GI tract are well positioned to detect local changes in gut peptides and relay these signals to the caudal hindbrain. Indeed, GLP-1 activates vagal neurons, and chemogenetic or optogenetic activation of vagal GLP-1R+ neurons robustly decreases food intake, even in fasted animals (Bai et al., 2019; Borgmann et al., 2021). Thus, the vagus nerve is an intriguing locus of action for hormonal changes after bariatric surgery and gut peptide therapeutics. However, vagal GLP-1R expression is not necessary for liraglutide-induced weight loss in mice (Sisley et al., 2014), but partially necessary for the reduced weight loss from another GLP-1RA, dulaglutide (Varin et al., 2019). The genetic crosses for these studies were similar (GLP-1R floxed mice crossed with Phox2b-Cre mice that targets vagal and hindbrain GLP-1R), so the inconsistent results could be driven by the different agonists or some other confounding factor. In a different strategy, acute inhibition of vagal-GLP-1R+ neurons using chemogenetics blunted the anorectic effect of liraglutide in mice (Borgmann et al., 2021). In rats, higher doses of exendin-4 and liraglutide were needed to reduce food intake after vagal deafferentation (Kanoski et al., 2011). Thus, more work is needed to clarify these contradictory reports, but there is evidence that vagal afferent neurons at least partially mediate reductions in food intake by GLP-1RAs.

Whether or not vagal neurons contribute to weight loss after bariatric surgery is unclear. Generally, vagal deafferentation blunted the acute satiety effect of bariatric surgery, but did not mediate its ability to induce long term weight loss (Hao et al., 2014). However, the use of surgical or chemical vagotomy to test the necessity of vagal neurons in weight loss is difficult to interpret. Vagotomy itself paradoxically leads to central compensation that results in similar benefits as bariatric surgery: enhanced satiety and altered meal patterning (Gautron, 2021; Hao et al., 2014). As previously mentioned, multiple parallel and redundant mechanisms regulate long-term regulation of feeding and energy balance, and it appears that compensatory feeding circuits are engaged after disruption of the vagus nerve. Like nuclei within specific regions of the CNS, sensory neurons in the vagus nerve have been found to cluster into diverse populations with different regulatory actions (Bai et al., 2019; Egerod et al., 2018; Kupari et al., 2019; Williams et al., 2016). Thus, future studies utilizing more acute and targeted manipulations of specific vagal afferent populations will likely improve our understanding of the ways in which vagal neurons contribute to the benefits of bariatric surgery and/or gut peptide therapeutics. For instance, viral nodose injections and dual Cre/Dre targeting systems have suggested that GLP-1R+ vagal afferents are important in in regulating satiety and glucose homeostasis (Bai et al., 2019; Borgmann et al., 2021).

Hypothalamus

Complex interactions between millions of neurons in different hypothalamic nuclei are essential for governing energy balance. Genome-wide association studies (GWAS) implicate genes important for hypothalamic control of energy balance in the pathogenesis of monogenetic obesity (Loos and Yeo, 2022), and environmental contributions to obesity clearly impact hypothalamic signaling (Berthoud et al., 2017; Thaler et al., 2012). Thus, it is no surprise that many lines of evidence support altered hypothalamic signaling as a key mechanism behind weight loss interventions.

The GLP-1R is expressed in the arcuate nucleus of the hypothalamus, and GLP-1RAs have been found to reach this region of the brain with some subtle differences in penetrance between semaglutide and liraglutide (Gabery et al., 2020). However, while chronic exendin-4 or liraglutide administration induced weight loss in lean hypothalamic GLP-1R KO animals, this effect was blunted in dietary induced obese mice suggesting an interaction of diet and GLP-1R function in the brain (Burmeister et al., 2017a). Additionally, mice with genetic deletion of the GLP-1R in the paraventricular hypothalamus (Ghosal et al., 2017), or the ventromedial nucleus (Burmeister et al., 2017b), had the expected suppression of feeding in response to peripherally administered GLP-1RAs. These data suggest that whole regions of the hypothalamus may not be required for the ability of GLP-1RAs to induce weight loss but does not rule out specific neuronal populations within each region.

Within the arcuate nucleus of the hypothalamus, the melanocortin system is critical for body weight regulation. In this system, the melanocortin 4 receptor (MC4R) signaling is modulated by peptides released from pro-opiomelanocortin (POMC) and agouti-related peptide (AGRP) neurons, to reduce or increase feeding, respectively. GLP-1RAs have been found to inhibit the activity of orexigenic AGRP neurons and activate anorexigenic POMC neurons (Secher et al., 2014). Furthermore, semaglutide increased expression of cocaine-amphetamine regulated transcript, and suppressed the upregulation of orexigenic peptides neuropeptide Y (NPY) and AGRP that normally follows weight loss (Gabery et al., 2020). However, a genetic KO of GLP-1R in POMC neurons still induced weight loss in response to exendin-4 or liraglutide (Burmeister et al., 2017a).

Various lines of evidence demonstrate that alterations in hypothalamic signaling mediate weight loss after bariatric surgery, as well. For instance, patients with obesity caused by tumor-induced hypothalamic damage do not lose as much weight after bariatric surgery (Dischinger et al., 2023). In a case report of a family with monogenetic forms of obesity due to a homozygous MC4R mutation, bariatric surgery failed to produce sustained weight loss in three out of the four children (Grinbaum et al., 2022). In line with this, mice with genetic deletion of MC4R lost less weight after RYGB (Hatoum et al., 2012). However, in another study, VSG was equally effective in rats with MC4R deficiency and in humans with MC4R mutations (Mul et al., 2012). It is possible that RYGB uniquely modulates melanocortin signaling to produce weight loss, and this should be further investigated. However, reports of the impact of POMC, NPY, and AGRP neurons on surgical outcome are inconsistent (Barkholt et al., 2016; Frank et al., 2016; Liu et al., 2017). Overall, more work is needed to substantiate the link between altered melanocortin signaling and weight loss by bariatric surgery.

Reward circuits

The limbic (or mesolimbic) system is the primary reward circuit comprised of several dopaminergic pathways that aim to promote positive behaviors, such as mating and feeding. Neurons in the ventral tegmental area (VTA) project to the striatum, amygdala, hippocampus, and prefrontal cortex. These connections are critical for establishing and maintaining feeding habits, and genetic and environmental disruptions to the limbic system are suggested to be major contributors to the obesity pandemic (Berthoud et al., 2017). For instance, highly palatable and calorically dense food are more readily available than ever before (Swinburn et al., 2011), and these foods engage hedonic circuits that promote feeding in the absence of hunger, overpowering satiation signals that would normally limit excess caloric intake (Hall et al., 2014). fMRI studies showed that individuals with obesity exhibit stronger reward activation to food cues, but reward responses to actual consumption of the food was reduced (Stice et al., 2008), thus increasing drive to eat while diminishing the pleasure received by eating. Another example of this is that individuals with obesity exhibit blunted striatal dopamine release in response to intragastric nutrients, a mode of nutrient ingestion that bypasses taste (Van Galen et al., 2023).

A critical question is whether bariatric surgery or GLP-1RAs modulate these rewards systems to induce weight loss. Interestingly, weight loss induced by caloric restriction, bariatric surgery, or gut peptide analogs all reduce hedonic hunger (Aukan et al., 2022; Blundell et al., 2017). However bariatric surgery and GLP-1RAs appear more effective in limiting compensatory increases in appetite that drive weight regain (Hall, 2023) (Figure 1). For instance, diet-induced weight loss fails to restore striatal dopamine signaling in response to intragastric infusion of nutrients (Van Galen et al., 2023) while semaglutide has been found to enhance dopaminergic neuron activity in the VTA (Kooij et al., 2023). Similarly, bariatric surgery restored obesity-induced disruptions in VTA dopamine and opioid signaling (Faulconbridge et al., 2016; Karlsson et al., 2021). Thus, meal-related reward signaling is disrupted in obesity and cannot be restored simply through dietary weight loss, while GLP-1RAs and bariatric surgery may increase diet adherence by restoring nutrient-induced reward, allowing some individuals to avoid weight regain. However, whether the neurocircuitry and molecular mechanisms that mediate these changes to the reward system are similar between these two interventions remains to be seen.

Figure 1: Central mechanisms mediating sustained weight loss after bariatric surgery and gut peptide therapeutics.

Bariatric surgery (gold) and GLP-1 receptor agonists (GLP-1RAs; blue) work via GLP-1R independent and dependent mechanisms, respectively, to alter multiple dynamics of feeding behavior to reduce food intake. Reduced food intake and subsequent weight loss is often met by compensatory increases in appetite, which are mitigated by these weight loss strategies to prevent weight regain and sustain weight loss maintenance. These weight loss therapies lead to distinct complications such as hypoglycemia and increased substance abuse after bariatric surgery and GLP-1RA-induced visceral illness.

Differing complications

Given that gut peptide therapeutics and bariatric surgery are very different approaches for weight loss, it is not surprising that they cause different complications. Some of these issues will be discussed here.

Nausea and conditioned taste aversion

Common side effects of GLP-1RAs are nausea, vomiting, and anorexia. For GLP-1RAs, nausea presents a barrier to patient compliance for many individuals. Although some hypothesize that nausea drives reduced food intake and consequently weight loss with GLP-1RAs, the nausea wanes over time and weight loss persists. Further, there are clear regions of the CNS where administration of GLP-1 inhibited food intake without conditioned taste aversion and vice versa (Kinzig et al., 2002). Rodent models suggest that GLP-1R activation in the AP is a particular region that mediates nausea associated with GLP-1RA (Zhang et al., 2021), and interestingly, this is prevented with a dual GLP-1/GIPR agonist (Borner et al., 2021); an effect that may be mediated via GABAergic neurons in AP (Zhang et al., 2021; Zhang et al., 2022). However, other rodent studies with GLP-1 administered directly into the CNS, demonstrated that GLP-1R within the central amygdala also induced conditioned taste aversion (Kinzig et al., 2002). Generally, deletion of GLP-1R within all glutamatergic neurons prevents aversion induced by liraglutide (Adams et al., 2018). Interestingly, women have a higher frequency of adverse events and particularly GI-related adverse events with gut peptide therapy compared to men (Rentzeperi et al., 2022), which requires further mechanistic study.

Unlike with GLP-1RAs, chronic nausea due to conditioned aversion is not a side effect of bariatric surgery. Some patients experience dumping syndrome, which is a cluster of symptoms including nausea, diarrhea, fatigue, and light-headedness (Salehi et al., 2018). but this is likely caused by the increased gastric emptying rate after surgery. In contrast, gastric emptying rate is slowed by GLP-1RA. The key takeaway is that nausea is a side effect of GLP-1RA that distinguishes it from the impact of bariatric surgery.

Substance use disorders

In 2013, a concerning study found that individuals were at an increased risk of alcohol use disorder after RYGB (Conason et al., 2013). One possibility is that bariatric surgery causes individuals to shift from compulsive eating to other addictive behaviors such as alcohol consumption (Blum, 2013). Another is that there is aberrant physiological signaling from the gut to the brain driving increased incidence of alcohol use disorder. However, no such shift has been seen with GLP-1RAs. Intriguingly, GLP-1RAs have been shown to have the opposite effect by reducing alcohol consumption (Aranäs et al., 2023; Quddos et al., 2023). Rodent and human studies suggest that GLP-1 attenuates rewarding aspects of alcohol by reducing striatal dopamine release (Aranäs et al., 2023; Jerlhag, 2020; Klausen et al., 2022). This difference in risk for substance use is particularly interesting given that both weight loss strategies apparently work in part by modulating dopaminergic reward systems and may be a relevant factor when making clinical decisions based on genetic and familial predispositions. For this review, we highlight these data as another example of how bariatric surgery and GLP-1RA have distinct effects on gut-brain signaling. However, these also highlight the need for continued understanding of how GLP-1RAs and bariatric surgery differentially impact alcohol consumption, and how dopaminergic reward pathways may mediate these effects are critical future steps.

Postbariatric hypoglycemia

While there is a small increase in hypoglycemia incidence with GLP-1RAs, approximately one third of individuals experience bouts of hypoglycemia, particularly postprandially, after both RYGB and VSG. Surgery-induced increases in GLP-1 have been suggested to partially contribute to hypoglycemia (Salehi et al., 2018). However other factors such as increased insulin sensitivity, blunted counterregulatory responses, genetic predisposition, as well as dysregulation of other hormones such as glucagon and fibroblast growth factor 19 (Patti and Goldfine, 2014) likely contribute. Intriguingly, one study found that after RYGB, postsurgical hypothalamic glucose uptake was reduced during hypoglycemia as compared to presurgical values, which may contribute to impaired counterregulation (Almby et al., 2021). Future work should assess how bariatric surgery alters CNS nutrient sensing and whether this directly blunts the initiation of counterregulatory responses to hypoglycemia. Currently, there are no treatments for postbariatric hypoglycemia, however a more stable glucagon analog, dasiglucagon (Nielsen et al., 2022) and a long-acting GLP-1R antagonist (Craig et al., 2021) have both shown promise.

Altogether, the body of literature suggests that bariatric surgery and GLP-1RAs both induce weight loss primarily by reducing food intake. Similarities and differences between the two interventions are summarized in Figure 1. Preclinical and clinical data suggest the brain drives the changes in feeding and reduction in body weight with both interventions. Importantly, for most patients, the compensatory increase in appetite that occurs with diet-induced weight loss is mitigated by both interventions. Although postprandial GLP-1 increases 10-fold, genetic mouse models and pharmacological studies do not support that this increase is necessary for the success of surgery. Additionally, differing complications between each intervention further supports distinct mechanisms of action.

Other considerations with these treatments

Along with the discussion of the neuronal mechanisms underlying current treatments for obesity, it is important to contextualize whether these data could inform personalized therapies. While a deep dive on this subject is beyond the scope of this review, two important personalized factors to mention are sex and age. Despite no difference in obesity rates between men and women in the United States (Hales et al., 2017), a disproportionate number of bariatric surgery patients are female (~80%) (Benoit et al., 2014). Early data in the bariatric surgery field suggested that there were sex differences in post-operative weight loss, but reports were conflicting (Khaykis et al., 2007; Tymitz et al., 2007) and were likely influenced by a higher frequency and greater severity of obesity-associated medical problems in men undergoing bariatric procedures compared to women (van Olst et al., 2023). Controlling for baseline variability, a recent study found that men and women lost similar amounts of weight at 12, 24, and 36 months postoperatively (Mousapour et al., 2021). Recent meta-analysis data also suggest that resolution of co-morbidities was similar between sexes (Risi et al., 2022). Regardless, preclinical works shows a synergistic interaction between estradiol and satiation from GLP-1 and CCK, increasing weight loss after RYGB (Asarian et al., 2012). Future work dissecting these interactions of sex hormones and gut peptides is of high interest. In contrast to surgery, consistently across several clinical trials, female sex has been identified as a baseline predictor for greater total weight loss with GLP-1RA (exenatide (Buysschaert et al., 2010), liraglutide (Wilding et al., 2016), and semaglutide (Jensterle et al., 2023). Further work is needed to understand the underlying mechanisms of sex differences in the response to these drugs and additionally should evaluate any role of sex hormones that may impact treatment considerations during reproductive age, perimenopause, and following menopause.

When considering obesity treatment by age, both bariatric surgery and GLP-1RAs are safe and effective tools for the treatment of obesity in children and adults (Elmaleh-Sachs et al., 2023; Hampl et al., 2023; Telci Caklili et al., 2023). In terms of efficacy, when comparing similar trials, placebo-subtracted change in body weight with semaglutide was similar between children and adults with similar safety profiles (Weghuber et al., 2022; Wilding et al., 2021). For bariatric surgery, while percent weight change five years after surgery was similar between adolescents and adults (−26%, −29% respectively), adolescents had a higher probability of remission of cardiometabolic comorbidities (T2D, hypertension) compared to adults (Inge et al., 2019).

Lastly, most of the work we have reviewed here has focused on preclinical models. Despite important species differences, we feel the data on animal models provide critical insights into the neurobiology of the responses to these interventions, something that will inform more targeted clinical studies.

Conclusion

In conclusion, the global rise in obesity emphasizes the urgent need for diverse and affordable anti-obesity treatments. Bariatric surgery and gut peptide therapeutics prove highly effective for sustained weight loss by influencing multiple aspects of feeding behavior. By activating neuronal circuits in the hindbrain, hypothalamus, and reward centers, these therapies prevent compensatory increases in food intake that are normally engaged after weight loss and allow individuals with obesity to maintain a lower body weight. Despite engaging similar downstream circuits to control feeding, these two weight loss therapies work via distinct initial mechanisms. Mono and dual-agonists (with other poly agonists in the pipeline showing promise as well) are highly effective for weight loss, but bariatric surgery has wider ranging physiological effects and is thus essentially a compound intervention. Continued research to deepen our understanding of neurobiological mechanisms mediating weight loss will be instrumental in refining and developing more effective, personalized anti-obesity therapies for diverse populations. Additionally, more work is needed to consider how sex and age may alter the efficacy of different weight loss strategies and inform clinical decisions.

Highlights.

• Bariatric surgery and GLP-1 agonists target the gut-brain axis.

• The brain is critical for mediating decreases in feeding and weight loss with both interventions.

• These two strategies work via distinct mechanisms to drive weight loss.