Role of the vagus nerve in the development and treatment of diet‐induced obesity

Abstract

This review highlights evidence for a role of the vagus nerve in the development of obesity and how targeting the vagus nerve with neuromodulation or pharmacology can be used as a therapeutic treatment of obesity. The vagus nerve innervating the gut plays an important role in controlling metabolism. It communicates peripheral information about the volume and type of nutrients between the gut and the brain. Depending on the nutritional status, vagal afferent neurons express two different neurochemical phenotypes that can inhibit or stimulate food intake. Chronic ingestion of calorie‐rich diets reduces sensitivity of vagal afferent neurons to peripheral signals and their constitutive expression of orexigenic receptors and neuropeptides. This disruption of vagal afferent signalling is sufficient to drive hyperphagia and obesity. Furthermore neuromodulation of the vagus nerve can be used in the treatment of obesity. Although the mechanisms are poorly understood, vagal nerve stimulation prevents weight gain in response to a high‐fat diet. In small clinical studies, in patients with depression or epilepsy, vagal nerve stimulation has been demonstrated to promote weight loss. Vagal blockade, which inhibits the vagus nerve, results in significant weight loss. Vagal blockade is proposed to inhibit aberrant orexigenic signals arising in obesity as a putative mechanism of vagal blockade‐induced weight loss. Approaches and molecular targets to develop future pharmacotherapy targeted to the vagus nerve for the treatment of obesity are proposed. In conclusion there is strong evidence that the vagus nerve is involved in the development of obesity and it is proving to be an attractive target for the treatment of obesity.

Abbreviations

BAT

brown adipose tissue

BMI

body mass index

CCK

cholecystokinin

FDA

Food and Drug Administration

GLP1

glucagon‐like peptide 1

KHFAC

kilohertz frequency alternating current

MMC

migrating motor complex

Nav1.8

voltage gated sodium channel 1.8

NTS

nucleus tractus solitarii

OEA

oleoylethanolamide

PPAR‐α

peroxisome proliferator‐activated receptor α

VBLOC

vagal blockade

VNS

vagal nerve stimulation

Anatomy of the vagus nerve

The vagus nerve is one of 12 pairs of cranial nerves. It is a complex mixed nerve comprising three fibre types: highly myelinated, low activation large‐diameter A‐fibres; lightly myelinated intermediate‐diameter B‐fibres; and unmyelinated, small‐diameter, high activation thresholds C‐fibres (Woodbury & Woodbury, 1990). A‐fibres can be further broken down into four subgroups (α to δ) with slightly different diameters, myelination and conduction velocities. Vagal fibres provide bidirectional information between the brain and peripheral organs. The majority of the fibres are vagal afferent fibres that carry sensory information from visceral organs to the brain (predominantly Aδ‐ and C‐fibres); the remainder are vagal efferent fibres that convey motor information to control peripheral organ function. Anatomical studies along with pharmacological and surgical techniques have provided evidence that the vagus nerve innervates most of the organs in the thorax and abdomen and plays a role in modulating the physiological functions associated with these organs (Fig. 1).

Figure 1. The vagus nerve enables bidirectional communication between peripheral organs and the brain .

The vagus nerve innervates a plethora of peripheral organs; sensory information is communicated from each of these organs to the brain via vagal afferent fibres (red arrows), and the brain controls motor function of each organ by signalling through vagal efferent fibres (black arrows). As is indicated by the thicker red line, vagal afferent fibres outnumber vagal efferent fibres. The vagus nerve has been implicated in mediating an extensive range of physiological functions. Highlighted are some of the different functions that are attributable to the vagus nerve innervating different peripheral organs, although not all have been well characterized. This figure is not meant to represent an exhaustive list of all the functions of the vagus nerve, but rather to emphasize the complexity of vagal innervation and provide a flavour of the extensive functions the vagus nerve is involved in regulating.

Afferent fibres vastly outnumber efferent fibres within the vagus nerve (Berthoud & Neuhuber, 2000). These sensory and motor fibres branch at the level of the jugular foramen; the thicker afferent rootlets terminate in the nucleus tractus solitarii (NTS) and lie underneath the efferent rootlets that project from their cell bodies in the dorsal motor nucleus and the nucleus ambiguous (Norgren & Smith, 1994). As the vagus nerve leaves the skull, the superior (jugular) and inferior (nodose) ganglia contain the afferent cell bodies. Since all vagal afferent fibres originate from these ganglia, and all vagal efferent fibres originate from the hindbrain, the cervical trunk of the vagus nerve consists of fibres innervating all organs of the body. Neural tracing studies demonstrated that branches leaving the cervical vagus in the thorax innervate the skin, acoustic meatus, pharynx, larynx, trachea, bronchi, lung, aortic arch, heart and oesophagus (Berthoud & Neuhuber, 2000). As the ventral (left) and dorsal (right) vagal trunks enter into the abdomen through the diaphragm, they divide into multiple branches with slightly different organ innervation patterns (Berthoud & Neuhuber, 2000). Both the ventral and dorsal gastric branches innervate the stomach and proximal duodenum; the dorsal accessory celiac branch has weak innervation to the liver but predominantly joins the ventral celiac branch innervating the distal duodenum, jejunum, ileum, cecum and colon (Berthoud & Neuhuber, 2000). The common hepatic branch exclusively derives from the ventral trunk. As its name suggests this branch supplies the hepatic triads, bile duct and portal vein in the liver, but also branches off to innervate lower parts of the stomach, pyloric sphincter, pancreas and proximal duodenum (Berthoud & Neuhuber, 2000). Thus, while branches that leave the cervical vagal trunks predominantly supply organs associated with cardiorespiratory function, the branches that leave the subdiaphragmatic vagal trunks innervate organs associated with metabolism.

Role of the vagus nerve in metabolism

Vagal afferent sensing from the gut in the control of food intake

Despite the blunt tools available to distinguish the function of afferent and efferent signalling to different organs, it is clear that subdiaphragmatic vagal afferent fibres provide negative‐feedback signals in the control of meal size. Within the gastrointestinal tract, mechanosensitive vagal afferent terminals suppress meal size in response to distension, and chemosensitive vagal afferent terminals suppress meal size in response to nutrient type and quantity (Ritter, 2004) (Fig. 2). Anterograde tracers injected into the nodose ganglia have identified multiple different types of vagal afferent terminals in the gastrointestinal tract that support the conclusion that the stomach is predominantly involved in meal termination as a result of mechanically sensing distension, while vagal afferent terminals in the proximal small intestine are involved in chemosensitively induced satiation (Berthoud, 2008 a).

Figure 2. The role of the vagus nerve in gut–brain signalling .

In the presence of food, enteroendocrine cells will release anorectic hormones that inhibit food intake. In the absence of food, different enteroendocrine cells of the gut release orexigenic hormones that stimulate food intake. Vagal afferents neurons (red), located in the nodose ganglia, express chemoreceptors on their terminals in the gut that sense these hormones and mechanoreceptors that sense distension. These satiating signals are conveyed to the nucleus tractus solitarii (NTS) in the hindbrain. NTS neurons (1) are activated and reduce meal size and duration, (2) signal to higher order neurons in the forebrain to regulate reward or energy homeostasis, and/or (3) signal to the vagal efferent neurons in the dorsal motor nucleus (DMN) in the hindbrain. Vagal efferent fibres (blue) activate stimulatory or inhibitory postganglionic neurons to control digestion and absorption.

Vagal afferent innervation of the stomach in satiation

In the stomach, the largest proportion of vagal afferent terminals are intraganglionic laminar endings that predominantly innervate the myenteric plexus, and intramuscular arrays that project to the circular and longitudinal muscle layers (Berthoud, 2008 a). Their location makes them ideally situated to detect contraction and distension. Their response to tension and stress was confirmed using a combination of electrophysiology, pharmacology and tracing experiments in vivo (Zagorodnyuk et al. 2001). Another less abundant vagal afferent fibre in the stomach terminates in the gastric mucosa. These are insensitive to stretch and tension, but respond to stroking (Page et al. 2002) and at least a proportion respond to pH changes (Clarke & Davison, 1978). The exact role of these gastric mucosal afferent terminals is unclear, but they have been implicated in signalling high concentrations of gastric acid to the brainstem (Michl et al. 2001). Furthermore their position within the villi in close apposition to the lamina propria situates them ideally to respond to hormones released from enteroendocrine cells of the stomach. Gastric hormones can be broken down into two categories: (1) those that regulate gastric acid secretion (histamine, gastrin and somatostatin) and (2) those involved in appetite control (ghrelin, nesfatin‐1 and leptin). Ghrelin is released in the absence of nutrients in the stomach (Al Massadi et al. 2010) and is thought to promote food intake by preventing negative feedback from the gut. Ghrelin inhibits vagally mediated satiating effects of cholecystokinin (CCK; de Lartigue et al. 2007; de Lartigue et al. 2010 b) and reduces vagal afferent sensitivity to gastric tension (Kentish et al. 2012). Gastric leptin is released by food intake (Cammisotto & Bendayan, 2012), and is thought to regulate meal size and duration by increasing the sensitivity of vagal afferent neurons to CCK (de Lartigue et al. 2010 b) and mucosal stroking (Kentish et al. 2013).

In addition to the anatomical and electrophysiological data, there are physiological and behavioural data supporting the concept that the stomach provides negative feedback on food intake based on the volume rather than the type of food ingested. Gastric loads volume‐dependently (1) activate mechanosensitive vagal afferent fibres (Schwartz et al. 1991), (2) increase neuronal activity in the nucleus of the solitary tract where vagal afferents terminate (Willing & Berthoud, 1997), and (3) suppress meal size (Phillips & Powley, 1996). Mechanosensitive vagal afferent fibre firing increased dose‐dependently irrespective of whether saline, carbohydrate, protein or fat was infused (Mathis et al. 1998).

Vagal afferent innervation of the small intestine in satiation

The full array of vagal afferent terminals can also be found in the small intestine, although the intramuscular terminals are scarce (Berthoud, 2008 a). Intraganglionic laminar endings that terminate within the myenteric plexus (Berthoud et al. 1995) presumably have similar function to those described terminating in the stomach. The myenteric plexus can regulate motility independently of central input, but it may be that these vagal afferent terminals provide information to the brain about the contractile pattern of the gut to coordinate motility. The mucosal terminals are most abundant in the proximal duodenum, and become less profuse in the distal small intestine (Berthoud, 2008 a). These fibres form branches that innervate crypts and villi in the submucosa, and terminals have been demonstrated in close proximity to enteroendocrine cells in the epithelium (Berthoud et al. 1995; Berthoud & Patterson, 1996). Receptors for many micronutrients are expressed by vagal afferent neurons (Duca et al. 2013 a), but the terminals do not cross into the lumen, suggesting that they do not play a role in sensing luminal content. Instead, enteroendocrine cells are capable of sensing nutrients in the lumen of the gut and release hormones into the lamina propria that activate vagal afferent terminals (Raybould, 2010). Initial studies infusing nutrients directly into the duodenum resulted in activation of vagal afferent fibres with different nutrients producing different patterns of activity (Schwartz & Moran, 1998). Subsequently, these responses to luminal nutrient infusion were demonstrated to be dependent on the release of hormones from enteroendocrine cells (Lal et al. 2001; Williams et al. 2009).

Different enteroendocrine cells express receptors for different nutrients and this produces a complex pattern of neuroendocrine signalling that can be sensed by vagal afferent terminals and provides negative feedback to the brain to initiate meal termination (Raybould, 2010) (Fig. 2). The best example of a gastrointestinal hormone acting via this pathway is CCK. It is released from enteroendocrine I cells in response to proteins or fat, and activates CCKA receptors expressed on vagal afferent terminals in the gut (Dockray, 2009). Exogenous CCK dose dependently activates duodenal and gastric vagal afferent neurons, activates neurons of the nucleus of the solitary tract, promotes satiation, slows motility, and releases gastric acids, pancreatic enzymes and bile involved in digestion (Dockray, 2009). All of these effects are abolished by surgical and chemical lesioning of the vagus nerve (MacLean, 1985; Le Sauter et al. 1988; Raybould & Tache, 1988).

To date 33 gastrointestinal hormones have been identified and demonstrated to play a role in digestion, absorption, and/or satiation (for review see Rehfeld, 2014). Recently, cytoplasmic processes from enteroendocrine cells in the small and large intestine have been demonstrated to project towards, and transsynaptically connect with, afferent fibres (Bohorquez et al. 2015). Although the exact provenance of the sensory fibres, the speed and frequency at which this occurs in vivo, and the relative importance of this mechanism in modulating food intake are yet to be determined, the discovery provides a putative temporally precise mechanism for nutrient feedback from the gut to the brain via vagal afferents.

Vagal afferent innervation of the liver in satiation

A role for vagal afferent innervation of the liver in the control of food intake has been proposed (Yi et al. 2010). The hypothesis is that nutrients absorbed from the small intestine travel to the liver via the portal vein, and are sensed by vagal afferent fibres innervating this region. This concept is supported by anatomical evidence identifying vagal fibres terminating on hepatic triads and portal vein, which receives absorbed nutrients directly from the gut, and the innervation of the bile duct, which is important in fat digestion (Berthoud et al. 1992). Furthermore, portal infusion of nutrients (d‐glucose, l‐lysine and glutamate) has been reported to activate afferent fibres of the hepatic vagal branch in situ (Niijima, 1982, 2000; Torii & Niijima, 2001). However, interpretation of studies performing common hepatic branch vagotomy are complicated by the fact that the majority of the fibres in this branch also supply the antrum, pyloric sphincter, proximal duodenum and pancreas (Berthoud et al. 1992). Thus, it is difficult to attribute the physiological phenotype of common hepatic branch vagotomy to the vagus nerve innervating the liver (Yi et al. 2010).

Postsynaptic vagal afferent terminals in the NTS in satiation

Postprandial mechano‐ and chemosensitive vagal afferent stimulation results in expression of c‐Fos, a marker of neuronal activation, in neurons of the NTS, the site of vagal afferent fibre termination in the hindbrain (Ritter, 2004) (Fig. 2). Decerebrated rats, in which hindbrain–forebrain communication is abolished, are capable of terminating a meal (Grill & Norgren, 1978; Hayes et al. 2008), suggesting that the caudal hindbrain is sufficient to suppress food intake and gastric emptying rates. Multiple different subtypes of NTS neurons have been demonstrated to be responsive to meal‐induced signalling (Rinaman et al. 1998; Appleyard et al. 2005; Maniscalco & Rinaman, 2013) and to play a role in satiation (Rinaman et al. 1998; Barrera et al. 2011; Zhan et al. 2013). Thus, the relay of peripheral information from subdiaphragmatic vagal afferent neurons to the NTS is sufficient to regulate meal size and duration. However, anatomical and functional studies suggest that vagal afferent activation of NTS neurons can be indirectly conveyed to higher order forebrain neurons involved in energy homeostasis (Monnikes et al. 1997) and reward such as nutrient conditioned learning (Tellez et al. 2013) (Fig. 2).

Vagal afferent plasticity

Vagal afferent neurons are more than a simple relay system of gastrointestinal information to the brain: these neurons integrate information about nutritional status (Dockray & Burdyga, 2011). In response to a meal, receptors and neuropeptides associated with inhibiting food intake are expressed (de Lartigue et al. 2007, 2010 a; Burdyga et al. 2008). Fasting reduces expression of these proteins (de Lartigue et al. 2007; Burdyga et al. 2008), and promotes synthesis of receptors and neuropeptides that are associated with increasing food intake (Burdyga et al. 2006 a,2006 b, 2010) (Fig. 3). CCK is necessary and sufficient for expression of a satiety phenotype, and downregulation of the orexigenic phenotype in vagal afferent neurons (Burdyga et al. 2006 a, 2008, 2010; de Lartigue et al. 2007), although leptin potentiates and ghrelin inhibits this CCK‐induced phenotypic switch (Burdyga et al. 2006 b; de Lartigue et al. 2010 b) (Fig. 3). In cultured vagal afferent neurons, we found that the same neuron can express both phenotypes and rapidly changes expression from orexigenic to anorectic in response to CCK (de Lartigue et al. 2007). This plasticity enables vagal afferent neurons to change their sensitivity to chemical and mechanical signals depending on the feeding state of the animal, and provide nuanced information about nutrient availability to the brain (reviewed in Dockray & Burdyga, 2011; de Lartigue, 2014).

Figure 3. Plasticity in vagal afferent neurons .

Approximately 40% of vagal afferent neurons innervate the gut. These neurons express two different neurochemical phenotypes that reflect the nutrient status. In response to nutrients, there is distension of the stomach and release of the satiating hormone cholecystokinin. Circulating leptin enhances the sensitivity of vagal afferent neurons to these peripheral signals, promotes vagal afferent neuron expression of receptors and neuropeptides associated with inhibiting food intake (anorexigenic phenotype – red neurons), and inhibits expression of receptors and neuropeptides associated with stimulating food intake (orexigenic phenotype – green neurons). In the absence of food, ghrelin and cannabinoids are released and inhibit expression of the anorexigenic phenotype in preference for the orexigenic phenotype in vagal afferent neurons. Release of anorexigenic neuropeptides from vagal afferent neurons to the NTS shortens the duration of meals and reduces their size, while the release of orexigenic neuropeptides prolongs meals and increases their size.

Mechanosensitivity of gastric vagal afferents also adapts depending on feeding status (Kentish & Page, 2014). Post‐prandial sensitivity of gastric vagal afferents to tension increases (Kentish et al. 2012). The response of these mechanosensitive fibres can be modulated by the hormonal milieu. Ghrelin reduces sensitivity to stretch in both fed and fasted conditions (Kentish et al. 2012), leptin reduces sensitivity to stretch exclusively in fasted conditions (Kentish et al. 2013), while neuropeptide W reduces sensitivity to stretch solely in fed conditions (Li et al. 2013).

Vagal efferent control of digestion and absorption

Vagal efferent cell bodies that control gastrointestinal function are located in the dorsal motor nucleus (Shapiro & Miselis, 1985; Norgren & Smith, 1988). The fibres are bundled in the vagus nerve along with the afferent fibres, but provide information from the brain to the periphery. Vagal efferent neurons innervate abdominal organs along the length of the alimentary canal from the anterior oesophagus to the transverse colon (Teff, 2008), with the highest density of projections terminating in the stomach (Berthoud et al. 1991). Although vagal efferent neurons release exclusively acetylcholine, they activate two different populations of postganglionic neurons that can mediate both stimulatory and inhibitory control of organ function neurons to control digestion and absorption of food (Browning & Travagli, 2011) (Fig. 2). This is highlighted by vagotomy studies in which gastric secretion and motility are profoundly altered. Under normal conditions vagal efferent input to the enteric plexus initiates acid release from parietal cells and release of proteases from chief cells. Both of these effects are severely blunted in a vagotomized stomach (Skak‐Nielsen et al. 1988). The motility effects are more complicated, with vagotomy increasing fundic tone while depressing antral motility (Takahashi & Owyang, 1997). Thus, the loss of vagal input reduces reservoir accommodation and prevents appropriate sieving. As a consequence, vagotomy increases retention of solids while fluids are dumped into the duodenum.

Vagal efferent innervation of the bile duct and pancreas regulates secretion of bile acids and pancreatic juices (Masuda et al. 1994) important in nutrient breakdown and maintaining appropriate pH levels within the small intestine. Although most of the small intestinal contractions are initiated by neurons of the enteric nervous system (Schemann, 2005), there is important involvement of vago‐vagal reflexes in coordinating motility (Rogers et al. 1995). Thus vagal efferent activation is implicated in the regulation of (1) digestion, both by temporal control of digestive enzyme release and by control of antral contractions important in mechanical breakdown, and also (2) absorption, by controlling the rate of gastric emptying and propulsion through the intestine.

Vagal efferent innervation of the gut plays a role in the control of the migrating motor complex (MMC). The MMC is a cyclical motor pattern made up of four phases initiated by the stomach and/or the duodenum that runs the length of the small intestine (Deloose et al. 2012). The MMC lasts approximately 2 h and occurs exclusively during fasting. The large wave of contraction (phase III) is thought to clear undigested loose content from the small intestine into the large intestine, stopping immediately in response to food intake. Vagal cooling studies, in which cervical skin loops transiently block the vagus nerve using cooling jackets, demonstrated that an intact vagus nerve is necessary for stomach MMCH, preserves MMC during feeding, and regulates levels of motilin, gastrin and pancreatic polypeptide required for MMC initiations (Hall et al. 1986; Chung et al. 1992).

The role of the vagus nerve in obesity

Disruption of the vagus nerve has been implicated in the short‐term control of feeding. Total subdiaphragmatic vagotomy (abolishing both afferent and efferent abdominal fibres; Fox, 2013), capsaicin (ablating all transient receptor potential cation channel, subfamily V, member 1 (TRPV1)‐expressing neurons, along with many vagal efferent neurons; Browning et al. 2013 a), or subdiaphragmatic deafferentation (unilateral resection of afferent branches in the jugular foramen, with a unilateral subdiaphragmatic vagotomy that removes 100% of afferent while leaving 50% of efferent fibres; Norgren & Smith, 1994) alters meal patterns, but results in underwhelming effects on body weight and daily food intake (Fox, 2013). Thus, while ablation of the vagus nerve leads to increases in meal size, there is a compensation in meal frequency that prevents long‐term changes in food intake or body weight.

It is important to note the considerable limitations of these techniques. As already alluded to these approaches do not fully distinguish between efferent and afferent signalling, which may act in opposition to each other, thereby masking an important effect of one or the other in energy homeostasis. Importantly, none of these approaches selectively target vagal fibres innervating individual organs, and since different organs may be sending opposing signals, ablating all signals from multiple organs may mask the importance of a specific organ. Finally, these techniques indiscriminately inhibit anorectic and orexigenic signals, which may mask the importance of one signal over another. Despite this substantial list of limitations, the lack of perceptible change in body weight in the absence of vagal signalling has led to the view that the vagus nerve is involved in the short‐term but not the long‐term control of food intake (Konturek et al. 2004; Berthoud, 2008 b). However, there is now growing evidence that disruption of vagal signalling is important in diet‐induced weight gain and may be an attractive target to reverse obesity.

Disrupted vagal signalling in obesity

Models of diet‐induced obesity, in which animals are given free access to palatable, energy‐dense foods, provide substantial evidence that sensitivity of vagal afferent neurons to peripheral signals becomes blunted. The immediate early gene, c‐Fos, used as a marker of neuronal activation, is significantly lower in the mid and caudal NTS following a meal in obese rats compared with lean rats (Covasa et al. 2000 a,b). Since the NTS is the first central site of integration of vagal afferent inputs, reduced neuronal activation in this region suggests that obesity is associated with reduced afferent signalling. This has been confirmed in a number of different studies. Vagal afferent firing in response to increasing tension‐baring loads is reduced in obese mice compared with lean mice (Daly et al. 2011; Kentish et al. 2012), suggesting reduced sensitivity to distension. Vagal afferent sensitivity to gastrointestinal hormones (CCK, bombesin and serotonin) is significantly reduced in rodents chronically fed a high‐fat diet compared with low‐fat‐fed mice (Covasa & Ritter, 1998, 2000; Covasa et al. 2000 a; Savastano & Covasa, 2005; Swartz et al. 2010; Daly et al. 2011; de Lartigue et al. 2012; Duca et al. 2013 b), suggesting reduced sensitivity to satiating hormones. The satiating effects of intestinal nutrients are reduced in obesity (Covasa et al. 2000 b; Covasa & Ritter, 2001; Duca et al. 2012, 2013 a). The biophysical properties of vagal afferent (Daly et al. 2011) and vagal efferent (Browning et al. 2013 b) neurons are impaired in diet‐induced obesity, with a general reduction in excitability and decreased ability to fire action potentials. In addition, there is evidence that vagal afferent neurons develop leptin resistance early in the development of obesity (de Lartigue et al. 2011, 2012), and that leptin resistance in vagal afferent neurons coincides with the onset of hyperphagia (de Lartigue et al. 2012). Together these data support the idea that disruption of vagal sensing of peripheral signals develops in obesity and that it may account for overconsumption of food (Fig. 4).

Figure 4. Loss of plasticity in vagal afferent neurons in response to chronic ingestion of a high‐fat diet .

A, in lean animals, leptin (released from adipocytes or the gut) increases expression of the immediate early gene, early growth response‐1 (EGR‐1), which in the absence of gastrointestinal signals is localized in the cytoplasm. Postprandially, CCK is released and activates CCK‐A receptors on vagal afferent neurons, resulting in translocation of EGR‐1 to the nucleus. EGR‐1 induces the synthesis of the anorectic neuropeptide cocaine and amphetamine‐regulated transcript (CART), and the Y2 receptor that binds the satiating hormone PYY3‐36. Release of CART from vagal afferent neurons into the NTS activates NTS neurons resulting in meal termination. B, in animals chronically fed a high‐fat diet, circulating levels of leptin are high, but vagal afferent neurons become insensitive to it. In response to a meal, the stomach is distended and CCK is released, but leptin‐resistant vagal afferent neurons have reduced sensitivity to these peripheral signals and vagal afferent neurons and remain stuck in a fasted phenotype, expressing the cannabinoid 1 receptor (CB1) and melanin concentrating hormone‐1 receptor (MCH1R) along with melanin concentrating hormone (MCH). Release of MCH from vagal afferent neurons into the NTS increases meal size and duration. C, similarly, knocking down leptin receptors in vagal afferent neurons (VAN) is sufficient to drive hyperphagia in chow‐fed animals. In the absence of leptin signalling, there is reduced sensitivity to CCK and plasticity is lost.

Disruption of vagal afferent signalling is sufficient to promote hyperphagia and obesity

Taking advantage of new genetic tools, we have recently provided clear evidence that disruption of vagal afferent signalling is sufficient to cause overconsumption of food and weight gain (de Lartigue et al. 2014 a). By crossing a voltage‐gated sodium channel 1.8 (Nav1.8) cre mouse with leptin receptor floxed mice, we generated a conditional knockout mouse lacking the leptin receptor in vagal afferent neurons. Although Nav1.8 is a marker of sensory neurons, and is not selective for vagal afferent neurons, co‐expression of leptin receptor and Nav1.8 occurs in high abundance almost exclusively within vagal afferent neurons; thus, crossing Nav1.8‐cre mice with leptin receptor floxed mice resulted in an extensive 93% knockdown of leptin receptor selectively in vagal afferent neurons (de Lartigue et al. 2014 a). Similar to rats that develop leptin resistance when fed a high‐fat diet (de Lartigue et al. 2011, 2012), the conditional knockout mice weighed 10–15% more on a chow diet compared with wild‐type mice as a result of increased adiposity (de Lartigue et al. 2014 a). There was no change in energy expenditure, but daily food intake increased as a result of an uncompensated increase in meal size during the dark phase (de Lartigue et al. 2014 a). Therefore, chronic ingestion of palatable diets reduces vagal afferent sensitivity to peripheral signals at least in part as a result of loss of leptin signalling, leading to chronic increases in meal size and duration that drive hyperphagia and obesity (Fig. 4). These data support the concept that vagal afferent signalling regulates short‐term control of food intake, but provide evidence that chronic disruption of meal patterns can lead to long‐term changes in energy homeostasis.

Loss of vagal afferent plasticity in obesity

In response to chronic ingestion of palatable diets, there is a loss of both mechanical (Kentish & Page, 2014) and neurochemical (de Lartigue et al. 2012) plasticity of vagal afferent neurons. Postprandial expression of receptors and neuropeptides that promote satiation is blunted, while orexigenic receptors and neuropeptides remain constitutively expressed (Fig. 4). The neurochemical switch is also absent in chow‐fed mice lacking leptin receptors (de Lartigue et al. 2014 a) (Fig. 4). Loss of vagal afferent plasticity was also reported in rats chronically administered low‐dose intraperitoneal lipopolysaccharide, a breakdown product of Gram‐negative bacteria found in higher circulating levels in obesity, which gain weight and are hyperphagic (de La Serre et al. 2015). Thus, in three different animal models of weight gain, the ability of vagal afferent neurons to integrate information about the nutritional state was lost. We have hypothesized that increased sensitivity to peripheral orexigenic signals and reduced satiety signalling to the brain would promote overconsumption of food and subsequently lead to obesity (de Lartigue et al. 2014 b). In support of this concept, we now have unpublished data that genetically blunting expression of anorectic neuropeptide signalling from vagal afferent neurons increases cumulative food intake and body weight in rats. In summary, there are now multiple converging lines of evidence that loss of the neurochemical plasticity in vagal afferent neurons may be an important mechanism in the development of obesity.

Targeting the vagus nerve with neuromodulation for the treatment of obesity

In the last section evidence was provided that vagal afferent neurons lose sensitivity to peripheral signals and inappropriately signal to the brain in response to chronic ingestion of palatable food. Disrupting vagal afferent neurons that mimic these effects leads to hyperphagia and weight gain. There is also growing evidence that neuromodulation of the vagus nerve can be used therapeutically in obesity. Neuromodulation has been demonstrated to be safe; surgical risks are minimal, infections manageable with antibiotics, and side‐effects are classified as mild (O'Reardon et al. 2006). Currently two different types of neuromodulation methods targeting the vagus nerve exist: vagal nerve stimulation and vagal blockade. Both involve electrical currents being applied to the vagus nerve, but while vagal nerve stimulation uses low‐frequency currents to activate the vagus nerve, vagal blockade utilizes high‐frequency alternating currents to inhibit vagal signalling. Interestingly both have been demonstrated to improve weight and food intake in obesity. In this section, the implantation protocols, pre‐clinical and clinical outcomes, and known mechanisms of vagal nerve stimulation and vagal blockade will be discussed.

Vagal nerve stimulation

Vagal nerve stimulation (VNS) involves implanting electrodes on the vagus nerve and using electrical pulses to generate firing potentials. A number of different stimulation parameters can be controlled including current intensity, pulse width, pulse frequency, and duration of the ON and OFF periods of stimulation. VNS has been approved by the Food and Drug Administration (FDA) since 1997 as an adjunctive therapy in reducing the frequency of seizures in adults and adolescents who are refractory to anti‐epileptic medications (Yuan & Silberstein, 2015). A lot of work has been done studying VNS in epilepsy models, and while it has been demonstrated to be safe and efficacious, the mechanisms remain unclear (Yuan & Silberstein, 2015). Since receiving FDA approval for the treatment of epilepsy, VNS has also been approved for treatment of treatment‐resistant depression (Shah et al. 2014). There is growing evidence that VNS may prove useful in a multitude of medical conditions, including obesity (Ogbonnaya & Kaliaperumal, 2013) (Fig. 5).

Figure 5. Vagal nerve stimulation effect on body weight .

A, in preclinical studies, vagal nerve stimulation (VNS) of the abdominal organs reduced body weight in lean animals and prevented weight gain in obese animals, and reduced caloric intake and increased satiation in all animals. Non‐invasive transcutaneous auricular VNS increased energy expenditure. VNS increased c‐Fos expression in the NTS, suggesting that vagal afferent neurons were activated. Gastric secretion was reduced, and gastric motility increased with VNS in a subset of lean animals. B, VNS is FDA‐approved for the treatment of refractory epilepsy and depression. Unilateral left cervical VNS with a range of different parameters reduced weight and sweet cravings in patients treated for depression or epilepsy, and this correlated to BMI. Energy expenditure was also found to be increased with left cervical VNS.

VNS control of body weight and food intake in animal models

Lean animals

In preclinical studies, electrodes implanted unilaterally on the ventral vagal trunk, or bilaterally on the ventral and dorsal vagal trunks, below the diaphragm using low‐frequency currents ranging from 0.01 to 30 Hz led to weight loss, increase satiation and reduced food intake, reduced sweet cravings, and increased energy expenditure (Fig. 5). In an initial VNS study in lean mongrel dogs, acute bilateral thoracic vagal trunk stimulation increased latency to eat and chronic stimulation resulted in significant weight loss after 4–5 months of treatment (roslin, 2001). Concurrently, in lean rats that were still growing, unilateral subdiaphragmatic VNS blunted weight gain and food intake (Krolczyk et al. 2001). Reduced weight gain and food intake has since been reported using a variety of different electrode implantation protocols and stimulation parameters in growing rabbits, minipigs and rats (Sobocki et al. 2001, 2002, 2006; Laskiewicz et al. 2003; Matyja et al. 2004; Krolczyk et al. 2005; Bugajski et al. 2007). One early study in lean rats found that bilateral subdiaphragmatic stimulation had more pronounced effects on food intake and body weight than unilateral subdiaphragmatic stimulation (Laskiewicz et al. 2003). Many subsequent studies continued to use unilateral stimulation of the left subdiaphragmatic vagal trunk, based on an early report in which right cervical stimulation in dogs led to greater cardiac events (Woodbury & Woodbury, 1990), despite the absence of any innervation to the heart from subdiaphragmatic vagal trunks.

High‐fat‐fed animals

In rats, unilateral subdiaphragmatic VNS was sufficient to prevent the onset of diet‐induced obesity. VNS (0.01–10 Hz) in rats fed a high‐fat diet for 2–4 weeks prior to surgery reduced excess weight gain, adiposity, and hyperphagia when the animals were kept on a high‐fat diet compared with unstimulated control rats (Bugajski et al. 2007; Gil et al. 2009, 2011 a,b, 2012, 2013; Ziomber et al. 2009). To date only one study has examined the role of VNS in reversing obesity. Minipigs were fed a Western diet for 15 months; weight stabilized after 5 weeks of vagal nerve stimulation, while control animals continued to gain weight, but no weight loss was observed despite food intake being markedly reduced (Val‐Laillet et al. 2010). It is possible that offering a choice of lower calorie diets in combination with VNS may have resulted in weight loss since carbohydrate and fat consumption was reduced while balanced diet was increased in a three‐diet preference test; however, this remains to be tested. Thus, to date VNS has successfully prevented excessive weight gain in high‐fat‐fed animals without leading to weight loss.

VNS control of body weight and food intake in human studies

To date, six clinical studies have been performed, looking at the effect of chronic cervical vagus nerve stimulation on metabolism (summarized in Fig. 5 B). The caveats with these studies is that they were performed in patients being treated for epilepsy or depression and therefore are not representative of the obesity population, the parameters between patients varied enormously, and the stimulation site was not in an optimal location for weight loss; nevertheless, the results are promising. The initial observation was that one patient receiving VNS for the treatment of epilepsy lost a substantial amount of weight following VNS, and this prompted a retrospective study of 32 patients treated for refractory epilepsy in which over 60% of the patients lost weight with VNS, and the remaining patients had stable weight post‐VNS with no weight gain. Interestingly the patients with higher VNS frequency settings lost the greatest weight (Burneo et al. 2002). Another report in 14 patients treated with left cervical VNS for treatment‐resistant depression for over 2 years found that weight loss was positively correlated with initial body mass index (BMI); the more severe the obesity, the greater the weight loss (Pardo et al. 2007). It should be highlighted that there have also been reports of VNS having no effect on body weight in depressed or epileptic patients (Rush et al. 2005; Koren & Holmes, 2006). Although these negative reports had much greater patient numbers, there was also a very large range of parameters used and there was no attempt to determine the body weight changes in subgroups of patients with similar parameters. Thus, large controlled clinical trials are still necessary.

In an acute prospective study, a group of depressed patients treated with VNS experienced greater preference changes to sweet food images in response to the device being turned on compared with when it was turned off. Half of the patients with the VNS device turned on had reduced preference for images of sweet food, while the other half had increased preference (Bodenlos et al. 2007 b). This discrepancy could not be explained by a single factor, but correlated with VNS current output parameters, the amount of time the device was used, emotional eating when depressed, and BMI. The same group reported that during acute stimulation of the vagus nerve there were differences in emotional response to images of sweet foods (Bodenlos et al. 2007 a). Higher BMI and higher sweet cravings were significantly correlated to elevated arousal levels in response to sweet food images when the device was turned on. Although this suggests that sweet cravings may be reduced with VNS in obesity, in a commentary on these two studies, Gibson & Mohiyeddini (2008) suggest that the increased preference for sweet food images may be a reflection on appetite and desire to eat rather than cravings. This is supported by the fact that the changes in response to sweet images are in response to the question, ‘how much would you like to eat the pictured food right now?’ – which is usually an indication of desire to eat – rather than in response to the question, ‘would you be able to resist tasting the pictured food?’ – which is a better indication of craving. Nevertheless, the changes in appetite or craving for sweet foods suggest that VNS may be useful in reducing consumption of sweet foods, and this is supported by studies in obese minipigs treated with VNS in which consumption of sweet foods was reduced (Val‐Laillet et al. 2010).

A recent study in VNS‐treated epilepsy patients demonstrate that turning off VNS for a few hours had a small but significant decrease on energy expenditure. The change in energy expenditure correlated with a reduction in brown adipose tissue (BAT) activity, suggesting that VNS could increase energy expenditure (Vijgen et al. 2013). This is supported by an animal study in which obese rats stimulated with transcutaneous auricular VNS (taVNS), that non‐invasively stimulates the auricular branch of the vagus nerve, prevented weight gain by increasing BAT weight, and β‐adrenoreceptors and uncoupling protein 1 mRNA expression in BAT (Han et al. 2015). The absence of excess weight gain during stimulation occurred without any change in food intake, suggesting that taVNS may be preventing weight gain by increased BAT‐induced thermogenesis (Han et al. 2015).

Mechanisms

Control of body weight

Multiple mechanisms by which VNS controls body weight have been described. Some studies report that food intake is significantly reduced, while others identified increases in energy expenditure. Pair feeding experiments that may provide insight into the relative importance of either one for body weight following VNS treatment, have not yet been performed. It is interesting to note that stimulation of vagal trunks selectively innervating the abdominal organs has only been reported to reduce food intake, that selectively stimulating transcutaneous auricular vagus nerve appears to exclusively increase thermogenesis, and that stimulating the cervical vagus (that innervates both abdominal and auricular branches) affects both food intake and thermogenesis. Further work delineating the neural circuits involved at different activation sites may provide insight into the mechanisms of VNS in the control of body weight.

Vagal afferent vs. vagal efferent pathways

There has been some discordance in the literature about whether VNS activates a vagal afferent and/or efferent mechanism in the control of body weight. Theoretically, at low‐frequency stimulation, afferent C‐fibres should be preferentially activated since they are unmyelinated and thus require a lower threshold for activation. The idea that afferent signalling is involved in mediating VNS effects is supported both by brain imaging studies (discussed in more detail below) and by studies demonstrating that increase in c‐Fos expression in the NTS in response to VNS correlates with reduced weight gain (Gil et al. 2011 b, 2013). As discussed above, activation of hindbrain neurons is sufficient to regulate meal size and duration (Grill & Norgren, 1978). Thus, a very likely mechanism of VNS‐induced weight loss and satiation is by reducing the amount of calories consumed during each meal throughout the day by activating a vagal afferent–NTS‐dependent pathway. The crucial experiment testing meal patterns in animals treated with VNS has not yet been performed. Furthermore, VNS activation of a vagal afferent pathway could increase the sensitivity of vagal afferent neurons to peripheral signals. As discussed above in the section ‘The role of the vagus nerve in obesity’, this would be particularly important in obesity, where baseline activity of vagal afferent fibres and vagal afferent response to peripheral stimuli are reduced. VNS may increase the responsiveness of vagal afferent neurons to distension and/or satiating hormones in obesity.

A number of studies also support the idea that VNS activates a vagal efferent pathway. VNS increased the amplitude of gastric contractions, increased gastric emptying, and decreased acid output, bradycardia and apnoea in a frequency‐dependent manner (Krolczyk et al. 2001; Sobocki et al. 2002; Matyja et al. 2004; Osharina et al. 2006), all controlled by efferent signals. Irrespective of whether VNS activates efferent fibres directly, or indirectly by a vago‐vagal reflex, increased efferent output could be a mechanism for VNS‐induced reduction in food intake. Increased gastric motility and gastric emptying may reduce the breakdown of macronutrients thereby reducing absorption. It should be noted that there has been no evidence of malabsorption in patients treated with VNS, so this remains an unlikely mechanism. Alternatively, increased motility could both (1) reduce nutrient interaction with enteroendocrine cells in the proximal small intestine, which would diminish release of gastrointestinal hormones from the upper small intestine in response to a meal, and (2) increase release of gastrointestinal hormones from the distal small intestine by increasing the levels of micronutrients reaching the lower part of the gut. The levels of circulating gastrointestinal hormones have not been well characterized in lean or obese models post‐VNS to address this possibility. In rats fed a high‐fat diet, VNS reduces circulating leptin levels (Ziomber et al. 2009; Gil et al. 2012), presumably as a result of reduced adiposity; ghrelin levels remain unchanged (Gil et al. 2012), while nesfatin‐1 levels were elevated (Gil et al. 2013).

While there is still some debate, with one study finding no effect of VNS on gastric emptying (Biraben et al. 2008), it appears that VNS can induce both afferent and efferent pathways. It is likely that VNS directly activates vagal afferent fibres and that this leads to secondary activation of both central pathways and a vago‐vagal reflex, which in turn activates vagal efferent neurons that modulate motor function. Until vagal afferent signalling can be selectively abolished in freely behaving animals treated with VNS, it will not be possible to fully address the question of whether VNS activates just afferent fibres, or both afferent and efferent fibres.

Imaging studies

Imaging studies suggest that cervical VNS in subjects treated for depression or epilepsy compared with unstimulated baseline conditions results in altered activation of multiple regions associated with feeding behaviour, including regions involved in reward, decision making, taste and homeostatic control of food intake. In addition to the hindbrain, the hypothalamus is also involved in appropriately matching food intake to metabolic need (Coll & Yeo, 2013) and altered signalling in the hypothalamus is an important mechanism in obesity (Suzuki et al. 2012). Unilateral stimulation of the cervical vagus increases activity within the hypothalamus (Henry et al. 1998). Although it is not possible to determine the exact area or neurons that are activated based on the imaging data, it suggests that VNS may activate at least two brain sites associated with modulating homeostatic feeding states. Non‐homeostatic appetitive eating, which is food intake driven by food palatability (taste, smell and visualizing food), wanting or liking, and other cognitive factors (i.e. habitual, sociocultural, emotional and economic) is influenced by multiple different brain regions particularly the dopaminergic limbic and prefrontal reward areas (Berthoud, 2011). Key areas involved in appetitive behaviour include the insula, orbitofrontal cortex, amygdala and hippocampus and these are all modulated by VNS (Ko et al. 1996; Henry et al. 1998). The insula is a region important in establishing salience (Menon & Uddin, 2010) and activity in this region is increased in lean fasted subjects in response to food images and highly correlated with ratings of appetite (Porubská et al. 2006). Activity of the anterior insula is increased in response to cervical VNS. Orbital frontal cortex activity is found to be elevated during hunger or with VNS, and is associated with perception of pleasantness of food (Fuhrer et al. 2008). The amygdala is associated with cue‐driven food consumption (Holland et al. 2002; Petrovich et al. 2002); in support of this, in human imaging studies, non‐obese male subjects asked to inhibit their craving for food during exposure to food cues had decreased metabolic activity in the amygdala (Wang et al. 2009). VNS reduces amygdala activity (Henry et al. 1998), suggesting that VNS may be involved in reducing cue‐driven consumption. Hippocampal activity was also reduced in response to cervical VNS (Henry et al. 1998); this region is associated with memory and recall function, and is also activated in response to wanting components of food reward (Frank et al. 2015). Finally activity in the putamen, which includes the dorsal striatum, is up‐regulated in response to VNS (Ko et al. 1996). In mice, dopamine release from the dorsal striatum in response to nutrient infusion into the gut has been demonstrated to be mediated by the vagus nerve (Tellez et al. 2013). This vagally mediated activation of the dorsal striatum plays an important role in post‐ingestive nutrient‐derived reward. The changes discussed above in activity of areas of the brain associated with reward correlated with changes in food preferences reported both in VNS‐treated humans and obese minipigs, and support the idea that VNS may alter taste perception and/or reward to prevent weight gain by preventing selection of calorific foods.

Imaging studies of left cervical VNS in animals largely corroborated the sites of metabolic activity described above in humans (Dedeurwaerdere et al. 2005; Reyt et al. 2010). It appears that the same basic brain regions involved in homeostasis and reward are modulated by bilateral subdiaphragmatic VNS in lean pigs (Val‐Laillet et al. 2015). It would be interesting to make a detailed comparison of brain activity in response to different sites of vagal nerve stimulation; however, it is difficult to conclude too much from these cervical and subdiaphragmatic VNS studies because the pig and rat brain are different. The main difference is that there is activation of the olfactory bulb after subdiaphragmatic VNS that is not present after cervical VNS, and there may be subtle differences in activation of subnuclei within the limbic regions. If the site of stimulation on the vagus nerve is identified as an important factor in promoting weight loss in obesity, it will be important to make detailed comparisons of brain activity in response to different sites of stimulation in the same animal model using the same stimulation parameters.

In humans receiving cervical VNS for treatment of depression or epilepsy, acute or chronic stimulation produces different activation patterns in imaging studies (Chae et al. 2003), suggesting that the brain adapts to VNS over time. This is an important observation and suggests that efficacy of obesity treatment may improve with increased VNS treatment. Many of the studies in rodents lasted 4–8 weeks, which may not have been sufficiently long to promote neural changes that lead to weight loss. In the longest study to date using bilateral subdiaphragmatic VNS, weight gain in obese minipigs continued for the first 5 weeks post‐stimulation and subsequently stabilized for the remaining 9 weeks of the study (Val‐Laillet et al. 2010). Therefore prolonged stimulation had beneficial effects on weight management. The effect of bilateral subdiaphragmatic VNS beyond 14 weeks remains unknown, but weight loss in patients treated with cervical VNS for depression or epilepsy was predominantly reported after 6–12 months of treatment.

Optimizing stimulation parameters

VNS has potential for the treatment of obesity, but the optimal parameters to promote weight loss are still unclear. There is evidence from both animal and human studies that higher frequency may improve outcome. In rats, frequency‐dependent improvements have been reported (Laskiewicz et al. 2003; Ziomber et al. 2009; Gil et al. 2013), but no study in any animal has used higher than 30 Hz. In minipigs there was no evidence of Wallerian degeneration at 30 Hz indicating that higher frequencies could be tested. Continuous nerve stimulation with 50 Hz can result in scattered degeneration of large myelinated fibres (Cohen & Georgievskaya, 2011), which might argue against the use of higher frequency stimulation, but fibre damage can be progressively reduced with stimulation at 50 Hz using intermittent signalling, or lowering the current to below 400 μA (Cohen & Georgievskaya, 2011). These observations suggest current, frequency, duration and on–off patterns of stimulation all interact, and it may be possible to further optimize these parameters to maximize the beneficial effects of VNS in obesity. It may be particularly important to stimulate the vagus nerve at 50 Hz, since physiological spike frequency in response to a meal increased to 50 Hz in a rat model (Krolczyk et al. 2001).

Moving forward

It will be important to determine both the optimal stimulation parameters, and the most appropriate site to stimulate (cervical vs. auricular vs. abdominal). The subdiaphragmatic vagal branches, stimulated bilaterally, may be a more appropriate site to target in a clinical setting based on our understanding of the vagal gut–brain axis in the control of food intake; however, further meticulous animal experiments are still required. Most importantly, although VNS has been reported to prevent excess weight gain in animals fed a high‐fat diet, there is currently no evidence that VNS can be an effective approach for weight loss in obesity. Demonstrating that reversing excess weight gain can occur with VNS would be a necessary proof of principle experiment before VNS could be seriously considered for clinical use. Understanding the mechanism for VNS‐induced hypophagia and thermogenesis may provide insight to optimize the protocol and achieve weight loss in obesity.

Vagal blockade

In vagal blockade, electrodes are placed on the anterior and posterior vagal trunks near the oesophago‐gastric junction along with a subcutaneously implanted neuroregulator. Reversible vagal inhibition is achieved by applying kilohertz frequency alternating current (KHFAC) directly to the nerve to block localized electrical conduction (Naples et al. 1988; Foldes et al. 2011). In both preclinical and clinical studies KHFAC has been demonstrated to be safe. In clinical trials, vagal blockade (VBLOC) treatment induces significant weight loss, earlier satiation during meals, prolonged fullness after meals, improved glycaemic control and liver function, and reduced blood pressure in obese subjects (Fig. 6).

Figure 6. Vagal blockade promotes weight loss in obese subjects .

A, in clinical trials of obese subjects, intermittent vagal blockade (VBLOC) significantly improved symptoms of obesity. VBLOC inhibits both afferent signalling as demonstrated by altered brain imaging and efferent function indicated by reduced pancreatic secretion and gastric contractions. In diabetic subjects VBLOC reduced body weight, glucose‐bound haemoglobin, and fasting plasma glucose levels. B, electrodes are placed on the anterior and posterior vagal trunks near the oesophago‐gastric junction along with a subcutaneously implanted neuroregulator. C, a mechanism for VBLOC‐induced weight loss is proposed. In obesity, vagal afferent neurons constitutively express the orexigenic neuropeptide melanin concentrating hormone (MCH). MCH prolongs meals, resulting in hyperphagia and weight gain. VBLOC may prevent release of MCH and MCH‐induced hyperphagia.

Preclinical studies

There have not been any animal studies demonstrating a role of KHFAC‐induced vagal blockade on weight loss. Initial preclinical work in rats demonstrated that KHFAC was capable of blocking vagus nerve signalling (Waataja et al. 2011). In a preparation consisting of an isolated abdominal vagus, stimulating electrodes were placed on the nerve in the abdomen, and whole nerve‐firing compound action potentials were measured with recording electrodes placed rostrally on the thoracic end. A 5 kHz biphasic waveform was applied for 1 min between the stimulating and recording electrodes, and immediately following blockade with KHFAC the vagus nerve was stimulated. Both fast myelinated Aδ‐fibres and slow unmyelinated C‐fibres were inhibited in an amplitude‐dependent manner with higher amplitudes resulting in longer inhibition of 5–6 min in C‐fibres and in excess of 10 min for Aδ‐fibres (Waataja et al. 2011). This experiment formed the basis for the 5 min ON, 5 min OFF algorithm for vagal blockade.

Importantly, preclinical data demonstrated the safety and efficacy of chronic intermittent ON–OFF 5 kHz application to the vagus nerve for 12 h daily over 55 days in a porcine model (Tweden, 2006 a,b). Inhibition of vagal function was confirmed quantifiably with both pancreatic exocrine secretion and gastric contraction decreased in response to KHFAC stimulation in vivo. Pancreatic exocrine secretion was restored within 5–30 min post blockade indicating that the physiological effects of vagal inhibition are reversible. Importantly there was no anatomical or histological damage to the vagus nerve in these animals after 55 days (Tweden, 2006 a,b).

Clinical studies

Pilot trials testing safety and efficacy of VBLOC for treating obesity

To date, four clinical trials have been performed by Enteromedics using VBLOC for the treatment of obesity. The first two clinical trials applied KHFAC to intermittently block subdiaphragmatic vagal signalling in a small population of patients (31 and 27, respectively) with refractory obesity, defined as individuals with a BMI in excess of 31 (mean BMI ∼40 kg m^–2) that had not achieved satisfactory or sustained weight loss with diet, behavioural intervention and/or pharmacotherapy (Camilleri et al. 2008, 2009). Both of these trials reported significant loss of excess weight in the absence of any serious adverse events as part of the 6‐month evaluation (Camilleri et al. 2008, 2009). Most impressively vagal blocking markedly reduced total caloric intake and fat consumption very early after turning on the device, and these benefits remained for the duration of the study.

In the second trial an improved second generation device was used, and the duration of therapeutic ‘ON’ time was changed from continuous to intermittent. Having retrospectively determined from the first clinical trial that an ‘ON’ time of 90–150 s was more effective than the original 300 s ‘ON’ protocol (Camilleri et al. 2009), a new ‘ON’ protocol consisting of two 120 s 5000 Hz stimulation separated by a 60 s off period was initiated. This optimized device and electrical algorithm in the second trial improved excess weight loss by nearly 60% compared with the initial trial without additional adverse events and was subsequently used in larger trials. Both of these preliminary trials reported similar improvements in meal termination and hunger compared with baseline pre‐implantation. In the second trial waist circumference and liver function were also found to be reduced and improved, respectively, after 6 months (Camilleri et al. 2009).

Large clinical trials

A large randomized, double‐blind clinical trial of VBLOC (EMPOWER) used the same parameters as above (2 min ON, 1 min OFF, 2 min ON, 5000 Hz waveform with amplitudes ranging from 3 to 8 mA, followed by 5 min with no impulses, for 9–16 h a day) in 192 subjects with refractory obesity across 15 centres (Sarr et al. 2012). Despite the fact that the 102 control subjects received much less current (two 26 ms bursts of 1000 Hz at 3 mA during the 5 min ON cycle, followed by 5 min with no impulse), both VBLOC and sham controls exhibited significant weight loss, reduced hunger and increased satiation compared with post‐op measurements. At the 1 year mark, VBLOC treatment reduced excess body weight by 7% from initial weight (17% excess weight loss), while control subjects had lost 6% of their initial weight (16% excess weight loss) (Sarr et al. 2012), a weight loss far in excess of previous trial placebo control groups both with (3.5% loss) and without (1–2% loss) lifestyle counselling (Wadden et al. 2011). The fact that the control group had such marked reduction in body weight, together with the observation that increased duration of device usage in both control and VBLOC groups translated into improved weight loss, led to the hypothesis that the low current delivered to the vagus nerve may be sufficient to have long‐term clinical and physiological benefits. These data also led to the recommendation to promote devise usage to greater than 12 h per day.

In order to address the possibility that the reduced current unintentionally modulated vagal activity, and reveal a statistically significant weight loss with VBLOC compared with control, another large‐scale randomized, double‐blind trial (ReCHARGE) was initiated (Ikramuddin et al. 2014). Recruitment of 239 subjects occurred across 10 sites, with 162 obese subjects receiving VBLOC and 77 control subjects implanted with a neuroregulator without leads. In this trial, after 1 year, mean percentage excess weight loss was significantly greater with VBLOC compared with the control group (24.4% vs. 15.9%), as was mean total loss of body weight (9.2% vs. 6.0%). Surprisingly, weight loss in the sham group without leads still resulted in a higher than expected weight loss, likely due to the powerful placebo effect caused by undergoing a sham surgery. The hypothesis that weight loss is a result of a surgical placebo is supported by follow‐up reports in which the sham group continued to exhibit marked weight loss up to 15 months post‐surgery after which time regain of body weight coincided with being unblinded (Shikora et al. 2015 b). At 18 months, the VBLOC group had sustained weight loss without additional adverse events, while the sham group regained 40% of the weight back, magnifying the difference between the groups (Shikora et al. 2015 b). As a direct result of these positive findings, the FDA have approved VBLOC for the treatment of obesity pending sustained weight loss during the ongoing 5‐year follow‐up.

In a subgroup of ReChARGE trial patients that presented with both obesity and type 2 diabetes mellitus, glycaemic control and elevated blood pressure were improved with VBLOC (Shikora et al. 2013). Haemoglobin A1c (HbA1c), a marker of glucose bound haemoglobin, is maintained at below 6% of the total haemoglobin in healthy individuals (Greci et al. 2003). In type 2 diabetes the goal is to keep HbA1C levels under 7% to minimize risk of complications, and two large‐scale studies, UK Prospective Diabetes Study (UKPDS) and the Diabetes Control and Complications Trial (DCCT), established that a 1% reduction in HbA1c in patients with type 2 diabetes cuts the risk of microvascular complications by 25% (King et al. 1999; Gubitosi‐Klug, 2014). VBLOC treatment markedly reduced HbA1C levels within the first few weeks, and from 3 months reduced the mean starting levels from 7.8% to 6.8% with the most pronounced effects observed in the subjects with the highest starting levels (Shikora et al. 2013). Similarly high fasting blood glucose levels in the diabetic subjects were markedly reduced from week 1, and the improved glucose tolerance continued throughout the study (Shikora et al. 2013). Similarly to sleeve gastrectomy (Myronovych et al. 2014), the improvements in glycaemic index were independent of weight loss (Shikora et al. 2013). A recent follow‐up report suggests that the beneficial effects of 12‐month VBLOC treatment on glycaemic control are maintained at 24 months (Shikora et al. 2015 a).

The statistically significant weight loss between VBLOC and sham groups, and FDA approval for VBLOC notwithstanding, it is important to note that the weight loss is far less pronounced than that seen after gastric banding (52% excess weight), vertical sleeve gastrectomy (64% excess weight loss), or roux‐en‐Y gastric bypass (66% excess weight loss) at the 2‐year follow‐up (Puzziferri et al. 2014). Furthermore VBLOC only resulted in an additional 50% weight loss compared with sham effects alone, and thus two‐thirds of the weight loss observed with VBLOC is accounted for by the surgical placebo effect. However, nearly a quarter of the patients treated with VBLOC had greater than 50% excess weight loss (Ikramuddin et al. 2014), suggesting that further optimization could increase the effectiveness of vagal blockade to resemble effects seen with bariatric surgery. Understanding the mechanisms will be an important step to maximizing the benefit achievable with this approach.

Adverse events of VBLOC therapy

The reported adverse events of VBLOC therapy are less severe than those associated with bariatric surgery. Crucially there have been no deaths from the 591 subjects implanted with the device (179 of whom are controls). However, the number of mild to moderate adverse effects are notable (Camilleri et al. 2008, 2009; Sarr et al. 2012; Ikramuddin et al. 2014). The most common complaint was mild or moderate pain at the site of the neuroregulator implant, which occurred in approximately 40% of subjects in both VBLOC and sham groups. Pain other than at the implant site was far more prominent in the VBLOC group, along with heartburn, nausea, dysphagia, cramps and bloating.

Mechanisms

KHFAC mechanism

KHFAC has been demonstrated to inhibit firing in multiple species in numerous peripheral nerves over a range of nerve diameters (Kilgore & Bhadra, 2014). The exact mechanism remains unclear but disruption of voltage‐gated sodium and/or potassium channels has been proposed. Modelling studies suggest that either fast potassium currents overwhelm inward sodium currents, or that inward sodium currents are too slow to track high‐frequency changes in membrane potential (Kilgore & Bhadra, 2014). Irrespective of the biophysical mechanism, KHFAC robustly blocks action potentials in nerve fibres ranging from large motor fibres (Bhadra & Kilgore, 2005) to small unmyelinated fibres (Joseph & Butera, 2011). Conduction blockade is maintained in afferent and efferent fibres as long as the KHFAC waveform is applied to the nerve, and is reversible upon KHFAC termination (Kilgore & Bhadra, 2014). A potential concern for translating KHFAC‐induced conduction blockade clinically has been the initial transient neural activation that occurs every time KHFAC is delivered to a nerve. This event, known as onset response, ranges from a single depolarization 2–3 times stronger than a single action potential to a prolonged strong activation lasting many seconds. While onset responses may be minimized with very high frequencies (>20 kHz) (Bhadra & Kilgore, 2005), they cannot be completely eliminated.

The cause of the mild to moderate pain with VBLOC was not discussed, but likely stemmed from the onset response that leads to unwanted vagal efferent muscle contraction and nociceptive vagal afferent signalling to the brain. Data on the effect of VNS on pain perception have led to the hypothesis that acute stimulation may lead to repeated periodic hypersensitivity to painful stimuli (Borckardt et al. 2005). Thus the two onset responses during each 5 min ON phase may hypersensitize VBLOC patients to pain and account for the large percentage of complaints of pain. It would be of interest to determine if minimizing the onset response could attenuate or minimize these adverse effects. This could subsequently help improve outcomes by enabling higher amplitudes to be tolerated by patients.

Another important question that remains to be addressed regarding the mechanism of KHFAC‐induced vagal blockade regards the timing. Increasing our understanding of both (1) when VBLOC should be applied during the day and (2) at what point during the ON–OFF cycle KHFAC inhibits vagal signalling may help with optimization of the outcomes. While it may be assumed, based on previous work in other peripheral nerves, that conductance blockade occurs during application of the KHFAC, experimental evidence only indicates that temporary loss of conduction in the vagus nerve occurs for 5 min after cessation of delivery of KHFAC before returning to baseline. It remains unclear if there is any vagal inhibition during current application. If vagal signalling is also inhibited during the application of the KHFAC, this would imply that VBLOC is in fact chronically inhibiting the vagus nerve for the duration of the time the device is active during the day. Increasing the OFF time may be desirable to enable some vagal signalling during the day. In addition although longer device application times correlate with improved outcomes, it would be interesting to determine if selectively blocking vagal signalling around meals would be more effective.

Weight loss

VBLOC leads to significant weight loss in obese individuals, but the mechanisms have not been fully tested. Energy expenditure by direct or indirect measurement has not yet been performed; however, caloric intake is significantly reduced and is likely to account for at least part of the weight loss. Total caloric intake is reduced rapidly post‐implantation, and is maintained for at least 2 years (the most recent follow‐up). Using visual analogue scale questionnaires to assess hunger and appetite, it was found that VBLOC increased satiation and reduced hunger. Therefore the reduction in caloric intake is likely to be driven by reduced food intake during and between meals. The effect of VBLOC on food preferences has not yet been determined.

Vagal afferent vs. vagal efferent pathways

Vagal signalling rostrally from the abdomen was inhibited by KHFAC in initial studies indicating that afferent signalling is inhibited (Waataja et al. 2011). Early studies using high‐frequency alternating current with different nerves in a range of animals resulted in reversible inhibition of muscle contraction (Bugnard & Hill, 1935; Cattell & Gerard, 1935), suggesting that efferent signals can also be abolished. The relative importance of afferent or efferent pathways in VBLOC‐induced weight loss in obesity remains unclear.

Irrespective of whether vagal afferent or vagal efferent pathways are inhibited, it is difficult to conceive of a mechanism by which inhibiting negative feedback would increase satiation, reduce hunger and lower body weight. While the onset response that occurs each time the current is applied may have some stimulatory effect, this is unlikely to account for the beneficial effects of VBLOC because subdiaphragmatic trunkal vagotomy, which also prevents any afferent or efferent signalling also results in weight loss with improved satiation and hunger in obese patients (Kral, 1978, 1979, 1980, 1989; Kral & Gortz, 1981; Gortz et al. 1990). One hypothesis to explain the mechanisms by which body weight and caloric intake are reduced in response to VBLOC or vagotomy is that they prevent aberrant orexigenic vagal afferent signalling in obesity (Fig. 6). As discussed above, obesity is characterized by a loss of neurochemical plasticity. In response to chronic high‐fat feeding there is constitutive expression of orexigenic neuropeptides and receptors associated with promoting food intake. Thus, preventing aberrant orexigenic signalling in obesity may reduce meal size and hunger, and subsequently reduce body weight (Fig. 6).

Targeting the vagus nerve with pharmacotherapy for the treatment of obesity

Based on the growing evidence described above and its situation in an accessible peripheral location, the vagus nerve is a promising pharmacological target for treating obesity (see Abstract Figure). Vagal afferent neurons and their terminals, as well as vagal efferent preganglionic terminals, are localized outside of the blood–brain barrier. Drugs designed to target the vagus nerve that are too large to cross the blood–brain barrier could be effective without unwanted side‐effects caused by off‐target interactions at central sites. Addressing disrupted vagal afferent signalling has the potential to lead to multifaceted benefits including reducing satiation, impulsitivity and wanting, and increasing thermogenesis. Finally the plasticity of vagal neurons makes them inherently rewireable, with changes in vagal afferent and efferent neurons having been demonstrated to be reversible (Browning & Travagli, 2011; Kentish et al. 2014). Therefore, although no pharmacological agents have been designed to selectively target the vagus nerve, this could be an attractive target site for future drug development.

It should be noted that although none of the drugs on the market selectively target the vagus nerve, some of the anti‐obesity medications may be working in part on the vagus nerve. Research has largely focused on the central mechanisms of anti‐obesity drugs; however, the extent to which their effects on weight loss are mediated via a vagal pathway remains unclear. The currently approved nicotinic acetylcholine receptor antagonist bupropion (tradename Contrave®), in addition to acting at other sites, will inhibit acetylcholine signalling by vagal efferent neurons and subsequently affect gastrointestinal function. The long‐lasting glucagon‐like peptide 1 (GLP1) receptor agonist liraglutide (trade name Saxenda®) as well as endocannabinoid receptor antagonist rimonabant (Acomplia®) could directly interact with GLP1 receptors and cannabinoid receptors, respectively, expressed on vagal afferent neurons to increase satiation. Likewise the lipase inhibitor orlistat (trade name Xenical® and Alli®) that prevents fat digestion (Yanovski & Yanovski, 2014), or the dipeptidyl peptidase 4 (DPP‐IV) inhibitors vildagliptin (Galvus®) and sitagliptin (Januvia®), which prevent the breakdown of circulating GLP1 (Yanovski & Yanovski, 2014), may indirectly affect gastrointestinal hormone release and consequently vagal afferent signalling. It is difficult to know how much of the weight loss effect of these drugs is mediated by a vagal pathway. However, there is evidence from knocking down GLP1 receptor expression in vagal afferent neurons that a vagal afferent pathway mediates GLP1 effects on satiation, gastric emptying, and thermogenesis (Krieger et al. 2016), suggesting that at least the drugs targeting GLP1 receptors could be acting in part through a vagal afferent pathway to promote weight loss.

Drug development

One advantage of developing drugs that target a peripheral site like the vagus nerve is the option of preventing brain penetration. This could reduce the severity of side‐effects by reducing off‐target effects of unwanted binding at central sites. Theoretically this could enable the use of higher doses and/or combining of multiple therapies. However, restricting a drug to the periphery alone is not sufficient to ensure the absence of side‐effects. Not only could a drug have off‐target effects by interacting with visceral organs, but the vagus nerve itself also modulates a multitude of physiological functions including cardiorespiratory control. Many obesity medications have been withdrawn because of valvular heart defects and pulmonary hypertension (Kang & Park, 2012), which could be vagally mediated. Thus, an important challenge will be to selectively target drugs to vagal afferents innervating abdominal organs, thereby reducing the risk of cardiorespiratory side‐effects. A possible mechanism to achieve cell type specificity includes developing antibody–drug conjugates similar to those used to target chemotherapy drugs to cancerous cells. In this case the antibody would bind a receptor expressed on vagal afferent neurons innervating the gut and present a hormone mimetic to either agonize or antagonize a second receptor. This approach would allow sensitive discrimination between vagal afferent neuron subpopulations, and would reduce activation of other cell types that do not express the appropriate receptor combination. An example of a successful antibody–drug conjugate is Ado‐trastuzumab emtansine (T‐DM1) consisting of an anti‐human epidermal growth factor receptor 2 (HER2) antibody, trastuzumab, joined to a stable linker to maytansinoid, DM1, a potent microtubule‐disrupting agent. The antibody targets HER2 expressing cancer cells, while DM1 initiates cell death. T‐DM1 has received FDA approval for the treatment of patients with HER2‐positive metastatic breast cancer (Lambert & Chari, 2014).

An important unanswered question in designing drugs to target the vagus nerve is whether the drug would only be effective when binding at afferent terminals or whether binding to the cell bodies and axon produces similar effects. Vagotomy studies provide an indirect answer to this question. Total abdominal vagotomy abolishes the satiating effects of 2 and 4 μg kg^–1 intraperitoneal injections of CCK (Joyner et al. 1993), suggesting that terminals are essential. However, total abdominal vagotomy has less pronounced effects on the satiating effects of higher doses of CCK (Joyner et al. 1993), and CCKA receptor antagonist prevents satiation in both vagotomized and intact pigs (Ripken et al. 2015). It remains to be determined whether this is because at higher doses the circulating levels of CCK are high enough that they are able to activate receptors on vagal afferent neurons and axons beyond the terminals, or whether a vagally independent mechanism is activated. Nevertheless, a drug that is administered orally and absorbed by the small intestine would presumably activate vagal afferent terminals and at higher doses could also activate the neuronal bodies in the nodose ganglia.

Molecular targets

The effectiveness of bariatric surgery in the treatment of morbid obesity has highlighted the potential for gut‐derived hormones to deliver metabolic improvements without the complications of surgery. Notably, circulating levels of the gastrointestinal incretin GLP1 are rapidly elevated following bariatric surgery (Morinigo et al. 2006), and liraglutide has been successfully used for weight loss in obesity (Davies et al. 2015). With vagal afferent neurons expressing the whole gamut of gut hormone receptors, targeting agonists for satiating hormones and inverse agonists/antagonists against orexigenic hormones to the vagus nerve is a promising direction for the future. As described above, vagal afferent neurons have reduced sensitivity to gastrointestinal hormones in obesity, and therefore increasing baseline activity of vagal afferent neurons by increasing circulating levels of agonists could reduce the threshold for endogenous gastrointestinal hormones to activate vagal afferent neurons.

Another approach would be to develop drugs aimed at resetting the sensitivity of vagal afferent neurons. Leptin resistance in vagal afferent neurons is associated with reduced sensitivity of these neurons to leptin, CCK and peptide YY (de Lartigue et al. 2012, 2014 a); indirect evidence suggests that loss of leptin signalling could also result in reduced sensitivity of vagal afferents to gastric distension (Kentish et al. 2013). Thus, a drug developed to reverse leptin resistance could provide multiple benefits. The mechanisms by which palatable diets initiate leptin resistance in vagal afferent neurons are likely to be multifaceted. However, chronic low‐dose lipopolysaccharide (LPS) has been demonstrated to promote leptin resistance in vagal afferent neurons (de La Serre et al. 2015), and therefore an antagonist of the LPS receptor, toll‐like receptor 4, could at least partially restore vagal afferent sensitivity to peripheral satiation signals.

Lipid messengers targeting the nuclear receptor peroxisome proliferator‐activated receptor α (PPAR‐α) may be developed into an appealing class of drug that could target multiple underlying mechanisms important in obesity. The PPAR‐α endogenous ligand, oleoylethanolamide (OEA), causes marked reduction in food intake (Rodriguez de Fonseca et al. 2001), and PPAR‐α knockout mice chronically consumed greater calories over 24 h (Schwartz et al. 2008). OEA increases in the small intestine are absent in animals fed a high‐fat diet (Tellez et al. 2013). Although PPAR‐α knockdown in vagal afferent neurons has not yet been performed, there is evidence that OEA may activate a vagal afferent pathway. Importantly, PPAR‐α receptors are expressed on vagal afferent neurons (Mansuy‐Aubert et al. 2015). The exact subpopulation of vagal afferent neurons that express PPAR‐α receptors are unknown, and the role of these receptors on vagal afferent neurons remains to be fully characterized. However, OEA‐induced satiation is blunted in capsaicin‐treated rats, and in vagally intact rats, intraperitoneal OEA elevates c‐Fos expression in the NTS (Rodriguez de Fonseca et al. 2001). In addition to its role in satiation, OEA acting via PPAR‐α receptors mediates dopamine‐dependent nutrient‐conditioned learning in rodents (Tellez et al. 2013). Dorsal striatum dopamine efflux was observed in response to feeding to promote reward (Berridge, 1996). Intraperitoneal OEA administration in mice robustly increased dopamine efflux in the dorsal striatum, and this effect was abolished with capsaicin or vagotomy (Tellez et al. 2013). In human subjects OEA decreased impulsivity (van Kooten et al. 2016), and could be important in obesity as impulsivity is associated with lack of perseverance and increased reward seeking (Mobbs et al. 2010). Thus, PPAR‐α agonists targeted to vagal afferents could become effective obesity medications by reducing satiation and impulsivity, while restoring the nutrient‐induced reward.

Conclusion

The vagus nerve provides bidirectional information between the brain and peripheral organs. Vagal efferent innervation of the gut is involved in digestion and absorption, while vagal afferent fibres innervating the gut provides negative feedback about volume and type of nutrients to the brain to inhibit food intake. Vagal afferent neurons do not simply relay signals to the brain, but also integrate these peripheral signals to inform the decision about previous meals. Vagal afferent neurons change their neurochemical phenotype in response to nutrient availability to incorporate information about previous meals into the decision about current meal size and duration.

The absence of a substantial effect on body weight or food intake in response to surgical or chemical ablation of the vagus nerve in lean animals has led to the view that the vagus nerve is involved in the short‐ but not long‐term control of food intake. However there is now substantial evidence that disruption of vagal afferent signalling can lead to long‐term changes in food intake. In obesity, there is reduced sensitivity of afferent neurons to peripheral meal‐induced signals, such as gastrointestinal hormones, gastric distension and leptin. Leptin resistance in vagal afferent neurons prevents appropriate change in the neurochemical expression in response to a meal; instead vagal afferent neurons constitutively express receptors and neuropeptides that promote food intake. This loss of plasticity and sensitivity to peripheral signals within vagal afferent neurons is sufficient to promote overconsumption of food and weight gain. Thus, vagal afferent signalling is a key mediator of short‐term control of food intake, but in response to chronic ingestion of calorie‐rich diets dysregulation of vagal afferent neurons that constitutively secrete orexigenic signals to the brain leads to increases in meal size, and subsequently hyperphagia and obesity.

Importantly, neuromodulation of the vagus nerve has been demonstrated to be a possible target for the treatment of obesity. In 2015 vagal nerve blockade was clinically approved by the FDA for the treatment of moderate to severe obesity. Reversible electrical inhibition of vagal signalling for at least 12 h per day results in significant weight loss in obese subjects, and this effect is maintained for at least 2 years. Interestingly, activation of the vagus nerve with VNS has also shown promise in preventing weight gain in animals fed a high‐fat diet. Further optimization and characterization of VNS would be required before it could be used clinically. The big challenge moving forward will be to elucidate the mechanisms of action by which both inhibition and stimulation of the vagus nerve can have similar therapeutic effects in obesity.

The disruption in vagal afferent neurons in obesity and the positive outcomes of neuromodulation in obesity highlight the possible effectiveness of developing pharmacological agents that selectively target the abdominal vagus nerve in treating obesity. As a result of their peripheral location, vagal fibres are amenable to drugs that are prevented from crossing the blood–brain barrier; thus drugs developed to target the abdominal vagus nerve could have the combined benefit of being an orally administered, multifaceted therapy with minimal side‐effects. This is an understudied area of research that has the potential for important therapeutic benefits.

Based on our current understanding of the vagus nerve in obesity, further studies are warranted. Future work should revolve around the development of better tools to study subpopulations of afferent and efferent neurons, along with more detailed investigation to determine the extent of the role these subpopulations play in the development of obesity, and how to restore normal electrophysiological and neurochemical function in obesity. Whether targeting the vagus nerve alone could be sufficient to reverse obesity remains an open question, but because of the low level of side‐effects when using selective vagal therapies like neuromodulation, and at least in theory pharmacology, these could be used in combination with other therapies.

Additional information

Competing interests

The author declares no conflict of interest.

Funding

This work was supported by National Institutes of Health Grant R00DK094871.

Biography

Guillaume de Lartigue is an Assistant Fellow at the John B. Pierce Laboratory and an Assistant Professor in the Molecular and Cellular Physiology Department at Yale Medical School. His research is focused on understanding the physiology and pathophysiology of gut–brain signalling in food intake and obesity.