Coming Full Circle: Contributions of Central and Peripheral Oxytocin Actions to Energy Balance

Abstract

The neuropeptide oxytocin has emerged as an important anorexigen in the regulation of energy balance. Its effects on food intake have largely been attributed to limiting meal size through interactions in key regulatory brain regions such as the hypothalamus and hindbrain. Pharmacologic and pair-feeding studies indicate that its ability to reduce body mass extends beyond that of food intake, affecting multiple factors that determine energy balance such as energy expenditure, lipolysis, and glucose regulation. Systemic administration of oxytocin recapitulates many of its effects when administered centrally, raising the questions of whether and to what extent circulating oxytocin contributes to energy regulation. Its therapeutic potential to treat metabolic conditions remains to be determined, but data from diet-induced and genetically obese rodent models as well as application of oxytocin in humans in other areas of research have revealed promising results thus far.

The physiological effect of oxytocin on uterine contraction has been recognized since the early 20th century, inspiring its Greek name, ὼκυτοκíνη, or “quick birth” (1). Its classical role as a neurohypophysial hormone extends to other functions such as lactation and osmoregulation, and the discovery of local actions within the central nervous system (CNS) has unveiled further involvement in social and reproductive behaviors (2–4). Oxytocin also participates in the regulation of energy balance, and this rapidly growing area of oxytocin physiology has generated intense interest in its relatively brief history. This review provides an overview of aspects of energy regulation in which oxytocin has been implicated as well as brings to focus some major outstanding questions yet to be addressed. Our aim is not to catalog each avenue through which oxytocin may affect energy balance—the complexity of oxytocin physiology and its interactions with other physiological systems span far beyond that which can be accomplished in a short review—but to instead give some direction to move the field toward better understanding this complexity.

Oxytocin secretion into circulation has long been known, but its central release, identified today as originating from parvocellular cells of the paraventricular nucleus of the hypothalamus (PVN) and somatodendrites of magnocellular cells in the PVN and supraoptic nucleus of the hypothalamus (SON) (5, 6), suggests additional functions. In 1989, Arletti and colleagues (7) first demonstrated the ability of oxytocin to reduce food intake in rats. Since then, numerous studies have associated alterations in oxytocin signaling with changes in energy balance. For example, peak circulating oxytocin levels across a 24-hour period generally correspond to typical patterns of food intake in mice (8), and hypothalamic oxytocin mRNA expression is reduced with fasting and restored upon refeeding (9). Ingestion of food also induces expression of c-Fos protein in hypothalamic oxytocin neurons in rats (10) and elevates their plasma oxytocin levels (11). In addition, deficient oxytocin mRNA and protein in the PVN appears to explain the hyperphagic obesity observed in single-minded 1 (Sim1) haploinsufficient mice (9) such that the amount of oxytocin mRNA deficiency correlates with the severity of hyperphagic obesity, and these deficits can be rescued with oxytocin treatment (12). Mutations of SIM1 are also associated with severe obesity in humans (13, 14), and in patients with Prader-Willi syndrome, a disease characterized by extreme hyperphagia and obesity, the number and size of PVN oxytocin neurons are reduced (15).

Although only one isoform of oxytocin receptor (OTR) has been identified, associations between OTR gene variants and social psychopathologies (16) raise the question of whether OTR polymorphisms might also affect metabolic phenotype. OTRs are typical G protein-coupled receptors primarily coupled to phospholipase C-β via G_q-binding proteins, triggering production of inositol trisphosphate and 1,2-diacylglycerol that lead to release of intracellular Ca²⁺ and phosphorylation of target proteins, respectively (17). In smooth muscle cells, increases in Ca²⁺ can lead to formation of Ca²⁺-calmodulin complexes that trigger myosin light-chain kinase activity to initiate smooth muscle contraction (18). In neurosecretory cells, elevated Ca²⁺ levels alter gene transcription and protein synthesis as well as modulate cellular excitability, firing patterns, and transmitter release. Thus, genetic variation in OTR carries the potential for widespread effects that may include energy regulation.

Contributions to Energy Regulation

Although oxytocin's ability to reduce food intake is now well established, the mechanisms underlying these particular actions are less defined. Studies in rodents indicate oxytocin regulates food intake by limiting meal size, particularly through interactions with regulatory processes in the hindbrain. For example, administration of an oxytocin antagonist alone stimulates intake of chow (7, 8, 19) as well as of glucose (20) and sucrose meals (21), resulting in increased meal sizes (19, 20, 22). These findings suggest that under normal conditions, endogenous oxytocin inhibits additional food intake that would result in consumption of larger meals. In addition, functional and anatomical data provide evidence for an interaction between oxytocin and hindbrain processes that regulate meal size. Oxytocin fibers comprise approximately 6% of PVN projections to the dorsal vagal complex (DVC), which includes the nucleus of the solitary tract (NTS) (23, 24), a region that integrates satiety signals generated by the presence of food in the gastrointestinal tract, such as cholecystokinin (CCK) and gastric distention, to limit meal size (25–30) (Figure 1). Plasma oxytocin levels are elevated in response to these same signals that activate the NTS (11, 31–33, but see Ref. 34), and release of oxytocin from descending PVN-NTS projections appears to enhance visceral afferent transmission to the NTS (35). In fact, knife cuts that sever PVN-hindbrain projections results in hyperphagia and obesity (36). Moreover, the anorectic effects of exogenously administered CCK-8 are blunted when oxytocin signaling in the hindbrain is impaired, demonstrating that oxytocin contributes to CCK-induced satiety (37, 38).

Figure 1.

A schematic of selected circuitry involved in oxytocin regulation of energy balance. Oxytocin release within the CNS and from the neurohypophysis stimulates further release in a positive-feedback manner (shown in black arrows) and modulates metabolic processes in both central and peripheral tissues (shown in gray) that result in reduced food intake and body mass. Abbreviation: DVC, dorsal vagal complex.

Considered together, these findings strongly suggest that oxytocin regulation of food intake is mediated at least in part through reductions in meal size, likely by enhancing the effectiveness of satiety signals such as CCK. It is important to note that oxytocin antagonism has been shown to decrease latency to feeding in fasted rats (7), indicating that oxytocin does not require ingestion of food to induce satiety. In fact, a role for oxytocin in maternal hyperphagia has been attributed to reduced central dendritic release and excitability of oxytocin neurons in midpregnancy, suggesting that tonic oxytocin signaling maintains normal levels of food intake (39). Furthermore, OTRs are found in mesocorticolimbic circuits and activation of these receptors can interfere with reward-related processes (40–43). Given oxytocin's modulatory role in motivated behaviors (eg, social attachment and drug addiction), oxytocin may also alter the incentive for food intake. Whether oxytocin affects feeding by modulating the salience and rewarding properties of food stimuli remains to be tested, but the presence of OTRs in reward circuits suggests oxytocin is capable of inhibiting feeding by altering appetitive, in addition to consummatory, components of ingestive behavior.

Oxytocin also reduces body mass, but not necessarily through suppression of food intake. Chronic oxytocin treatment in rodents has been observed to produce sustained reductions in body mass despite relatively transient changes in food intake (44, 45). Consistent with these findings, mice lacking oxytocin or OTRs develop late-onset obesity without considerable changes to daily intake of chow (46–48), and pair-feeding studies corroborate the finding that reductions in food intake only partially account for oxytocin's ability to decrease body mass (44, 49). Interestingly, reductions in body mass have been reported in mice in which most of their forebrain OTRs were deleted postnatally (50), but the extent to which OTRs were deleted from energy-regulatory areas (eg, hypothalamus) was not discussed, and it is unclear whether compensatory up-regulation of OTRs in non-forebrain regions such as the hindbrain and periphery contributed to the lean phenotype. Nonetheless, the absence of a hyperphagic phenotype in such genetic models is inconsistent with pharmacological studies demonstrating oxytocin effects on food intake. Genetic studies inducing global loss of ligands or their receptors do well to elucidate contributions of endogenous signaling to gross phenotypes, but they offer little spatial resolution to locate major sites of action and often fail to account for compensatory mechanisms throughout development. Indeed, OTRs are expressed throughout the CNS, offering widespread means through which global genetic knockdown may result in nonspecific effects. In rats, this includes regions of the olfactory system, cortex, basal ganglia, limbic system, thalamus, hypothalamus, brainstem, and spinal cord, and expression patterns have been found to vary greatly throughout development and across species (17). Thus, future efforts combining timed and localized ablation and/or re-expression of OTRs will allow for better interpretation of pharmacological effects that may be supraphysiological but nonetheless support a role for oxytocin in regulating food intake.

Regardless, the above findings indicate reductions in body mass may be achieved through mechanisms aside from food intake, such as changes in energy expenditure or assimilation efficiency. Although direct effects of oxytocin on energy assimilation are presently unknown and effects of oxytocin on gastric motility have been contradictory (51–57), several lines of evidence suggest changes in energy expenditure contribute to oxytocin's ability to reduce body mass. Oxytocin increases heart rate, body temperature, and oxygen consumption in mice (8, 58, 59), and daily oxytocin treatment in rats prevents the decrease in energy expenditure associated with weight loss (49). Although mice lacking OTRs demonstrate no change in spontaneous motor activity relative to wild-type mice, they exhibit a greater decrease in core body temperature after cold exposure (48). Morphology of interscapular brown adipose tissue also suggests hypoactivity in this tissue. These findings, coupled with the existence of direct PVN oxytocin projections to spinal cord (24) and polysynaptic connections to brown adipose tissue (60) and the stellate ganglion (61), lend support to a role of oxytocin in sympathetic regulation of thermogenesis and cardiac activity (62–64). Oxytocin is itself regulated by other factors such as glucocorticoids, sex steroids, and synaptotagmin-4, an atypical modulator of synaptic exocytosis. Mice lacking synaptotagmin-4 demonstrate increased oxygen consumption and resistance to developing diet-induced obesity (DIO) (58), and oxytocin interactions with the hypothalamo-pituitary-adrenal axis and steroid hormones during pregnancy and lactation (65) likely contribute considerably to maternal metabolism.

Central vs Peripheral Effects

The above findings indicate that at least a portion of oxytocin's effects on energy balance are mediated within the CNS, but to what extent? Oxytocin affects gastric motility, but there is evidence for both central and peripheral mediation (51, 66, 67). Energy expenditure may also be regulated upon activation of either central or peripheral receptors. At the very least, in vitro studies suggest circulating oxytocin can regulate adiposity in the absence of sympathetic innervation. OTRs are found in primary and cultured adipocytes (68–70), and oxytocin directly activates isolated adipocytes and human multipotent adipose-derived stem cells (71, 72). Oxytocin has been shown to negatively regulate adipogenesis in the latter, and application of oxytocin to differentiated 3T3-L1 adipocytes increases mRNA expression for proteins involved in lipolysis and fatty acid β-oxidation as well as the anorectic lipid signaling molecule oleoylethanolamide (44). Consistent with these effects, fat mass and adipocyte size are reduced in mice under chronic systemic treatment with oxytocin (45), whereas OTR-deficient mice exhibit increased fat mass and adipocyte size relative to wild-type (48). These data, along with those from pair-feeding studies (44), indicate that effects on adiposity are independent of changes in food intake and are mediated in part by a peripheral mechanism.

In addition, there are data to support a role for circulating oxytocin in glucose regulation. Mice lacking oxytocin exhibit decreased insulin sensitivity and impaired glucose tolerance (47), and daily injections of peripheral oxytocin improves glucose tolerance in mice maintained on a high-fat diet (8, 45). Importantly, oxytocin can directly stimulate insulin and glucagon release from mouse islets (73) and rat pancreas (74), which argues in favor of a peripheral contribution to glucose regulation (although oxytocin may also promote insulin secretion via vagal cholinergic neurons innervating β-cells) (75).

Certainly, oxytocin decreases food intake and body mass when administered peripherally (7, 8, 44, 45, 49, 76, 77), but it is unclear whether the relevant sites of action for these effects are in the brain, periphery, or both. The ability to address this question is complicated by the fact that oxytocin is one of the few hormones that acts in a positive feedback manner to stimulate its own release. Circulating oxytocin may trigger further release from the pituitary, but this does not preclude stimulation of local release within the brain, and central oxytocin may itself be self-stimulating. In fact, there is evidence to support each of these possibilities. Systemically administered oxytocin elicits release of oxytocin from ex vivo slices of PVN, indicating that circulating oxytocin can stimulate central release (8). Moreover, peripherally administered oxytocin can mimic its central actions. Intraperitoneal injection of oxytocin increases c-Fos immunoreactivity in oxytocin-synthesizing neurons in the PVN and SON (43, 78) as well as neurons in the NTS (8, 45, 49, 78) and ventromedial hypothalamus (8), two brain regions directly innervated by oxytocin neurons. When administered centrally, oxytocin dose-dependently increases oxytocin mRNA expression in rat hypothalamus as well as plasma oxytocin levels (44), and electrical stimulation of the PVN likewise leads to increased oxytocin in the blood and cerebrospinal fluid (79).

Given what is known of oxytocin circuitry, a substantial component of oxytocin's effects on food intake is likely mediated centrally. Oxytocin projections from the PVN to NTS provide the sole source of oxytocin to this region (80), and reduction of hindbrain oxytocin signaling blunts the ability of CCK-8 to decrease food intake in rats (37, 38). Reduced PVN oxytocin signaling via intra-PVN injections of oxytocin shRNA lentivirus increases the daily amount of food consumed in mice (58), supporting a role of oxytocin in regulating food intake. Central oxytocin signaling also contributes to the effects of other anorexigens. Leptin, a cytokine that among other functions regulates energy balance, activates parvocellular PVN oxytocin neurons (22) and increases expression of oxytocin mRNA in PVN (81) through what appears to be a melanocortin-dependent mechanism (9, 82). Third ventricular administration of an OTR antagonist blunts leptin's anorectic effects and abolishes its ability to enhance CCK-8 activation of the NTS (22), as well as reverses the anorectic effects of α-MSH, a natural ligand to melanocortin receptors (83).

Although centrally released oxytocin appears critical to its anorectic effects, systemically administered oxytocin is nonetheless able to reduce food intake as well. This, coupled with the observation that gastric distention (33) or administration of CCK (11) elicits plasma oxytocin secretion, indicates that circulating oxytocin is somehow linked to cessation of meals. The rise in plasma oxytocin may be required for execution of its peripheral actions or for further stimulation of central pathways. Given that oxytocin concentrations are over 3-fold higher in rat hypothalamus than in heart and uterus, and oxytocin levels in pituitary extracts are over 250 times higher than in heart perfusate (84), the primary source of circulating oxytocin likely derives from the neurohypophysis. Although circulating oxytocin may have restricted entrance into the brain, some studies indicate oxytocin does cross the blood-brain barrier (85, 86), and regions where the blood-brain barrier is leaky or absent (eg, median eminence and area postrema) may serve as sites of oxytocin uptake. Regardless, the contribution of circulating oxytocin to energy balance is largely unknown and warrants further investigation. Given recent advances in the ability to produce designer transgenic models, the role of peripheral oxytocin in regulating energy balance can be elucidated further through genetic manipulations, such as timed conditional deletion of OTRs in select peripheral tissues and re-expression of OTRs in neurons and select neuronal populations in oxytocin-null mice.

The Next Obesity Drug?

Thus far, oxytocin offers an intriguing potential strategy to circumvent leptin resistance and maintain appropriate levels of downstream regulation. Oxytocin treatment remains effective in reducing food intake and body mass in rodents with DIO-elicited impairment of leptin signaling (8, 45, 49) in fatty Zucker and Koletsky rats with genetically defective leptin signaling (49, 87) and in obese mice with Sim1 haploinsufficiency (9). In fact, DIO mice and rats were reported to be more sensitive to oxytocin treatment than lean controls. A lower dose of oxytocin was required to obtain significant reductions of body mass in DIO rats, and these animals responded to oxytocin treatment with greater induction of c-Fos protein in hindbrain neurons that regulate feeding (49). In addition, mice maintained on a high-fat diet demonstrate prolonged anorectic responses to oxytocin treatment relative to chow-fed controls (45). Interestingly, OTRs are observed in high- and low-affinity states, and cholesterol is an essential component for high-affinity oxytocin binding (17). Addition of cholesterol to media containing solubilized OTRs increases high-affinity oxytocin binding and stabilizes the receptor in a high-affinity state (88, 89). This raises an intriguing, albeit speculative, question of whether high cholesterol levels explain why oxytocin treatment is more effective in DIO rodents.

As with most other G protein-coupled receptors, OTRs may undergo rapid desensitization after persistent stimulation (90), raising the risk of developing tolerance, but studies so far indicate oxytocin treatment remains effective when administered somewhat chronically (ie, 7–17 days) (44, 45, 49). Further investigation is needed to determine whether oxytocin remains effective when administered long-term and the extent to which findings in rodents translate to human application. Clearly, many precautions must be taken prior to its use as a human therapeutic. For example, oxytocin affects blood pressure (91, 92) and cardiac muscle contractility (92), has analgesic effects (93), and generally would not be advised for use in pregnant women. However, prelabor uterine sensitivity to oxytocin results from a more than 150-fold increase in myometrial OTRs (94), and this induction of OTRs is primed by high estrogen levels just prior to term (95). Thus, administration of oxytocin in nonpregnant populations has little potential of causing inappropriate uterine contractions. Importantly, successful use of intranasal oxytocin therapy in humans has already been demonstrated in both sexes, with positive outcomes such as increased trust, altruism, encoding of positive social memories, and improvement of psychiatric symptoms and cardiac control (96–100). Such findings provide encouraging support for the use of oxytocin in clinical treatment of metabolic conditions.

Conclusions

The contribution of oxytocin in energy regulation is a burgeoning area of research that has made considerable progress in a relatively short period of time. Oxytocin is clearly capable of altering energy balance through several means, including food intake, energy expenditure, and adiposity levels. However, the mechanisms of action for many of these effects remain unclear, and the relative contribution of centrally and peripherally mediated effects requires further clarification. Additional studies in humans should provide much needed insight for oxytocin's potential as a therapeutic agent in treating metabolic conditions. Oxytocin physiology is indeed complex, and as parallels among aspects of oxytocin physiology (eg, social bonding and drug reward) and interactions between other systems (eg, glucocorticoids and sex steroids) come to light, an integrative reflection on what is known of oxytocin will be necessary to fully understand the nature and extent to which oxytocin affects energy balance.