Oxytocin and appetite

The Warrior's Secret

Erectile Dysfunction Holistic Treatments

Get Instant Access

Gareth Leng*, Tatsushi Onaka, Celine Caquineau, Nancy Sabatier, Vicky A. Tobin and Yuki Takayanagi

Centre for Integrative Physiology, The University of Edinburgh, Edinburgh UK and Jichi Medical University,

Tochigi, Japan

Abstract: Oxytocin has potent central effects on feeding behaviour, as well as on social and sexual behaviours, and one likely substrate for its anorectic effect is the ventromedial nucleus of the hypothalamus. This nucleus expresses a high density of oxytocin receptors, but contains very few oxytocin-containing fibres, hence it is a likely target of 'neurohormonal' actions of oxytocin, including possibly oxytocin released from the dendrites of magnocellular oxytocin neurones. As oxytocin release from dendrites is regulated independent of electrical activity and of secretion from the neurohypophysis, exactly how this release is regulated by metabolic and reproduction-related signals remains to be established fully. Intriguingly though, it looks as though this central release of oxytocin from magnocellular neurons might be instrumental in a fundamental shift in motivational behaviour — switching behaviour from being driven by the need to find and consume food, to the need to reproduce.

Keywords: supraoptic nucleus; ventromedial hypothalamus; prolactin-releasing peptide; dendritic release

Introduction

Classically, feeding is regulated by an alternation between 'hunger' signals, which activate specific hunger centres in the hypothalamus, and 'satiety signals', which activate specific satiety centres. The hunger and satiety centres interact with each other, and are modulated according to the energy stores and momentary metabolic requirements; these are signalled by both circulating factors (such as plasma concentrations of leptin, insulin and ghrelin) and by neurally mediated signals, some of which arise from the gastro-intestinal tract. How these signals are processed depends on the animal's internal state,

* Corresponding author. Tel.: 0044 131 650 2869; Fax: 0044 131 650 2872; E-mail: [email protected]

reflecting varying motivational drives, and is sensitive to diverse environmental cues, including those arising from photoperiod.

A striking example of the complex motivational effects on appetite regulation is the mutually exclusive nature of feeding-related behaviours and sexual behaviours. Put most simply, for all animals, the most important drives are to eat and to reproduce. As it is important that each is satisfied efficiently and effectively, animals commit their behaviour to achieving either sex or food, rather than risk being both hungry and frustrated by failing to give their undivided attention to either goal. As a well-fed animal may be better able to compete for a mate and then to copulate with sustained enthusiasm, the natural course of events is that the first priority of hungry animals is to eat, and when sated they turn to sex. This seems to be achieved by a reciprocal regulation of sexual and ingestive behaviours, in which, for mammals hypothalamic oxytocin appears to play an interesting and possibly important part.

The oxytocin and vasopressin peptide lineages arose by a gene duplication that probably happened early in vertebrate evolution. Most invertebrates express a single peptide of this family, expressed in multifunctional neurons that regulate various aspects of reproduction, electrolyte balance and energy intake and utilization. In mammals, oxyto-cin neurons include magnocellular neurons that project to the neurohypophysis and which regulate lactation and parturition by oxytocin released into the peripheral circulation, and parvocellular oxytocin neurons that project to the brainstem and spinal cord, and which are involved in both control of food intake and control of penile erection. However, magnocellular neurons also release very large amounts of oxytocin within the hypothalamus, where oxytocin is reciprocally involved in the motivation for sexual behaviour (in both males and females) and in appetite.

The ventromedial nucleus of the hypothalamus (VMH) is a likely focus of some of these central motivational effects of oxytocin. It has a density of oxytocin receptors that is one of the highest in the brain, yet appears virtually devoid of innervation by oxytocin-containing fibres. This nucleus is a likely target of oxytocin released from magnocellular neurons, either by oxytocin released in an activity-dependent way from the axons of magnocellular neurons that circumnavigate the VMH, passing through areas penetrated by the dendrites of VMH neurons, or by the oxytocin that is released in massive amounts by the more distant dendrites of the magnocellular neurons. We end this review by pointing out that the massive amounts of oxytocin contained within each neurosecretory vesicle, and the relative rarity of release events, mean that central release of oxytocin cannot plausibly be organized as a signal with temporal and anatomical specificity, instead, oxytocin release in the brain is intrinsically and inevitably a hormone-like signal.

Oxytocin and oxytocin-like peptides throughout evolution

Oxytocin and vasopressin are produced by separate genes, but the genes are so closely related that they are clearly derived by duplication of a common ancestor gene at some point in our evolutionary history. At first sight, oxytocin appears to be a quintessentially mammalian hormone; released from the neurohypophysis in response to suckling, it appears to be absolutely indispensable for suckling-induced milk let-down, as well as being important for parturition (see Russell et al., 2003). All eutherian mammals make oxytocin; most marsupials (metatheria) make mesotocin, which is important in both lactation and parturition in marsupials as in eutherian mammals. Mesotocin differs from oxytocin by a single amino acid, is equally potent at the oxytocin receptor and is thought to reflect a neutral mutation.

However, it is not just mammals that make an oxytocin-like peptide; all vertebrates except for the cyclostomes (hagfishes and lamprey), have two neurohypophysial hormones, one closely related to vasopressin and one to oxytocin. In birds, meso-tocin replaces oxytocin, acting through the oxyto-cin receptor, and intracerebroventricular injections of oxytocin cause a dose-dependent decrease in food intake, feeding time and pecking frequency (Jonaidi et al, 2003).

By contrast, most invertebrates have just one oxytocin/vasopressin related peptide, with the curious exception of the mollusc cephalopods, which have both an oxytocin-like peptide (cepha-lotocin) and a vasopressin-like peptide (octopres-sin) (Takuwa-Kuroda et al., 2003). Thus oxytocin and vasopressin evolved via duplication of an ancestral gene at least 400 million years ago, before the evolution of the fishes, but possibly even earlier (hagfishes and lampreys evolved over 550 million years ago).

The arginine vasotocin/isotocin neuroendocrine system in fishes

The arginine vasotocin/isotocin neuroendocrine system regulates many aspects of fish physiology and behaviour, including circadian and seasonal biology, responses to stress, metabolism, reproduction, cardiovascular function and osmoregula-tion. Isotocin and vasotocin, the fish homologues of oxytocin and vasopressin, are similarly expressed in magnocellular neurons that project to the neurohypophysis, and are best known for their roles in oviposition (egg-laying) and osmoregulation. The neuron-specific expression of the oxytocin and vasopressin gene families, and the mechanisms mediating this cell-specificity, evolved before the divergence of the fish and mammalian lineages, as was elegantly demonstrated when Gilligan et al. (2003) generated a transgenic mouse bearing overlapping Fugu cosmids that contained the isotocin and/or vasotocin genes: the Fugu isotocin and vasotocin genes were found to be expressed specifically in the mouse oxytocin and vasopressin neurons, respectively.

Both peptides have behavioural effects as well as these hormone-like actions. In territorial fish species, vasotocin has a role in aggressive behaviour analogous to that described for vasopressin in the prairie vole (see Young et al., 2005). Isotocin also has roles in reproductive behaviour, and there is evidence for its involvement in behaviours that depend critically on social interaction. In the teleost fish Lythrypnus dalli, when a dominant male is removed from the social group, the dominant female will change sex to male, and this change is associated with a marked loss of isotocin immunoreactivity in the preoptic area of the brain (Black et al., 2004). In eels, isotocin and vasotocin are also involved in ingestive behaviour, as fibres containing these peptides innervate the esophageal sphincter muscles (Watanabe et al., 2007); here the two peptides are mutually antagonistic, and it is thought that alternating drives from the two neuronal populations promote sequential muscle contraction and relaxation during ingestion.

In cytclostomes, vasotocin (the presumed direct common ancestor of vasopressin and oxytocin) is expressed in neurons of the preoptic nucleus of the hypothalamus. These project to the neurohypo-physis, but also to other brain areas, notably to the reticular formation (Nozaki and Gorbman, 1983). The physiological role of vasotocin in cytclostomes is unclear, but it seems to have a role in energy metabolism as well as in osmoregulation (John et al., 1977).

Roles of oxytocin/vasopressin like peptides in invertebrates

In modern invertebrates, peptides related to oxytocin or vasopressin are expressed in neurons and are involved in reproduction and osmoregula-tion, but also in feeding and energy utilization. Seven such vasopressin/oxytocin-related peptides have so far been isolated in different invertebrate phyla. Annetocin was first characterized in the earthworm Eisenia foetida, and is mainly expressed in anterior neurons, in the subesophageal ganglia, and in areas involved in the regulation of reproduction (Satake et al., 1999). Injections of annetocin induce egg-laying behaviour in earthworms and in leeches, but also produce a concomitant loss of body-weight. As annetocin also stimulates gut contraction in the earthworm (Ukena et al., 1995), it may have a role in energy intake/utilization.

Conopressin is abundantly expressed in the central nervous system of most gastropods, and is also present in the penis nerve and in fibres that innervate the vas deferens; conopressin induces muscular contractions of the vas deferens and inhibits central neurons that control female reproductive behaviour. In the simultaneous hermaphrodite snail Lymnaea stagnalis, conopressin is present in most neurons of the anterior lobe of the right cerebral ganglion, most of which project to the male copulatory organ (the penile complex). During copulation, when the male copulatory organ is completely everted, these neurons are electrically active, and as stimulation of these neurons can induce eversion (De Boer et al., 1997), it seems that these neurons control the snail equivalent of penile erection.

Interactions between feeding and sex

Feeding and sexual behaviour are mutually exclusive goal-orientated behaviours. For example, consider the invertebrate C. elegans. To mate, the adult male must first find an appropriate partner (a hermaphrodite) and will wander about its environment to locate one. If an adult male is left isolated on a food source it will eventually leave the food to search for a hermaphrodite. This mate-searching behaviour is called 'leaving behaviour' (Lipton et al., 2004). In food-deprived animals, this 'leaving' behaviour is reduced, suggesting that it is responsive to nutritional status, and that a lower nutritional status will decrease sex drive. Interestingly, this mate-searching behaviour is under the control of a single gene, and this gene is regulated by an insulin-response element.

When an appropriate partner is found, mating begins when the tail of the male contacts his mate. Several subsequent behaviours are needed for successful copulation. One crucial step is the insertion of the male's two copulatory spicules into its mate, where they must remain in position until sperm transfer. Each spicule is attached to retractor muscles that maintain the spicule in the male's tail, and protractor muscles that eject the spicule during mating and keep it inside the mate until copulation is completed. The male spicule muscles are regulated by pharyngeal neurons. If the pharynx is pumping abnormally, for example in the absence of food, the pharyngeal neurons suppress the protraction of the spicules. Garcia and Sternberg (2003) found that a mutation of a gene coding for an ERG- like K+ channels will induce spontaneous protraction of spicules in the absence of any hermaphrodite cues; this spontaneous spicule protraction is food dependent. During contact with its mate, the pharynx in males will eventually stop pumping but it will continue in males where the pharyngeal neurons have been ablated. Importantly, this suggests that the same neurons that regulate pharyngeal activity (and thus feeding) also regulate mating.

Although feeding behaviour varies between herbivorous and carnivorous gastropods, in all gastropod species, feeding behaviours depend on the internal state of the animal, its current environment, and its past experience. The herbivorous Aplysia spends most of its time either feeding or mating. Simply removing food increases mating behaviour (Susswein, 1984), but isolation from potential mates inhibits feeding (Botzer et al., 1991). Conversely, pheromones from potential mates can facilitate feeding behaviour (Nedvetzki et al., 1998), suggesting that, if a potential mate is present, the animal will eat vigorously until sated, allowing it to spend the rest of the day mating, without interference from distracting hunger pangs.

Regulation of feeding in mammals

Thus, throughout evolution, there is a close reciprocal organization of appetites for food and sex. In mammals, central actions of oxytocin inhibit feeding and stimulate sexual arousal, suggesting that oxytocin may play an important role in orchestrating these behaviours. In mammals, energy intake and utilization is regulated by several interconnected populations of neurons in the hypothalamus and caudal brainstem. These express many different neuropeptides that have orexigenic or anorexigenic effects when injected centrally. Among the most important, are neurons in the arcuate nucleus that make neuropeptide Y (NPY), a potent orexigen; these are activated by circulating ghrelin (secreted from the empty stomach), and inhibited by circulating leptin (secreted from adipocytes in proportion to the body's stores of fat). Within the arcuate nucleus, NPY neurons innervate pro-opiomelanocortin (POMC) neurons (see Cone, 2005); this inhibitory projection is mediated mainly by GABA which is co-expressed with NPY. The POMC neurons in turn produce a-melanocyte-stimulating hormone (a-MSH), which has potent anorectic effects mediated mainly by hypothalamic MC4 receptors. a-MSH appears to be essential for normal regulation of food intake, as mutations of either the POMC gene or of the MC4 receptor are conspicuously associated with hyperphagia and adiposity. However, a-MSH is equally involved in sexual arousal, being a potent promoter of penile erections, again through its actions at MC4 receptors. The NPY neurons also synthesize agouti-related protein (AgRP), an endogenous antagonist at the MC4 receptors. The POMC neurons are activated by leptin and inhibited by ghrelin, both directly and indirectly via the NPY/AgRP neurons. Both POMC neurons and NPY/AgRP neurons project densely to the para-ventricular nucleus of the hypothalamus (PVN), which has several roles in the regulation of metabolism, including: (i) through its control of the sympathetic nervous system; (ii) through regulation of thyroid stimulating hormone secretion; (iii) through the regulation of ACTH secretion and production of the catabolic glucocorticoid hormones and (iv) through descending projections to the gastric regulatory centres of the caudal brainstem.

The arcuate neuronal populations are regulated not only by circulating factors, but also by ascending signals from the brainstem that mediate signals from the gut. Among the ascending projections, noradrenergic neurons of the A2 cell group in the nucleus tractus solitarii (NTS) are activated by gastric distension and by peripheral CCK, a satiety-signalling peptide secreted from the gut in the presence of food. Several peptides are co-expressed in subpopulations of the noradrener-gic neurons, and these neurons are also functionally diverse, but other peptide-expressing neurons in the NTS that do not express noradrenaline also project to the hypothalamus, and apparently also carry feeding-related signals that arise from the gut. Thus several peptide messengers act in the hypothalamus to mediate signals from the gastrointestinal tract; one of the most important being prolactin-releasing peptide (PrRP), which we will talk more of later. Blood-borne signals related to nutrient status are also detected by neurons in the area postrema; a circumventricular organ adjacent to the NTS that is outside the blood-brain barrier, and which is densely interconnected with the NTS.

Oxytocin and feeding behaviour

Within the brain, oxytocin is released from centrally projecting parvocellular neurons of the PVN, and from the soma and dendrites of magnocellular neurons of the PVN and SON. The magnocellular oxytocin regions of both nuclei are innervated by fibres containing a-MSH, and the oxytocin cells densely express mRNA for the MC4 receptor (see Sabatier, 2006). Parvocellular regions of the

PVN are innervated by both NPY- and a-MSH-containing fibres from the arcuate nucleus.

Oxytocin-containing nerve endings are present in many brain areas, but are especially dense in the brainstem and spinal cord. The first suggestion that oxytocin might be involved in regulating feeding behaviour, came from Verbalis and colleagues (e.g. Olson et al., 1991a, b) who noted that increases in plasma oxytocin concentrations are generally accompanied by reductions in food intake. Peripheral injections of oxytocin have no effect on food intake, but i.c.v. injections of oxytocin or oxytocin agonist potently inhibit feeding, and these effects are prevented by an oxytocin antagonist (e.g. Arletti et al., 1990; Lokrantz et al., 1997; see Sabatier, 2006 and Douglas et al., 2007).

There is clear evidence of a role of parvocellular oxytocin neurons in feeding. These neurons densely project to the NTS, where they innervate NTS neurons that are activated by CCK during feeding, supporting the hypothesis that oxytocin potentiates the inhibitory action of CCK on meal size, and leptin modulates this action of oxytocin (e.g. Blevins et al., 2003, 2004).

However, there is also evidence that magno-cellular oxytocin neurons are involved in regulating appetite. Fos expression is increased in the supraoptic nucleus (SON) in rats after food intake (Johnstone et al., 2006); Fos expression and oxytocin secretion are also strongly activated by gastric distension or by systemic injections of CCK (see Sabatier, 2006), and fasting reduces the expression of nitric oxide synthesis in magnocellular neurons — an effect reversed by leptin treatment (Isse et al., 1999). As nitric oxide synthase is expressed in both oxytocin cells and vasopressin cells and is produced in an activity-dependent manner, this probably reflects an inhibitory effect of fasting on magnocellular neuronal activity, though this has not been studied directly. Interestingly, a-MSH inhibits the electrical activity of magnocellular oxytocin neurons, and so reduces oxytocin secretion into the systemic circulation, but potently stimulates oxy-tocin release from neuronal dendrites. This central release of oxytocin is mediated by MC4 receptors, and the release is a consequence of peptide-induced mobilization of intracellular calcium stores. Thus it seems that a-MSH selectively stimulates central, rather than peripheral oxytocin release from magnocellular neurons (Sabatier et al., 2003; Sabatier and Leng, 2006). The dendrites of magnocellular neurons are by far the largest store of oxytocin in the brain. Each neuron has typically two or three large dendrites, together accounting for about 80% of the total cell volume of a magnocellular neuron — and, in the SON, the dendrites contain >85% of the total oxytocin content of the SON. The hypothalamus contains about 9000 magnocellular oxytocin neurons; each of their dendrites contains several thousand of the large neurosecretory vesicles in which oxytocin is packaged, and each vesicle contains about 85,000 molecules of oxytocin. These dendrites release large amounts of oxytocin semi-indepen-dently of the electrical activity of the cell bodies; although some peptides can trigger release of oxytocin from the dendrites, electrical activity of the cells does not normally evoke release, but some peptides can 'prime' the dendritic stores of peptide, making them available for release in response to electrical activity (Ludwig and Leng, 2006).

Oxytocin and sexual behaviour

Both peripheral and central oxytocin are involved in erectile function in rats (see Leng et al., 2005). During copulation, oxytocin is secreted from the neurohypophysis into the peripheral circulation with a large pulse released at ejaculation, and receptors in the rat penis seem to mediate contractility in vivo. In addition, oxytocin is released within the brain (Waldherr and Neumann, 2007) and has major effects on sexual arousal; i.c.v. injections induce penile erection and yawning in rats, whereas injections of oxytocin antagonist decreased mounting and intromission frequencies and abolished ejaculation. Parvocellular oxytocin neurons project to the spinal cord, and this is one important site at which oxytocin controls penile erection, but intrahypothalamic actions are also important; injection of oxytocin in the PVN induces penile erection, and lesions of the PVN impair oxytocin-induced penile erection. Central oxytocin also stimulates female sexual behaviour by increasing sexual receptivity and facilitating lordosis, thought to be at least in part by actions at the VMH.

The timing of central and peripheral release of oxytocin during mating is not clear, as this complex behaviour involves several distinct phases from motivation to ejaculation. Peripheral oxyto-cin is needed for the consummatory phase of sexual behaviour, but central oxytocin might also be important for the motivational phase; interestingly, it seems that the effects of a-MSH on sexual behaviour might be mediated in part by its actions upon magnocellular oxytocin neurons to induce dendritic oxytocin release (Caquineau et al., 2006).

Appetite in oxytocin-deficient mice

Oxytocin-deficient mice have normal body weight, but there are several abnormalities of ingestive behaviour in these mice. In rats, stimuli that increase oxytocin release peripherally are generally accompanied by a reduction in food intake — dehydration for example, which is a potent stimulus for oxytocin secretion as well as for vasopressin secretion, is accompanied by a profound depression of appetite. Such dehydration-induced anorexia is attenuated in oxytocin-deficient mice (Rinaman et al., 2005), while consumption of solutions that contain NaCl is enhanced (Puryear et al., 2001; Amico et al., 2003; Vollmer et al., 2006). This is not solely attributable to any specific effect on sodium appetite, as oxytocin-deficient mice will also over-consume palatable sucrose solutions (Miedlar et al., 2007), and both sweet and non-sweet carbohydrate solutions (Sclafani et al., 2007).

While mice lacking oxytocin appear generally normal in body weight, male mice lacking the oxytocin receptor (oxytocin receptor-KO mice) have an overt, though mild, late-onset obesity, being slightly heavier than wild-type mice. However, we have found no significant difference in the daily total food intake of wild-type and OTR-KO mice. As some parvocellular oxytocin neurons in the PVN project polysynaptically to brown adipose tissue (Oldfield et al., 2002), it is possible that oxytocin is involved in peripheral regulation of energy metabolism. We found that, when OTR-KO mice are placed in a cold environment (5°C), their body temperature decreases more than wild-type mice, so a reduced energy consumption might contribute to the late-onset obesity.

Because stressful stimuli activate both magno-cellular and parvocellular oxytocin neurons in the hypothalamus, we analyzed the effects of stress on food intake in male OTR-KO mice. Immobilization for 1 h reduced the amounts of food eaten over the following 24 h in both wild-type and OTR-KO mice, but there was no significant difference between the groups, suggesting that oxytocin receptors are not necessary to decrease food intake following stressful stimuli.

The total amounts of daily food intake in home cages are not significantly different between wildtype and OTR-KO mice, but OTR-KO mice eat significantly more at each meal than wild-type mice, although the meal frequency per day is not significantly different. CCK-A receptor KO mice show a very similar phenotype (Bi et al., 2004); as CCK is an important peripheral satiety-signalling peptide, it is thus possible that oxytocin receptors are involved in satiety signalling within the brain.

Although the precise neuronal source of the oxytocin that is involved in food intake remains to be determined conclusively, the magnocellular oxytocin neurons in the hypothalamus are the major source of oxytocin within the hypothalamus. Magnocellular oxytocin neurons receive a strong noradrenaline/PrRP projection from the medulla oblongata, so we examined whether this projection is involved in the activation of oxytocin neurons following stressful stimuli (Onaka, 2004) and CCK administration. We found that conditioned fear stimuli activated PrRP neurons identified retrogradely as projecting to the SON, and that neutralizing endogenous PrRP by monoclonal antibodies attenuated oxytocin secretion from the pituitary in response to conditioned fear; the increase in plasma oxytocin concentrations induced by conditioned fear was also blocked in PrRP-KO mice. In addition, application of PrRP induced somato-dendritic oxytocin release from the isolated SON in vitro. We also examined the role of PrRP following food intake. Food intake activated PrRP neurons in the medulla oblongata.

Oxytocin secretion following CCK administration was attenuated in PrRP-KO mice, suggesting that PrRP projections play a role in activating magno-cellular oxytocin neurons following food intake.

We thus examined food intake in PrRP-KO mice in more detail. PrRP-KO mice showed late-onset obesity and hyperphagia. The size of meals eaten, but not their frequency, was increased in PrRP-KO mice; essentially the PrRP-KO mice differ from wild-type mice in eating more during meals, but less between meals. Meal size is normally limited by acute signals arising from the distended stomach, and CCK is thought to be an important mediator of this, acting via CCK-A receptors on afferent gastric vagal nerve endings. CCK-induced anorexia was blocked in PrRP-KO mice, consistent with the conclusion that PrRP relays satiety signalling within the brain (Bechtold and Luckman, 2006). Furthermore, expression of PrRP mRNA is reduced in hyperphagic conditions such as lactation in rats (Bechtold and Luckman, 2007), and in streptozotocin-induced diabetic or Zucker diabetic rats (Mera et al., 2007).

The ventromedial nucleus of the hypothalamus (VMH)

The above data suggest that PrRP-oxytocin system is important for satiety signalling. The major brainstem centre for satiety signalling is the NTS; however, in the hypothalamus, the major site appears to be the VMH.

The VMH is mostly known for its involvement in the regulation of sexual and feeding behaviours in the rat, although it also regulates blood pressure and pain pathways. It is a heterogeneous nucleus that can be subdivided into dorsomedial, central and ventrolateral regions based on the expression of transcription factors, receptors and neuropep-tides (McClellan et al., 2006). For example, brain-derived neurotrophic factor (BDNF) and pituitary adenylate cyclase-activating polypeptide (PACAP) are expressed in discrete populations of neurons that are scattered throughout the nucleus, while nitric oxide synthase, somatostatin and enkephalin are expressed in neurons found mainly in the ventrolateral region of the nucleus.

The ventrolateral region of the VMH is thought to be mainly responsible for the facilitation of lordosis behaviour in female rats, and this is thought to be associated with the presence there of neurons that express estrogen receptors (Pfaff and Keiner, 1973; Pfaff and Sakuma, 1979). The dorsal VMH has been known as a satiety centre since 1940, when it was first reported that rats with lesions of the VMH display a 'voracious' appetite. One interesting subpopulation of VMH neurons, which delineates the more dorsal and central regions, uniquely expresses the nuclear receptor steroidogenic factor-1 (SF1). Deletion of SF1 results in abnormal VMH development and obesity in mice (Majdic et al., 2002), and mutations of SF1 are linked to obesity and type-2 diabetes in humans (Liu et al., 2006). Leptin excites SF1 neurons in vitro, and the actions of leptin in the VMH help to resist diet-induced obesity (Dhillon et al., 2006). Another population of VMH neurons is involved in glucose homoeo-stasis, and contains neurons that respond to changes in extracellular glucose (Song et al., 2001), these glucoreceptor neurons are also activated by leptin in vivo (Shiraishi et al., 2000).

The VMH projects to the anterior hypothalamus, the medial preoptic area and the PVN, and has the major extrahypothalamic projections to the amygdala, the bed nucleus of the stria terminalis and the periaqueductal grey. The ventrolateral VMH, which is particularly associated with sexual behaviour, projects to thalamic nuclei, the retrochiasmatic area, the raphe and the dorsomedial hypothalamus, whereas the dorsome-dial region, which is specifically important for appetite, projects to the bed nucleus of the stria terminalis, the nucleus accumbens and the medial prefrontal cortex. The medial VMH send strong excitatory inputs to proopiomelanocortin (POMC) neurons in the arcuate nucleus; this connection is dynamically regulated by nutritional state, as it is weakened by fasting (Sternson et al., 2005).

In the rat brain, the VMH is among the regions that contains the highest densities of oxytocin binding sites as shown by autoradiography (Freund-Mercier et al., 1987; Tribollet et al., 1988; Kremarik et al., 1995; Bale et al., 2001) and oxytocin receptor mRNA expression, as shown by in situ hybridization (Yoshimura et al., 1993; Bale et al., 1995; Bale and Dorsa, 1995), with particularly intense labelling in the ventrolateral region of the VMH. A similar distribution of oxytocin binding sites was found in guinea-pig (Tribollet et al., 1992), and vole (Young et al., 1996).

In the female rat, oxytocin receptors in the VMH are regulated by ovarian hormones estradiol and progesterone, and their activation by oxytocin facilitates lordosis (Schumacher et al., 1989). Estradiol treatment induces a four-fold increase in the number of oxytocin binding sites in the VMH (de Kloet et al., 1986; Coirini et al., 1989), and increases expression of oxytocin receptor mRNA (Bale and Dorsa, 1995). In estrogen-primed rats, progesterone also induced an expansion of the area occupied by oxytocin receptors into the periphery of the VMH, where dendrites of VMH neurons are found (Schumacher et al., 1989). Following focal electrolytic lesions of the VMH, oxytocin receptor binding is reduced by over 80% in the area surrounding the nucleus, indicating that most of the oxytocin receptors there are located on fibres originating within the VMH, most likely on the dendrites of VMH neurons which extend into the extranuclear region (Johnson et al., 1991). However, in the same study, the estradiol-induced increase in oxytocin binding was higher within the VMH than in the extra-nuclear region, suggesting that the cells bodies of VMH neurons are the most important site for the expression of functional oxytocin receptors.

The presence of functional oxytocin receptors in the VMH has been confirmed by electrophysiolo-gical studies. In extracellular recordings of single VMH neurons in hypothalamic slices, Kow and Pfaff (1986) reported that 58% of VMH neurons did not respond to oxytocin, 28% were excited and 14% were inhibited. In a similar study, these authors found that 94% of the oxytocin-respon-ding neurons recorded in the ventrolateral region of the VMH responded with an excitation of their electrical activity, and this excitatory response was enhanced by estradiol pre-treatment (Kow et al., 1991). Oxytocin applied on slices of guinea-pig hypothalamus excited about half of the VMH neurons tested, whereas none were inhibited (Inenaga et al., 1991).

There is still very little known about the electrophysiological behaviour of these neurons in vivo, and in particular, about the effects of appetite- and sex-related peptides such as oxytocin, on their firing pattern. We performed extracellular recordings of single VMH neurons in vivo in urethane-anaesthe-tized male rats and studied the effects of injecting 1-10 ng of oxytocin (in 1-2 ml) directly into the third ventricle. We found that 51% of 65 VMH neurons tested were excited by oxytocin (Fig. 1), while 21% were inhibited and 28% were not affected.

The abundance of functional oxytocin receptors on cell bodies and dendrites of VMH neurons contrasts strikingly with the absence of oxytocin-containing fibres and synapses innervating the nucleus. The VMH contains very few fibres that show any immunoreactivity for either oxytocin or vasopressin (Caldwell et al., 1988; Jirikowski et al.,

1988; Schumacher et al., 1989; Fig. 2), and it is not known whether the few oxytocin fibres there are 'stray' axons or dendrites of magnocellular neurons, or come from parvocellular neurons of the PVN. By marked contrast to the virtual absence of oxytocin fibres in the VMH, the region ventrolateral to the VMH is densely permeated by both vasopressin-and oxytocin-containing axons of magnocellular neurons on their way from the PVN and SON to the neurohypophysis (Daniels and Flanagan-Cato, 2000), and by the axons of parvocellular neurose-cretory neurons on their way to the median eminence. The magnocellular neuronal axons do not have any collateral branches in this region, so give rise to no conventional nerve endings and synapses. This does not however mean that no oxytocin is released from these axons; the large dense-cored vesicles can be released from any part

Interspike intervals (s) Time (s) Time (s)

Fig. 1. (A) The ventral approach for electrophysiological recordings of VMH neurons in vivo. In urethane-anaesthetized rats, the ventral surface of the brain is exposed transpharyngeally; a glass electrode is placed in the VMH for extracellular recordings of single VMH neurons, and an i.c.v. cannula is implanted into the third ventricle for central injection of oxytocin. (B) In vivo recording of the mean firing rate in a single VMH neuron. Note the increase in firing rate after i.c.v. injection of 10 ng oxytocin. (C) Interspike interval distribution of the neuron shown in (B) constructed in 10-ms bins from 5 min of stationary spontaneous discharge activity. (D) hazard function constructed from the interspike interval distribution shown on (B). The hazard function was normalized to the total hazard over the first 500 ms. (E) Event correlation histogram for the neuron recorded in (B). 3 V, third ventricle; ARC, arcuate nucleus; VMH, ventromedial hypothalamus.

Interspike intervals (s) Time (s) Time (s)

Fig. 1. (A) The ventral approach for electrophysiological recordings of VMH neurons in vivo. In urethane-anaesthetized rats, the ventral surface of the brain is exposed transpharyngeally; a glass electrode is placed in the VMH for extracellular recordings of single VMH neurons, and an i.c.v. cannula is implanted into the third ventricle for central injection of oxytocin. (B) In vivo recording of the mean firing rate in a single VMH neuron. Note the increase in firing rate after i.c.v. injection of 10 ng oxytocin. (C) Interspike interval distribution of the neuron shown in (B) constructed in 10-ms bins from 5 min of stationary spontaneous discharge activity. (D) hazard function constructed from the interspike interval distribution shown on (B). The hazard function was normalized to the total hazard over the first 500 ms. (E) Event correlation histogram for the neuron recorded in (B). 3 V, third ventricle; ARC, arcuate nucleus; VMH, ventromedial hypothalamus.

Was this article helpful?

0 0
Breaking Bulimia

Breaking Bulimia

We have all been there: turning to the refrigerator if feeling lonely or bored or indulging in seconds or thirds if strained. But if you suffer from bulimia, the from time to time urge to overeat is more like an obsession.

Get My Free Ebook


Post a comment