Divergent Neuronal Circuitries Underlying Acute Orexigenic Effects of Peripheral or Central Ghrelin: Critical Role of Brain Accessibility

Ghrelin is an octanoylated peptide hormone that potently and rapidly increases food intake. The orexigenic action of ghrelin involves the hypothalamic arcuate nucleus (ARC), which is accessible to plasma ghrelin and expresses high levels of the ghrelin receptor. Local administration of ghrelin in a variety of other brain nuclei also increases food intake. It is currently unclear, however, whether these non‐ARC ghrelin brain targets are impacted by physiological increases of plasma ghrelin. Thus, the present study aimed to clarify which ghrelin brain targets participate in the short‐term orexigenic actions of ghrelin. First, c‐Fos induction into mouse brains centrally or peripherally treated with ghrelin was analysed. It was confirmed that peripherally administered ghrelin dose‐dependently increases food intake and mainly activates c‐Fos in ARC neurones. By contrast, centrally administered ghrelin activates c‐Fos in a larger number of brain nuclei. To determine which nuclei are directly accessible to ghrelin, mice were centrally or peripherally injected with a fluorescent ghrelin tracer. It was found that peripherally injected tracer mainly accesses the ARC, whereas centrally injected tracer reaches most brain areas known to express ghrelin receptors. Subsequently, the effects of ghrelin were tested in ARC‐ablated mice and it was found that these mice failed to increase food intake in response to peripherally administered ghrelin but fully responded to centrally administered ghrelin. ARC‐ablated mice showed patterns of ghrelin‐induced c‐Fos expression similar to those seen in control mice with the exception of the ARC, where no c‐Fos was found. Thus, peripheral ghrelin mainly accesses the ARC, which is required for the orexigenic effects of the hormone. Central ghrelin accesses a variety of nuclei, which can mediate the orexigenic effects of the hormone, even in the absence of an intact ARC.     

Diet-induced obesity attenuates fasting-induced hyperphagia

Obesity impairs arcuate (ARC) neuropeptide Y (NPY)/agouti-releated peptide (AgRP) neuronal function and renders these homeostatic neurones unresponsive to the orexigenic hormone ghrelin. In the present study, we investigated the effect of diet-induced obesity (DIO) on feeding behaviour, ARC neuronal activation and mRNA expression following another orexigenic stimulus, an overnight fast. We show that 9 weeks of high-fat feeding attenuates fasting-induced hyperphagia by suppressing ARC neuronal activation and hypothalamic NPY/AgRP mRNA expression. Thus, the lack of appropriate feeding responses in DIO mice to a fast is caused by failure ARC neurones to recognise and/or respond to orexigenic cues. We propose that fasting-induced hyperphagia is regulated not by homeostatic control of appetite in DIO mice, but rather by changes in the reward circuitry.     

Inter-relation between Interleukin (IL)-1, IL-6 and Body Fat Regulating Circuits of the Hypothalamic Arcuate Nucleus

Interleukin (IL)-1 and IL-6 are immune modulating cytokines that also affect metabolic function because both IL-1 receptor I deficient (IL-1RI^−/−) and IL-6 deficient (IL-6^−/−) mice develop late-onset obesity and leptin resistance. Both IL-1 and IL-6 appear to target the central nervous system (CNS) to increase energy expenditure. The hypothalamic arcuate nucleus (ARC) is a major relay between the periphery and CNS in body fat regulation (e.g. by being a target of leptin). The present study aimed to investigate the possible mechanisms responsible for the effects exerted by endogenous IL-1 and IL-6 on body fat at the level of the ARC, as well as possible interactions between IL-1 and IL-6. Therefore, we measured the gene expression of neuropeptides of the ARC involved in energy balance in IL-1RI^−/− and IL-6^−/− mice. We also investigated the interactions between expression of IL-1 and IL-6 in these mice, and mapped IL-6 receptor α (IL-6Rα) in the ARC. The expression of the obesity promoting peptide neuropeptide Y (NPY), found in the ARC, was increased in IL-1RI^−/− mice. The expression of NPY and agouti-related peptide (AgRP), known to be co-expressed with NPY in ARC neurones, was increased in cold exposed IL-6^−/− mice. IL-6Rα immunoreactivity was densely localised in the ARC, especially in the medial part, and was partly found in NPY positive cell bodies and also α-melanocyte-stimulating hormone positive cell bodies. The expression of hypothalamic IL-6 was decreased in IL-1RI^−/− mice, whereas IL-1ß expression was increased in IL-6^−/− mice. The results of the present study indicate that depletion of the activity of the fat suppressing cytokines IL-1 and IL-6 in knockout mice can increase the expression of the obesity promoting neuropeptide NPY in the ARC. Depletion of IL-1 activity suppresses IL-6 expression, and IL-6Rα-like immunoreactivity is present in neurones in the medial ARC, including neurones containing NPY. Therefore, IL-6, IL-1 and NPY/AgRP could interact at the level of the hypothalamic ARC in the regulation of body fat.     

Leptin Responsive and GABAergic Projections to the Rostral Preoptic Area in Mice

The adipocyte‐derived hormone leptin plays a critical role in the control of reproduction via signalling in hypothalamic neurones. The drivers of the hypothalamic‐pituitary‐gonadal axis, the gonadotrophin‐releasing hormone (GnRH) neurones, do not have the receptors for leptin. Therefore, intermediate leptin responsive neurones must provide leptin‐to‐GnRH signalling. We investigated the populations of leptin responsive neurones that provide input to the rostral preoptic area (rPOA) where GnRH cell bodies reside. Fluorescent retrograde tracer beads (RetroBeads; Lumafluor Inc., Naples, FL, USA) were injected into the rPOA of transgenic leptin receptor enhanced green fluorescent protein (Lepr‐eGFP) reporter mice. Uptake of the RetroBeads by Lepr‐eGFP neurones was assessed throughout the hypothalamus. RetroBead uptake was most evident in the medial arcuate nucleus (ARC), the dorsomedial nucleus (DMN) and the ventral premammillary nucleus (PMV) of the hypothalamus. The uptake of RetroBeads specifically by Lepr‐eGFP neurones was highest in the medial ARC (18% of tracer‐labelled neurones Lepr‐eGFP‐positive). Because neurones that are both leptin responsive and GABAergic play a critical role in the regulation of fertility by leptin, we next focussed on the location of these populations. To address whether GABAergic neurones in leptin‐responsive hypothalamic regions project to the rPOA, the experiment was repeated in GABA neurone reporter mice (Vgat‐tdTomato). Between 10% and 45% of RetroBead‐labelled neurones in the ARC were GABAergic, whereas uptake of tracer by GABAergic neurones in the DMN and PMV was very low (< 5%). These results show that both leptin responsive and GABAergic neurones from the ARC project to the region of the GnRH cell bodies. Our findings suggest that LEPR‐expressing GABA neurones from the ARC may be mediators of leptin‐to‐GnRH signalling.     

Endogenous and exogenous ghrelin enhance the colonic and gastric manifestations of dextran sodium sulphate-induced colitis in mice

Abstract Ghrelin is an important orexigenic peptide that not only exerts gastroprokinetic but also immunoregulatory effects. This study aimed to assess the role of endogenous and exogenous ghrelin in the pathogenesis of colitis and in the disturbances of gastric emptying and colonic contractility during this process. Dextran sodium sulphate colitis was induced for 5 days in (i) ghrelin^+/+ and ghrelin^?/? mice and clinical and histological parameters were monitored at days 5, 10 and 26 and (ii) in Naval Medical Research Institute non‐inbred Swiss (NMRI) mice treated with ghrelin (100 nmol kg^?1) twice daily for 5 or 10 days. Neural contractility changes were measured in colonic smooth muscle strips, whereas gastric emptying was measured with the ¹⁴C octanoic acid breath test. Inflammation increased ghrelin plasma levels. Body weight loss, histological damage, myeloperoxidase activity and IL‐1β levels were attenuated in ghrelin^?/? mice. Whereas absence of ghrelin did not affect changes in colonic contractility, gastric emptying in the acute phase was accelerated in ghrelin^+/+ but not in ghrelin^?/? mice. In agreement with the studies in ghrelin knockout mice, 10 days treatment of NMRI mice with exogenous ghrelin enhanced the clinical disease activity and promoted infiltration of neutrophils and colonic IL‐1β levels. Unexpectedly, ghrelin treatment decreased excitatory and inhibitory neural responses in the colon of healthy but not of inflamed NMRI mice. Endogenous ghrelin enhances the course of the inflammatory process and is involved in the disturbances of gastric emptying associated with colitis. Treatment with exogenous ghrelin aggravates colitis, thereby limiting the potential therapeutic properties of ghrelin during intestinal inflammation.     

Ghrelin acts on leptin-responsive neurones in the rat arcuate nucleus

Leptin decreases food intake and increases energy expenditure in rodents by inhibiting neurones in the hypothalamic arcuate nucleus. The growth hormone secretagogue (GHS) ghrelin is known to stimulate food intake and to be the endogenous ligand for the GHS-receptor, which is strongly expressed in the arcuate nucleus, like the leptin receptor (Ob-R). In this study, we analysed the effect of systemic ghrelin administration on Fos expression in the arcuate nucleus on neurones expressing Ob-R. Injection of ghrelin (0.2 mg/kg, i.p) significantly increased the number of neurones expressing Fos protein in the ventromedial arcuate nucleus. Fifty-seven percent of all Fos-positive cells in the ventromedial arcuate nucleus were also positive for Ob-R staining. Furthermore, we investigated electrophysiologically the effect of ghrelin and leptin on the activity of arcuate neurones in an in-vitro slice preparation. Ghrelin stimulated the electrical activity dose-dependently in 80% of all cells tested (n=49) with a threshold concentration of 10(-11) M; only 8% were inhibited and 12% did not respond. The effect of ghrelin (10(-7) M) was weakly antagonized by the peptidic GHS-receptor antagonist (D-Lys3)-GHRP-6 (10(-4) M), which also showed a much weaker affinity (IC(50), 0.9 x 10(-6) M) to the GHS-receptor than ghrelin (IC(50), 0.3 x 10(-9) M). Ghrelin increased the electrical activity in 76% of all cells which were inhibited by leptin (n=17). These data show that ghrelin interacts with the leptin hypothalamic network in the arcuate nucleus. The opposite effect of leptin and ghrelin on neurones in the arcuate nucleus may serve as a neurophysiological correlate of the orexigenic and anorectic effects of ghrelin and leptin.     

Effect of Deletion of Ghrelin‐O‐Acyltransferase on the Pulsatile Release of Growth Hormone in Mice

Ghrelin, a gut hormone originating from the post‐translational cleavage of preproghrelin, is the endogenous ligand of growth hormone secretagogue receptor 1a (GHS‐R1a). Within the growth hormone (GH) axis, the biological activity of ghrelin requires octanoylation by ghrelin‐O‐acyltransferase (GOAT), conferring selective binding to the GHS‐R1a receptor via acylated ghrelin. Complete loss of preproghrelin‐derived signalling (through deletion of the Ghrl gene) contributes to a decline in peak GH release; however, the selective contribution of endogenous acyl‐ghrelin to pulsatile GH release remains to be established. We assessed the pulsatile release of GH in ad lib. fed male germline goat^?/? mice, extending measures to include mRNA for key hypothalamic regulators of GH release, and peripheral factors that are modulated relative to GH release. The amount of GH released was reduced in young goat^?/? mice compared to age‐matched wild‐type mice, whereas pulse frequency and irregularity increased. Altered GH release did not coincide with alterations in hypothalamic Ghrh, Srif, Npy or Ghsr mRNA expression, or pituitary GH content, suggesting that loss of Goat does not compromise canonical mechanisms that contribute to pituitary GH production and release. Although loss of Goat resulted in an irregular pattern of GH release (characterised by an increase in the number of GH pulses observed during extended secretory events), this did not contribute to a change in the expression of sexually dimorphic GH‐dependent liver genes. Of interest, circulating levels of insulin‐like growth factor (IGF)‐1 were elevated in goat^?/? mice. This rise in circulating levels of IGF‐1 was correlated with an increase in GH pulse frequency, suggesting that sustained or increased IGF‐1 release in goat^?/? mice may occur in response to altered GH release patterning. Our observations demonstrate that germline loss of Goat alters GH release and patterning. Although the biological relevance of altered GH secretory patterning remains unclear, we propose that this may contribute to sustained IGF‐1 release and growth in goat^?/? mice.     

Direct and Indirect Effects of Cannabinoids on in vitro GABA Release in the Rat Arcuate Nucleus

Within the hypothalamic arcuate nucleus, two neuronal subpopulations play particularly important roles in energy balance; neurones expressing neuropeptide Y (NPY), agouti‐related peptide (AgRP) and GABA are orexigenic, whereas neurones expressing pro‐opiomelanocortin and CART are anorexigenic. The pivotal role of these neuropeptides in energy homeostasis is well‐known, although GABA may also be an important signal because targeted knockout of the GABA transporter in NPY/AgRP/GABA neurones results in a lean, obesity‐resistant phenotype. In the present study, we describe an in vitro model of K⁺‐evoked GABA release from the hypothalamus and determine the effects of cannabinoid receptor activation. K⁺‐evoked GABA release was sensitive to leptin, insulin and PYY(3‐36), indicating that GABA was released by arcuate NPY/AgRP/GABA neurones. In the presence of tetrodotoxin (TTX), the cannabinoid CB1 receptor agonist WIN 55,212‐2 inhibited K⁺‐evoked GABA release. This was prevented by the CB1 receptor inverse agonist rimonabant. Rimonabant had no effect when applied alone. In the absence of TTX, however, the opposite effects were observed: WIN 55,212‐2 had no effect while rimonabant inhibited GABA release. This indicates that GABA release can involve an indirect, TTX‐sensitive mechanism. The most parsimonious explanation for the inhibition of GABA release by a CB receptor inverse agonist is via the disinhibition of an cannabinoid‐sensitive inhibitory input onto GABAergic neurones. One local source of an inhibitory neurotransmitter is the opioidergic arcuate neurones. In our in vitro model, K⁺‐evoked GABA release was inhibited by the endogenous opioid peptide β‐endorphin in a naloxone‐sensitive manner. The inhibitory effect of rimonabant was also prevented by naloxone and a κ‐opioid receptor selective antagonist, suggesting that GABA release from arcuate NPY/AgRP/GABA neurones can be inhibited by endogenous opioid peptides, and that the release of opioid peptides is sensitive to cannabinoids.     

Des‐Acyl Ghrelin Directly Targets the Arcuate Nucleus in a Ghrelin‐Receptor Independent Manner and Impairs the Orexigenic Effect of Ghrelin

Ghrelin is a stomach‐derived octanoylated peptide hormone that plays a variety of well‐established biological roles acting via its specific receptor known as growth hormone secretagogue receptor (GHSR). In plasma, a des‐octanoylated form of ghrelin, named des‐acyl ghrelin (DAG), also exists. DAG is suggested to be a signalling molecule that has specific targets, including the brain, and regulates some physiological functions. However, no specific receptor for DAG has been reported until now, and, consequently, the potential role of DAG as a hormone has remained a matter of debate. In the present study, we show that DAG specifically binds to and acts on a subset of arcuate nucleus (ARC) cells in a GHSR‐independent manner. ARC cells labelled by a DAG fluorescent tracer include the neuropeptide Y (NPY) and non‐NPY neurones. Given the well‐established role of the ARC in appetite regulation, we tested the effect of centrally administered DAG on food intake. We found that DAG failed to affect dark phase feeding, as well as food intake, after a starvation period; however, it impaired the orexigenic actions of peripherally administered ghrelin. Thus, we conclude that DAG directly targets ARC neurones and antagonises the orexigenic effects of peripherally administered ghrelin.     

GPR39, a receptor of the ghrelin receptor family, plays a role in the regulation of glucose homeostasis in a mouse model of early onset diet-induced obesity

GPR39, which may function as a Zn²⁺ sensor, is a member of the G protein‐coupled receptor family that also includes the receptor for the hunger hormone ghrelin. The down‐regulation of GPR39 mRNA in adipose tissue of obese type 2 diabetic patients suggests that GPR39 may contribute to the pathogenesis of the disease. The present study aimed to investigate the role of GPR39 in the regulation of energy balance and glucose homeostasis in wild‐type (GPR39^+/+) and GPR39 knockout mice (GPR39^?/?) with obesity‐related type 2 diabetes. GPR39 mRNA levels in adipose tissue of fasted GPR39^+/+ mice fed a high‐fat diet (HFD) for 30 weeks were reduced and correlated positively with blood glucose levels. Body weight, fat percentage and energy intake were increased in the HFD group but did not differ between both genotypes. Within the HFD group, blood glucose levels were lower in GPR39^?/? than in GPR39^+/+ mice, despite significant reductions in prandial plasma insulin levels. The latter may not be a result of changes in β‐cell hyperplasia because immunohistochemical staining of pancreata of mice on a HFD showed no differences between genotypes. The lower blood glucose levels may involve alterations in insulin sensitivity as revealed by glucose tolerance tests and respiratory quotient measurements that showed a preference of obese GPR39^?/? mice for the use of carbohydrates as metabolic fuel. The increase in plasma ghrelin levels in GPR39^?/? mice fed a HFD may contribute to the alterations in glucose homeostasis, whereas changes in gastric emptying or intestinal Zn²⁺ absorption are not involved. The results obtained in the present study suggest that GPR39 plays a role in the pathogenesis of obesity‐related type 2 diabetes by affecting the regulation of glucose homeostasis.     

Acute Inhibition of Central c‐Jun N‐terminal Kinase Restores Hypothalamic Insulin Signalling and Alleviates Glucose Intolerance in Diabetic Mice

The hypothalamus has been identified as a main insulin target tissue for regulating normal body weight and glucose metabolism. Recent observations suggest that c‐Jun‐N‐terminal kinase (JNK)‐signalling plays a crucial role in the development of obesity and insulin resistance because neuronal JNK‐1 ablation in the mouse prevented high‐fat diet‐induced obesity (DIO) and increased energy expenditure, as well as insulin sensitivity. In the present study, we investigated whether central JNK inhibition is associated with sensitisation of hypothalamic insulin signalling in mice fed a high‐fat diet for 3 weeks and in leptin‐deficient mice. We determined whether i.c.v. injection of a pharmacological JNK‐inhibitor (SP600125) improved impaired glucose homeostasis. By immunohistochemistry, we first observed that JNK activity was increased in the arcuate nucleus (ARC) and the ventromedial hypothalamus (VMH) in both mouse models, relative to normoglycaemic controls. This suggests that up‐regulation of JNK in these regions is associated with glucose intolerance and obesity, independent of leptin levels. Acute i.c.v. injection of SP600125 ameliorated glucose tolerance within 30 min in both leptin‐deficient and DIO mice. Given the acute nature of i.c.v. injections, these effects cannot be attributed to changes in food intake or energy balance. In a hypothalamic cell line, and in the ARC and VMH of leptin‐deficient mice, JNK inhibition by SP600125 consistently improved impaired insulin signalling. This was determined by a reduction of phospho‐insulin receptor substrate‐1 [IRS‐1(Ser612)] protein in a hypothalamic cell line and a decline in the number of pIRS‐1(Ser612) immunoreactive cells in the ARC and VMH. Serine 612 phosphorylation of IRS‐1 is assumed to negatively regulate insulin signalling. In leptin‐deficient mice, in both nuclei, central inhibition of JNK increased the number of cells immunoreactive for phospho‐Akt (Ser473) and phospho‐GSK‐3β (Ser9), which are important markers of insulin signalling. Collectively, our data suggest that the acute inhibition of central JNK improves impaired glucose homeostasis and is associated with sensitisation of hypothalamic insulin signalling.     

Phosphodiesterase‐3B‐cAMP Pathway of Leptin Signalling in the Hypothalamus is Impaired During the Development of Diet‐Induced Obesity in FVB/N Mice

The phosphodiesterase‐3B (PDE3B)‐cAMP pathway plays an important role in transducing the action of leptin in the hypothalamus. Obesity is usually associated with hyperleptinaemia and resistance to anorectic and body weight‐reducing effects of leptin. To determine whether the hypothalamic PDE3B‐cAMP pathway of leptin signalling is impaired during the development of diet‐induced obesity (DIO), we fed male FVB/N mice a high‐fat diet (HFD: 58% kcal as fat) or low‐fat diet (LFD: 6% kcal as fat) for 4 weeks. HFD fed mice developed DIO in association with hyperphagia, hyperleptinaemia and hyperinsulinaemia. Leptin (i.p.) significantly increased hypothalamic PDE3B activity and phosphorylated (p)‐Akt levels in LFD‐fed but not in HFD‐fed mice. However, basal p‐Akt levels in hypothalamus were increased in DIO mice. Additionally, amongst six‐microdissected brain nuclei examined, leptin selectively decreased cAMP levels in the arcuate nucleus (ARC) of LFD‐fed mice but failed to do so in HFD‐fed mice. We next tested whether both the PBE3B and Akt pathways of leptin signalling remained impaired in DIO mice on the HFD for 12 weeks (long‐term). DIO mice were hyperinsulinaemic and hyperleptinaemic in association with impaired glucose and insulin tolerance. Although, in LFD‐fed mice, leptin significantly increased PDE3B activity and p‐Akt levels in the hypothalamus, it failed to do so in HFD‐fed mice. Also, basal p‐Akt levels in the hypothalamus were increased in DIO mice and leptin had no further effect. Similarly, immunocytochemistry showed that leptin increased the number of p‐Akt‐positive cells in the ARC of LFD‐fed but not in HFD‐fed mice, and there was an increased basal number of p‐Akt positive cells in the ARC of DIO mice. These results suggest that the PDE3B‐cAMP‐ and Akt‐pathways of leptin signalling in the hypothalamus are impaired during the development of DIO. Thus, a defect in the regulation by leptin of the hypothalamic PDE3B‐cAMP pathway and Akt signalling may be one of the mechanisms of central leptin resistance and the development of DIO.     

Onion component,isoalliin, stimulates feeding and activates the arcuate nucleus neuropeptide Y,ghrelin- and Ninjin'yoeito-responsive neurons

Appetite loss or anorexia substantially decreases the quality of life in patients with cancer, depression and gastrointestinal disorders, and can lead to sarcopenia and frailty. Foods that restore appetite have been sought-for but are not currently available. Historically, onion intake was adopted to treat a variety of diseases with reduced appetite including cancer and gastrointestinal disturbances. While isoalliin is a core component of onion, the effects of isoalliin on feeding behavior and feeding centers remain unknown. Neuropeptide Y (NPY) and ghrelin are the most potent central and peripheral inducers of appetite. A Japanese kampo medicine Ninjin'yoeito activates ghrelin-responsive NPY neurons in the hypothalamic arcuate nucleus (ARC) and counteracts anorexia induced by an anti-cancer drug cisplatin. This study explored the effects of isoalliin on feeding behavior and activities of ARC neurons in mice. Isoalliin, injected intraperitoneally, dose-dependently increased food intake during dark phase (DP) and daily without altering light phase (LP) food intake. We measured cytosolic Ca²⁺ concentration ([Ca²⁺]_i) in single ARC neurons including NPY neurons identified by GFP fluorescence. Isoalliin increased [Ca²⁺]_i in 10 of 18 (55.6%) NPY neurons, a majority of which also responded to ghrelin with [Ca²⁺]_i increases, indicating that the ARC ghrelin-responsive NPY neuron is the major target of isoalliin. Isoalliin also increased [Ca²⁺]_i in the ARC neurons that responded to Ninjin'yoeito. These results indicate that isoalliin enhances feeding at the active period and activates ARC ghrelin-responsive NPY neurons and Ninjin'yoeito-responsive neurons. These abilities of isoalliin to stimulate DP feeding and activate ARC orexigenic neurons provide scientific evidence for the health beneficial effects of onion experienced historically and globally.     

Leptin Regulation of Agrp and Npy mRNA in the Rat Hypothalamus

Agouti-related protein (AGRP) is synthesized in the same neurones in the arcuate nucleus as neuropeptide Y (NPY), another potent orexigenic peptide. AGRP antagonizes the action of alpha-melanocyte stimulating hormone, a derivative of pro-opiomelanocortin (POMC) at the hypothalamic MC4 receptor to increase food intake. Although leptin has been shown to regulate Agrp/Npy and Pomc-expressing neurones, there are differences with respect to Agrp regulation in leptin receptor-deficient mice and rats. Unlike the obese leptin receptor-deficient db/db mouse, which exhibits upregulation of Agrp mRNA expression in the medial basal hypothalamus (MBH) compared to lean controls, the obese leptin receptor-deficient (faf; Koletsky) rat does not exhibit upregulation of Agrp expression. To determine whether this represents a general difference between leptin receptor-deficient mice and rats, neuropeptide gene expression was analysed in the MBH of lean and obese rats segregating for a different leptin receptor mutation, Leprfa (Zucker). Fasting in lean rats (+/fa) for 72 h significantly increased Agrp and Npy mRNA expression, and decreased Pomc mRNA expression as detected by a sensitive solution hybridization/S1 nuclease protection assay. Npy mRNA levels were significantly increased in fed obese fa/fa compared to lean rats, and further increased in the obese animals after fasting. In contrast, Agrp mRNA levels did not differ between fed lean and fed obese rats, and fasting did not significantly change Agrp levels in obese rats. To determine whether the change in Agrp expression that occurs with food deprivation in lean rats could be prevented by leptin replacement, Sprague-Dawley rats were fasted and infused via subcutaneous osmotic micropumps for 48 h with either saline or recombinant mouse leptin. Fasting significantly increased Agrp and Npy, and decreased Pomc mRNA levels. Leptin infusion almost completely reversed these changes such that there was no significant difference between the levels in the fasted rats and those that were fed ad libitum. Thus, in fasted lean rats, Agrp and Npy are upregulated in parallel when leptin levels fall and are downregulated by leptin infusion. By contrast, the absence of a functional leptin receptor results in the upregulation of Npy but not Agrp mRNA.