Oxytocin (Oxt), a neuropeptide produced in the hypothalamus, is implicated in regulation of feeding. Recent studies have shown that peripheral administration of Oxt suppresses feeding and, when infused subchronically, ameliorates hyperphagic obesity. However, the route through which peripheral Oxt informs the brain is obscure. This study aimed to explore whether vagal afferents mediate the sensing and anorexigenic effect of peripherally injected Oxt in mice. Intraperitoneal Oxt injection suppressed food intake and increased c-Fos expression in nucleus tractus solitarius to which vagal afferents project. The Oxt-induced feeding suppression and c-Fos expression in nucleus tractus solitarius were blunted in mice whose vagal afferent nerves were blocked by subdiaphragmatic vagotomy or capsaicin treatment. Oxt induced membrane depolarization and increases in cytosolic Ca2+ concentration ([Ca2+]i) in single vagal afferent neurons. The Oxt-induced [Ca2+]i increases were markedly suppressed by Oxt receptor antagonist. These Oxt-responsive neurons also responded to cholecystokinin-8 and contained cocaine- and amphetamine-regulated transcript. In obese diabetic db/db mice, leptin failed to increase, but Oxt increased [Ca2+]i in vagal afferent neurons, and single or subchronic infusion of Oxt decreased food intake and body weight gain. These results demonstrate that peripheral Oxt injection suppresses food intake by activating vagal afferent neurons and thereby ameliorates obesity in leptin-resistant db/db mice. The peripheral Oxt-regulated vagal afferent neuron provides a novel target for treating hyperphagia and obesity.
- nodose ganglion
- food intake
oxytocin (Oxt) is a neurohypophysial hormone produced in the neurons located in the paraventricular nucleus (PVN) and supraoptic nucleus (SON) of hypothalamus (6). Oxt released from the nerve terminals in the pituitary into peripheral circulation plays an important role in mammalian labor and lactation (24). Oxt is also released within various brain regions from axonal terminals, dendrites, and cell bodies of Oxt neurons, and these central Oxt induce various functions, including social behavior (12), stress responses (37), and promotion of learning and memory (48).
It has been established that Oxt signaling within the central nerves system is implicated in the regulation of food intake. Intracerebroventricular injection of Oxt suppresses food intake, whereas that of Oxt antagonist increases it (2, 3). Meal-related factors, including food intake (22), cholecystokinin (CCK) (36, 43), gastric distension (43), and activation of gastric vagal afferents (50), activate Oxt neurons in the hypothalamus and increase plasma Oxt levels (34, 50, 51). Inversely, fasting reduces Oxt mRNA expression in the hypothalamic PVN (27, 49). Knockdown of Oxt in PVN neurons results in increased food intake and body weight in mice fed normal and high-fat diets (55). The neural pathway downstream of anorexigenic Oxt involves proopiomelanocortin (31) and glucagon-like peptide-1 neurons (44) in the nucleus tractus solitarius (NTS).
Recently, Oxt draws much attention for its therapeutic potential to treat hyperphagia and obesity, as well as autism. Central subchronic infusion of Oxt decreases food intake and body weight in high-fat-diet-induced obese (DIO) mice and rats (10, 55). Interestingly, peripheral subchronic administration of Oxt also reduces food intake and weight gain in DIO mice and rats (10, 30, 33). Thus peripherally administered Oxt mimics the effects of centrally administered Oxt. Here, an important issue is the pathway that conveys the peripheral Oxt's information to the brain for inhibit feeding. The half-life of circulating Oxt is very short, around 1∼2 min (15). Furthermore, it has been reported that the transfer of circulating Oxt to the brain is tightly restricted by the blood-brain barrier (BBB), and that only 0.002% of peripherally injected Oxt reaches the central nervous system (32). Previous reports show that intraperitoneal (ip) Oxt administration immediately reduces food intake and induces c-Fos expression in areas of the hindbrain, including NTS and area postrema (AP), that are linked to the control of meal size (30, 33). The NTS and AP are the area where BBB is leaky and Oxt receptors are expressed (54). Therefore, it is plausible that peripheral Oxt could act directly on the neurons in the NTS and AP. On the other hand, NTS is the projection site of vagal afferents (16). The vagal afferent nerves sense the satiety hormones, such as CCK, glucagon-like peptide-1, and peptide YY3–36, and transmit their signals to the brain, thereby decreasing food intake (21). Moreover, vagal afferent neurons express the Oxt receptor mRNA and protein (52). Hence, we hypothesized that peripheral Oxt directly interacts with the vagal afferents, and that this interaction is relayed to signaling to the critical area of the brain and inhibition of feeding behavior.
The present study aimed to clarify whether peripheral Oxt interacts with vagal afferent neurons and, consequently, suppresses food intake. We also explored the therapeutic potential of this interaction for treatment of obesity. We investigated in mice the effects of ip Oxt injection on feeding behavior and c-Fos expression in the NTS, the area innervated by vagal afferents. Furthermore, we examined whether these effects of Oxt on feeding and c-Fos expression were counteracted by subdiaphragmatic vagotomy and systemic capsaicin (CAP) treatment. The effect of Oxt on the membrane potential and cytosolic Ca2+ concentration ([Ca2+]i) in vagal afferent neurons isolated from the nodose ganglion (NG) were measured. Finally, we examined whether peripheral Oxt administration could activate vagal afferent pathway and ameliorate hyperphagia and obesity in db/db mice with leptin resistance, a condition characteristically associated with human obesity.
RESEARCH DESIGN AND METHODS
CCK octapeptide (26–33, sulfated form, CCK-8) and Oxt were purchased from Peptide Institute (Osaka, Japan). Oxt receptor antagonist [d(CH2)51,Tyr(Me)2,Orn8]-oxytocin (H4928) was obtained from Bachem, mouse leptin from R&D Systems, and CAP from Sigma.
Male C57BL/6J mice, aged 8∼12 wk, ICR mice aged 6∼12 wk (Japan SLC, Shizuoka, Japan), and db/db and wild-type BKS mice aged 9∼12 wk (CLEA Japan, Tokyo, Japan) were used. The animals were housed in individual cages for at least 1 wk under conditions of controlled temperature (23 ± 1°C), humidity (55 ± 5%), and lighting (light on at 730 and off at 1930). Food and water were available ad libitum. Mice were sufficiently habituated to handling before the experiment of feeding and c-Fos expression. Animal experiments were carried out under approval by the Institutional Animal Experiment Committee of the Jichi Medical University, and in accordance with the Institutional Regulation for Animal Experiments and Fundamental Guideline for Proper Conduct of Animal Experiment and Related Activities in Academic Research Institutions, under the jurisdiction of the Ministry of Education, Culture, Sports, Science and Technology.
CAP treatment and subdiaphragmatic vagotomy.
To impair CAP-sensitive sensory nerves, systemic CAP treatment was performed as described (14, 47). In brief, C57BL/6J mice were anesthetized with tribromoethanol (200 mg/kg ip), followed by subcutaneous (sc) administration of CAP at 50 mg/kg body wt (5 ml/kg, solution composition: 10% ethanol, 10% Tween 80, and 80% saline). A second CAP (75 mg/kg sc) injection was performed 2 days later with the same protocol. Finally, CAP (5 mg/kg ip) was injected into the conscious mice 2 days later.
Bilateral subdiaphragmatic vagotomy was performed as described (25). In brief, a midline incision was made to provide wide exposure of the upper abdominal organ in C57BL/6J mice anesthetized with tribromoethanol. The bilateral subdiaphragmatic trunks of vagal nerves along the esophagus were exposed and cut. In the sham operation group, these vagal trunks were exposed but not cut. Vagotomized and sham-operated mice were maintained on a nutritionally complete liquid diet for a human baby (Chilmil, Morinaga, Tokyo, Japan). One to two weeks after the operation, the experiments of feeding or c-Fos expression were performed.
To confirm that CAP treatment and subdiaphragmatic vagotomy were successfully carried out, food intake was measured after ip injection of CCK-8, the substance established to inhibit food intake by activating vagal afferent nerves (46). CCK-8 (4 μg/kg) was ip injected at the light phase (1000) in 16 h-fasted mice, followed by measurement of food intake at 30 min after injection. Anorexigenic effect of CCK-8 was blunted in mice receiving CAP treatment and vagotomy compared with control and sham-operated mice, respectively [means ± SE (number) by unpaired t-test]: in the control group, saline 0.50 ± 0.020 g (9) vs. CCK 0.23 ± 0.039 g (8), P < 0.01; in CAP-treated group, saline 0.42 ± 0.016 g (6) vs. CCK 0.34 ± 0.035 g (7), not significant; in sham group, saline 1.17 ± 0.13 g (12) vs. CCK 0.28 ± 0.11 g (12), P < 0.01; in vagotomized group, saline 1.33 ± 0.20 g (10) vs. CCK 1.02 ± 0.10 g (10), not significant. In contrast, body weight was not different between CAP-treated and control mice and between vagotomized and sham-operated mice (data not shown). These results confirmed that both the CAP treatment and surgical vagotomy successfully blocked the vagal afferent function, while keeping the body's condition sound.
Measurements of food intake after single administration of Oxt.
Mice were deprived of food at 1730 with free access to water, ip injected with saline or Oxt dissolved in saline at 1920, and given standard chow (CE-2, CLEA Japan) at 1930 when the dark phase started. Vagotomized and sham-operated C57BL/6J mice were fasted overnight (16 h), ip injected with saline or Oxt at 950, and provided with the liquid diet at 1000.
Conditioned taste aversion test.
To accustom mice to a water deprivation schedule, C57BL/6J mice were allowed access to two water bottles for 2 h (1000∼1200) for 5 days. On the 6th day, mice were given 0.15% saccharine instead of water for 0.5 h and then injected ip with saline (10 ml/kg), Oxt (200 μg/kg, 10 ml/kg), or lithium chloride (0.15 M, 20 ml/kg). The 7th day was the rest day when mice were given 2 h of normal water access. The 8th day was the test day when two-bottle preference (0.15% saccharine vs. water) test was performed for 0.5 h. Conditioned taste aversion was determined as saccharine preference ratio, saccharine intake/total intake.
Measurements of c-Fos expression in medial NTS.
Oxt at 200 μg/kg was ip injected in the vagotomized and sham-operated C57BL/6J mice. At 90 min after injection, these mice were transcardially perfused with 4% paraformaldehyde under anesthesia. The brains were collected, postfixed in the same fixative for overnight at 4°C, and incubated in phosphate buffer containing 30% sucrose for 48 h. Coronal sections (40 μm) of hindbrain were cut using a freezing microtome, collected at 120-μm intervals, and processed for c-Fos immunoreactivity, as described (26). Anti-c-Fos antisera (sc-52, 1:10,000, Santa Cruz Biotechology) were used as the primary antibody. Color was developed with a nickel-diaminobenzidine solution. Neurons immunopositive to c-Fos in medial NTS (bregma −7.32 to −7.76 mm) were counted.
Preparation of single neurons from nodose ganglia.
Single neurons were isolated from mouse NGs as described (19). Briefly, NGs were treated 20 min at 37°C with 0.1∼0.5 mg/ml collagenase Ia (Sigma), 0.4∼0.6 mg/ml dispase II (Roche, Basel, Swiss), 15 μg/ml DNase II type IV (Sigma), and 0.75 mg/ml bovin serum albumin (Sigma) in HEPES-buffered Krebs-Ringer bicarbonate buffer (HKRB) composed of (in mM) 4.7 KCl, 1.2 KH2PO4, 129 NaCl, 5 KaHCO3, 1.2 MgSO4, 1.8 CaCl2, and 10 HEPES, with pH adjusted at 7.4 using NaOH supplemented with 5.6 glucose. Single neurons were cultured in Eagle's minimal essential medium containing 5.6 mM glucose supplemented with 10% fetal bovine serum, 100 μg/ml streptomycin, and 100 U/ml penicillin for 12∼24 h.
Measurements of [Ca2+]i in NG neurons.
Measurements of [Ca2+]i in primary cultured NG neurons from ICR, db/db, and BKS mice were carried out as described (19). Briefly, following incubation with 2 μM fura 2-AM (DOJINDO, Kumamoto, Japan) for 30 min at 37°C, the cells were mounted in a chamber and superfused with HKRB containing 5.6 mM glucose at 1.3 ml/min at 30°C. Fluorescence ratio images at 510 nm due to excitation at 340 and 380 nm were produced by an Aquacosmos version 2.5 (Hamamatsu Photonics, Shizuoka, Japan). For repeated administration of Oxt, 4-min pulses of Oxt at 10−11 M to 10−6 M were sequentially applied with washing periods of 8 min or longer.
To exclude possible artifactual data in [Ca2+]i experiments, the following criteria for responses were used. When [Ca2+]i changed within 5 min after addition of agents and their amplitudes were at least twice larger than the spontaneous fluctuations of the baseline, they were considered responses. Only the neurons that responded to 55 mM KCl at the end of recordings were analyzed, to secure the soundness of recorded neurons. The amplitude of [Ca2+]i response to test reagents was normalized to 55 mM KCl in each NG neuron.
Patch-clamp experiments in NG neurons.
Perforated whole cell currents recordings in single NG neurons from C57BL/6J mice were performed at room temperature (25°C) using a pipette solution containing (in mM) 40 K2SO4, 50 KCl, 5 MgCl2, 0.5 EGTA, 10 HEPES, and amphotericin B (150 μg/ml, Sigma) at pH 7.2 with KOH, as described (20). Electrodes with resistance of 3∼5 MΩ were used. Membrane potentials were recorded using an amplifier (Axopatch 200B; Molecular Devices, Foster, CA) in a computer using pCLAMP 9.2 software. Only the neurons that showed stable baseline of membrane potential for more than 3 min were analyzed. When the membrane potential changed with amplitude larger than 1.5 mV and greater than twofold standard deviations of the mean during 2 min before application, it was considered response.
Immunocytochemical identification of cocaine- and amphetamine-regulated transcript neurons.
After [Ca2+]i measurements, the cells were fixed with 4% paraformaldehyde for 2 h at room temperature and processed for immunocytochemistry for cocaine- and amphetamine-regulated transcript (CART), as described (20). Anti-CART (55–102) antibody (H-003–62, 1:10,000, Phoenix Pharmaceuticals) was used. The neurons in which [Ca2+]i was recorded were correlated with their corresponding immunocytochemical results based on the phase-contrast photographs of the neurons taken right after [Ca2+]i measurements and the photographs of the neurons after immunostaining (20).
Chronic treatment with Oxt in diabetic db/db mice.
Diabetic db/db mice aged 11 wk received sc infusion of Oxt at 1,600 μg·kg−1·day−1 or saline over 2 wk using osmotic minipumps (Alzet, model 2002). Body weight gain and food intake were measured until the 12th day after Oxt infusion.
All data were shown as means ± SE. Statistical analysis was performed by one-way ANOVA, followed by Tukey's multiple-comparison tests, unpaired or paired t-test, or χ2 test using the Prism 5 (GraphPad Software). P < 0.05 was considered significant.
IP Oxt injection decreases food intake via vagal afferents without aversive behavior.
IP injection of Oxt (200 μg/kg and 400 μg/kg) in mice decreased food intake during 0.5–6 h after injection in a dose-dependent manner (Fig. 1A), confirming previous report (30). The anorexigenic effect of Oxt at 200 μg/kg was abolished in the mice pretreated with CAP that denervates CAP-sensitive sensory nerves, including vagal afferents (Fig. 1B). In contrast, the anorexigenic effect of 400 μg/kg Oxt during 0.5–1 h after injection was markedly impaired, while that during 3–6 h remained partially in CAP-tread mice (Fig. 1B). These results suggested that ip Oxt at 200 μg/kg inhibits feeding primarily via sensory nerves, including vagal afferents, while, at a higher dose of 400 μg/kg, it recruits an additional route, which operates during 3–6 h. In the rest of our study, we used 200 μg/kg Oxt that inhibits feeding via vagal afferents.
IP Oxt injection (200 μg/kg) reduced liquid-diet intake at 0.5 and 1 h in the control mice that received sham operation (Fig. 1C). In vagotomized mice, in contrast, the anorexigenic effect of Oxt disappeared completely (Fig. 1D). Furthermore, ip Oxt injection, unlike lithium chloride, did not influence saccharine preference (Fig. 1E), indicating that Oxt did not induce conditioned taste aversion. These results demonstrate that peripheral Oxt administration decreases food intake via interacting with vagal afferents.
IP Oxt induces c-Fos expression in NTS in a vagal afferent-dependent manner.
IP injection of 200 μg/kg Oxt into sham-operated mice induced c-Fos expression in medial NTS, to which vagal afferents project (16) (Fig. 2, A, B, and E). In the vagotomized mice, in contrast, ip Oxt injection had no significant effect on the c-Fos expression in medial NTS (Fig. 2, C, D, and E). These results are consistent with the hypothesis that peripheral Oxt activates the neurons in medial NTS via a vagal afferent-dependent mechanism.
Oxt induces membrane depolarization, action potential firings, and [Ca2+]i increases in NG neurons.
We measured direct effects of Oxt on membrane potential and [Ca2+]i in the single neurons isolated from the NG, which is the assembly of cell bodies of all visceral vagal afferent neurons. Administration of Oxt at 10−7 M induced the membrane depolarization and action potential firings in a single NG neuron (Fig. 3A). Among 36 NG neurons examined, Oxt evoked membrane depolarization (3.00 ± 1.72 mV) in 7 neurons (19.4%) (Fig. 3B) and action potential firings (before 0.0639 ± 0.033 Hz vs. during Oxt 0.204 ± 0.102 Hz, P < 0.05 by paired t-test) in 4 neurons (11.1%). Oxt at 10−10 to 10−6 M increased [Ca2+]i in NG neurons, while at 10−11 M it had no effect (Fig. 3, C–E). Incidence of the [Ca2+]i response increased as a function of Oxt dose from 10−10 to 10−7 M and took a plateau value (around 15%) at 10−7 and 10−6 M (Fig. 3D). Thus Oxt increased [Ca2+]i in NG neurons in a concentration-dependent manner. The [Ca2+]i response to Oxt was markedly suppressed in the presence of an Oxt receptor antagonist [d(CH2)51,Tyr(Me)2,Orn8]-oxytocin (H4928) (Fig. 3, F–H), suggesting that Oxt acts on NG neurons via Oxt receptor.
Oxt activates NG neurons that respond to CCK-8 and CAP and that contain CART.
It is well known that systemic administration of CCK decreases food intake via directly interacting with vagal afferents (21, 46), and that CAP activates C-type (CAP-sensitive) vagal afferent neurons (13). We examined whether Oxt-responsive NG neurons are distinct from or overlap with CCK-8- or CAP-responsive neurons. Oxt (10−7 M), CCK-8 (10−8 M), and CAP (10−7 M), administered sequentially, evoked [Ca2+]i increases in 42 (16.4%), 107 (41.8%), and 166 (64.8%) of 256 NG neurons that responded to KCl, respectively (Fig. 4, A and B). Among 42 NG neurons that responded to Oxt, 41 (95%) responded to CCK-8, and 39 (93%) responded to CAP (Fig. 4, A and B). In accordance with these [Ca2+]i results, patch-clamp experiments showed that, among seven NG neurons that responded to Oxt (10−7 M) with depolarization, six (85.7%) responded to CCK-8 (10−8 M). These results suggest that Oxt primarily targets CCK- and CAP-sensitive NG neurons.
It was reported that the majority of CCK-1 receptor-expressing NG neurons co-express CART (7), suggesting that the CART neuron is involved in CCK-induced inhibition of food intake. We examined whether Oxt interacts with NG neurons containing CART. As shown in Fig. 4, C and D, 19 of 24 neurons (79.2%) that responded to Oxt with [Ca2+]i increases were shown to be immunoreactive to CART. The incidence of the CART-immunoreactive neurons was significantly (P < 0.01 by χ2 test) higher in Oxt-responsive neurons (79.2%) than in total neurons (68 of 177 neurons, 38.4%), indicating that Oxt preferentially targets CART neurons.
Oxt activates NG neurons and decreases food intake and body weight in obese diabetic db/db mice.
It is considered that peripherally administered leptin decreases food intake partly via interacting with vagal afferents (39, 41), a process possibly involved in homeostatic regulation of body weight. Leptin resistance is often associated with obesity. We investigated the effects of leptin and Oxt on vagal afferent neurons in obese Type 2 diabetic db/db mice with mutated leptin receptor and resultant leptin resistance. In NG neurons from wild-type mice, both leptin (10−8 M) and Oxt (10−7 M) evoked [Ca2+]i increases (Fig. 5A). Leptin increased [Ca2+]i in 9 out of 285 NG neurons (Fig. 5C), and all of these leptin-responsive neurons also responded to Oxt. In NG neurons isolated from db/db mice, by contrast, leptin failed to increase [Ca2+]i (Fig. 5B). Notably, Oxt evoked [Ca2+]i increases in these NG neurons of db/db mice (Fig. 5B), and the incidence and amplitude of [Ca2+]i responses to Oxt were indistinguishable between db/db and wild-type mice (Fig. 5, C and D). Thus Oxt activated leptin-resistant vagal afferent neurons of db/db mice.
Single ip injection of 200 μg/kg Oxt into db/db mice significantly suppressed food intake for 0.5–3 h after injection to an extent similar to that in normal mice (Fig. 6A vs. Fig. 1A). We examined whether subchronic administration of Oxt for 2 wk using osmotic minipumps ameliorates hyperphagia in db/db mice. The surgery and implantation of osmotic minipumps markedly and transiently decreased daily food intake to a level around 2 g/day in saline and Oxt injection groups (Fig. 6B). The daily food intake largely recovered at day 2 after the surgery/implant and thereafter became stable through day 12 at the level around 6∼8 g/day for saline and 4∼6 g/day for Oxt (Fig. 6B). Thus subchronic Oxt treatment, compared with saline treatment, significantly reduced daily food intake for the period of day 2∼12 and cumulative food intake for 12 days (Fig. 6C). Body weight gain at day 12 was −1.8 ± 0.48 g in saline-treated group and −3.2 ± 0.29 g in Oxt-treated group, showing a significant reduction of body weight by Oxt (Fig. 6D). The minus value of body weight gain at day 12 is suggested to be due to the transient decrease of daily food intake after surgery/implant. Thus peripheral Oxt injection activated vagal afferents, decreased food intake, and ameliorated obesity in obese db/db mice with leptin resistance.
In the present study, ip administration of Oxt (200 μg/kg) suppressed food intake without evoking aversive behavior and induced c-Fos expression in NTS, and these effects were blunted in the mice treated with CAP or that received vagotomy. Oxt evoked membrane depolarization, action potential firings, and [Ca2+]i increases in the single NG neurons isolated from vagal afferents. The majority of the Oxt-responsive NG neurons also responded to CCK-8 and contained CART, both of which are known to inhibit feeding through vagal afferents. In Type 2 diabetic db/db mice, a model of leptin-resistant obesity, Oxt activated NG neurons, while leptin failed to do so. The results indicate that Oxt is able to activate NG neurons under leptin resistance, a condition characteristically associated with obesity. Moreover, peripheral subchronic Oxt infusion ameliorated hyperphagia and obesity in db/db mice. These results demonstrate that peripheral administration of Oxt, at least in pharmacological doses, suppresses food intake by activating vagal afferent NG neurons and subsequent signaling to the brain, and suggest that this peripheral Oxt-regulated vagal afferent route provides a novel tool to treat hyperphagia and obesity.
It has previously been shown that a fraction of NG neurons express the Oxt receptor mRNA and protein (52). In the present study, Oxt increased [Ca2+]i in isolated NG neurons, and this response was inhibited by Oxt receptor antagonist, indicating that Oxt activates vagal afferents via Oxt receptor. Approximately 15% of NG neurons responded to Oxt. This value (15%) of the incidence of Oxt-responsive vagal afferent neurons may be reasonable, due to the following considerations. First, vagal afferents innervate diverse tissues, forming many distinct subpopulations. A particular subpopulation that innervates a particular tissue could have a distinct ability of sensing specific factors. Second, peptide YY3–36, pancreatic polypeptide, and nesfatin-1, the hormones known to decrease food intake via interacting with vagal afferents, increase [Ca2+]i in ∼10% of NG neurons (18, 19). Taken together, it is suggested that, among the highly heterogeneous NG neurons innervating diverse tissues and controlling a variety of functions, its particular subpopulation comprising 10∼15% of total NG neurons has the ability to respond to a particular hormone.
The present study showed that Oxt induces action potential firings and increases [Ca2+]i in a subfraction of NG neurons that respond to CCK-8 and leptin. Previous studies showed that the NG neuron responses to CCK-8 and leptin are implicated in inhibition of feeding. CCK, released from enteroendocrine I cells in response to dietary nutrients (11), directly activates vagal afferent neurons via CCK-1 receptor (28, 45) and thereby rapidly decreases food intake (46). Leptin is produced by the stomach epithelial cells in addition to adipocytes (4). Leptin, released from the stomach in response to food and CCK (4), activates NG neurons of vagal afferents (40) and thereby inhibits food intake for a short time (41). Moreover, leptin receptor is co-expressed with CCK-1 receptor in NG neurons (29), and CCK and leptin synergistically activate NG neurons (40) and decrease food intake (5). These findings by us and others, taken together, suggest that the Oxt activation of CCK- and leptin-responsive NG neurons is linked to inhibition of food intake. Whether Oxt and leptin also cooperatively activate NG neurons and inhibit feeding remains to be studied.
We also found that Oxt activates NG neurons containing CART, a well-known neurotransmitter of vagal afferents (7). CART is co-expressed with CCK-1 and leptin receptors in NG neurons (7, 17), and CCK-8 and leptin stimulate expression and secretion of CART in NG neurons (9, 17). The anorexigenic effect of coadministration of CCK and leptin is blunted in the rats treated with CART small interfering RNA (17), suggesting that CART serves as a neurotransmitter in the NG neurons that respond to CCK and leptin. Accordingly, the Oxt-targeted, CCK- and leptin-responsive CART neurons in vagal afferents may play a role in regulation of feeding.
The anorexigenic effect of ip Oxt at 400 μg/kg in the early phase (0.5∼1 h) was blunted by CAP treatment, whereas that in the later phase (3∼6 h) significantly remained, indicative of vagal afferent-independent route. It has recently been reported that ip injection of Oxt (450∼600 μg/kg) results in increased Oxt levels in the brain, including the hippocampus and amygdala (35), suggesting the transport of Oxt through BBB. Furthermore, the NTS, AP, and hypothalamic arcuate nucleus, the feeding-related areas, are considered to have leaky BBB and express Oxt receptors (54). Therefore, Oxt administered at a high dose (400 μg/kg) might significantly pass through BBB and act on the brain and thereby reduce feeding in the later period. However, further study is definitely needed to verify this possibility.
It remains unclear whether the anorexigenic effect of Oxt via vagal afferent pathway is pharmacological and/or plays a role in the normal regulation of daily food intake. Oxt is synthesized in the neurons in the hypothalamic PVN and SON, and released from the posterior pituitary to circulation. Plasma Oxt concentration is around 0.01∼0.1 nM in rodents and humans (23, 38, 56) and is elevated by several times during delivery and lactation (24) and postprandial periods (51). In the present study using single NG neurons, the effective concentration of Oxt to activate vagal afferents was around 0.1∼100 nM (Fig. 1, A–D), ranging from the plasma Oxt level to its 1,000-fold higher level. Hence, the plasma Oxt, when elevated during lactation and postprandial periods, may act on a portion of vagal afferent neurons. It has been reported that ip administration of Oxt in a pharmacological dose (12 μg/mouse, around 450∼600 μg/kg) results in plasma Oxt concentration around 1.5 nM (35), a dose capable of activating vagal afferent neurons in our study. Therefore, in our study, Oxt ip injection in pharmacological doses (200 and 400 μg/kg) may have reduced food intake, at least in part, via interacting with vagal afferents. More importantly, vagal afferents often terminate at the vicinity of hormone-secreting cells and sense higher concentrations of hormones in a paracrine fashion, as was previously proposed for CCK and insulin (20, 21, 42). This could also be the case for Oxt, in light of previous reports that Oxt is present in the uterus, testis, heart, pancreas, intestinal epithelium, and enteric nervous system (1, 24, 52). Moreover, ip injection of Oxt receptor antagonist was shown to induce hyperphagia during daytime in mice, suggesting that Oxt could regulate diurnal meal patterns (56). These findings by us and others suggest that endogenous Oxt in the periphery possibly interacts with the vagal afferents and regulate feeding. However, further studies are definitely required to verify the potential physiological role of the afferent vagal nerve in sensing Oxt in the circulation and/or peripheral tissues and in regulating feeding.
Obesity induces leptin resistance, a key factor that worsens obesity and leads to metabolic syndrome. Under leptin-resistant conditions, not only leptin itself, but also leptin-dependent agents, lose their ability to induce anorexigenic effect. Recent reports show that diet-induced obesity leads to leptin resistance in vagal afferents (8). Importantly, peripheral Oxt was shown to reduce food intake and body weight in DIO rats and mice (30, 33). The present study used the leptin-resistant obese db/db mice and examined the subchronic effect of Oxt released from implanted osmotic minipumps. The surgery and implantation of osmotic minipumps markedly and transiently decreased daily food intake. Daily food intake largely recovered at day 2 and thereafter became stable. Oxt infusion resulted in significant reductions in the daily food intake during the stable phase and in body weight gain at day 12. The present study shows that peripheral subchronic infusion of Oxt decreases food intake and body weight, without aversive behavior. Previous reports show that peripheral Oxt induces no adverse effects on the locomotor activity (30) and blood pressure (30). Our present and previous results indicate that the activation of vagal afferents by peripheral Oxt injection provides a promising tool to treat hyperphagia and obesity. In addition, the vagal afferent-mediated pathway could possibly serve as a signaling route of Oxt for treatment of autism (53), although further study is required to address this issue.
This work was supported by Grant-in-Aid for Young Scientist (B) (22790218, 24790221) from Japan Society for the Promotion of Science (JSPS), Jichi Medical University Young Investigator Award, and The Naito Foundation to Y. Iwasaki. A part of this study was supported by Grant-in-Aid for Scientific Research (C) (24591341) and Scientific Research on Innovative Areas (23126523) from JSPS, Memorial Foundation for Female Natural Scientists, and Kowa Life Science Foundation to Y. Maejima. This work was supported by Grant-in-Aid for Scientific Research (B) (23390044) and for Challenging Exploratory Research (26670453) from JSPS, Strategic Research Program for Brain Sciences (10036069) by the Ministry of Education, Culture, Sports, Science and Technology of Japan (MEXT), MEXT-Supported Programs for Strategic Research Foundation at Private Universities 2011–2015 (Cooperative Basic and Clinical Research on Circadian Medicine) and 2013–2017, Health Labor Sciences Research Grants from the Ministry of Health, Labor, and Welfare, Japan, a grant from Japan Diabetes Foundation, and a grant from Salt Science Research Foundation (no. 1434) to T. Yada. This study was subsidized by JKA through its promotion funds from KEIRIN RACE to T. Yada.
No conflicts of interest, financial or otherwise, are declared by the author(s).
Author contributions: T.Y., Y.I., and Y.M. conception and design of research; Y.I., Y.M., S.S., M.Y., T.A., K.K., P.K., and M.K. performed experiments; T.Y., Y.I., Y.M., S.S., and M.K. analyzed data; T.Y., Y.I., Y.M., H.N., and M.K. interpreted results of experiments; T.Y., Y.I., Y.M., and M.K. prepared figures; T.Y., Y.I., and S.S. drafted manuscript; T.Y., Y.I., S.S., and M.K. edited and revised manuscript; T.Y. and Y.I. approved final version of manuscript.
The authors thank Dr. Yukari Date at University of Miyazaki and Dr. Shuichi Koda at Asubio Pharma for advice on vogotomy. We thank Kaori Tsubonoya, Chizu Sakamoto, Minako Warashina, Seiko Ookuma, Miyuki Kondo, Megumi Motoshima, Atsumi Shinozaki, and Yuka Hobo at Jichi Medical University for technical assistance.
- Copyright © 2015 the American Physiological Society