Mechanism of Ghrelin-Induced Gastric Contractions in Suncus murinus (House Musk Shrew): Involvement of Intrinsic Primary Afferent Neurons

Here, we have reported that motilin can induce contractions in a dose-dependent manner in isolated Suncus murinus (house musk shrew) stomach. We have also shown that after pretreatment with a low dose of motilin (10−10 M), ghrelin also induces gastric contractions at levels of 10−10 M to 10−7 M. However, the neural mechanism of ghrelin action in the stomach has not been fully revealed. In the present study, we studied the mechanism of ghrelin-induced contraction in vitro using a pharmacological method. The responses to ghrelin in the stomach were almost completely abolished by hexamethonium and were significantly suppressed by the administration of phentolamine, prazosin, ondansetron, and naloxone. Additionally, N-nitro-l-arginine methylester significantly potentiated the contractions. Importantly, the mucosa is essential for ghrelin-induced, but not motilin-induced, gastric contractions. To evaluate the involvement of intrinsic primary afferent neurons (IPANs), which are multiaxonal neurons that pass signals from the mucosa to the myenteric plexus, we examined the effect of the IPAN-related pathway on ghrelin-induced contractions and found that pretreatment with adenosine and tachykinergic receptor 3 antagonists (SR142801) significantly eliminated the contractions and GR113808 (5-hydroxytryptamine receptor 4 antagonist) almost completely eliminated it. The results indicate that ghrelin stimulates and modulates suncus gastric contractions through cholinergic, adrenergic, serotonergic, opioidergic neurons and nitric oxide synthases in the myenteric plexus. The mucosa is also important for ghrelin-induced gastric contractions, and IPANs may be the important interneurons that pass the signal from the mucosa to the myenteric plexus.


Introduction
Ghrelin is an orexigenic peptide isolated from the stomach as an endogenous ligand for growth hormone secretagogue receptors (GHS-Rs) [1]. Ghrelin is mainly involved in the stimulation of growth hormone secretion [1] and food intake [2] and forms a peptide family with motilin owing to their similarities in not only peptides but also receptors [3]. Moreover, endogenous ghrelin is generally considered to regulate gastric contractions in the interdigestive state [4]; administration of ghrelin stimulates gastric contractions in rats [5], mice [6], and humans [7] and may be an alternative to motilin for gastrointestinal (GI) motility in motilinlacking rodents [3]. However, the mechanism underlying the role played by ghrelin in the regulation of fasted motor activities in the GI tract is not fully understood.
Functional analyses of ghrelin-induced GI motor-stimulating actions both in vivo and in vitro have suggested 2 main mechanisms for these responses. Ghrelin stimulates fasted intestinal motor activity in rats through ghrelin receptors on vagal afferent nerves [8]. Moreover, the gastric motor-stimulating action of ghrelin in rats shows vagovagal sensitivity [5,9]. The expression of ghrelin receptors in the nodose ganglion [10,11] and the capability of ghrelin to modify the discharges of afferent vagal neurons [12] also support the essential role of a vagovagal reflex pathway in ghrelininduced responses. In addition to this reflex pathway is a mechanism via direct activation of the enteric nervous system in ghrelinstimulated contraction. In rats and mice, the gastroprokinetic activity of ghrelin is observed in vitro as an increase in neuronally mediated contractions evoked by electrical field stimulation (EFS) [9,13,14,15,16,17], and a ghrelin-induced fasted motor pattern has also been observed in vagotomized rats [8]. Together, these results suggest that at least one of the target sites of ghrelin in rodents is the enteric nervous system. However, the phenotypes of ghrelin-sensitive enteric nerves have not been clearly described to date.
One explanation for the gap in information is that the effects of ghrelin activity have thus far been investigated using EFS systems in the case of smooth muscle preparations [9,13,14,15,16,17]. Therefore, the presence of a complete neural package in the stomach has not been studied. Moreover, the actions of ghrelin are species dependent, similar to those of the ghrelin-related peptide motilin. For example, ghrelin does not stimulate canine and rabbit GI motility [7,18] but induces gastric contractions in rats, mice, and humans, and although motilin stimulates GI motility in rabbits [19], dogs [20], and humans [21], it has no effect in mice and rats. To address these dissimilarities, we used Suncus murinus (house musk shrew) in an organ bath study. S. murinus belongs to the order Insectivora, family Soricidae, and this order of animals is considered one of the key groups for understanding the origin of mammals [22,23]. We have already identified the complementary DNA sequences of suncus motilin and ghrelin in S. murinus using polymerase chain reaction cloning [24,25]. We have also identified GHS-R and G protein-coupled receptor 38 genes in S. murinus [26]. Moreover, we have studied the contractile properties of the stomach in conscious, free-moving S. murinus as well as in organbath experiments and found that S. murinus has GI motility that is almost identical to that in humans and dogs [24,27]. We have also published the mechanism of motilin-induced gastric contractions in the S. murinus stomach [28].
Recently, we demonstrated that ghrelin can induce gastric contractions after pretreatment with a low dose of motilin, and this coordination of motilin and ghrelin may be necessary for the initiation of phase III contractions [29]. However, the mechanism and neural pathway of that synergistic effect in the enteric nervous system is unknown. To clarify this point, we investigated the mechanism of ghrelin-induced contractions in vitro using the whole stomach of S. murinus.

Animals
The experiments were conducted using female suncuses (weight, 45-75 g) older than 5 weeks, from an outbred KAT strain established from a wild population in Kathmandu, Nepal [30]. Animals were housed individually in plastic cages equipped with empty cans for nest boxes. The animals were given food (trout pellets; Nippon Formula Feed Manufacturing Co. Ltd., Yokohama, Japan) and water ad libitum. The animal room was maintained at 21uC to 24uC and the light-dark cycle was controlled to change every 12 h (lights on from 0800 to 2000 h). All procedures were approved by and performed in accordance with the guidelines of the Saitama University Committee on Animal Research. All efforts were made to minimize animal suffering and reduce the number of animals used in the experiment.

Preparation of Suncus Isolated Stomach
After being deeply anesthetized with diethyl ether, the animals were killed via decapitation, and their stomachs were immediately placed into freshly prepared Krebs solution (composition in mM: NaCl, 118; KCl, 4.75; CaCl 2 , 2.5; MgSO 4 , 1.2; NaH 2 PO 4 , 1.8; NaHCO 3 , 25; and glucose, 11.5; pH 7.2) after laparotomy. The mesentery attachments and fatty tissues were removed, and the inside of each stomach was washed with Krebs solution through a small incision in the gastric fundus. The stomachs were then mounted in 10-mL water-jacketed organ baths and initially loaded with weight totaling approximately 1.0 g. The temperature of the Krebs solution was maintained at 37uC 60.5uC, and the solution was aerated continuously with a mixture of 95% O 2 and 5% CO 2 .

Gastric Contractility Study
The contractile activities of the stomach in response to ghrelin treatment were monitored using an isometric force transducer (UM-203; Iwashiya Kishimoto Medical Instruments, Kyoto, Japan) and software (PicoLog for Windows, Pico Technology Ltd., St. Neots, UK). To normalize the contractions in this experiment, we added acetylcholine (ACh; 10 25 M) to the organ bath twice before the cumulative administration of ghrelin (after pretreatment with 10 210 M motilin) in the absence and presence of an antagonist. At the end of the experiment, ACh (10 25 M) was added once again to the organ bath, and the percentage of maximal contractions was calculated by averaging the tonic response induced by these 3 administrations. Note that in each case, the ACh administration evoked almost the same tonic gastric contractions. Then the effects of acyl ghrelin in the absence or presence of antagonists were expressed as a percentage of the control contractions. Concentration-response curves were obtained through cumulative addition of acyl ghrelin with or without antagonists or an inhibitor at appropriate intervals to the organ bath.

Damaging the Mucosa
The mucosa was damaged mechanically. In brief, the ghrelininduced (with low-dose motilin pretreatment) and motilin-induced gastric contractions were recorded as a control, the stomach was removed from the organ bath, and the mucosa was damaged through gentle pressing with cover-glass forceps. These damaged stomachs were loaded into the organ bath and washed several times with Krebs buffer. The stomachs were again mounted in 10-mL water-jacketed organ baths, and the contractile properties were measured using ghrelin and motilin.

Statistical Analysis
The results of experiments have been expressed as mean 6 standard error of the mean values of more than 4 separate experiments using whole stomachs. One-way analysis of variance followed by the Student's t-test was used for the statistical analysis of data. P,0.05 and P,0.01 were considered significant.

Cholinergic Pathway
We have previously shown that ghrelin induces gastric contractions after pretreatment with a low dose of motilin (10 210 M) [42]. We have also reported that atropine, a muscarinic receptor antagonist, completely abolishes the response to ghrelin in the stomach and suppresses spontaneous contractile activity [42]. Similarly, hexamethonium, a nicotinic receptor antagonist, almost completely eliminates contraction induced by ghrelin

Adrenergic Pathway
We also examined the involvement of adrenergic neurons. Phentolamine, an a receptor antagonist, markedly inhibited ghrelin-induced contraction ( Figure 2A) and decreased spontaneous contraction (data not shown). Ghrelin-induced tonic contraction was also significantly suppressed by pretreatment with prazosin, an a 1 receptor antagonist ( Figure 2B). Phentolamine and prazosin reduced the stimulatory effect of 10 27 M dose of ghrelin by 74.166.3% and 55.666.5%, respectively. Conversely, yohimbine, an a 2 receptor antagonist, and timolol, a b receptor antagonist, did not affect ghrelin stimulatory contraction ( Figure 2C, D).

Serotonergic Pathway
In the present study, ritanserin, a 5-hydroxytryptamine (HT) 2 receptor antagonist, did not decrease ghrelin-induced contractions at any concentration ( Figure 3A) but markedly suppressed spontaneous contractile activity (data not shown). In contrast to ritanserin, ondansetron, a 5-HT 3 receptor antagonist, significantly suppressed the contractions induced by ghrelin ( Figure 3B) and the rate (%) of reduction by the ondansetron on the stimulatory effects of 10 27 M dose of ghrelin was 58.163.2%.

Other Receptor Antagonists and Nitric Oxide (NO) Synthase Inhibitor
The effects of other receptor antagonists and NO synthase inhibitors were also investigated for further characterization of the ghrelin response. Naloxone, an opiate receptor antagonist, potentially suppressed the contractions induced by ghrelin ( Figure 3C) but did not affect spontaneously occurring phasic contractions (data not shown) and the inhibitory rate (%) of the naloxone on the effect of 10 27 M ghrelin was 62.966.1%. Neither spontaneously occurring contractions nor motilin-induced contractions were decreased by FK888, a neurokinin (NK) 1 receptor antagonist ( Figure 3D). N-nitro-L-arginine methylester, an inhibitor of NO synthase, potentiated the contractions induced by 10 210 and 10 29 M ghrelin but did not significantly change contractions induced by other ghrelin concentrations ( Figure 3E).

Importance of Mucosa in Ghrelin-induced Contraction
Because ghrelin-producing cells have been observed in the gastric mucosa of S. murinus [25], we predicted that the mucosa plays an important role in the ghrelin stimulatory pathway. Figure 4A depicts the synergistic effect of motilin and ghrelin with and without damage to the gastric mucosa. Interestingly, after the breakdown in mucosa, ghrelin-induced gastric contraction was completely abolished (see Figure 4A, B), whereas motilin-induced contraction was preserved (see Figure 4A, C), suggesting that mucosa has a central role in triggering ghrelin-induced contractions. Effects of Adenosine, SR142801, and GR113808 on Ghrelin-induced Contraction Intrinsic primary afferent neurons (IPANs) reportedly play a pivotal role in controlling the motility in the GI tract through synaptic connections with other neurons in the myenteric, submucosal, and mucosal ganglia [43,44]. Therefore, we predicted that IPANs play a significant role in the ghrelin regulatory pathway. IPAN is the primary transmitter of ACh and tachykinin [39,44]. The 5HT 4 receptor is also expressed on IPANs [45]. To confirm the involvement of these factors, we used pretreatment with GR113808 (5HT 4 receptor antagonist), adenosine, an agonist of adenosine receptors (to inhibit the release of ACh from IPANs through A1 receptors), and SR142801 (NK 3 receptor antagonist), and found that GR113808 completely eliminated ghrelin-induced contractions ( Figure 5A, B). The concentration-response curve showed that adenosine pretreatment significantly abolished ghrelin-induced gastric contraction ( Figure 5C). Pretreatment with SR142801 almost completely inhibited ghrelin-induced contractions at ghrelin doses of 10 210 and 10 29 M and significantly suppressed it at a concentration of 10 28 M ghrelin ( Figure 5D). GR113808 and adenosine reduced the effects of 10 27 M dose of ghrelin by 92.662.5% and 39.9618.7%, respectively, and SR142801 reduced the effects of 10 29 M ghrelin by 82.167.3%.

Discussion
Since the discovery of ghrelin, the mechanisms of ghrelin activity with respect to GI motility have been increasingly investigated; however, clear insights into these mechanisms have been elusive thus far. In the present study, we used the stomach of S. murinus, a new model animal for the study of motilin/ghrelin, and revealed mechanisms of ghrelin-induced contractions with a pharmacological in vitro method. To investigate the response of the neural network to ghrelin in the enteric nervous system, we examined the effects of various receptor antagonists and a NO synthase inhibitor on ghrelin-induced contractions and characterized the pharmacological properties in the suncus stomach in vitro.
We have already reported the novel observation that ghrelin (10 210 to 10 27 M) can induce contraction in isolated S. murinus stomach in a dose-dependent manner when pretreated with a low concentration of motilin (10 210 M) [29]. Ghrelin-induced S. murinus gastric contractions have also been confirmed to operate in a vagus-independent manner [29]. In the present study, hexamethonium, a ganglion-blocking agent, almost completely suppressed the action of ghrelin and, as we have reported in a previous study,  atropine also completely inhibits ghrelin-induced gastric contractions [29]. Several functional in vivo and in vitro studies have reported that the cholinergic system may be the dominant motor pathway in ghrelin-induced contractions [5,17]. These results  (Table S1), presynaptic cholinergic activation plays a much more prominent role than that of motilin in ghrelin-induced gastric contraction [28].
The notable inhibitory effects of phentolamine and prazosin indicate that a receptors, specifically a 1 receptors, but not a2 and b receptors, are also involved in ghrelin-induced contraction. Conversely, our previous study clearly showed that a2 receptors are important for motilin-induced gastric contraction [28]. Taken together, these results suggest that different neural pathways exist between ghrelin-and motilin-induced gastric contractions.
The significance of 5HT neurons has been observed in intestinal motility with vagus dependency [46,47]. By contrast, we showed that 5HT3 but not 5HT2 neurons are involved in ghrelin-induced contractions with vagus independence. Similarly, for the first time, we observed that the opioidergic pathway has a vital role in ghrelin-stimulated contraction. Moreover, compared with motilininduced gastric contraction in S. murinus [28], ghrelin-induced gastric contractions is strongly suppressed by ondansetron and naloxone pretreatment (see Table S1). Bassil et al. [16] have reported an EFS study in which ghrelin-induced contractions was evoked with multiple neurotransmitters. Ghrelin increases EFSevoked contraction mediated by a combination of cholinergic and non-cholinergic excitatory activity in rats [15]. These results suggest that not only pre-ganglionic cholinergic neurons but also the involvement of 5HT3 and opioidergic neurons are likely of similar importance for ghrelin-induced contraction in S. murinus stomach.
Another significant observation is that motilin is essential for ghrelin-induced gastric contractions [42]. Ghrelin administration alone, despite the administration of high doses (10 27 M), failed to stimulate any gastric contractions; however, pretreatment with a low dose of motilin restored sensitivity to ghrelin, implying that motilin opens the gate of the ghrelin pathway. Two possibilities explain this gate-opening property: (1) motilin may synergistically stimulate the ghrelin pathway, or (2) motilin may produce inhibitory effects on some strong suppression mechanism. Considering the differing properties between the independent administration of ghrelin (1000 times higher doses) and motilinpretreated ghrelin, which can induce gastric contractions (i.e., Apart from this, the most interesting finding in this study is the significance of the mucosa for ghrelin-induced, but not motilininduced, gastric contractions. We have reported that ghrelinproducing cells are found in suncus gastric mucosa [25] and that GHS-R messenger RNA expression is found in both the mucosa and the muscle layers of S. murinus stomach [26]. Therefore, some specialized neurons may responsible for the transfer of signals from the mucosa to the myenteric plexus in ghrelin-induced gastric contraction. Mucosal stimulation is thought to activate IPANs in both the submucosal [48,49,50,51] and the myenteric plexuses [39,52,53], and this distinctive shape is known as Dogiel type II morphology [44]. Therefore, we assumed that the IPANs may act as interneurons in ghrelin stimulatory action. The primary transmit component of IPANs are ACh and tachykinergic neurons [44]. Moreover, adenosine activates adenosine receptors, and most afterhyperpolarizing/Dogiel type II neurons (about 85%) are hyperpolarized by adenosine treatment-largely through A1 receptors [54] via inhibition of ACh release in the myenteric plexus [55]-and many studies have shown that the cell bodies of myenteric IPANs bear NK3 receptors [56,57,58,59]. In the present study, we found that pretreatment with adenosine significantly suppressed ghrelin-induced gastric contractions and that an NK3 receptor antagonist (SR142801) almost completely abolished the contraction at doses of 10 210 and 10 29 M. Conversely, we also observed that both adenosine and SR142801 had no effect on motilin-induced gastric contractions ( Figures S1A, B). Although GR113808 (5HT4 receptor antagonist) treatment completely inhibited ghrelin-induced contraction, motilin-induced contraction was partially suppressed by GR113808 treatment (Figure S2). 5HT 4 receptors are reportedly present in IPANs and in some other neuron types in the myenteric plexus [45]. Recently, Takahashi et al. [60] observed that 5HT 4 receptors are important in regulating migrating motor complexes through IPANs. Taken together, these data indicate that a different stimulatory pathway exists between the motilin and the ghrelin effects and that IPANs may be implicated in the effects of ghrelin on GI motility.
IPANs with cell bodies in myenteric ganglia have been shown to transmit via slow excitatory postsynaptic potentials (EPSPs) to interneurons and motor neurons, and indirect evidence suggests that they also transmit via fast EPSPs [39]. Slow EPSPs in IPANs are mimicked by the NK3 receptor agonist senktide [61] and partially blocked by the NK3 receptor antagonist SR142801 [41]. ACh, acting via muscarinic receptors, also elicits slow depolarizing responses in myenteric IPANs [62]. However, a major component of the transmission by IPANs appears to occur via cholinergic fast EPSPs because the application of an antagonist of nicotinic ACh receptors, hexamethonium, at the site of the primary afferent to interneuron synapses reduces ascending reflexes in response to mucosal distortion [63]. Moreover, 5HT 4 receptor activation facilitates ACh release in the myenteric plexus [64] and may initiate enteric reflexes by acting on the ending IPANs [65]. Therefore, the inhibitory effects of atropine [29], GR113808, adenosine, and SR142801 on ghrelin-induced gastric contractions suggest the importance of IPANs in the regulation of ghrelin activity in motility response. However, we cannot make a firm conclusion because the pharmacological properties of IPANs have been elucidated only for the intestines and not for the stomach thus far. Therefore, a future line of investigation will be to confirm the morphological presence of IPANs in the S. murinus stomach with co-localization of receptors for ghrelin, ACh, NK 3 , and 5HT 4 .
Finally, we concluded that the effect of ghrelin-induced gastric contractions is mediated through postganglionic and preganglionic cholinergic receptors, 5-HT 3 receptors, a and a 1 receptors, and opioid receptors; NO synthesis is also involved, and the final Figure 6. Predicted mechanism of the ghrelin-stimulatory pathway in Suncus murinus stomach. Considered with the previous results that motilin (10 29 to 10 27 M) induces gastric contractions by activating the myenteric plexus [28], the present findings suggest that a very low dose of motilin (10 210 M) may initiate an ''open gate'' effect and restore ghrelin activity via the myenteric plexus. Consequently, ghrelin may transfer signals from the mucosa to the myenteric plexus through involvement of intrinsic primary afferent neurons (IPANs) and regulate gastric contractions through muscarinic cholinergic receptors. doi:10.1371/journal.pone.0060365.g006 mediators of these ghrelin-induced gastric contractions are myenteric cholinergic neurons [28,29]. Moreover, the mucosa plays an important role in the neural pathway of ghrelin-induced gastric contraction. The reason why pretreatment with a low dose of motilin (10 210 M) is needed to induce ghrelin contractions remains unknown. Our predicted mechanism for ghrelin-induced gastric contractions is demonstrated in Figure 6; motilin stimulates gastric contractions through the myenteric plexus [28], whereas ghrelin may stimulate muscarinic cholinergic receptors by passing signals from the mucosa to the myenteric plexus through the stimulation of 5HT 4 , ACh, and NK 3 release from IPANs. Administration of low doses of motilin may open a gate so that ghrelin can induce gastric contractions in the S. murinus stomach. The confirmation of this mechanism may be an interesting finding in the near future. Moreover, the synergistic effects of motilin and ghrelin observed in the present study are probably speciesdependent, and clarifying the potential clinical application of these motilin/ghrelin-derived compounds would be interesting.