Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Ghrelin, a 28 amino acid octanoylated peptide, was recently isolated from the stomach of rat, and was identified as the endogenous ligand for the growth hormone secretagogue receptor (GHS-R) (Kojima et al., 1999). Before this discovery, the GHS-R was known to be activated by small synthetic peptides, growth hormone secretagogues (GHS), such as H-His-D-Trp-Ala-Trp-D-Phe-Lys-NH₂, also named growth hormone-releasing peptide-6 (GHRP-6). Ghrelin is now recognized as an important regulator of growth hormone (GH) secretion and energy homeostasis. Ghrelin is predominantly produced by the stomach, and the concentration of circulating ghrelin is influenced by acute and chronic changes in the nutritional status (for review, see Muccioli et al., 2002). Thus, ghrelin probably also regulates diverse processes of the digestive system.

Ghrelin is structurally related to motilin. In fact, the peptide was also identified by another group who named it "motilin-related peptide" (Tomasetto et al., 2000). They found that motilin and motilin-related peptide had a complementary expression pattern (endocrine cells of the intestine and the stomach, respectively) and suggested that motilin-related peptide may also function as a gastric hormone. Because this group deduced the amino acid sequence from the nucleotide sequence of the precursor, they did not identify the octanoylation that is now known to be crucial for the biological activity of ghrelin. For this reason, it seems indicated to use the name ghrelin.

Recent studies show that indeed the spectrum of biological activities of both peptides shows some striking similarities. Thus, although ghrelin was discovered via its effect on growth hormone release (Kojima et al., 1999), it was reported almost two decades ago that motilin, be it at high concentrations, stimulated growth hormone release in vitro and in vivo (Samson et al., 1982, 1984). Ghrelin and motilin may also act as an anabolic signal molecule during energy depletion because the levels of both peptides are increased by fasting, decreased by feeding and after an oral glucose load (Christofides et al., 1979; Peeters et al., 1980; Tschop et al., 2000), and both peptides stimulate insulin secretion (Suzuki et al., 1998; Date et al., 2002).

The effects of ghrelin on food intake confirm a role for ghrelin in the regulation of energy balance. Indeed, i.c.v. injection of ghrelin strongly stimulates feeding in rats and increases body weight by interacting with the pathways of neuropeptide Y and agouti-related protein in the arcuate nucleus (Nakazato et al., 2001). Also peripheral administration of this peptide alters food intake in mice and rats (Tschop et al., 2000) and intravenous administration enhances appetite and food intake in humans (Wren et al., 2001). An increase in food intake has also been observed after peripheral and central administration of motilin in rats and mice (Garthwaite, 1985; Rosenfeld and Garthwaite, 1987; Asakawa et al., 1998).

Hunger after ghrelin administration has been reported as a "side effect" in a clinical study analyzing GH release (Arvat et al., 2000). Motilin has also been called a "hunger hormone" because the increased plasma motilin levels during the fasted state trigger phase 3 gastric contractions during this period. Thus, the major physiological role of motilin has been attributed to its contractile effects on the gastrointestinal tract where it plays a key role in the regulation of the interdigestive motility pattern. Furthermore motilin and the motilin agonists erythromycin and its derivatives have gastroprokinetic activity (for review, see Peeters, 1993). Recent studies indicate that ghrelin has motor effects as well. Ghrelin has been reported to stimulate gastric motility in rats (Masuda et al., 2000), accelerate gastric emptying in mice (Asakawa et al., 2001), and resolve gastric postoperative ileus in rats (Trudel et al., 2002).

In addition to the structural similarities of the peptides themselves, 36% amino acid identity, and an identical precursor organization, their receptors, too, show a marked sequence homology with an overall identity of 52%, rising to 87% in the transmembrane regions. Therefore, at least some of the overlap in the spectrum of biological activities could be due to cross-interaction. To investigate this hypothesis, we compared the potency of motilin with that of ghrelin and some related peptides such as GHRP-6, a synthetic agonist of the GHS-R, in receptor binding and contractility studies in the rabbit gastric antrum, the classical in vitro model for motilin (Peeters and Depoortere, 1994). Furthermore, the mechanism of action of motilin and GHRP-6 and the receptor specificity of their contractile effects was studied in the presence of selective antagonists.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Chemicals

Norleucine¹³-porcine-motilin (motilin) was purchased from Eurogentec (Namur, Belgium). The motilin antagonist Phe-cyclo[Lys-Tyr(3-tBu)-Ala-] trifluoroacetate (GM-109) was a gift from Dr. N. Takanashi (Chugai Pharmaceutical Company, Gotemba, Japan). Ghrelin (1-28)-OH octanoylated in Ser³ [ghrelin (1-28) oct] was from Phoenix (Belmont, CA). The growth hormone-releasing peptide GHRP-6 and its antagonist D-Lys³-GHRP-6 were purchased from Bachem (Bubendorf, Switzerland).

Octanoylated ghrelin (1-5) amide [ghrelin (1-5) oct], octanoylated ghrelin (1-23)-OH [ghrelin (1-23) oct], and nonoctanoylated ghrelin (1-23)-OH [ghrelin (1-23) nonoct] were custom synthesized by solid phase methodology using the 9-fluorenylmethoxycarbonyl strategy. Octanoylation was performed according to Bednarek et al. (2000). The neurokinin (NK₁) antagonist Sanofi Research 140333 (SR140333) and the NK₂ antagonist SR48968 were a gift from Dr. X. Emonds-Alt (Sanofi Research, Montpellier, France).

Tissue Preparation

Adult rabbits of either sex were sacrificed by a blow on the neck. The stomach was removed and rinsed with saline. All procedures were approved by the Ethical Committee for Animal Experiments of the University of Leuven.

Motilin Receptor Binding Studies

Membrane Preparation. The antrum was dissected free from the mucosa, minced, and homogenized in sucrose buffer (50 mM Tris-HCl buffer, pH 7.4, 250 mM sucrose, 25 mM KCl, and 10 mM MgCl₂) with inhibitors (1 mM iodoacetamide, 1 µM pepstatin, 0.1 mM phenylmethylsulfonyl fluoride, 0.1 g/l trypsin inhibitor, and 0.25 g/l bacitracin). Homogenates were centrifuged at 1000g for 15 min, washed four times, and finally resuspended in 0.9% NaCl.

Displacement Curves. A competition binding assay was performed by incubating membranes (1 mg of protein) with ¹²⁵I-motilin (specific activity ± 1000 cpm/fmol, final concentration 50 pM) and increasing concentrations of motilin, ghrelin (1-5) oct, ghrelin (1-23) oct, ghrelin (1-23) nonoct, ghrelin (1-28) oct, GHRP-6, or D-Lys³-GHRP-6 for 60 min at 30°C. The reaction was stopped by adding cold buffer, and membrane-bound motilin was separated by centrifugation at 1000g. All data were corrected for nonspecific binding determined in the presence of an excess (10⁶ M) of unlabeled motilin. The dissociation constant (K_d) was calculated from the displacement curves fitted to the equation of Akera-Cheng by computer (Akera and Cheng, 1977).

Contractility Studies. Circular strips, freed from mucosa (0.2 × 2.5 cm) were cut from the antrum and suspended along their circular axis in a tissue bath filled with Krebs' buffer (120.9 mM NaCl, 2.0 mM NaH₂PO₄, 15.5 mM NaHCO₃, 5.9 mM KCl, 1.25 mM CaCl₂, 1.2 mM MgCl₂, and 11.5 mM glucose) gassed with 95% O₂, 5% CO₂. After equilibration at optimal stretch, electrical field stimulation (EFS) was applied via two parallel platinum rod electrodes using a Grass S88 stimulator. Frequency spectra (1, 2, 4, 8, and 16 Hz) were obtained by pulse trains (pulse 1 ms, train 10 s, 5 V). Voltage was kept at 5 V using a Stimu-Splitter II (Med Lab, Loveland, CO). Each consecutive pulse train was followed by a 90-s interval. Contractions were measured using an isometric force transducer/amplifier (Harvard Appartus, Inc., South Natick, MA), recorded on a multicorder, and sampled for digital analysis using the Windaq data acquisition system and a DI-2000 PGH card (Dataq Instruments, Akron, OH).

9

5

5

6

5

7

7

6

6

RT-PCR for the Ghrelin Receptor (GHS-R)

Total RNA was prepared from rabbit antral smooth muscle strips using the TRIzol reagent (Invitrogen, Carlsbad, CA). Single-stranded cDNA was synthesized using a random hexameric primer (10 µM) and 200 units of Superscript II RNase H reverse transcriptase (Invitrogen). The obtained cDNA served as a template for the polymerase chain reaction, consisting of 38 cycles of amplification (95°C for 1 min, 55°C for 1 min, 72°C for 2 min with a final elongation of 10 min at 72°C) using 0.5 U of TaqDNA polymerase (Amersham Biosciences AB, Uppsala, Sweden) and 0.5 µM primers. The primers GHS-R.for (forward: 5'-GGA CCA GAA CCA CAA GCA RA-3') and GHS-R.rev (reverse: 5'-TGA GGT AGA AGA GGA CAA AGG A-3') were selected in conserved regions found after alignment of sequences published for human (GenBank no. U60179), rat (GenBank no. U94321), and swine (GenBank no. U60178) GHS-R mRNA. This PCR product was subjected to a nested PCR using as primers GHS-R2.for (5'-CMG TGA ARA TGC TKG CTG TG-3') and GHS-R2.rev (5'-TGG CTG ATC TGA GCY ATC TC-3') or GHS-R.for and GHS-R2.rev. This resulted in a PCR product of, respectively, 124 or 144 bp. PCR products were analyzed on a 1.8% agarose gel and visualized with ethidium bromide.

Statistical Analysis

Data are represented as mean ± S.E.M. The modulation of the response to electrical field stimulation by pharmacological agents at the individual frequencies was analyzed by Student's paired t test. Comparison of the response to motilin and GHRP-6 in the absence and presence of specific blockers or antagonists was analyzed by two-way ANOVA. A value of P < 0.05 was considered statistically significant.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Interaction of Ghrelin and GHRP-6 with Motilin Receptor in Gastric Antrum of Rabbit.

Binding Studies. As shown in Fig. 1, ghrelin, GHRP-6, and related peptides weakly displaced membrane-bound ¹²⁵I-motilin. The order of affinity (pK_d) was motilin (9.13 ± 0.03) > GHRP-6 (5.54 ± 0.08) > D-Lys³-GHRP-6 (4.66 ± 0.04) > ghrelin (1-28) oct (4.23 ± 0.07) > ghrelin (1-23) oct (4.06 ± 0.08) > ghrelin (1-5) oct (3.75 ± 0.06). GHRP-6 had the highest affinity (5.54), which was reduced when the Ala in position 3 was replaced by D-Lys, a compound that is known to be a GHS-R antagonist. The affinity of GHRP-6 was 20.4-fold higher than that of ghrelin (1-28) oct. The weak interaction of ghrelin is due to the N terminus, because the potencies of ghrelin (1-28), ghrelin (1-23), and ghrelin (1-5) were not markedly different. The octanoyl residue, which is crucial for interaction with the ghrelin receptor, also increases affinity for the motilin receptor, because the interaction of nonoctanoylated ghrelin (1-23) was virtually zero.

View larger version (30K): [in this window] [in a new window]   Fig. 1.   Displacement of 125I-motilin bound to a crude membrane preparation from the rabbit antrum by unlabeled motilin, GHRP-6, D-Lys3-GHRP-6, ghrelin (1-28) oct, ghrelin (1-23) oct, ghrelin (1-23) nonoct, and ghrelin (1-5) oct.

Contractility Data. Nonstimulated strips. It is known that motilin contracts smooth muscle strips from the rabbit antrum, via a direct smooth muscle effect. The pEC₅₀ value for this interaction is 7.48 ± 0.11 (Van Assche et al., 1997). However, no contractile effect was observed with GHRP-6 or ghrelin (1-23) oct up to 10⁵ M.

Electrical field stimulation. Electrical field stimulation evoked muscle twitch on- and off-responses that increased in amplitude with an increase in frequency from 0.55 ± 0.14 g/mm² (1 Hz) to 8.56 ± 0.32 g/mm² (16 Hz) for the on-responses and from 0.70 ± 0.15 g/mm² (1 Hz) to 5.51 ± 0.19 g/mm² (16 Hz) for the off-responses. Tetrodotoxin (3 µM) abolished all contractions over the entire frequency spectrum, demonstrating that all responses were neurogenic (Fig. 2).

View larger version (10K): [in this window] [in a new window]   Fig. 2.   Effect of tetrodotoxin (TTX) on EFS-induced responses. Muscle strips were stimulated at increasing frequencies in the absence (top) and presence of TTX (3 µM, bottom). All responses were neurogenic because they could be abolished by TTX.

9

5

5

6

View larger version (23K): [in this window] [in a new window]   Fig. 3.   Effect of ghrelin (1-23) oct, GHRP-6, and motilin on EFS-induced on-responses (top) and off-responses (bottom) in the rabbit antrum. Muscle strips were stimulated at increasing frequencies in the absence and presence of ghrelin (1-23) oct (105 M) (n = 5), GHRP-6 (105 M) (n = 21), or motilin (109 M) (n = 14), and the effect on the tension of the on- and off-responses was measured. Results are represented as the change in tension. *, P < 0.05; **, P < 0.001; ***, P < 0.0005; ****, P < 0.00005 indicate significant changes in tension compared with the response in the absence of agents in the same strip preparation.

Pharmacological analysis of neurally mediated effects. The effect of motilin on both the on- and off-responses was blocked by the motilin antagonist GM-109 (10⁶ M), which by itself did not affect EFS-induced responses (Fig. 4). For GHRP-6 only the effect on the off-responses was blocked by GM-109, but the enhancement of the on-responses was preserved (Fig. 5).

View larger version (21K): [in this window] [in a new window]   Fig. 4.   Effect of a motilin antagonist, NANC conditions, and an NK1 and NK2 antagonist on motilin-induced neural on-responses (top) and off-responses (bottom) in the rabbit antrum. Results represent the change in tension (grams per square millimeter) induced by motilin (109 M) in the absence (n = 14) and presence of, respectively, GM-109 (106 M) (n = 10), atropine (5 × 106 M) + guanethidine (3 × 106 M) (n = 21), or SR140333 (5 × 107 M) + SR48968 (5 × 107 M) (n = 8). *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001 indicate significant changes in tension compared with the response in the presence of antagonists without application of motilin in the same strip preparation. Two-way ANOVA revealed an overall significance for the response to motilin in the presence of GM-109 (on, P < 0.0001; off, P < 0.0001), in the presence of atropine and guanethidine (on, P < 0.0001; off, P = 0.297), and in the presence of SR140333 + SR48968 (on, P < 0.0001; off, P < 0.001) compared with the response of motilin in the absence of antagonists.

View larger version (21K): [in this window] [in a new window]   Fig. 5.   Comparison of the effect of a motilin antagonist, NANC conditions, and an NK1 and NK2 antagonist on neural responses induced by GHRP-6. Results represent the change in tension (grams per square millimeter) induced by GHRP-6 (105 M) in the absence (n = 21) and presence of GM-109 (106 M) (n = 13), atropine (5 × 106 M) + guanethidine (3 × 106 M) (n = 11), or SR140333 (5 × 107 M) + SR48968 (5 × 107 M) (n = 13). Top, effect on the on-responses. Bottom, effect on the off-responses. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.00005 indicate significant changes in tension compared with the response in the presence of antagonists without application of GHRP-6 in the same strip preparation. Two-way ANOVA revealed an overall significance for the response to GHRP-6 in the presence of GM-109 (on, P = 0.163; off, P < 0.0001), in the presence of atropine and guanethidine (on, P < 0.0001; off, P = 0.158), and in the presence of SR140333 + SR48968 (on, P < 0.0001; off, P < 0.0001) compared with the response of GHRP-6 in the absence of antagonists.

View larger version (11K): [in this window] [in a new window]   Fig. 6.   Representative tracing showing the effect of motilin on EFS-induced responses at increasing frequency of stimulation under NANC conditions. The muscle strip was electrically stimulated under normal conditions and resulted in on- and off-responses (top). Under NANC conditions, only off-responses occurred at higher frequencies of stimulation (middle). Addition of motilin (109 M) under NANC conditions enhanced off-responses over the entire frequency spectrum (bottom).

View larger version (20K): [in this window] [in a new window]   Fig. 7.   Effect of the GHS-R antagonist D-Lys3-GHRP-6 on neuronal responses induced by GHRP-6. Strips were stimulated with D-Lys3-GHRP-6 (105 M) followed by stimulation with GHRP-6 (106 M). Results show the change in tension (grams per square millimeter) induced by GHRP-6 in the absence and presence of D-Lys3-GHRP-6 (n = 16) and of D-Lys3-GHRP-6 alone (n = 22). Top, effect on the on-responses. Bottom, effect on the off-responses. *, P < 0.05; **, P < 0.01; ***, P < 0.005; and ****, P < 0.00005 indicate significant changes in tension compared with the control response or the response in the presence of D-Lys3-GHRP-6 in the same preparation. Two-way ANOVA analysis revealed an overall significance for the response to GHRP-6 in the presence of D-Lys3-GHRP-6 (on, P < 0.296; off, P < 0.804) compared with the response of GHRP-6 in the absence of antagonist.

Expression of Growth Hormone Secretagogue Receptor in Rabbit Gastric Antrum

RT-PCR studies with cDNA prepared from total RNA isolated from rabbit antral smooth muscle strips revealed a GHS-R transcript product corresponding to the predicted size, based upon the set of primers used. Figure 8 shows an agarose gel electrophoresis of the nested PCR products obtained by subjecting first-round PCR products, with GHS-R.for and GHS-R.rev as primers, to a nested PCR with either GHS-R2.for and GHS-R2.rev (lane 1, expected size 124 bp) or GHS-R.for and GHS-R2.rev (lane 2, 144 bp) as primers.

View larger version (51K): [in this window] [in a new window]   Fig. 8.   GHS-R expression in the rabbit antrum. Ethidium bromide staining of agarose gel electrophoresis of the nested PCR products of cDNA prepared from the rabbit antrum. First-round PCR products obtained with GHS-R.for and GHS-R.rev as primers were subjected to a second PCR using either GHS-R2.for and GHS-R2.rev (124 bp, lane 1) or GHS-R.for and GHS-R2.rev (144 bp, lane 2) as primers. Lane 3, molecular weight marker giving bands of 100-bp spacing. A scheme indicating the primer positions in respect to human GHS-R 1a mRNA is shown.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

The pulsatile release of GH from the pituitary somatotrophs is regulated by two hypothalamic neuropeptides: growth hormone-releasing hormone and somatostatin that, respectively, stimulate and inhibit GH secretion. Release of GH from the pituitary somatotrophs can also be controlled by growth hormone secretagogues synthetic peptidyl (e.g., GHRP-6 and hexarelin) and nonpeptidyl molecules (e.g., MK 06777), whose actions are mediated by a specific GHS receptor (Howard et al., 1996), which is distinct from the GHRH receptor. Ghrelin was recently discovered as the natural ligand for the GHS receptor (Kojima et al., 1999). However, the principal site of ghrelin synthesis is the stomach, and the density of ghrelin binding sites is higher in several peripheral tissues than in the pituitary (Papotti et al., 2000), indicating that the role of ghrelin is not limited to the regulation of GH secretion. In this article, we explored the possibility that ghrelin could stimulate motor activity via an interaction with motilin receptors. The study was therefore performed in the rabbit gastric antrum, the classical model for the study of motor effects of motilin and motilin agonists (Peeters and Depoortere, 1994).

We found that ghrelin and ghrelin analogs interact very weakly with the motilin receptor in binding studies, whereas GHRP-6 showed some affinity (pK_d = 5.54). Thus, although GHRP-6 has virtually no sequence similarity with motilin, it has a stronger affinity for the motilin receptor than ghrelin. In contrast, ghrelin does have sequence similarity (36%) with motilin but the pharmacophore of ghrelin, the [1-5] fragment and in particular the octanoyl residue on Ser³ (Bednarek et al., 2000), and the pharmacophore of motilin, the [1-7] fragment and in particular Phe¹, Val², Ile⁴, and Tyr⁷ (Macielag et al., 1992; Peeters et al., 1992), show little if any overlap. Interestingly, the weak interaction that we observed seems to be related to the octanoyl group because ghrelin (1-28) oct and ghrelin (1-5) oct have a comparable affinity, which is completely lost when the octanoyl group is removed. Ghrelin is the first peptide isolated from natural sources in which the hydroxyl group of one of its serine residues is acylated by n-octanoic acid, and this unique post-translational modification seems to be necessary for the GH-releasing potency of both human and rat ghrelin (Kojima et al., 1999). Apparently, it also creates a structural feature, unrelated to the amino acid sequence, which allows a weak interaction with the motilin receptor.

The weak affinity for the motilin receptor is unable to induce a contractile effect in our model. Therefore, the motor effects that have been reported for ghrelin are probably due to other pathways involving GHS receptors. In vivo, the effect of intravenous administration of ghrelin on gastric contractions is mediated through the vagus (Masuda et al., 2000), but the concentrations of ghrelin used in this study, from 4 to 20 µg/kg, are considerably higher than for motilin, 50 ng/kg (Boivin et al., 1997; Masuda et al., 2000). In contrast, the stimulation of growth hormone secretion and the stimulation of food intake in rats are achieved at lower doses of ghrelin (growth hormone secretion 10 µg, i.v.; food intake 33 ng, i.c.v.) than for motilin (growth hormone secretion 100 µg, i.v.; food intake 1 µg, i.c.v.) (Samson et al., 1984; Rosenfeld et al., 1987; Nakazato et al., 2001; Tolle et al., 2001). Thus, although both peptides and their receptors are related to each other, the major physiological roles of motilin are its effect on intestinal motility, whereas for ghrelin, the most remarkable activities remain to be elucidated because other effects beside its effect on motility, GH secretion, and food intake have been reported. These include the cardioprotective actions of ghrelin (Nagaya et al., 2001) mediated by its antiapoptotic activity (Baldanzi et al., 2002) and the antiproliferative effects of ghrelin observed in neoplastic cell lines (Cassoni et al., 2001). In the latter two cases, the nonoctanoylated ghrelin also is effective, implicating that it might be mediated by another subtype of the ghrelin receptor. Overlapping effects at pharmacological concentrations may be due to cross-interaction but are probably physiologically not important.

GHRP-6, a ghrelin receptor agonist, had a higher affinity for the motilin receptor and did affect the response to electrical field stimulation. This effect is partially mediated through interaction with the motilin receptor, because it could be blocked by the motilin antagonist GM-109, and partially via another receptor, possibly a subtype of the GHS-R receptor with a low affinity for ghrelin. The transcripts detected by our RT-PCR studies could be transcripts of this GHS-R subtype. It is interesting to note that similarly to our study, in nonendocrine tissues, ghrelin has a much lower potency (100-fold) than hexarelin, a GHRP-6 analog, to displace bound ¹²⁵I-Tyr-Ala-hexarelin than in endocrine tissues (5-fold) (Papotti et al., 2000). Another indication for the involvement of a GHS-R subtype is the observation that the GHRP-6 antagonist D-Lys³-GHRP-6 was an agonist in our model. Moreover, this compound had weak excitatory effects on the on-responses but no effect on the off-responses, in agreement with its lower affinity for binding to the motilin receptor.

We have previously shown that in the rabbit antrum motilin may act on motilin receptors on both smooth muscle cells and myenteric neurons (Van Assche et al., 1997). In this study, we further explored the neurally mediated effect of motilin and obtained evidence for motilin-activated cholinergic and tachykininergic pathways. Especially, the effect of motilin on the on-response is mediated by a motilin receptor on cholinergic nerves because it is blocked under NANC conditions. This is in agreement with a previous in vitro study from Kitazawa et al. (1993) that indicated that motilin can induce the release of acetylcholine from enteric neurons in the rabbit duodenum. Because the effect of motilin on the on-responses was also reduced by NK₁ and NK₂ antagonists, it is plausible to assume that tachykinins synergize with acetylcholine in the transmission process. Substance P was originally characterized as an atropine-resistant stimulant of contraction of rabbit small intestine (von Euler and Gaddum, 1931). In guinea pig stomach, human ileum, and rat colon both NK₁ and NK₂ receptors mediate NANC neuromuscular transmission (Zagorodnyuk and Maggi, 1997; Zagorodnyuk et al., 1997; Serio et al., 1998). We provided evidence that also in the rabbit antrum, the effect of motilin on the off-responses is mediated by a tachykininergic pathway that acts independent of ACh release.

In vivo studies in humans indicated that the effect of motilin and the motilin agonist erythromycin on antral motor activity is mediated through the activation of cholinergic neurons (Boivin et al., 1997; Coulie et al., 1998). Coulie et al. (1998) noted that atropine failed to block the prolonged rhythmic antral contractile activity induced by high doses of erythromycin and suggested that it might be due to activation of smooth muscle motilin receptors, which have a lower affinity (Van Assche et al., 1997). However, as the present study indicates, it might also reflect the activation of a noncholinergic neural pathway. This is in agreement with in vivo studies in dogs with the erythromycin derivative EM-523, which found that the contraction-stimulating activity of EM-523 in the stomach in the postprandial period is partially mediated through the cholinergic pathway; and partially through a noncholinergic pathway involving 5-hydroxytryptamine₃ and NK₁ receptors (Shiba et al., 1995).

In conclusion, our study indicates that in the rabbit antrum ghrelin is not able to have a meaningful interaction with its most related family member, the motilin receptor. Because we found no evidence that ghrelin stimulates contractile activity via local pathways. Our study, therefore, supports previous observations that ghrelin affects gastric motility via the vagal nerve. On the contrary, at high concentrations the ghrelin receptor agonist GHRP-6 acts on motilin receptors located on tachykininergic nerves that act independent of acetylcholine release to stimulate neural contractile responses. Moreover, GHRP-6 also activates another receptor that is probably a GHS-R subtype located on cholinergic neurons that corelease tachykinins. Our data therefore suggest that both the GHS-R and motilin receptor, and their ligands ghrelin and motilin, may be part of a larger family with yet-to-be-discovered peptides and receptors. The divergence in response of these peptides opens the road for the development of new gastroprokinetic drugs.