The overlooked background of gastric MMC III: the signalling pathways of motilin receptors inducing canine left gastric artery relaxation

Background: In diabetic patients with gastroparesis, the gastric blood supply is often decreased and delayed gastric emptying associating MMCIII absence is the main symptom. Under physiological conditions, motilin has been shown to induce a sustained increase in left gastric artery (LGA) blood ow and initiate MMC phase III simultaneously. The study aimed to elucidating the signal transduction pathways of motilin receptors (MLNRs) in the relaxation of LGA. Methods: MLNR expression in the LGA was analysed by immunohistochemistry. Motilin-induced relaxation of the LGA was tested in a multi-wire myograph system. Effects of inhibitors or blockers in the signal transduction pathway were observed. Results: Immunohistochemical and immunouorescence staining showed that the MLNRs were on the membranes of endothelial cells. Motilin relaxed U46619 pre-contracted canine LGA rings in a concentration-dependent manner, with an EC50 value of 9.010 ± 0.789 × 10 −8 M. Motilin’s effect was inhibited by denuded endothelium but not by muscarinic receptor inhibitors. The effect was selectively and competitively inhibited by Phe-cyclo[Lys-Tyr(3-tBu)-Ala-]•triuoroacetate (GM-109; MLNR antagonist) and completely or partially inhibited by inhibitors of the G protein–phospholipase C– inositol trisphosphate (G pr–PLC–IP 3 ) and nitric oxide synthase–nitric oxide–soluble guanylyl cyclase (NOS–NO–sGC) signal transduction pathway, inhibitors of cyclooxygenase and myoendothelial gap junction, blockers of the potassium channel and low/free Ca 2+ Krebs solutions, but not by inhibitors of protein kinase C, protein kinase A or L-type voltage-operated Ca 2+ channel. Conclusions: MLNRs were on the membranes of endothelial cells of canine LGA. The main intracellular signal transduction pathway was motilin– MLNR–G pr–PLC–IP 3 –NOS–NO–sGC–cGMP. results clearly demonstrated the signal transduction mechanism that motilin regulates gastric blood ow in dogs in the inter-digestive phase under physiological conditions and may provide a new theoretical basis for the research on diabetic gastroparesis.

The research on how the gastric arteries periodically deliver motilin from proximal small intestine to the targeted MLNRs on the stomach are uncommon. In the present study, considering the complexity of signalling pathways of artery relaxation, which simultaneously involves both the endothelial cells and smooth muscle cells, freshly isolated intact LGAs from healthy dogs were mounted in a multi-wire myograph system. Thereafter, inhibitors or blockers of the intracellular signalling transduction pathways were used to investigate the intracellular mechanisms of motilin-induced LGA relaxation. The ultimate purpose of understanding these mechanisms is to prevent or delay the progression of diabetic vascular complications and improve the quality of life in these patients.

Preparation of Canine LGA
After an overnight fast, 96 mongrel adult healthy dogs of either sex, weighing 15-30 kg, obtained from the Section of Surgical Teaching, Jilin University, were sacri ced by rapid exsanguination from the common carotid artery under deep anaesthesia with sodium pentobarbital (30 mg kg −1 , i.v.). The connective tissues and fat were carefully dissected under a dissecting microscope (SZ61, Olympus, Japan) avoiding over-pulling and clamping.
Immunostaining of MLNR Immunohistochemical staining of MLNR The LGAs were procured from six dogs. The para n-embedded tissues were cut into 4-μm sections, depara nised, rehydrated, boiled in retrieval solution and washed in PBS. After blocking in normal horse serum for 30 min, the sections were incubated with rabbit anti-dog MLNR antibody (1:100; RaQualia Pharma Inc., Taketoyo, Japan) overnight at 4°C [14]. The next day, the sections were incubated with biotinylated goat anti-rabbit antibody (Sigma, St. Louis, MO, USA) and horseradish peroxidase-conjugated avidin (Sigma, Shanghai, China) for 30 min at room temperature. The stained sections were then visualised using 3,3′-diaminobenzidine (Sigma, St. Louis, MO, USA). For negative controls, the primary antibody was replaced with a nity-puri ed pre-immune IgG.
Immuno uorescence staining of MLNR The LGAs were procured from three dogs. The 6-μm para n-embedded sections were depara nised, incubated in 0.1% Triton X-100 at 4°C for 1 h and washed three times with PBS. After 5-min digestion with proteinase K, the sections were washed with PBS and blocked in 5% normal goat serum for 30 min. Subsequently, the sections were incubated with rabbit anti-dog MLNR antibody (1:100, RaQualia Pharma Inc., Taketoyo, Japan) and mouse antidog cluster of differentiation 34 (CD34) antibody (1:100; Affymetrix, Santa Clara, CA, USA) overnight at 4°C. After three 10-min washes with PBS, the sections were incubated with uorescein isothiocyanate-labelled (green) goat anti-rabbit immunoglobulin (Ig)G (1:50; Abbkine, California, USA) and Cy3-labelled (red) goat anti-mouse IgG (1:100; Abbkine, California, USA) for 1 h at room temperature. The sections were washed again three times with PBS for 10 min, incubated with Hoechst 33342 (Sigma, Shanghai, China) for nuclear staining for 5 min, washed twice with PBS for 5 min each and nally mounted. The slides were visualised and photographed under a confocal microscope (FV1000; Olympus, Tokyo, Japan). For negative controls, the primary antibody was replaced with a nity-puri ed pre-immune IgG. All images were processed using FV10-ASW 1.7 software (Olympus, Tokyo, Japan).

Record of Isometric Vascular Tone
Each LGA was approximately 18-30 mm long and was cut into six to ten 3-mm length rings which were immediately mounted between two L-shaped stainless-steel hooks (300 μm in diameter) in the organ bath of the multi-wire myograph system (DMT620, Demark Subsequently, U46619 (a thromboxane A 2 analogue) (5 × 10 −8 M) was used. Once a sustained contract tension was reached, one concentration of motilin was examined to avoid tachyphylaxis [31]. Endothelium-denuded rings were made by gently scraping the endothelial cells with a pair of pointed metal forceps before mounting. In the inhibitor groups, the rings were rst incubated with different inhibitors for 15-40 min before adding U46619; rings from the same LGA incubated with saline were considered as the control group. In the Ca 2+ -free Krebs solution, Ca 2+ was replaced with 1 mM EGTA. The endothelium integrity or functional removal was veri ed by acetylcholine (10 −5 M) at the end of each test, with the relaxation rate (RR) >80% or <10%, respectively.
The RR was expressed as a percentage decrease in the tension induced by U46619 and/or high-potassium solution according to the following formula: In the calculation of the inhibition rate (IR), the tension was normalised to the corresponding values of the control group: where T is a sustained tension, L is the lowest tension, i represents the inhibitor group and c represents the control group.
Concentration-response curves were analysed by nonlinear regression analysis with variable slopes in GraphPad Prism 6 (GraphPad Software, San Diego California, USA), from which the ECx (x% maximal effective concentration) and HillSlope were obtained. The pA2 value was calculated according to the Van Rossum equation: where, A is the molar concentration of the antagonist and CR is the ratio of the EC50 value (EC50 with an antagonist/EC50 without an antagonist) [32].

Measurement of NO and cGMP Levels in LGA
Tissue collection and homogenisation were conducted based on the methods provided by Schachter et al [33].
LGAs from three dogs were pooled together to provide su cient tissue for analysis at each time. When NO was detected, seven groups were created. Three groups were rst incubated When cGMP was detected, the groups were the same as those in which NO was detected, except that the ODQ (10 −5 M) group was included. U46619 (5 × 10 −8 M) was the negative group (blank column), motilin (9 × 10 −8 M) added after U46619 (5 × 10 −8 M) was the positive control group (motilin column), inhibitors added before U46619 and motilin were the experimental groups and acetylcholine (10 −5 M) was used to con rm the activity of endothelial cells (acetylcholine group).
Tissues were homogenised in the following solution: 10.0 mM Tris-HCL, 0.1 mM EDTA-2Na, 10 mM sucrose and 136.7 mM NaCl (pH 7.4) in 1:9 ratio of weight (g) to volume (mL). The homogenate mixture was then centrifuged at 4°C, 2.4 g (5000 rpm) for 10 min, and the supernatant was used for detection. We strictly adhered to the protocol for the detection process. The total protein concentration was assayed using a bicinchoninic acid (BCA) total protein assay kit (A045-3; Jiancheng, Nanjing, China) and bovine serum albumin was used as the standard and expressed in μmol mL −1 . The cGMP concentration was assayed using a canine cGMP ELISA kit (Cat No. ela05471Ca, SANCHEZ, Colorado, USA) and expressed in picomoles of cGMP per milligram of protein (pmol mg −1 ). The NO concentration was assayed using a nitrate reductase system kit (S0023; Beyotime, Haimen, China) and expressed in micromoles of NO per gram of protein (μmol g −1 ).

Statistical Analysis
All numerical data are presented as means ± standard error of the mean (SEM) with n equal to the number of dogs in the multi-wire myograph system and N equal to the number of repetitions in other experiments. Statistical analysis was performed using GraphPad Prism 6 (GraphPad Software). Data analyses were performed using a paired or unpaired t -test for paired or unpaired data, and one-way analysis of variance followed by Dunnett's posthoc test for multiple comparisons. P < 0.05 was considered statistically signi cant.
Immunohistochemical staining revealed brown linear staining along the endothelium of LGA, but no brown staining in the negative control group (N = 6) (Fig. 1a). Immuno uorescence double staining was performed using the CD34 (antibody of endothelial cell marker) and MLNR antibody after LGA sectioning, which showed that CD34 and MLNR completely overlapped (N = 3) (Fig. 1b).
Effects of Endothelial Denudation on the Motilin-or Acetylcholine-induced Vasorelaxation of LGA Endothelial denudation abolished the motilin-induced vasorelaxation. We also found that acetylcholine did not relax the preparations without endothelium. In summary, motilin-induced and acetylcholine-induced relaxation of LGA were both endothelium-dependent, as shown in Table 1.
Representative images are shown in Fig. 3. Table 1 The effects of endothelium-denudation and atropine on motilin-and acetylcholine-induced relaxation in canine LGA rings

Effects of a Muscarinic Receptor (MR) Inhibitor on the Motilin-or Acetylcholineinduced Vasorelaxation of LGA
After the pre-treatment of endothelium-intact rings with atropine (an MR inhibitor), motilin-induced relaxation did not decrease. As expected, atropine abolished the acetylcholine-induced vasorelaxation signi cantly. The results indicate that motilin-induced relaxation of LGA did not occur through muscarinic receptor. Data are summarised in Table 1  IR values > 78.12% ± 4.82% (Fig. 4c).We also found that GM-109 at 10 − 5 M had no effect on the baseline, U46619-induced contraction and acetylcholine-induced relaxation in LGA rings (Fig. 4d).
Roles of the G pro-PLC-IP 3   L-NAME (a NOS inhibitor) and ODQ (a sGC inhibitor) markedly attenuated the motilin-induced vasorelaxation of LGA. Indomethacin, a cyclooxygenase inhibitor, slightly but signi cantly decreased motilin-induced relaxation. MEGJs and potassium channels [29] are two important factors that determine EDH. Pre-treatment of LGA rings with 18α-GA, a MEGJ inhibitor, showed a partial attenuation of motilin-induced vasorelaxation. The high K + solution containing 3 × 10 − 2 M KCl, a nonspeci c depolarising agent, markedly decreased the relaxation. TEA (a nonspeci c K + channel blocker) partially attenuated the motilin-induced vasorelaxation. Glibenclamide, an ATP-sensitive K + channel (K ATP ) blocker, increased the relaxation rate. The results are shown in Table 2, and representative original images are shown in Fig. 6.
The levels of NO and cGMP in LGA tissues (each N = 3) were tested. As shown in Fig. 7a,  Table 2, and the representative images are shown in Fig. 8.

Discussion
In this study, we not only demonstrated that MLNRs were expressed on membranes of endothelial cells of canine LGA but also revealed that motilin functioned primarily through the motilin-MLNR-G pr-PLC-IP 3 -Ca 2+ -NOS-NO-sGC-cGMP signalling pathway to relax LGA. In addition, PGI 2 , EDHF and extracellular Ca 2+ had roles. The signalling pathways of MLNR in inducing the relaxation of LGA is summarised in the model diagram (Fig. 9).
MLNRs were identi ed to express on the endothelial cell membrane of LGA in dogs by our immune and endothelial denudation experiments, which is a site that had only been speculated previously [12]. Motilin induced increased gastric blood ow [12] or LGA relaxation are both dose depended. We also demonstrated that GM-109 was the selective and competitive antagonist of endothelial MLNRs. The results that the antagonism parameter (pA2) of GM-109 for motilin (7.638) in LGA preparations was close to that demonstrated in the rabbit duodenum (7.37) [34] suggest that GM-109 has similar antagonistic activity on the MLNRs in the both sites, which is indicating that these MLNRs may have similar binding sites and/or features. We can conclude that MLNRs are no doubt the basic molecular structures required for motilin's effects on gastric arteries both in vivo [12] and in vitro experiments. Thus, there signal transduction pathway after MLNR in LGA must be the same.
Motilin plays a synergic and e cient role in regulating multiple physiological activities in the gastrointestinal tract. Although these effects are all MLNR and G pr depended, the following mechanism are quite different. In human and dogs, motilin's effects on gastrointestinal smooth muscles contraction [26,27,35,36], hunger sensations and feeding signals [37] are all cholinergically facilitated and acetylcholine is thought to be the nal mediator.
However, the activated motilin-MLNR-G pr-PLC-IP 3 signal transduction pathway in in endothelial cells is much more directly. It is not related to acetylcholine, as its effect can't be inhibited by anticholinergic agents. It directly induced the release of vasorelaxing substances in endothelial cells.
The NOS-NO-sGC-cGMP pathway plays a critical role in motilin-induced relaxation of canine LGA. We proved that NO is promoted only by motilin and released from vascular endothelial cells, but not from the intestinal neurons promoted by acetylcholine [35,36] or motilin [20]. In addition to NO system, prostaglandins (mainly PGI 2 ) and EDHF also play roles. However, the e cacies of the inhibitors (although in a high concentration) of the two pathways were relatively low, indicating that PGI 2 and EDHF may not be the primary vasodilators in motilin-induced relaxation of canine LGA.
In fact, the increase of the intracellular Ca 2+ is related to. the production and release of all three vasorelaxing substances Motilin induced a two-phase increase in Ca 2+ in endothelial cells of the porcine aortic valves: the initial increase is related to the release of Ca 2+ from intracellular Ca 2+ storage, whereas the maintenance of a stable Ca 2+ concentration was associated with extracellular Ca 2+ in ow [24]. The release of Ca 2+ from intracellular Ca 2+ storage can be explained by the well-known G pr-PLC-IP3 signalling pathway. However, the in ow of extracellular Ca 2+ was found to be related to the extracellular Ca 2+ concentration and SOCC, and the effects of VOCC were excluded. This mechanism is very different from that in smooth muscle cells (excitable cells) [38], because endothelial cells are non-excitable cells and the resting potential may induce failure of > 50% of VOCC activation [39].
Considering patients with diabetic gastroparesis always lost gastric phase III activities [3], MLNR agonists are often the candidate drugs for improving delayed gastric emptying or mimicking gastric MMC III in the treatment of diabetic gastroparesis [40,41]. However, diabetic microangiopathy pathologically consists of damaged endothelial cells and capillary basement membrane thickening, in which endothelial dysfunction mainly manifests as the reduced production or release of NO [42]. It is just the main MLNR's signal transduction pathway inducing LGA relaxation discovered by us. This suggests a link between decreased ability of motilin to regulate gastric blood supply and the occurrence of diabetic gastroparesis. Furthermore, lack of a fundamental improvement in blood supply of gastric wall will no doubt contradict the effectiveness of motilin agonists. Thus, more measures related to protect the endothelial functions or improve the gastric blood should be taken account. Further research is warranted in a transgenic or geneknockout animal model of diabetic gastroparesis to characterise the more pathological roles of motilin and MLNRs in endothelial cells.

Consent for publication
Not applicable Availability of data and materials The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.

Competing interests
The authors declare that they have no competing interests.  The original traces of motilin-induced relaxation of LGA rings pre-treated by endothelium denudation or atropine With endothelium (blue) or atropine (10−5 M) (green) or without (red), the double slash (//) indicates that the blocking process was prior to the addition of U46619. MTL: motilin; ACh: acetylcholine; E: endothelium-denudation; LGA: left gastric artery    LGA tissues (N = 3 replicates). In Fig. 4a and b, compared with the blank or motilin column respectively, * or # P < 0.05 with Dunnett's post-hoc test. LGA tissues (N = 3 replicates). In Fig. 4a and b, compared with the blank or motilin column respectively, * or # P < 0.05 with Dunnett's post-hoc test.   The signalling pathways of motilin receptor in inducing the relaxation of canine LGA. Green arrows indicate the signalling pathways in endothelial and smooth muscle cells during motilin-or acetylcholine-induced relaxation of canine LGA. Red arrows with at heads indicate the blocking effect on the signalling pathways. The blockers or inhibitors shown in red indicate that the medicines inhibited motilin-activated signalling pathways and those shown in blue indicate no inhibitory effect on motilin-activated signalling pathways.

Figure 9
The signalling pathways of motilin receptor in inducing the relaxation of canine LGA. Green arrows indicate the signalling pathways in endothelial and smooth muscle cells during motilin-or acetylcholine-induced relaxation of canine LGA. Red arrows with at heads indicate the blocking effect on the signalling pathways. The blockers or inhibitors shown in red indicate that the medicines inhibited motilin-activated signalling pathways and those shown in blue indicate no inhibitory effect on motilin-activated signalling pathways.