Mechanism of bombesin-induced tonic contraction of the porcine lower esophageal sphincter

Gastroesophageal reflux disease (GERD) is a disorder that is related to an incompetent lower esophageal sphincter (LES). Previous studies showed that bombesin could increase LES pressure in humans and opossums. The aim of the present study was to characterize the effects of bombesin on porcine LES contraction. We used the selective agonists, neuromedin B (NMB), gastrin-releasing peptide (GRP), and [D-Tyr6,Apa-4Cl11,Phe13,Nle14]bombesin-(6-14) (DTACPN-BN), as well as receptor antagonists of bombesin receptor subtype 2 (BB2), and 3 (BB3) for ex vivo contraction studies. Atropine, nifedipine, tetrodotoxin, and ω-conotoxin GVIA were used to explore the agonist-induced LES contraction mechanism. Reverse transcription polymerase chain reaction and immunohistochemistry were applied to detect bombesin receptor expression. Our results indicate that GRP and DTACPN-BN, but not NMB, induced tonic contractions of the porcine LES in a dose-dependent manner, and the contractions were inhibited with selective BB2 and BB3 antagonists. The GRP-induced contraction is mainly caused by L-type Ca2+ channel-mediated Ca2+ influx. However, DTACPN-BN-induced contractions are associated with neuronal conduction. RT-PCR and immunohistochemistry revealed that BB2 and BB3 were expressed in the porcine LES. Bombesin-induced tonic contraction of the LES is mediated through BB2 and BB3. Bombesin, BB2, and BB3 agonists might have the potential to treat GERD.

PPI therapy 2 . Additionally, long-term PPI use may cause some side effects, such as micronutrient absorption, drug interference and microscopic colitis 3 . Surgical procedures, such as Nissen fundoplication, increase the LES resting tone and can lead to an improved esophageal contraction amplitude 4 . However, symptom recurrence may occur after fundoplication. Medications to strengthen the lower esophageal sphincter, such as baclofen, which reduces transient LES relaxation, are few in number. Baclofen may cause severe neurological side effects, such as weakness, confusion, tremors, and dyskinesia 5 . Therefore, drugs that enhance LES contractions in order to avoid acid or other irritant reflux are being developed.
Bombesin is a 14-amino acid peptide and originates from the skin of the European fire-bellied toad (Bombina bombina) 6,7 . There are three bombesin receptor subtypes, including bombesin receptor subtype 1 (BB 1 ), bombesin receptor subtype 2 (BB 2 ), and bombesin receptor subtype 3 (BB 3 ) 7,8 . Bombesin can stimulate G cells in the stomach to release gastrin. Bombesin is also considered to be involved in satiety and appetite regulation, as well as in stress-induced anorexia and pressure-induced obesity 9,10 . These phenomena may be related to a specific distribution of bombesin receptors in the edge region of the cerebral cortex 11 . Furthermore, an increase in bombesin receptor activity may be indicative of a number of tumors, such as small cell lung cancer, gastric cancer, and neuroblastoma 12 .
Bombesin is a potent LES contractile agent in opossums. The contraction mechanism involves direct effects on smooth muscle and indirect postganglionic neuron stimulation 13 . Bombesin increases the LES pressure in humans and can affect esophageal motility by increasing peristaltic wave amplitude and duration. Neither the effects of vagal cholinergic agents nor the release of gastrointestinal hormones caused by bombesin are considered to be related to the mechanism of action of bombesin on esophageal motility in humans 14 .
Although previous studies found that bombesin can increase LES contraction in opossums and LES pressure in humans, the bombesin receptor subtypes involved and the underlying mechanism are still not clear 13,14 . Here, we used porcine LES to investigate which of the bombesin receptor subtypes, BB 1 , BB 2 , and BB 3 , are involved in bombesin-induced LES contraction. The results of this study support the bombesin drug potential in GERD treatment.

Results
The effects of bombesin and the bombesin receptor agonists on porcine LES sling and clasp muscle strip tension. The structure of the porcine LES is similar to that of the human LES. It is composed of sling muscle fibers at the greater curvature (GC) and clasp muscle fibers at the lesser curvature (LC). Bombesin receptor agonists including a BB 1 agonist, neuromedin B (NMB), a BB 2 agonist, gastrin-releasing peptide (GRP), and a BB 3 agonist, [D-Tyr 6 ,Apa-4Cl 11 ,Phe 13 ,Nle 14 ]bombesin-(6-14) (DTACPN-BN), were used to elucidate which bombesin receptor subtype or subtypes are involved in bombesin-induced LES contraction. Bombesin, GRP, and DTACPN-BN all elicited sustained and dose-dependent contraction of the isolated porcine LES sling and clasp muscle strip (Fig. 1A,B). Figure 2A shows that bombesin caused detectable contraction of the LES sling muscle strips at 1 nM and maximal contraction at 3 μ M. The maximal tension caused by 3 μ M bombesin was 36 ± 5% of the tension caused by 1 μ M carbachol (n = 4). In the LES clasp muscle strips, bombesin caused detectable contraction at 1 nM and maximal contraction at 1 μ M (Fig. 2B). The maximal tension caused by 1 μ M bombesin was 22 ± 5% of the tension caused by 1 μ M carbachol (n = 4). Similarly, GRP caused detectable contraction of the LES sling muscle strips at 1 nM, half-maximal contraction at 0.18 ± 0.04 μ M, and maximal contraction at 3 μ M ( Fig. 2A). The maximal tension caused by 3 μ M GRP was 38 ± 4% of the tension caused by 1 μ M carbachol (n = 8). In the LES clasp muscle strips, GRP caused detectable contraction at 1 nM, half-maximal contraction at 42 ± 10 nM, and maximal contraction at 3 μ M (Fig. 2B). The maximal tension caused by 3 μ M GRP was 26 ± 2% of the tension caused by 1 μ M carbachol (n = 7). DTACPN-BN, furthermore, caused detectable contraction of the LES sling muscle strips at 1 nM and maximal contraction at 3 μ M ( Fig. 2A). The maximal tension caused by 3 μ M DTACPN-BN was 42 ± 7% of the tension caused by 1 μ M carbachol (n = 4). In the LES clasp muscle strips, DTACPN-BN caused detectable contraction at 1 nM and maximal contraction at 3 μ M (Fig. 2B). The maximal tension caused by 1 μ M DTACPN-BN was 20 ± 2% of the tension caused by 1 μ M carbachol (n = 4). Both DTACPN-BN and GRP induced dose-dependent tonic contractions in the isolated porcine LES sling and clasp muscle strips. In contrast, NMB did not induce a contraction ( Fig. 2A,B). The effects of BB 1 , BB 2 and BB 3 antagonists on bombesin-induced porcine LES sling and clasp muscle strip contractions. To further examine whether the bombesin-induced LES contraction is mediated through BB 1, BB 2 and/or BB 3 , LES muscle strips were treated with BB 1 (PD 168368), BB 2 and BB 3 antagonists, respectively, before the bombesin stimulation. PD 168368 (1 μ M) and Bantag-1 (1 μ M) alone did not inhibit bombesin-induced contraction of both the porcine LES sling (Fig. 3E) and clasp (Fig. 3F) muscle strips (p > 0.05, n ≥ 4/group). However, the BB 2 antagonist DPDM-BN EA (1 μ M) shifted the bombesin concentration-response curve to the right but the difference between the EC 50 levels of bombesin alone and EC 50 levels of bombesin plus DPDM-BN EA was not significant (p > 0.05, n = 4/group). In contrast, the BB 2 antagonist DPDM-BN EA combining with BB 3 antagonist Bantag-1 (both 1 μ M) significantly shifted the bombesin concentration-response curve to the right, with ten-and sixty-fold changes at the EC 50 levels for the LES sling and clasp muscle strips, respectively. Specifically, the EC 50 levels of bombesin alone were 72 ± 21 nM (sling) and 10 ± 0.51 nM (clasp) (n = 4/group). The EC 50 levels of bombesin plus 1 μ M DPDM-BN EA combining with Bantag-1 were 0.72 ± 0.22 μ M (sling) and 0.69 ± 0.14 μ M (clasp), respectively (both p < 0.05, compared with bombesin alone; n = 6 and 5/ group, respectively). Thus, it is likely that the bombesin-induced LES contraction is mediated through BB 2 and BB 3 .

The effects of BB
The effects of tetrodotoxin (TTX), atropine, and nifedipine on GRP-induced porcine LES sling muscle strip contractions. Sling muscle strips exhibited greater contraction forces than the clasp muscle strips. The maximal contraction forces that were induced by GRP on the sling muscle strips were one and a half times greater than the clasp muscle strips, and the maximal contraction forces induced by DTACPN-BN on the sling muscle strips were two times greater than the clasp muscle strips (Fig. 2). Therefore, we used the sling muscle strips to investigate the mechanism underlying the GRP-induced contractions. The muscle strips were pretreated with TTX, atropine, and nifedipine. As shown in Fig. 4A, the GRP-induced dose-dependent contraction was almost unaffected by TTX (p > 0.05, compared with GRP alone, n = 4). Similarly, atropine had no significant effect on the GRP-induced contraction (p > 0.05, compared with GRP alone, n = 4). In contrast, the GRP-induced contraction of the LES sling muscle fibers was significantly inhibited by the L-type Ca 2+ channel blocker, nifedipine (p < 0.05, compared with GRP alone, n = 4), at a concentration of 1 μ M. Reverse transcription polymerase chain reaction (RT-PCR) analysis of BB 1 , BB 2 , and BB 3 transcript levels in the LES sling and clasp muscle fibers. Using BB 2 -specific primers, we could generate a PCR product of 487 bp (Fig. 5) from both the sling and clasp muscle fibers. Likewise, a BB 3 -specific PCR product of 373 bp was obtained from both the sling and clasp muscle fibers. The size of the β -actin PCR product (the internal control) was 148 bp, as predicted. No PCR product corresponding to the NMB receptor (BB 1 ) was detected in the cDNA that was generated from sling or clasp muscle fibers (n = 3).  Immunohistochemistry (IHC). Immunohistostaining of the neuromedin B receptor (BB 1 ) with a specific BB 1 antibody showed negative results in both the sling and clasp muscles. However, there was positive BB 1 immunostaining on the positive control tissues of Sprague Dawley rat testes (as shown in Supplementary Fig. S1). In contrast, IHC results revealed that the GRP receptor (BB 2 ) was ubiquitous in the LES sling (Fig. 6A) and clasp (Fig. 6B) muscle fibers (n = 3). In addition, immunosignals that were specific for the BB 3 were found in myenteric ganglia of the LES sling (Fig. 7A) and clasp (Fig. 7B) muscles (n = 3).

Discussion
Bombesin receptors are a family of G protein-coupled receptors, including the NMB receptor (BB 1 ) as well as GRP receptors (BB 2 ) and BB 3 . BB 2 is widely present in the pancreas, central nervous system, and gastrointestinal tract. BB 2 plays a role in many gastrointestinal functions, including gastrointestinal motility regulation, insulin stimulation, pancreatic and gastric acid secretion, colonic ion transport stimulation and the secretion of various hormones 15 . GRP, a BB 2 agonist, plays a major role in gastric emptying, small bowel transit, and gallbladder contraction 16 . In this study, we showed that GRP can induce LES sling and clasp muscle fiber contractions in pigs and that the selective BB 2 antagonist, DPDM-BN EA, can significantly inhibit GRP-induced LES contraction, indicating that GRP causes LES contraction via BB 2 .
The contraction of porcine LES smooth muscle fibers is modulated by multiple factors, especially autonomic neural innervations. The excitatory postganglionic vagal nerves transmit signals to nerve endings. Acetylcholine is released from the presynaptic nerve endings and stimulates postsynaptic muscarinic receptors. After stimulation of the muscarinic receptor, extracellular Ca 2+ flows into the intracellular space, principally through L-type Ca 2+ channels. Increased intracellular Ca 2+ binds to calmodulin, which subsequently leads to Ca 2+ /calmodulin-dependent activation of myosin light chain kinase (MCLK), which in turn induces smooth muscle contraction 17 .
In the first step, we used TTX, a selective neuronal Na + channel blocker, to explore whether GRP acts on the nerve fibers or directly on the LES smooth muscle. We found that TTX did not inhibit GRP-induced contraction of the LES, suggesting that the contraction was not due to an effect of GRP on the LES neuronal fibers. Additionally, atropine, a nonselective muscarinic receptor antagonist, was applied to investigate whether GRP acts directly on the muscarinic receptor of the LES smooth muscle, resulting in contraction. The GRP-induced contraction was not inhibited by atropine, indicating that GRP does not affect the LES muscarinic receptors. In contrast, we found that nifedipine could block GRP-induced LES contractions. Nifedipine is an L-type Ca 2+ channel blocker that inhibits extracellular Ca 2+ influx, and LES smooth muscle contraction is generally related to an increase in the intracellular Ca 2+ concentration 18 . Our results show that the extracellular Ca 2+ influx through the L-type Ca 2+ channel was indeed a main GRP-induced contraction pathway of the porcine LES. However, the L-type calcium blocker may also decrease the intracellular calcium stores in muscle strips owing to blockade of calcium influx which is needed for calcium store refilling. Therefore, the role of intracellular calcium stores cannot be excluded completely. BB 3 was first cloned from the guinea pig uterus 19 . BB 3 plays a role in the regulation of energy homeostasis, glucose/insulin levels, lung development, satiety, and injury 15 . Previous studies showed that BB 3 is present in the myenteric and submucosal ganglia in the tunica muscularis of the gastrointestinal tract and in nerve fibers between the myenteric ganglia. Additionally, with specific BB 3 antibodies, it was also detected in interstitial cells of Cajal (ICC) 20 . To date, the function of BB 3 in the gastrointestinal tract is not clear, and it was proposed that BB 3 might be related to gastrointestinal motility regulation 20 . In this study, we showed that the selective BB 3 agonist, DTACPN-BN, could induce LES sling and clasp muscle fiber contractions in pigs and that the selective BB 3 antagonist, Bantag-1, could significantly decrease porcine LES sling and clasp muscle fiber contractions, suggesting that DTACPN-BN causes LES contraction via BB 3 .
The neuronal Na + channel blocker, TTX, and the neuronal Ca 2+ channel blocker, CTX, were used to explore the mechanism of BB 3 agonist-induced LES contraction in pigs. Our results revealed that both TTX and CTX significantly decreased BB 3 agonist-induced contraction of the porcine LES sling muscle fibers. This finding implies that nerve conduction was associated with the BB 3 agonist-induced LES contraction.
RT-PCR identified BB 2 and BB 3 transcripts in the LES sling and clasp muscles; however, BB 1 transcripts were not detected. This finding is in good agreement with the results we obtained for the BB 2 and BB 3 agonist-induced LES sling and clasp muscle fiber contractions. Immunohistochemistry with specific BB 2 and BB 3 antibodies showed that BB 2 was mainly expressed in the muscle, while BB 3 was principally present in the ganglia. Taken together, these findings reveal that BB 2 and BB 3 are localized in the LES sling and clasp muscles in pigs at both the mRNA and protein levels.
Clinically, proton pump inhibitor refractory GERD can be non-reflux or reflux related. The reflux-related causes include residual acid reflux due to inappropriate medical treatment or poor compliance, non-acid reflux such as duodeno-gastro-esophageal (bile) reflux, hypersensitivity to physiological amounts of acid, and persistent impairment of esophageal mucosal integrity 2 . We used selective bombesin receptor agonists and antagonists to demonstrate that both BB 2 and BB 3 , but not BB 1 agonists, could cause tonic LES contraction in a dose-dependent manner. Therefore, the BB 2 and BB 3 receptors were both involved in the bombesin-induced LES contraction. These results support the hypothesis that bombesin and its agonists are a new kind of potential drug for the enhancement of esophageal sphincter contraction in order to avoid irritant reflux. These BB 2 and BB 3 agonists will be helpful to treat or as adjuvant therapy for patients with proton pump inhibitor refractory GERD or those suffering from the side effects of long-term PPI use.
In conclusion, the present study showed that bombesin-induced LES contraction occurs through BB 2 and BB 3 . This study also indicates that BB 2 -mediated contraction is not due to the action of nerve and muscarinic receptors, but rather is caused by the influx of extracellular Ca 2+ into myocytes through L-type Ca 2+ channels. In contrast, the BB 3 -mediated tonic LES contraction was associated with neuronal conduction. Therefore, bombesin, as well as BB 2 and BB 3 agonists, might have the potential to treat GERD.

Materials and Methods
Porcine stomach, including the lower esophagus, was purchased from a local slaughterhouse. The stomachs, from pigs weighing approximately 110 Kg, were placed in ice-cold Kreb-Henseleit buffer solution that had the following composition (in mM): 15.5 NaHCO 3 , 122 NaCl, 4.7 KCl, 1.2 KH 2 PO 4 , 1. The effects of bombesin, NMB, GRP, and DTACPN-BN on porcine LES sling and clasp muscle tension. The porcine LES is often used as an animal model to study esophageal motility and GERD 22,23 . Fresh porcine esophagus and stomachs were cut longitudinally along a midline between the GC and LC of the stomach. The mucosa was removed with surgical scissors. The bigger semicircular sling and smaller semicircular clasp muscle bundles of the LES were identified and obtained as previously described 24 . The muscle strips, which were 3 mm wide and 10 mm long, were trimmed from the sling and clasp muscles, respectively, and suspended in 7 ml organ baths containing Krebs-Henseleit buffer solution, incubated at 37 °C, and continuously gassed with 95% O 2 + 5% CO 2 . Subsequently, the muscle strips were connected to isometric force transducers (FORT10g, World Precision Instruments Inc., Sarasota, FL, USA), which were connected to amplifiers and a computer recording system (BIOPAC Systems, CA, USA). The basal tension of the muscle strips was set at 1.0 g. After a 30 min equilibration period, carbachol (1 μ M) was added to the organ bath, the muscle strip contractions were measured, and the carbachol was washed out. Carbachol-induced contraction served as a reference (100%) for the contractile response to bombesin or the selective agonists 25  antagonist, and Bantag-1 is a selective BB 3 antagonist. To investigate whether BB 2 is involved in bombesin-induced LES contraction, the isolated sling or clasp muscle strips were treated with DPDM-BN EA (1 μ M) for 6 min, followed by the addition of GRP to the organ bath in a cumulative manner 26 . Similarly, to investigate whether BB 3 is involved in bombesin-induced LES contraction, the isolated sling or clasp muscle strips were treated with Bantag-1 (1 μ M) for 6 min, followed by the addition of DTACPN-BN to the organ bath in a cumulative manner 21 . In addition, to investigate which receptor is involved in bombesin-induced LES contraction, the isolated sling or clasp muscle strips were treated with PD 168368, DPDM-BN EA, Bantag-1 (1 μ M respectively) or DPDM-BN EA combining with Bantag-1 (both 1 μ M) for 6 min, followed by the addition of bombesin to the organ bath in a cumulative manner.
The effects of TTX, CTX, atropine, and nifedipine on GRP-or DTACPN-BN-induced porcine LES sling muscle strip contractions. TTX, atropine, and nifedipine are neuronal Na + channel, muscarinic receptor, and L-type Ca 2+ channel blockers, respectively. To investigate the mechanism of GRP-or DTACPN-BN-induced LES contraction in the pigs, the isolated sling muscle strips were treated with TTX (1 μ M, 15 min), atropine (1 μ M, 6 min), or nifedipine (1 μ M, 20 min). GRP or DTACPN-BN was added to the muscles strips that were treated with TTX. Additionally, GRP was added to the muscle strips that were treated with either atropine or nifedipine. Furthermore, the neuronal Ca + channel blocker, CTX (1 μ M), was used to pretreat the isolated sling muscle strips, and 15 min later, DTACPN-BN was added to the organ bath 27-29 . RT-PCR for detection of bombesin receptor mRNA in porcine LES. Sling and clasp muscle fibers that were obtained from the porcine LES were stored in RNAlater solution (Applied Biosystems Inc., Foster City, CA, USA) at 4 °C for two days to allow for tissue penetration, and then, excess liquid was removed and the tissue was stored at − 80 °C until use.
RNA was isolated from tissue samples with the guanidine isothiocyanate method, using the GeneJET RNA Purification Kit (Thermo Fisher Scientific Inc., Waltham, USA), according to the manufacturer's protocol. The quality and purity of the isolated RNA were measured with an UV/Vis spectrophotometer (DU800, Beckman Coulter, CA, USA).
After RNA concentration measurements, cDNA was synthesized using the extracted RNA and the High Capacity cDNA Reverse Transcription Kit (Applied Biosystems Inc., CA, USA) for RT-PCR. The RNA sample preparation for reverse transcription was performed according to the manufacturer's recommendation. A volume corresponding to 1 μ g of RNA was used for the cDNA synthesis. The cDNA samples were diluted to a volume of 100 μ l and prepared for subsequent PCR.
Immunohistochemistry. Formalin-fixed and paraffin-embedded porcine sling and clasp muscle fibers were prepared for IHC staining. The IHC stains were performed using standard reagents and techniques on a BOND-MAX Automated Staining System (Leica Microsystems). Briefly, the deparaffinized and rehydrated sections were subjected to heat-induced antigen restoration with bond epitope retrieval solution 1 (citrate-based pH 6.0 solution, Leica Microsystems). The staining procedure involved a peroxidase block with 3% hydrogen peroxide for 5 min, incubation with a 1:300 dilution of rabbit neuromedin B receptor polyclonal antibody, LS-A825 (Lifespan Biosciences, Seattle, WA, USA) for BB 1 , a 1:300 dilution of rabbit GRPR polyclonal antibody (Abnova, Taipei, Taiwan) for BB 2 , or a 1:200 dilution of rabbit anti-Brs3/bombesin receptor 3 polyclonal antibody (Bioss, MA, USA) for BB 3 for 30 min at room temperature. Then, the samples were subsequently incubated with an anti-rabbit horseradish peroxidase polymer for 10 min at room temperature. The samples were then treated with a chromogen, 3,3′ -diaminobenzidine tetrahydrochloride (DAB), for 10 min at room temperature and counterstained with hematoxylin for 5 min. The sling and clasp muscle sections were stained with normal rabbit IgG at equimolar concentrations as negative controls. Data analysis. These data are expressed as the mean ± standard error of the mean (SEM). GraphPad Prism 5 was used to determine the half-maximal contraction (EC 50 ) values. The statistical analysis of the results was performed with Student's t-tests or one-way ANOVA followed by Tukey's post hoc test. In all cases, the differences were considered significant when p < 0.05.