GASTROENTEROLOGY 2010;138:1091–1101 Sensory and Motor Innervation of the Crural Diaphragm by the Vagus Nerves RICHARD L. YOUNG,*,‡ AMANDA J. PAGE,*,‡,§ NICOLE J. COOPER,* CLAUDINE L. FRISBY,* and L. ASHLEY BLACKSHAW*,‡,§ BACKGROUND & AIMS: During gastroesophageal reflux, transient lower esophageal sphincter relaxation and crural diaphragm (CD) inhibition occur concomitantly. Modifying vagus nerve control of transient lower esophageal sphincter relaxation is a major focus of development of therapeutics for gastroesophageal reflux disease, but neural mechanisms that coordinate the CD are poorly understood. METHODS: Nerve tracing and immunolabeling were used to assess innervation of the diaphragm and lower esophageal sphincter in ferrets. Mechanosensory responses of vagal afferents in the CD and electromyography responses of the CD were recorded in novel in vitro preparations and in vivo. RESULTS: Retrograde tracing revealed a unique population of vagal CD sensory neurons in nodose ganglia and CD motor neurons in brainstem vagal nuclei. Anterograde tracing revealed specialized vagal endings in the CD and phrenoesophageal ligament—sites of vagal afferent mechanosensitivity recorded in vitro. Spontaneous electromyography activity persisted in the CD following bilateral phrenicotomy in vivo, while vagus nerve stimulation evoked electromyography responses in the CD in vitro and in vivo. CONCLUSIONS: We conclude that vagal sensory and motor neurons functionally innervate the CD and phrenoesophageal ligament. CD vagal afferents show mechanosensitivity to distortion of the gastroesophageal junction, while vagal motor neurons innervate both CD and distal esophagus and may represent a common substrate for motor control of the reflux barrier. Keywords: Crural Diaphragm; Vagus Nerve; Retrograde Tracing; Anterograde Tracing; Electrophysiology. W ith a prevalence of 10%–20% in the adult population and up to 50% in infants, gastroesophageal reflux disease is one of the most common diseases of the upper gastrointestinal tract.4,5 The major barrier against reflux is formed by the lower esophageal sphincter together with the partly superimposed crural diaphragm (CD). Combined pH-metric and manometric studies have documented that transient lower esophageal sphincter relaxations (TLESR) account for most reflux episodes in healthy subjects and gastroesophageal reflux disease patients (for review, see Mittal et al6). During TLESR, there is simultaneous relaxation of the lower esophageal sphincter (LES), profound inhibition of activity in the CD, and esophageal shortening, permitting reflux of gastric content into the esophagus. Reduction of TLESRs may consequently be an effective means to prevent reflux from occurring.7,8 TLESR is triggered via a vagal pathway involving activation of gastric vagal mechanoreceptors, which provide input to nuclei in the medulla oblongata of the central nervous system, which in turn coordinate episodic activation of vagal pathways to inhibitory enteric neurons supplying smooth muscle of the LES (for review, see Blackshaw9). Inhibition of the CD that accompanies TLESR has been thought, like diaphragmatic motor control in general, to be mediated via the phrenic nerves, such that medullary nuclei provide inhibitory signals to the phrenic motor nucleus in the spinal cord, which shut down rhythmic output via !-motor neurons to the CD—an independent descending pathway to that controlling the LES. However, a number of studies have questioned the sole involvement of the phrenic nerve in reflex control of CD inhibition.10 –12 Although gastric afferent fibers in the vagus nerves are important in triggering TLESR, and LES vagal efferents are important in smooth muscle relaxation, the role of the vagus in the sensory and motor innervation of the CD and adjacent structures is unknown. We hypothesized that, in order to demonstrate such tight coordination with the LES, the CD must receive direct motor control via the vagus nerve. It was important to test this hypothesis in a species that displays TLESR and gastroesophageal reflux, so we chose the ferret.13,14 We took a multifaceted approach, Abbreviations used in this paper: CD, crural diaphragm; ChAT, choline acetyltransferase; CTB, cholera toxin subunit B; DMN, dorsal motor nucleus; EMG, electromyography; HRP, horseradish peroxidase; LES, lower esophageal sphincter; NG, nodose ganglia; PFA, paraformaldehyde; TLESR, transient lower esophageal sphincter relaxation; VAChT, vesicular acetylcholine transporter; WGA, wheat germ agglutinin. © 2010 by the AGA Institute 0016-5085/10/$36.00 doi:10.1053/j.gastro.2009.08.053 BASIC– ALIMENTARY TRACT *Nerve-Gut Research Laboratory, Hanson Institute, Department of Gastroenterology and Hepatology, Royal Adelaide Hospital, Adelaide, South Australia, Australia; ‡ Discipline of Medicine, Faculty of Health Sciences and §Discipline of Physiology, School of Molecular and Biomedical Sciences, University of Adelaide, Adelaide, South Australia, Australia 1092 YOUNG ET AL mapping the connections and neurochemistry of vagal innervation of the reflux barrier, and investigating the functional role of the vagus nerve in both sensory and motor control in isolated preparations of the CD. GASTROENTEROLOGY Vol. 138, No. 3 C5 and C6 spinal cord (and dorsal root ganglia). Control experiments for tracer injection are detailed in Supplementary Material. Anterograde Tracing Methods Animal Preparation Experiments were performed on adult ferrets (0.6 –1.4 kg) of either sex. All studies were performed in accordance with the Australian code of practice for the care and use of animals for scientific purposes and with the approval of the Animal Ethics Committees of the Institute of Medical and Veterinary Science (Adelaide, Australia) and the University of Adelaide. Animals had free access to water and a standard carnivore diet and were fasted overnight prior to experimentation. Restricted Volume Retrograde Tracing BASIC– ALIMENTARY TRACT Ferrets were anesthetized with isoflurane (2%–3% in oxygen) and a midline laparotomy performed to access the dorsal surface of the left and right diaphragm. In 26 animals, the viscera were retracted, the diaphragm gently tensioned laterally, and 2 "L fluorescein isothiocyanate conjugate of cholera toxin-B subunit (CTB-fluorescein isothiocyanate; Sigma-Aldrich, Castle Hill, Australia) or Alexa Fluor 555 conjugate (CTB-AF555; Invitrogen, Glen Waverley, Australia) injected beneath the peritoneal lining of the CD using a Hamilton microsyringe equipped with a 34-gauge needle (total volume ! 4 "L). The injection sites were 5 mm bilateral and adjacent to the esophageal hiatus and parallel to the orientation of the left and right crus muscle fibers. In 14 of these ferrets, dual-site tracing was performed by injecting 2 "L alternate tracer into the peritoneal lining of the dorsolateral costal diaphragm 10 –15 mm bilateral of midline (n ! 6), or into the dorsal and ventral subserosa of the distal esophagus (n ! 8). After injection, injection sites were dried, the laparotomy incision closed with individual sutures, and antibiotic (Terramycin, Pfizer, West Ryde, Australia, 10 mg/kg) and analgesic (Metacam, Boehringer Ingelheim, North Ryde, Australia, 2.5 mg/kg) administered subcutaneously. All ferrets recovered well from surgery and were routinely monitored. After 4 days, ferrets were given a lethal dose of sodium pentobarbitone (180 mg kg"1 intraperitoneally), the stomach filled with 60 mL 4% paraformaldehyde (PFA) in 0.1 M phosphate-buffered saline, pH 7.4 at 4°C via an oral polyethylene tube, and the ferret perfused transcardially with 500 mL heparinized saline (1000 IU/L) followed by 1000 mL 4% PFA at 4°C. The nodose ganglia (NG), brainstem, and cervical spinal cord with attached dorsal root ganglia were then removed, fixed overnight at 4°C in PFA, then cryoprotected in 30% sucrose at 4°C for 24 – 48 hours. Frozen transverse sections were then cut serially at 20 "m through the rostrocaudal axis of the NG, brainstem, and Ferrets were anesthetized as noted, a midline incision made, and the left NG exposed by blunt dissection. A 30-gauge needle was used to pierce the ganglion sheath and a glass capillary, drawn to a 25-"m tip, was used to deliver 5 "L wheat germ agglutinin conjugated to horseradish peroxidase (WGA-HRP; Sigma Aldrich) via pneumatic injection. After injection, injection sites were dried, the incision closed, and antibiotic (Terramycin, 10 mg/ kg) and analgesic (Metacam, 2.5 mg/kg) administered subcutaneously. All ferrets recovered well from surgery and were routinely monitored. After 4 days, ferrets were given a lethal dose of pentobarbitone and perfuse-fixed as mentioned previously. The stomach and reflux barrier were removed, the stomach was divided into ventral and dorsal halves along the lesser and greater curvatures and pinned flat. The submucosa and mucosa were removed from the stomach, leaving the muscle layers and myenteric plexus intact, which were fixed in 4% PFA overnight at 4°C. WGA-HRP was then subject to tyramide signal amplification with a tyramide signal amplification-Biotin Tyramide kit (Perkin-Elmer Life Sciences; Boston, MA) followed by detection with streptavidin Alexa Fluor 546 conjugate (Invitrogen). Dorsal and ventral stomach specimens were then whole-mounted on gelatin-coated slides, dried overnight, and vagal afferent endings visualized by fluorescence microscopy as positive control for anterograde tracing. The reflux barrier was prepared for sectioning by reducing the CD surrounding the esophageal hiatus to a 2-cm square and fixing the specimen overnight at 4°C in 4% PFA with an angled support inserted through the residual esophagus to preserve barrier anatomy. The specimen was then cryoprotected in 30% sucrose at 4°C for 24 to 48 hours, then cryosectioned in transverse (20 "m) through the dorsal-ventral axis. WGAHRP in vagal endings was visualized by epifluorescence and confocal microscopy in alternate serial sections of the reflux barrier— one series using the tyramide signal amplification protocol described here, and the alternate series following direct immunohistochemistry for WGA. Immunohistochemistry, Visualization, and Quantification Immunoreactivity for choline acetyltransferase (ChAT) and neuronal nitric oxide synthase were detected in traced brainstem sections, while vesicular acetylcholine transporter (VAChT) and WGA were detected in traced sections of the reflux barrier. Details of assay conditions, visualization, and quantification are detailed in Supplementary Material. Isolated Ferret CD Vagal Afferent Preparation This preparation was based on our isolated gastroesophageal preparation, which has been described previously.15,16 Ferrets were given a lethal dose of sodium pentobarbitone (180 mg kg"1, intraperitoneally), exsanguinated, and thoracic organs and diaphragm removed together and placed in a modified Krebs solution comprising (in mM): 118.1 NaCl, 4.7 KCl, 25.1 NaHCO3, 1.3 NaH2PO4, 1.2 MgSO4.7H2O, 1.5 CaCl2, 1.0 citric acid, and 11.1 glucose, bubbled with 95% O2"5% CO2, pH 7.4 at 4°C. Thoracic organs were removed and the vagus nerve cleared of connective tissue. The diaphragm with residual esophagus was pinned out flat in one chamber of a 2-chamber organ bath, perfused with modified Krebs solution at 34°C. The L-type calcium channel antagonist nifedipine (1 "M) was then added to the superfusing solution to suppress smooth muscle activity. The vagus nerves were led into an adjacent chamber through a small hole, bathed in mineral oil, and dissected into strands for single-fiber electrophysiological recordings. The location of receptive fields of vagal afferent endings within the CD, phrenoesophageal ligament, and associated connective tissue was determined by mechanical stimulation with a brush, then high-threshold mechanical responses were classified using a blunt glass rod.17 Accurate quantification of mechanical responses of CD vagal afferents was measured by the focal movement of calibrated von Frey hairs across the receptive field without applying stretch; stretch was applied across receptive fields by the use of calibrated weights and a pulley-cantilever system.17 Single neurons were discriminated offline using Spike2 (Cambridge Electronic Design, Cambridge, UK) based on action potential shape, duration, and amplitude. Peristimulus time histograms and instantaneous frequency plots were generated from these neurons using Spike2 analysis. In some experiments, receptive fields were marked with carbon particles and the surrounding tissue removed, fixed overnight at 4°C in PFA, then immunohistochemistry performed by incubating with rabbit antiserum against the general neural marker protein gene product (PGP9.5), followed by detection with streptavidin Alexa Fluor 546 conjugate (Invitrogen). The specimen was then whole-mounted on gelatin-coated slides, dried overnight, and individual vagal afferent endings visualized by fluorescence microscopy. Isolated Ferret CD Vagal Efferent Preparation Tissue collection and preparation was identical to that for vagal afferent preparation, except that the phrenic nerves were preserved and also cleared of connective tissue. The diaphragm with residual esophagus was pinned out in a single-chamber organ bath and perfused with modified Krebs solution at 34°C. Vagus or phrenic nerve trunks were stimulated individually via hook elec- A VAGAL INNERVATION OF THE CRURAL DIAPHRAGM 1093 trodes using a pulse generator and a constant-current stimulus unit (JRAK, Melbourne, Australia); a 5-V electrical stimulus of 0.5 Hz and 0.5-ms pulse width was used to evoke electromyograms (EMG) in esophagus, CD, or costal diaphragm. Stimulation of the phrenic nerve trunk was used, as the CD branch was difficult to distinguish in ferrets. EMG responses were recorded via fine bipolar electrodes inserted into the CD parallel with the muscle fibers, 1–2 mm adjacent the esophageal hiatus. Bipolar intramuscular electrodes were also placed in the ventral wall of the distal esophagus or in the lateral portion of the costal diaphragm for validation of EMG responses. In 4 separate preparations, after EMG responses had been determined, a ganglioplegic drug, hexamethonium (100 "M), was added to the superfusing Krebs solution and allowed to equilibrate for 20 minutes, after which EMG responses in the CD were retested. Subsequently, a neuromuscular junction-blocking drug, pancuronium (5 "M), was added to the superfusing solution, and EMG responses retested after 20 minutes. This was performed to establish whether EMG responses in the CD were mediated through vagal postganglionic collateral innervation, the neuromuscular junction, or by direct muscle excitation. Time control experiments were also performed in which there was no change in EMG responses over a comparable duration. Details of sequential dissection experiments are presented in Supplementary Material. In Vivo Ferret CD Experiments The function of vagal innervation of the CD was also assessed in anesthetized ferrets subject to bilateral thoracic phrenicotomy and vagus nerve stimulation. Full details of these experiments are presented in Supplementary Material. Results Vagal Innervation of the CD To investigate the possibility of a vagal innervation of the CD, we observed labeling in vagal neurons after tracer injections into the CD (Figure 1A). We found unique labeling of 45 # 13 vagal afferent neurons per NG that were distributed throughout the ganglion (Figure 1B). This contrasted with earlier studies tracing from gastric muscle with larger volumes, in which 11.2% ($2600) of vagal afferents were labeled per NG.18,19 In brainstem, a total of 133 # 14 vagal motor neurons were labeled bilaterally from the CD in each animal; these were predominantly located rostral from the level of the obex within lateral regions of the dorsal motor nucleus (DMN; Figure 1E–1F). In contrast, tracer injected bilaterally into the dorsolateral costal diaphragm did not label any vagal neurons. As expected, tracer injection in the distal esophagus labeled a population of 123 # 9 vagal afferent neurons per NG with a general distribution, and 69 # 11 BASIC– ALIMENTARY TRACT March 2010 1094 YOUNG ET AL GASTROENTEROLOGY Vol. 138, No. 3 BASIC– ALIMENTARY TRACT Figure 1. Retrograde tracing of vagal nerve pathways from the crural diaphragm (CD) and distal esophagus in ferrets. (A) Abdominal surface view of the diaphragm showing injection sites of retrograde tracers cholera toxin subunit B (CTB)-AF555 (CD, red) and CTB-fluorescein isothiocyanate (distal esophagus, green). (B) Transverse view of left nodose ganglia (NG) showing a vagal afferent traced from the CD with CTB-AF555 (white). (C) Low magnification view of left NG showing afferents traced from the CD (CTB-AF555, red) and distal esophagus (CTB-fluorescein isothiocyanate, green). (D) High-magnification view of left NG showing vagal afferents traced separately, and a rare afferent dual-traced from the CD (CTB-AF555, red) and distal esophagus (CTB-fluorescein isothiocyanate, green). (E) Diagram of the ferret rostral brainstem adapted from Boissonade et al.20 (F) Transverse view of the rostral brainstem depicted in (E, boxed) showing a vagal motor neuron in the dorsal motor nucleus (DMN) traced from the CD by CTB-AF555 (white). (G) High-magnification view of the DMN showing vagal motor neurons traced separately and dual traced (yellow) from the CD (CTB-AF555, red) and distal esophagus (CTB-fluorescein isothiocyanate, green); weakly filled dendrites are evident within the nucleus centralis following distal esophagus tracing. Note: (B–D, and G) are intentionally imaged at lower contrast to identify nontraced neuron populations. Key for image (I): sg, subnucleus gelatinosus; mn, medial subnucleus of the nucleus tractus solitarius (NTS); ncom, commissural subnucleus of NTS; DMN, dorsal motor nucleus of the vagus; XII, hypoglossal nucleus; IC, nucleus intercalatus; cc, central canal; TS, solitary tract; d/dln, dorsal/dorsolateral subnucleus; v/vln, ventral/ventrolateral subnucleus; nI, intermediate subnucleus; ni, interstitial subnucleus. Scale bar (B, D, G) ! 50 "m, (C) ! 500 "m, (E, F) ! 200 "m. motor neurons in the DMN, predominantly rostral from the level of the obex. Esophageal motor neurons in the nucleus ambiguus were occasionally traced, as expected. Vagal motor neurons traced from the distal esophagus also showed weakly filled dendrites concentrated within the nucleus centralis of the nucleus tractus solitarius, the presumed terminal area of esophageal vagal afferents (Figure 1G).20,21 This was not apparent in vagal motor neurons traced from the CD. Vagal afferents were generally traced separately from CD or distal esophagus, although 8 # 6 double-labeled neurons were identified per NG (Figure 1C and 1D). In contrast, a large population of vagal motor neurons were double-labeled from the distal esophagus and CD (46 # 9 neurons, Figure 1G). Data on spinal innervation of the CD and on control experiments for tracer injection are presented in Supplementary Figure 1 and in Supplementary Material. Because there was strong evidence for a vagal sensory innervation of the CD, we sought to visualize the structure of their peripheral endings. This was achieved by anterograde tracing from the left NG with WGA-HRP. Strikingly, vagal fibers and endings were labeled within the CD and phrenoesophageal ligament. In contrast, no endings were labeled within the costal diaphragm. Vagal endings within the CD were evident as discrete regions of fine punctate labeling at high-power magnification, often with clear evidence of the entry of a single parent axon (Figure 2A). Vagal punctate endings were identified predominantly within connective tissue at the abdominal surface of the CD, ipsilateral to the injected ganglion. March 2010 A VAGAL INNERVATION OF THE CRURAL DIAPHRAGM 1095 They were also found in lower numbers within the phrenoesophageal ligament at the esophageal hiatus and in connective tissue at the thoracic surface of the CD. These endings were also found clustered in fascia of the CD muscle (Figure 2B), occasionally in direct contact with individual muscle fibers (Figure 2C). Vagal punctate endings showed a limited distribution in the CD, restricted laterally to within 400 "m of the esophageal hiatus and within the 5 mm dorsoventral span of the esophagus. Labeling by WGA-HRP tracer was subsequently confirmed by the WGA antiserum, which labeled vagal fibers (Figure 2D) and punctate endings in the CD; these data confirm the specificity of the tracing method. Following anterograde tracing, a second type of vagal ending was identified only within the phrenoesophageal ligament. Vagal laminar endings were labeled within the phrenoesophageal ligament, oriented in the oral-aboral axis adjacent to the tether point of the ligament with the esophageal serosa (Figure 2E). These laminar endings were frequently located in-series with each other and bilateral to the esophageal hiatus and were absent from lateral regions of the ligament near the insertion with CD Figure 3. Vagal motor neurons that innervate the crural diaphragm (CD) express markers of excitatory neurotransmission. (A) High magnification view of the dorsal motor nucleus showing vagal motor neurons traced from the CD (cholera toxin subunit B [CTB]-AF555, red) immunopositive for choline acetyltransferase (ChAT) (green). (B) A CD vagal motor neuron (CTB-AF555, red) immunonegative for neuronal nitric oxide synthase (green). Scale bar (A, B) ! 50 "m. BASIC– ALIMENTARY TRACT Figure 2. Anterograde tracing of vagal nerve pathways to the crural diaphragm (CD) and phrenoesophageal ligament in ferrets. (A) High magnification extended focal image (EFI) of a vagal punctate ending filled by wheat germ agglutinin– horseradish peroxidase (WGA-HRP) (with parent axon) located within connective tissue of the abdominal CD surface. (B) EFI image of vagal punctate endings in connective tissue of the abdominal CD surface, close to muscle fascia. (C) High magnification view of a punctate ending associated with CD muscle [M]. (D) Vagal fibers immunopositive for WGA in connective tissue of the abdominal CD surface. (E) Vagal laminar endings labeled by WGA-HRP in the phrenoesophageal ligament; these were located adjacent the anchor point with the thoracic distal esophagus and bilateral to the esophageal hiatus. ESO, esophagus; PEL, phrenoesophageal ligament (inset box depicts vagal ending shown in F). (F) 3-dimensional confocal projection view (Z ! 40 "m, 1 "m slice) of a labeled vagal laminar ending showing detailed tertiary structure and a parent axon entering at the apex. Scale bar (A, B, D) ! 10 "m, (C) ! 50 "m, (E) ! 100 "m, (F) ! 25 "m. 1096 YOUNG ET AL GASTROENTEROLOGY Vol. 138, No. 3 BASIC– ALIMENTARY TRACT Figure 4. Functional innervation of the crural diaphragm (CD) by vagal afferents. (A) Thoracic surface view of the diaphragm showing major structures and mechanosensory sites recorded within the CD in the single vagal fibers. Individual vagal afferent fiber responses recorded to high-threshold mechanical stimuli (blunt probing with rod, B) and low-threshold stimuli (10 mg von Frey hair, C) in the right crus muscle of the CD. (D) A “multimodal” vagal afferent fiber located in the left crus-phrenoesophageal ligament that responded to von Frey hair stimuli (70 –1000 mg, D) and to stretch applied across the receptive field. (E) High-magnification view of PGP9.5 immunoreactive tissue whole-mount corresponding to the marked receptive field recorded in (D); a putative vagal laminar ending in connective tissue possessing distinct arborizing laminar structures and a clear parent axon is evident. Scale bar (E) ! 25 "m. muscle. Confocal imaging of these endings revealed a distinct tertiary structure, comprised of profusely arborizing laminar structures similar to that described for intraganglionic laminar endings;22 a distinct entry point of a parent axon was visible at the apex of a number of these endings (Figure 2F). Additional images of vagal fibers and endings labeled within the CD and phrenoesophageal ligament are provided in Supplementary Figure 2. Neurochemistry of Vagal Innervation of the CD To identify the neurochemical phenotype of vagal motor neurons that innervate the CD, we performed immunolabeling in retrograde-labeled brainstem. We ob- served that vagal motor neurons in the DMN were frequently immunopositive for ChAT as expected, with the majority showing heavy immunolabeling for ChAT in cytoplasm and processes. In contrast, immunolabeling for neuronal nitric oxide synthase was sparse in the DMN, with only a few motor neurons labeled in rostrolateral regions. Vagal motor neurons retrogradely traced from the CD were exclusively ChAT immunopositive (Figure 3A) and neuronal nitric oxide synthase immunonegative (Figure 3B). Immunoreactivity for VAChT was detected at CD motor end-plates in anterograde-labeled sections of the reflux barrier, as expected (Supplementary Figure 2E); however, none of the identified vagal structures were immunopositive for VAChT. These findings indicate that vagal punctate endings in the CD and laminar March 2010 A VAGAL INNERVATION OF THE CRURAL DIAPHRAGM 1097 vious in vitro recordings of gastroesophageal mechanoreceptors in ferrets, vagal afferents with receptive fields in the CD or phrenoesophageal ligament were rare and represented %5% of all vagal afferents encountered in this preparation. To determine if mechanoreceptive sites in the phrenoesophageal ligament contained anatomically distinct endings, we performed immunolabeling for PGP9.5 in tissue specimens following electrophysiology. At marked mechanoreceptive sites in the phrenoesophageal ligament we observed endings in connective tissue that possessed distinct tertiary and arborizing laminar structures (Figure 4E) that were identical to anterograde-identified vagal laminar endings. Figure 5. Functional innervation of the crural diaphragm (CD) by vagal motor neurons. (A) Electromyography (EMG) responses recorded in the right crus muscle of the CD in response to stimulation of individual phrenic and vagus nerve trunks. (B) Single EMG waveform corresponding to responses in (A). (C) Frequency-dependent fatigue of EMG responses of right crus muscle to right vagus nerve trunk stimulation at frequencies of 2.5–50 Hz. endings in the phrenoesophageal ligament were of vagal afferent origin, and that vagal motor endplates were not filled from efferent fibers of passage during NG injections. Functional Innervation of the CD by Vagal Afferent Neurons To determine the function of vagal sensory innervation of the CD, we tested the mechanosensitivity of vagal afferent fibers in an in vitro diaphragm preparation. We observed receptive fields of vagal mechanoreceptors within the CD muscle (in association with muscle bundles of the right crus), within the phrenoesophageal ligament, and in connective tissue adjacent to the distal esophagus and CD (Figure 4A). Receptive fields were small (1–3 mm2) and distinct. Two types of vagal mechanoreceptor were identified in the CD; those that responded only to high-intensity blunt probing with a glass rod (Figure 4B), corresponding to high-threshold mechanoreceptors, and those that responded to focal compression with von Frey hairs with forces as low as 10 mg, corresponding to low-threshold mechanoreceptors (Figure 4C). These receptors did not respond to stretch up to 10 g, and were generally silent at rest. Two types of vagal mechanoreceptors were identified within the phrenoesophageal ligaments that were distinct from types in the CD. These responded either to calibrated von Frey hairs, as “dynamic mechanoreceptors,” or responded both to calibrated von Frey hairs and to stretch applied across the receptive field, as “multimodal afferents” (Figure 4D). In comparison with our pre- To determine the function of vagal motor innervation of the CD, we tested the responses of CD muscle to vagal stimulation in an in vitro diaphragm preparation and in vivo. Electrical stimulation of either phrenic nerve in vitro produced EMG responses within the dorsolateral costal diaphragm and in crus muscles of the CD, as expected (Figure 5A and 5B). EMG responses evoked by phrenic nerve stimulation comprised compound action potentials of amplitude in direct proportion to that of the stimulus, with ipsilateral predominance. Electrical stimulation of either vagus nerve produced EMG responses within the distal esophageal striated muscle, as expected, but also evoked EMG responses within crus muscles of the CD (Figure 5A and 5B). EMG responses evoked in the CD by stimulation of the dedicated phrenic nerve input were 2.9-fold larger than following vagal stimulation, as would be expected from the increased density of motor neurons traced from the CD in the phrenic motor nucleus than in the DMN (see Figure 1, Supplementary Figure 1). Vagal evoked EMG in the CD was free of artefact, showed frequency-dependent fatigue (&10 Hz) and was therefore characteristic of a muscle Figure 6. Direct innervation of crural diaphragm (CD) neuromuscular junctions by vagal motor neurons. Electromyography (EMG) responses recorded in right crus muscle in response to stimulation of the right phrenic and vagus nerve trunks in the absence (left), or presence of the autonomic ganglion blocker, hexamethonium (100 "M, middle) or neuromuscular junction blocker, pancuronium (5 "M, right). Drugs were administered to superfusing Krebs; responses recorded were after a 30-minute equilibration period. BASIC– ALIMENTARY TRACT Functional Innervation of the CD by Vagal Motor Neurons 1098 YOUNG ET AL GASTROENTEROLOGY Vol. 138, No. 3 Figure 7. Effects of bilateral phrenicotomy on electromyography (EMG) responses of the of crural diaphragm (CD) in vivo. (A) EMG responses recorded in left crus muscle in anesthetized, ventilated ferrets prior to (left), and following bilateral phrenicotomy (middle, right). (A, right) Expanded view of tonic EMG spikes present in CD (arrows) but absent from costal diaphragm (COSD) following phrenicotomy. Note: Cardiac electric interference is evident in both COSD and CD recordings. (B) EMG responses of COSD and CD following bilateral phrenicotomy and 20-V 0.5-ms pulse width vagal stimulation, showing short latency evoked EMG responses in the CD (filled arrowhead) and longer variable latency reflex EMG responses in CD (open arrowhead). BASIC– ALIMENTARY TRACT response (Figure 5C). These responses were not due to current spread from the esophagus, as phrenic stimulation did not produce EMG responses in the distal esophagus, nor did vagus stimulation produce EMG responses in dorsolateral costal diaphragm. Blockade of ganglionic transmission with hexamethonium (100 "M) failed to significantly alter vagal evoked EMG responses in the CD in this preparation, whereas pancuronium (5 "M) inhibited all evoked EMG responses in the CD by blockade at phrenic and vagal motor endplates (Figure 6). In in vivo experiments, bilateral phrenicotomy in anesthetized and ventilated ferrets led to a complete loss of EMG activity in the costal diaphragm, while spontaneous EMG activity continued in the CD (Figure 7A). Subsequent electrical stimulation of the intact vagus trunks evoked short-latency EMG spikes in the CD, and increased EMG frequency after a longer and more variable latency (Figure 7B). Full details of these experiments are provided in Supplementary Material. Discussion Our major findings of this study are that the vagus nerve provides a functional sensory and motor innervation of the CD. We have characterized high- and low-threshold vagal mechanoreceptors in CD muscle and dynamic and multimodal mechanoreceptors within the phrenoesophageal ligament. Our anatomical studies confirm that specialized vagal afferent endings exist in these locations. Furthermore, we have characterized the anatomy and function of a direct vagal motor innervation of the CD that is excitatory and tonically active in vivo. Our data suggest that vagal afferents may sense stimuli specific to the phrenoesophageal ligament and CD. We have also established that vagal motor neurons provide innervation of the LES and CD, and may represent common targets of a central pattern generator for TLESR and gastroesophageal reflux. A summary diagram depicting this functional neurocircuitry associated with motor control of the LES and CD is shown in Figure 8. Vagal Afferent Innervation of the CD Results of this study provide the first direct anatomical and functional evidence that a small population of vagal afferents directly and uniquely innervates the CD in ferrets. Vagal afferent cell bodies were not labeled after retrograde tracing from the costal diaphragm, nor were vagal endings anterogradely traced from the NG to the costal diaphragm, indicating that the CD is unique in receiving a dual sensory innervation from the phrenic and vagus nerves. Vagal afferents in the NG were occasionally double-labeled following tracing from the LES and CD, raising the intriguing possibility that single vagal afferents may receive convergent sensory input from the LES and CD, as is the case with those innervating other regions of the stomach.23 Vagal afferent laminar endings identified within the phrenoesophageal ligament in this study possessed a tertiary structure comparable with gastroesophageal mechanoreceptive intraganglionic laminar endings. Afferents Figure 8. Neurocircuitry of motor control of the lower esophageal sphincter (LES) and crural diaphragm (CD). A central pattern generator for transient lower esophageal sphincter relaxation (TLESR) coordinates premotor input to LES- and CD-projecting vagal motor neurons in dorsal motor nucleus (DMN), and to CD-projecting phrenic motor neurons in spinal cord. During a TLESR, this pattern generator provides premotor excitation to vagal motor neurons that excite postganglionic nonadrenergic, noncholinergic enteric neurons within the LES. Simultaneously, premotor inhibition is provided to three tonically active circuits: (i) to vagal motor neurons that excite postganglionic cholinergic enteric neurons within the LES, (ii) to vagal motor neurons that directly excite motor endplates within the CD, and (iii) to phrenic motor neurons via a yet to be defined pathway, possibly involving the DMN.37 In this manner, dual and synchronous output from this central pattern generator will relax the LES and inhibit the CD, creating a common cavity that favors reflux. Note: this simplified diagram does not show bilateral innervation of the LES or CD, or innervation of the costal diaphragm. recorded from this location in vitro also responded to calibrated von Frey hairs and to stretch, similar to esophageal tension-mucosal receptors reported in this species.15 These laminar endings may therefore sense axial distension of the LES or stretch in the phrenoesophageal ligament. In addition, the extraluminal location of vagal laminar and punctate endings within the reflux barrier and their relevant mechanical sensitivity suggest they respond to perturbed barrier anatomy or to changing posture— conditions known to strongly influence the frequency of TLESR.24 –26 Future experiments involving selective dennervation of these CD vagal afferents will reveal whether they serve as additional triggers proposed for TLESR.27 Vagal Motor Innervation of the CD Results of this study show that vagal motor neurons provide a direct innervation to the CD, but not the A VAGAL INNERVATION OF THE CRURAL DIAPHRAGM 1099 costal diaphragm in ferrets, a finding recently corroborated by another group.28 Therefore, the CD is equipped with innervation for both respiratory and anti-reflux barrier control, consistent with evidence of a dual functional role.29 –31 A population of nitrergic vagal motor neurons that directly innervate the LES in ferrets has been described previously within the rostrocaudal DMN.21 We confirmed this population of nitrergic motor neurons, but showed that CD vagal motor neurons were exclusively cholinergic and therefore excitatory, expressing ChAT centrally and VAChT at peripheral endings. Studies in the isolated diaphragm preparation confirmed that vagal motor innervation to the CD was functionally excitatory, and that EMG responses in crus muscles occurred, as expected, at a classical neuromuscular junction because responses to vagal stimulation were abolished by pancuronium. A similar innervation of striated muscle by vagal parasympathetic nerves has been described previously, in esophageal striated muscle of cats32 and pigs.33 Our finding that vagal motor innervation of the CD was tonically active and independent of cardiac or respiratory modulation in vivo indicates separate pathways of central control of the CD, as proposed by earlier work.10 In addition, vagal reflex activation of EMG responses in the CD in vivo provides strong evidence of a bilateral and multisynaptic central pathway that converges on DMN neurons that innervate the CD. Because of the significant contribution of cholinergic tone to basal LES pressure34 and tonic activity of CD vagal motor neurons revealed in the current study, a TLESR event is likely to involve dual central inhibition of excitatory vagal inputs to the LES and CD. This inhibition must occur simultaneously with inhibition of phrenic motor drive to the CD and with activation of inhibitory vagal inputs to the LES. Indeed, LES and CD projecting vagal motor neurons may represent an important point of convergence of a central pattern generator for gastroesophageal reflux. Although our data clearly show different mechanisms of excitatory input to the CD, it is important to note that our experiments do not address the central mechanism resulting in inhibition of this input during TLESR. This is an important subject for further study, which must now include both central vagal and phrenic pathways as candidates. In addition to the pathways we have described in this study, there is evidence for a peripheral mechanism of reflex inhibition of the CD. Liu and colleagues12,35 showed that distension of the distal esophagus in cats triggered relaxation of the esophagogastric junction via a pathway that was unaffected by bilateral phrenicotomy or vagotomy. It is important to note, however, that patterns of CD inhibition differ in duration and degree of inhibition during TLESR compared to those following esophageal distension.36 Our evidence of a functional sensory and motor vagal innervation of the CD provides strong support for a central vagal pathway that may exert BASIC– ALIMENTARY TRACT March 2010 1100 YOUNG ET AL dual control of the LES and CD during TLESR. The current data, however, does not refute the existence of a peripheral mechanism for CD inhibition in response to distension of the distal esophagus; indeed both mechanisms may operate at the CD under different conditions. In conclusion, these data provide important new clues to the role of the vagus and CD in the control of the antireflux barrier. Results here show that the CD and phrenoesophageal ligament are uniquely innervated by vagal afferent and motor neurons and that specialized vagal endings within these tissues correlate with sensory and motor responses recorded within the CD. It is likely, therefore, that a central pattern generator for TLESR acts within the DMN to differentially control vagal motor neurons innervating the LES and to inhibit vagal motor neurons innervating the CD, together with central inhibition of phrenic motor neurons. The identity of molecular targets specific to vagal innervation of the CD may provide new possibilities for the pharmacotherapy of gastroesophageal reflux disease. In addition, these data provide important new information on the consequences of surgical manipulation of the gastroesophageal junction and should help to refine and direct surgical strategies. BASIC– ALIMENTARY TRACT Supplementary Material Note: To access the supplementary material accompanying this article, visit the online version of Gastroenterology at www.gastrojournal.org, and at doi: 10.1053/j.gastro.2009.08.053. References 1. Young RL, Cooper NJ, Blackshaw LA. A novel innervation of the crural diaphragm by the vagus. Gastroenterology 2006; 130(Suppl):453. 2. Young RL, Cooper NJ, Blackshaw LA. A functional innervation of the crural diaphragm by the vagus. Gastroenterology 2007; 132(Suppl):1583. 3. Young RL, Cooper NJ, Blackshaw LA. Anatomy of vagal afferent endings in the crural diaphragm. Gastroenterology 2008; 134(Suppl):524. 4. Sonnenberg A, El-Serag HB. Clinical epidemiology and natural history of gastroesophageal reflux disease. Yale J Biol Med 1999; 72:81–92. 5. Nelson SP, Chen EH, Syniar GM, et al. Prevalence of symptoms of gastroesophageal reflux during infancy. A pediatric practicebased survey. Pediatric Practice Research Group. Arch Pediatr Adolesc Med 1997;151:569 –572. 6. Mittal RK, Holloway RH, Penagini R, et al. Transient lower esophageal sphincter relaxation. Gastroenterology 1995;109:601– 610. 7. Hirsch DP, Tytgat GN, Boeckxstaens GE. Transient lower oesophageal sphincter relaxations—a pharmacological target for gastrooesophageal reflux disease? Aliment Pharmacol Ther 2002;16: 17–26. 8. Lehmann A. Inhibitors of transient lower esophageal sphincter relaxations (reflux inhibitors) in the future treatment of GERD. Gastroenterol Hepatol Ann Rev 2006;1:109 –117. GASTROENTEROLOGY Vol. 138, No. 3 9. Blackshaw LA. New insights in the neural regulation of the lower oesophageal sphincter. Eur Rev Med Pharmacol Sci 2008; 12(Suppl 1):33–39. 10. Altschuler SM, Davies RO, Pack AI. Role of medullary inspiratory neurones in the control of the diaphragm during oesophageal stimulation in cats. J Physiol 1987;391:289 –298. 11. Oyer LM, Knuth SL, Ward DK, et al. Reflex inhibition of crural diaphragmatic activity by esophageal distention in cats. Respir Physiol 1989;77:195–202. 12. Liu J, Puckett JL, Takeda T, et al. Crural diaphragm inhibition during esophageal distension correlates with contraction of the esophageal longitudinal muscle in cats. Am J Physiol 2005;288: G927–G932. 13. Blackshaw LA, Staunton E, Dent J, et al. Mechanisms of gastrooesophageal reflux in the ferret. Neurogastroenterol Motil 1998; 10:49 –56. 14. Staunton E, Smid SD, Dent J, et al. Triggering of transient LES relaxations in ferrets: role of sympathetic pathways and effects of baclofen. Am J Physiol 2000;279:G157–G62. 15. Page AJ, Blackshaw LA. An in vitro study of the properties of vagal afferent fibres innervating the ferret oesophagus and stomach. J Physiol 1998;512:907–916. 16. Page AJ, Blackshaw LA. GABA(B) receptors inhibit mechanosensitivity of primary afferent endings. J Neurosci 1999;19:8597– 8602. 17. Page AJ, Martin CM, Blackshaw LA. Vagal mechanoreceptors and chemoreceptors in mouse stomach and esophagus. J Neurophysiol 2002;87:2095–2103. 18. Young RL, Cooper NJ, Blackshaw LA. Chemical coding and central projections of gastric vagal afferent neurons. Neurogastroenterol Motil 2008;20:708 –718. 19. Young RL, Cooper NJ, Blackshaw LA. Anatomy and function of group III metabotropic glutamate receptors in gastric vagal pathways. Neuropharmacology 2008;54:965–975. 20. Boissonade FM, Davison JS, Egizii R, et al. The dorsal vagal complex of the ferret: anatomical and immunohistochemical studies. Neurogastroenterol Motil 1996;8:255–272. 21. Hyland NP, Abrahams TP, Fuchs K, et al. Organization and neurochemistry of vagal preganglionic neurons innervating the lower esophageal sphincter in ferrets. J Comp Neurol 2001;430:222– 234. 22. Fox EA, Phillips RJ, Martinson FA, et al. Vagal afferent innervation of smooth muscle in the stomach and duodenum of the mouse: morphology and topography. J Comp Neurol 2000;428:558 –576. 23. Berthoud HR, Lynn PA, Blackshaw LA. Vagal and spinal mechanosensors in the rat stomach and colon have multiple receptive fields. Am J Physiol 2001;280:R1371–R1381. 24. Ireland AC, Dent J, Holloway RH. Preservation of postural control of transient lower oesophageal sphincter relaxations in patients with reflux oesophagitis. Gut 1999;44:313–316. 25. Kahrilas PJ, Shi G, Manka M, et al. Increased frequency of transient lower esophageal sphincter relaxation induced by gastric distention in reflux patients with hiatal hernia. Gastroenterology 2000;118:688 – 695. 26. Omari TI, Rommel N, Staunton E, et al. Paradoxical impact of body positioning on gastroesophageal reflux and gastric emptying in the premature neonate. J Pediatr 2004;145:194 –200. 27. Scheffer RC, Tatum RP, Shi G, et al. Reduced tLESR elicitation in response to gastric distension in fundoplication patients. Am J Physiol 2003;284:G815–G820. 28. Niedringhaus M, Jackson PG, Pearson R, et al. Brainstem sites controlling the lower esophageal sphincter and crural diaphragm in the ferret: a neuroanatomical study. Auton Neurosci 2008; 144:50 – 60. 29. Mittal RK, Balaban DH. The esophagogastric junction. N Engl J Med 1997;336:924 –932. 30. Pickering M, Jones JF. The diaphragm: two physiological muscles in one. J Anat 2002;201:305–312. 31. Shafik A, Shafik I, El Sibai O, et al. The effect of esophageal and gastric distension on the crural diaphragm. World J Surg 2006; 30:199 –204. 32. Collman PI, Tremblay L, Diamant NE. The central vagal efferent supply to the esophagus and lower esophageal sphincter of the cat. Gastroenterology 1993;104:1430 –1438. 33. Wu M, Majewski M, Wojtkiewicz J, et al. Anatomical and neurochemical features of the extrinsic and intrinsic innervation of the striated muscle in the porcine esophagus: evidence for regional and species differences. Cell Tissue Res 2003;311:289 –297. 34. Dodds WJ, Dent J, Hogan WJ, et al. Effect of atropine on esophageal motor function in humans. Am J Physiol 1981;240:G290 – G296. 35. Liu J, Yamamoto Y, Schirmer BD, et al. Evidence for a peripheral mechanism of esophagocrural diaphragm inhibitory reflex in cats. Am J Physiol 2000;278:G281–G288. 36. Pandolfino JE, Zhang QG, Ghosh SK, et al. Transient lower esophageal sphincter relaxations and reflux: mechanistic analysis using concurrent fluoroscopy and high-resolution manometry. Gastroenterology 2006;131:1725–1733. 37. Niedringhaus M, Jackson PG, Evans SR, et al. Dorsal motor nucleus of the vagus: a site for evoking simultaneous changes in crural diaphragm activity, lower esophageal sphincter pressure, and fundus tone. Am J Physiol 2008;294:R121–R131. A VAGAL INNERVATION OF THE CRURAL DIAPHRAGM 1101 Received December 16, 2008. Accepted August 19, 2009. Reprint requests Address requests for reprints to: Richard L. Young, PhD, Nerve-Gut Research Laboratory, Hanson Institute, Frome Road, Adelaide, South Australia 5000, Australia. e-mail: [email protected] Acknowledgments Early accounts of this work were presented at the Digestive Disease Week meetings of the American Gastroenterological Association in Los Angeles, CA (2006),1 Washington, DC (2007),2 and San Diego, CA (2008).3 Conflicts of interest The authors disclose no conflicts. Funding This work was supported by a grant from AstraZeneca. L. Ashley Blackshaw was supported by a National Health and Medical Research Council (Australia) Senior Research Fellowship. AstraZeneca was not involved in the study design or collection, analysis, or interpretation of data. BASIC– ALIMENTARY TRACT March 2010 1101.e1 YOUNG ET AL Supplementary Materials and Methods Control Experiments for Tracer Injection Control experiments were also performed to test for tracer spread in dual-injected animals and to assess tracer uptake within the intraperitoneal space. Two separate ferrets were injected with CTB-AF555 (CD) and CTB-fluorescein isothiocyanate (distal esophagus) and perfused at 2 hours following tracing. Transverse sections of the reflux barrier were then prepared to directly assess tracer distribution. In an additional 2 ferrets, 4 "L CTBAF555 was injected into the intraperitoneal space adjacent to the diaphragm to test for tracer leakage. Survival times and tissue processing were the same as those following diaphragm tracing. Immunohistochemistry Immunoreactivity for ChAT and neuronal nitric oxide synthase (nNOS) were detected in traced brainstem sections, while VAChT and WGA were detected in traced sections of the reflux barrier. ChAT, nNOS, and VAChT were detected using rabbit antisera from a number of suppliers; WGA was detected using a goat antiserum from Vector Laboratories (Burlingame, CA) (Supplementary Table 1). ChAT, nNOS, and VAChT primary antibodies were visualized using a goat secondary antibody conjugated to Alexa Fluor 488 (AF488 green; Invitrogen), while WGA was visualized using Alexa Fluor 546 (AF546 red; Invitrogen). Sections were air dried at room temperature for 10 minutes and rinsed twice in phosphatebuffered saline ' 0.2% Triton X-100 (PBS-T, pH 7.4; Sigma-Aldrich) to facilitate antibody penetration. Following 2-hour incubation at 37°C with primary antisera (1:200 in 10% PBS-T) unbound antibody was removed with PBS-T washes and slides incubated for 1 hour at room temperature with secondary antisera (1:200 in PBST). Slides were given final PBS-T washes, drained, and mounted with ProLong Antifade (Invitrogen). The specificities of these primary antisera for ChAT,1 nNOS,2 and VAChT3 have been well-characterized in previous studies, including homologous and heterologous preabsorption. In addition, we confirmed positive control labeling in target tissues for ChAT, VAChT (rat cerebellum), and nNOS (mouse intestinal myenteric neurons) prior to use in ferret tissues. Slides in which the primary antiserum was omitted showed no labeling and served as negative controls. Visualization and Quantification High-magnification epifluorescent images were obtained on a conventional epifluorescence microscope (BX-51; Olympus, Australia) equipped with specific excitation filters and images acquired on a Photometrics CoolSnapfx monochrome digital camera (Roper Scientific, Tucson, AZ). A differential interference contrast (Nomarski) stage was also used for imaging brainstem sections which were overlaid to align cytoarchitectural GASTROENTEROLOGY Vol. 138, No. 3 features of nucleus tractus solitarius subnuclei described in ferret with target immunoreactivity and retrograde tracer. All fluorescence images were imported unmodified into V'' imaging software v 4.0 (Digital Optics, Auckland, New Zealand), pseudocolored, and merged for composite images; luminance intensity was not adjusted. Anterogradely traced vagal endings in the reflux barrier were further imaged by confocal microscopy (Leica SP5) and 3-dimensional projection images generated using Leica Confocal Software (Leica Microsystems, Australia). Sequential Dissection Experiments in the Isolated Ferret CD Vagal Efferent Preparation In 2 preparations, a sequential dissection was performed on vagal nerve trunks to determine the route of innervation of the CD. In this experiment, a cut was made 1 cm above the gastroesophageal junction in the left vagus trunk and the caudal trunk dissected free of preparation. Vagal stimulation was performed as detailed above, and EMG responses were recorded in the left CD muscle. The vagal trunk was then cut sequentially at 1.5, 2, 2.5, and 3 cm above the gastroesophageal junction and EMG responses recorded at each intervention; the same protocol was performed on the right vagus nerve trunk with EMG responses recorded in the right CD muscle. In Vivo EMG Experiments Vagal innervation of the CD was also assessed in a whole-animal preparation. Ferrets were anesthetized with pentobarbitone sodium (50 mg/kg intraperitoneally), and the right carotid artery and jugular vein were cannulated for blood pressure recordings and administration of intravenous anesthetic, respectively. Supplemental doses of pentobarbitone were administered as required to abolish the hindlimb pinch-withdrawal reflex. The right and left vagus nerves were isolated and a tracheotomy performed; body temperature was maintained at 37–39°C via a heating pad. A midline laparotomy was performed to access the abdominal surface of the left diaphragm. In 6 animals, the viscera were retracted, the diaphragm gently tensioned laterally, and fine bipolar electrodes secured in the left CD 1–2 mm adjacent the esophageal hiatus; these were isolated using Parafilm (Pechiney Plastic Packaging, Chicago, IL). The viscera were then released and the laparotomy closed. A thoracotomy and median sternotomy were then performed in the 7– 8 rib space and ventilation commenced (15 mL, 30 breaths per minute). The left and right phrenic nerves were isolated and a snare attached for quick access. Fine bipolar electrodes were then placed in the thoracic surface of the left ventrolateral costal diaphragm. Blood pressure and EMG activity in the left CD and costal diaphragm were recorded simultaneously and analysed offline using Spike2 (Cambridge Electronic Design). After 10 minutes of stable baseline recording from CD, costal diaphragm, and blood pressure the thoracic phrenic March 2010 nerves were sectioned and the ensuing EMG activity recorded. Intact vagus nerve trunks were then individually stimulated via hook electrodes, using a pulse generator and a constant-current stimulus unit delivering a single 20-V electrical stimulus of 0.5-ms pulse width. Drugs and Solutions Hexamethonium bromide was obtained from Sigma-Aldrich. Pancuronium bromide was obtained from AstraZeneca (North Ryde, Australia). All drug dilutions were prepared and diluted in isotonic saline on the day of experimentation. Supplementary Results Spinal Innervation of the CD To determine the location of afferent and efferent cell bodies innervating the diaphragm we performed retrograde tracing by bilateral injection of CTB-fluorescein isothiocyanate or CTB-AF555 into the CD. As expected, this resulted in retrograde labeling of a small population of spinal afferent neurons in cervical dorsal root ganglion (DRG) and motor neurons in the ventromedial C5 and C6 ventral horn of the spinal cord (Supplementary Figure 1). Similar findings were made when tracer was injected into the costal diaphragm, but not when injected into the distal esophagus, as expected. When different tracers were injected in the CD and costal diaphragm in the same experiments, separate afferent neurons were labeled in DRG (Supplementary Figure 1A), and separate motor neurons were labeled in the ventral horn (Supplementary Figure 1B). Labeled motor neurons were intermingled within a compact cell group representing the phrenic motor nucleus. All of these exhibited large multipolar perikarya that gave rise to several dendrites that extended radially from the cell body. Double-labeled motorneurons were occasionally identified in these animals (Supplementary Figure 1C). Control Experiments for Tracer Injection In order to determine if tracer spread into adjacent regions could explain our findings, we observed peripheral tracer spread 2 hours after injection into respective target tissues. This showed CTB-AF555 injected into the CD was restricted to skeletal muscle of the CD, while CTB-fluorescein isothiocyanate injected into the esophagus was evident only in the serosa of the distal esophagus and in the inner portion of the phrenoesophageal ligament (data not illustrated). Control intraperitoneal injection of the same volume of either tracer weakly labeled a small number of NG neurons (%1%) and DMN neurons, presumably due to uptake by vagal motorneurons and afferents innervating abdominal viscera (data not illustrated). This weak tracing contrasted with the bright labeling seen with tracing from CD, costal diaphragm, or distal esophagus, and was at a level below A VAGAL INNERVATION OF THE CRURAL DIAPHRAGM 1101.e2 that used to define positive tracing in other experiments. Thus, we were confident that the labeling we observed in tracing experiment, from specific targets could not have arisen from nonselective uptake. Anterograde Tracing Positive Control Anterograde tracing labeled specialized vagal afferent endings within the myenteric plexus and muscular layers of the ferret stomach, corresponding to intraganglionic laminar endings (Supplementary Figure 2A) and intramuscular arrays, as observed in many species. Intraganglionic laminar endings were concentrated within the gastric fundus and body, with fewer labeled toward the greater curvature and were more abundant in the ventral specimen of the stomach. Tracing of gastric intraganglionic laminar endings thus provided a confirmation of effective anterograde tracing of afferent endings in individual animals. Sequential Dissection Experiments Sequential dissection experiments revealed that motor innervation of the CD by the vagus arose bilaterally via nerve branches that diverged from each trunk of the thoracic vagus 1–2 cm above the gastroesophageal junction and ran within muscle layers of the distal esophagus prior to innervating the CD (data not illustrated). In Vivo EMG Experiments Spontaneous EMG activity in the left CD and ventrolateral costal diaphragm were reliably recorded in these in vivo experiments (Figure 7A). Bilateral phrenicotomy led to a complete loss of EMG activity in the costal diaphragm, after which animals were artificially ventilated. Spontaneous EMG activity continued in the CD at a rate of 0.3 # 0.1 per second. This nonphrenic EMG activity was not phaselocked to cardiac electrical activity or ventilation, suggesting a separate central coordination. Stimulation of intact vagus nerve trunks in ventilated and phrenicotomized ferrets evoked short-latency EMG spikes in the left CD (but not costal diaphragm) with a latency consistent with input from motorneurons with conduction velocities of A# fiber (8.3 # 0.3 m/s). This stimulus also produced reflex increase in EMG activity after a longer and more variable latency. Stimulation of the right vagus increased the EMG frequency in the left CD immediately following the stimulus by 14-fold (0.3 # 0.1 vs 3.8 # 0.7 per second in 1-second poststimulation; Figure 7B), while stimulation of the left vagus increased EMG frequency in the left CD by 4-fold (0.3 # 0.1 vs 1.1 # 0.5 per second). These data provide strong evidence of a bilateral, multisynaptic central pathway converging on DMN neurons that innervate the CD. We also performed additional experiments to assess effects of ipsilateral and bilateral vagotomy on diaphragm EMG in phrenic-intact, spontaneous breathing ferrets. These data were inconclusive as the loss of the Hering- 1101.e3 YOUNG ET AL Breuer reflex upon vagotomy led to increased EMG activity (amplitude) in both compartments of the diaphragm. Supplementary References 1. Consonni S, Leone S, Becchetti A, et al. Developmental and neurochemical features of cholinergic neurons in the murine cerebral cortex. BMC Neurosci 2009;10:18. GASTROENTEROLOGY Vol. 138, No. 3 2. Christie AE, Edwards JM, Cherny E, et al. Immunocytochemical evidence for nitric oxide- and carbon monoxide-producing neurons in the stomatogastric nervous system of the crayfish Cherax quadricarinatus. J Comp Neurol 2003;467:293–306. 3. Weihe E, Schutz B, Hartschuh W, et al. Coexpression of cholinergic and noradrenergic phenotypes in human and nonhuman autonomic nervous system. J Comp Neurol 2005;492:370 – 379. March 2010 A VAGAL INNERVATION OF THE CRURAL DIAPHRAGM 1101.e4 Supplementary Figure 1. Retrograde tracing of spinal nerve pathways from the crural diaphragm (CD) in ferrets. (A) Transverse view of dorsal root ganglion (DRG) at C6 level showing spinal afferent neurons traced separately from CD (cholera toxin subunit B [CTB]-AF555, red) and costal diaphragm (CTB-fluorescein isothiocyanate, green). (B) Transverse high-magnification view of C6 spinal cord showing CD motorneurons and processes in the ventral horn traced with CTB-AF555 (white). (C) High-magnification view of C6 ventral horn showing motor neurons traced separately, and jointly from the CD (CTB-AF555, red) and costal diaphragm (CTB-fluorescein isothiocyanate, green). Scale bar (A, C) ! 100 "m, (B) ! 200 "m. Supplementary Figure 2. Anterograde tracing of vagal nerve pathways to the crural diaphragm (CD) and phrenoesophageal ligament in ferrets. (A) Vagal intraganglionic laminar ending;(IGLE) in whole mount of stomach filled by anterograde tracing from the left nodose ganglia (NG) with wheat germ agglutinin– horseradish peroxidase (WGA-HRP) (white). (B) High-magnification extended focal image (EFI) of vagal punctate endings filled by WGA-HRP within connective tissue of the abdominal CD surface. (C) Low-magnification view of a vagal punctate ending at the base of a CD muscle fascia near the esophageal hiatus. AB CT, connective tissue of the abdominal CD surface. (D) Vagal punctate endings immunopositive for WGA in the abdominal CD surface. (E) High-magnification view of a motor endplate in the CD immunopositive for vesicular acetylcholine transporter (VAChT); none of the identified vagal structures were immunopositive for VAChT. (F) EFI view of a vagal laminar ending in the phrenoesophageal ligament. Scale bar (A) ! 500 "m, (B, D, E) ! 10 "m, (C) ! 100 "m, (F) ! 25 "m. 1101.e5 YOUNG ET AL GASTROENTEROLOGY Vol. 138, No. 3 Supplementary Table 1. Antisera Used for Localization and Classification of Vagal Innervation of the Crural Diaphragm Antisera Target ChAT VAChT nNOS WGA PGP Dilution 1:100 1:500 1:200 1:1000, goat antibody 1:500, anti-human Supplier Millipore Phoenix Biotech Zymed Laboratories Vector Laboratories Category no. AB144P H-V007 61-7000 AS-2024 Ultraclone RA95101 NOTE. All primary antibodies were directed to C-termini and were rabbit anti-rat antibodies unless indicated. ChAT, choline acetyltransferase; nNOS, neuronal nitric oxide synthase; PGP, protein gene product 9.5; VAChT, vesicular acetylcholine transporter; WGA, wheat germ agglutinin.
© Copyright 2026 Paperzz