SEARCH

SEARCH BY CITATION

Keywords:

  • 5-hydroxytryptamine;
  • glucose;
  • vagus

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgments
  9. Funding
  10. Disclosure
  11. Author Contribution
  12. References

Background  Intestinal glucose induces gastric relaxation via vagally mediated sensory-motor reflexes. Glucose can alter the activity of gastrointestinal (GI) vagal afferent (sensory) neurons directly, via closure of ATP-sensitive potassium channels, and indirectly, via the release of 5-hydroxytryptamine (5-HT) from mucosal enteroendocrine cells. We hypothesized that glucose may also be able to modulate the ability of GI vagal afferent neurons to respond to the released 5-HT, via regulation of neuronal 5-HT3 receptors.

Methods  Whole-cell patch clamp recordings were made from acutely dissociated GI-projecting vagal afferent neurons exposed to equiosmolar Krebs’ solution containing different concentrations of d-glucose (1.25–20 mmol L−1) and the response to picospritz application of 5-HT assessed. The distribution of 5-HT3 receptors in neurons exposed to different glucose concentrations was also assessed immunohistochemically.

Key Results  Increasing or decreasing extracellular d-glucose concentration increased or decreased, respectively, the 5-HT-induced inward current and the proportion of 5-HT3 receptors associated with the neuronal membrane. These responses were blocked by the Golgi-disrupting agent Brefeldin-A (5 μmol L−1) suggesting involvement of a protein-trafficking pathway. Furthermore, l-glucose did not mimic the response of d-glucose implying that metabolic events downstream of neuronal glucose uptake are required to observe the modulation of 5-HT3 receptor mediated responses.

Conclusions & Inferences  These results suggest that, in addition to inducing the release of 5-HT from enterochromaffin cells, glucose may also increase the ability of GI vagal sensory neurons to respond to the released 5-HT, providing a means by which the vagal afferent signal can be amplified or prolonged.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgments
  9. Funding
  10. Disclosure
  11. Author Contribution
  12. References

Vagal neurocircuits are vitally important in the integration and co-ordination of homeostatic functions including the regulation of gastrointestinal (GI) reflexes. Blood glucose levels oscillate throughout the day and increase dramatically following food ingestion. Such acute changes in blood glucose levels, even within the physiological range, exert profound vagally mediated effects upon gastric motility and emptying to stabilize these fluctuations and maintain homeostasis.1,2 Appropriate and effective homeostatic reflexes require the integration and co-ordination of a wide range of visceral sensations with neuronal, hormonal, and metabolic signals. Vago-vagal sensory-motor reflexes are critically important in this regard as the majority of non-nociceptive information between the GI tract and the central nervous system (CNS) is relayed via the vagus nerve.

Vagal sensory neurons, the cell bodies of which lie in the paired nodose ganglia, serve the classic afferent functions, relaying sensory information from the GI tract to the brainstem via a glutamatergic synapse at the level of the nucleus of the tractus solitarius (NTS)3,4. Neurons of the NTS assimilate this vast volume of sensory information with inputs from other brainstem and higher CNS centers and send this integrated signal to, among other areas, the adjacent dorsal motor nucleus of the vagus (DMV). The DMV contains the preganglionic parasympathetic motoneurons that provide the motor output back to the GI tract via the efferent vagus nerve.4

Increases in both peripheral and central glucose levels inhibit gastric motility and tone,5–8 yet the precise mechanism responsible for these alterations still remains to be elucidated fully. A significant degree of glucose-mediated excitation of GI vagal sensory nerve fibers is due to an indirect, or paracrine, mechanism of action. Glucose induces the release of 5-hydroxytryptamine (5-HT) from enteroendocrine cells which, in turn, excites 5-HT3 receptors present on GI vagal sensory nerve terminals.9–11 Certainly, agonists at 5-HT3 receptors delay gastric emptying while antagonists accelerate gastric transit.10,12,13 Studies have also suggested, however, that the response of vagal afferent fibers to luminal glucose can be modulated by intravenous glucose, suggesting that glucose may also act directly upon vagal afferent neurons or fibers.14 More recent studies have demonstrated that some vagal afferent neurons are themselves glucose sensitive, increasing their excitability in response to increased extracellular glucose levels via actions at ATP-sensitive potassium channels, in a manner similar to the canonical model of pancreatic β-cells.15

We have demonstrated previously that glucose can modulate glutamate release from the central terminals of vagal afferent neurons via an alteration in the number of functional presynaptic 5-HT3 receptors.16,17 This phenomenon does not appear to be restricted to central vagal terminals, however, as vagal afferent neurons are also capable of altering their neurotransmitter phenotype in response to physiological conditions.18–20 As vagal afferent neurons also display functional 5-HT3 receptors21–24 and are capable of responding to circulating blood glucose levels, this study aimed to determine whether extracellular glucose can modulate the response of GI vagal afferent neurons to 5-HT via alterations in 5-HT3 receptor density and/or function.

Materials and Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgments
  9. Funding
  10. Disclosure
  11. Author Contribution
  12. References

Ethical approval

All experiments were performed in accordance with the Pennsylvania State University’s Institutional Animal Care and Use Committee, under an approved protocol.

Retrograde tracing

Experiments were performed in both male and female Sprague–Dawley rats. To label GI-projecting vagal afferent neurons, the fluorescent neuronal tracer 1,1′-dioctadecyl-3.3.3′,3′-tetramethylindocarbocyanine perchlorate (DiI; Invitrogen, Grand Island, NY, USA) was applied to the GI tract as described previously.25 Briefly, after induction of anesthesia with isofluorane (2.5% with air, 600 mL min−1), the abdominal area of 14-day old rats (= 73) was cleaned before exposing the stomach and proximal intestine via a laparotomy. The stomach and proximal intestine were isolated from the surrounding viscera, and crystals of DiI were placed onto the major curvature of either the fundus, the corpus, or the antrum/pylorus areas, or onto the antimesenteric border of the duodenum. The dye crystals were embedded in place using a fast-hardening epoxy resin; the surgical area was washed with sterile saline before closing the abdomen. Rats were allowed to recover for 7–15 days before experimentation.

In another series of experiments, crystals of DiI were applied to the superior laryngeal nerve (SLN) of 12 rats (28–35 days of age). Rats were anesthetized with a mixture of ketamine/xylazine/acepromazine in saline (80/1.6/5.0 mg kg−1, respectively); after a deep level of anesthesia was obtained (abolition of the foot-pinch withdrawal reflex) a ventral midline neck incision was made to expose the left cervical vagus nerve. The SLN was isolated from surrounding tissue by a small piece of parafilm. Crystals of DiI were apposed to the SLN and affixed in place with a fast-hardening epoxy resin. The surgical area was rinsed with sterile saline before the incision was closed. Rats were allowed to recover for 7–15 days prior to experimentation.

Dissociation of nodose ganglia

Rats were euthanized via anesthetic overdose (isoflurane, 5%) followed by administration of a bilateral pneumothorax. Nodose ganglia were removed under microscopic guidance and placed in chilled, oxygenated Hank’s balanced salt solution (HBSS; Gibco Laboratories, Life Technologies, Grand Island, NY, USA). The connective tissue surrounding the ganglia was removed before the ganglia were incubated in HBSS containing 10 mg mL−1 collagenase and 40 mg mL−1 dispase at 37 °C for 45 min. Ganglia were then placed in fresh HBSS and triturated through a pipette tip until tissue was no longer visible. Cells were centrifuged at 150 g for 5 min, placed in fresh HBSS and centrifuged at 500 g for 8 min. Neurons were resuspended in the plating medium (see below for composition), plated on poly-l-lysine coated coverslips, and incubated in a humid incubator (5% CO2) at 37 °C. After 2 h, cells were fed (see below for composition). Electrophysiological recordings and immunohistochemical studies were carried within 24 h of dissociation.

Electrophysiological recordings

Cover glasses with plated dissociated nodose ganglion neurons were placed in a custom-made perfusion chamber (volume 500 μL; Michigan Precision Instruments, Parma, MI, USA) and maintained at 35 °C by perfusion with warmed Krebs’ solution at a rate of 2.5–3.0 mL min−1. Retrogradely labeled neurons (DiI-filled) were identified using a Nikon E600FN microscope (Nikon Corporation, Tokyo, Japan) equipped with tetramethylrhodamine isothiocyanate (TRITC) epifluorescent filters. Once the identity of a DiI-filled neuron was confirmed, electrophysiological recordings were made under bright-field illumination.

Whole-cell recordings were made with patch pipettes of resistance 2–3 MΩ when filled with potassium gluconate intracellular solution (see below for composition). Data were acquired using a single electrode voltage-clamp amplifier (Axopatch 1D; Molecular Devices, Sunnyvale, CA, USA) at a rate of 10 kHz, filtered at 2 kHz, and digitized via a Digidata 1320 interface (Molecular Devices), before being stored and analyzed on a personal computer using pClamp9 software (Molecular Devices). Only those neurons with a series resistance of <15 MΩ were considered acceptable.

Drugs and Krebs’ solutions containing different equiosmolar concentrations of glucose were applied via perfusion through a series of manually operated valves for a period of time sufficient for the response to reach plateau. Neurons were allowed to recover fully between drug additions. A separate patch pipette was filled with 5-HT (100 μmol L−1) and connected to a picospritzer (Parker Hannifin Corp., Cleveland, OH, USA). Each neuron served as its own control and the response to picospritz application of 5-HT was assessed before, during, and after perfusion with equiosmolar Krebs’ solution containing different glucose concentrations using anova with repeated measures. Statistical significance was set at 5%. Results are expressed as mean ± SE.

Immunohistochemistry

Dissociated nodose neurons were exposed to feeding medium with varying concentrations of glucose (1.25–20 mmol L−1) for 30 min before fixation in 4% paraformaldehyde for 15 min. Cells were washed for 5 min in 50% ethanol in distilled water, then washed (3 × 5 min) in phosphate buffered saline (PBS; see below for composition) containing 0.1% bovine serum albumin (BSA). Slides were then placed in rabbit anti-5-HT3B receptor antibody (1 : 500; Abcam ab39629, Abcam, Cambridge, MA, USA; diluted in PBS containing BSA) and incubated overnight at room temperature. Slides were washed (3 × 5 min) with PBS containing BSA before incubation in secondary antibody (AlexaFluor488-conjugated goat anti-rabbit IgG, 1 : 100; Invitrogen) for 1 h at room temperature. Slides were washed (3 × 5 min) in PBS containing BSA and mounted in Fluoromount-G (Southern Biotech., Birmingham, AL, USA).

Control experiments included (i) omission of the primary antibody, (ii) omission of the secondary antibody, and (iii) use of an inappropriate secondary antibody (AlexaFluor488-conjugated goat anti-donkey IgG 1 : 100; Invitrogen, Life Technologies). 5-HT3 receptor immunoreactivity was not observed under any of these conditions.

To determine whether glucose induced the trafficking of 5-HT3 receptors, acutely dissociated neurons were exposed 5 μmol L−1 brefeldin-A, an inhibitor of protein transport, for 30 min prior to exposure to different concentrations of glucose (2.5–20 mmol L−1) containing brefeldin-A. Neurons were fixed subsequently and processed for 5-HT3 receptor immunoreactivity as above. To determine whether cellular or metabolic events following glucose uptake into nodose neurons was responsible for trafficking of 5-HT3 receptors, acutely dissociated neurons were also exposed to feeding medium containing 5 mmol L−1d-glucose + 15 mmol L−1l-glucose prior to fixation.

Confocal microscopy

Confocal photomicrographs of 5-HT3 receptor immunoreactive cells were taken on an Olympus IX81 microscope and captured with Olympus Fluoview 1000 software (Olympus America Inc., Center Valley, PA, USA). Images of individual cells were imported into Image J software (http://rsbweb.nih.gov/ij). The density of receptor on the membrane was quantified as described previously.26,27 Briefly, one confocal image (1 μm thick) across the maximum diameter of the neuron was analyzed. The total number of pixels, defined by setting the threshold for immunoreactivity above the threshold for background staining, was determined within the area of a cell. The cell membrane was defined by outlining the inner and outer edges of the plasma membrane, within 0.5 μmol L−1 of each other. The number of pixels inside both lines was measured using the same parameters and the percentage of membrane-associated 5-HT3 receptor density was calculated by subtraction. Data were analyzed by comparing the percentage of membrane-associated receptor density between different groups, using an anova followed by the Student’s t-test. Statistical significance was set at 5%. Results are expressed as mean ± SE. Linear regression analysis was used to determine the correlation between the glucose concentration and the percentage of membrane-associated 5-HT3 receptor density.

Materials

Supplemented minimal essential medium (sMEM): Minimal essential medium (Gibco, Life Technologies) supplemented with 0.15 g per 100 mL NaHCO3 and 29.2 mg l-glutamine, pH 7.4.

Plating medium: sMEM supplemented with fetal calf serum (10%), nerve growth factor (1%), penicillin–streptomycin (1%), and CaCl2 (4 mmol L−1).

Feeding medium: sMEM supplemented with fetal calf serum (10%), nerve growth factor (1%), and penicillin–streptomycin (1%).

Phosphate buffered saline (mmol L−1): 115 NaCl, 75 Na2HPO4, 7.5 KH2PO4.

Paraformaldehyde fixative: 4% paraformaldehyde in PBS.

Krebs’ solution (mmol L−1): 126 NaCl, 25 NaHCO3, 2.5 KCl, 1.2 MgCl2, 2.4 CaCl2, 1.2 NaH2PO4, and 5 d-glucose maintained at pH 7.4 by bubbling with 95%O2/5%CO2.

Intracellular pipette solution (mmol L−1): 128 K-gluconate, 10 KCl, 0.3 CaCl2, 1 MgCl2, 10 HEPES, 1 EGTA, 2 NaATP, and 0.25 NaGTP adjusted to pH 7.35 with KOH.

Dispase was purchased from Roche Diagnostics (Indianapolis, IN, USA); all other chemicals were purchased from Sigma Chemical Co. (St. Louis, MO, USA).

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgments
  9. Funding
  10. Disclosure
  11. Author Contribution
  12. References

Picospritz application of 5-HT induces a 5-HT3 receptor mediated current in GI vagal afferent neurons

To determine the type of receptor responsible for the inward current induced in response to picospritz application of 5-HT (100 μmol L−1), GI vagal afferent neurons were voltage clamped at −50 mV. Picospritz application of 5-HT induced an inward current of 1047 ± 286 pA (= 6). In the presence of the 5-HT2 receptor selective antagonist, ketanserin (1 μmol L−1), the response to re-application of 5-HT was unaffected (982 ± 251 pA, i.e., 96 ± 3.3% of control amplitude, > 0.05; data not shown). In contrast, in the presence of the 5-HT3 receptor selective antagonist, ondansetron (1 μmol L−1), the 5-HT-induced inward current was reduced to 179 ± 77 pA, i.e., to 16 ± 3.6% of control amplitude (< 0.05; data not shown).

These results demonstrate that the inward current induced in GI vagal afferent neurons in response to picospritz application of 5-HT is mediated exclusively via activation of 5-HT3 receptors.

The basic properties of GI vagal afferent neurones responsive to 5-HT differ from non-responsive neurons

The basic membrane properties of GI vagal afferent neurons were recorded as described previously.25 Basic membrane properties did not differ between GI regions (fundus, corpus, antrum/pylorus, and duodenum); the results were therefore combined. The input resistance of neurons responsive to 5-HT (289 ± 32 MΩ; = 33) did not differ from non-responsive neurons (298 ± 27 MΩ; = 34). The duration of the action potential at threshold was significantly shorter for GI vagal afferent neurons responsive to 5-HT as compared with non-responsive neurons (3.8 ± 0.28 ms vs 4.8 ± 0.32 ms, respectively, < 0.05). Furthermore, whereas the action potential afterhyperpolarization amplitude did not differ between GI responsive vs non-responsive neurons (18.1 ± 1.06 mV vs 17.9 ± 1.15 mV, respectively, < 0.05), the duration of the afterhyperpolarization was significantly shorter in responsive neurons (16.6 ± 1.90 ms vs 28.3 ± 3.94 ms, < 0.05).

Extracellular glucose concentration modulates the amplitude of the 5-HT-induced current in GI vagal afferent neurons

Previous studies have demonstrated that subpopulations of gastric-projecting vagal afferent neurons are glucose excited, i.e., increasing extracellular glucose concentration induces a membrane depolarization, whereas decreasing extracellular glucose concentration induces a membrane hyperpolarization.15 In this study, increasing extracellular glucose concentration from 5 to either 10 or 20 mmol L−1 induced an inward current of 43 ± 6 or 58 ± 21 pA, respectively, in 7 of 31 neurons (22%; = 3, 4, respectively). Conversely, decreasing extracellular glucose concentration from 5 to 2.5 mmol L−1 induced an outward current of 20 ± 5 pA in 4 of 22 neurons (18%).

In the presence of control concentrations of extracellular glucose (5 mmol L−1), picospritz application of 5-HT (100 μmol L−1) induced an inward current of 286 ± 50.1 pA amplitude in GI vagal afferent neurons voltage clamped at −50 mV (= 33). In 28 of these neurons (85%), the magnitude of the 5-HT-induced inward current was modulated by the extracellular glucose concentration; the 5-HT-induced current in the remaining 5 (15%) of neurons was unaffected. Perfusion with equiosmolar Krebs’ solution containing 1.25 or 2.5 mmol L−1d-glucose decreased the 5-HT-induced inward current to 58 ± 9.6% and 54 ± 5.3% of the control response, respectively (< 0.05; Fig. 1A). The change in amplitude of the 5-HT-induced current reached a steady state within 5–8 min and was stable thereafter. In contrast, increasing the extracellular glucose concentration to 10 or 20 mmol L−1 increased the magnitude of the 5-HT-induced inward current to 133 ± 8.9% and 187 ± 16.7% of control amplitude, respectively (< 0.05 for both; Fig. 1B).

image

Figure 1.  Extracellular glucose concentration modulates the response to 5-HT and the density of membrane-associated 5-HT3 receptors in GI-projecting vagal afferent neurons. (A) Electrophysiological traces illustrating the 5-HT-induced inward current in a single GI-projecting vagal afferent neuron voltage clamped at −50 mV. Superfusion with equiosmolar Krebs’ solution containing different concentrations of glucose modulated the amplitude of the 5-HT-induced inward current; decreasing the glucose concentration to 1.25 mmol L−1 decreased the magnitude of the inward current, whereas increasing the glucose concentration to 20 mmol L−1 increased the magnitude of the inward current. (B) Concentration–response curve illustrating the modulation of the 5-HT-induced inward current in GI vagal afferent neurons by glucose. Neurons were voltage clamped at −50 mV prior to picrospritz application of 5-HT; each neuron was exposed to equiosmolar Krebs’ solution containing at least three different concentrations of glucose. The 5-HT-induced inward current was normalized to the amplitude of the current induced in the presence of 5 mmol L−1 glucose. = 6–41 for each point. (C) Photomicrographs of representative confocal images (1 μm thick) of GI-projecting vagal afferent neurons (5-HT3 receptors – green; DiI – red). (a) An acutely dissociated vagal afferent neuron illustrating 5-HT3 receptor distribution (20 mmol L−1 glucose), (b) the same neuron as in (a) depicting DiI-labeling from the GI tract, (c) merged image. 5-HT3 receptor distribution in acutely dissociated GI-projecting vagal afferent neurons following exposure to 1.25 mmol L−1 (d), 5 mmol L−1 (e), or 20 mmol L−1 (f) glucose prior to fixation. Note that as the extracellular glucose concentration was elevated, membrane-associated 5-HT3 receptor density increased. Calibration bar = 20 μm. (D) Concentration–response curve illustrating the modulation of 5-HT3 receptor density in GI vagal afferent neurons induced by extracellular glucose concentration. 5-HT3 receptor density associated with the neuronal membrane was expressed as a proportion of that neuron’s total 5-HT3 receptor density. Note that an increase in glucose level was associated with a linear increase in 5-HT3 receptor density associated with the neuronal membrane (r2 = 0.97).

Download figure to PowerPoint

Extracellular glucose concentration modulates the density of 5-HT3 receptors on the neuronal membrane of GI vagal afferent neurons

Immunohistochemical studies were performed to assess the ability of the extracellular glucose concentration to modulate the proportion of 5-HT3 receptors associated with the membrane of GI vagal afferent neurons. The proportion of membrane-associated 5-HT3 receptors, expressed as a percentage of total 5-HT3 receptor density, was assessed following 30 min exposure to concentrations of extracellular glucose ranging from 1.25 to 20 mmol L−1 (Fig. 1C). At 1.25 mmol L−1 extracellular glucose, only 19.8 ± 1.58% of the total 5-HT3 receptor density was membrane associated (= 24), but this proportion increased linearly to 36.6 ± 2.62% at 20 mmol L−1 glucose (= 28; r2 = 0.97; Fig. 1D). The linear nature of this relationship suggests that even physiological changes in glucose level (4–8 mmol L−1) may alter the proportion of membrane-associated 5-HT3 receptors on GI vagal afferent neurons, although this needs to be examined in more detail. While the immunohistochemical data do not exclude the possibility that increasing extracellular glucose concentration induces degradation of intracellular 5-HT3 receptors (thus increasing the relative proportion of membrane-associated receptors), taken together with the electrophysiological data, these data suggest that extracellular glucose concentration rapidly and reversibly modulates both the response to activation of 5-HT3 receptors in GI-projecting vagal afferent neurons and the proportion of 5-HT3 receptors present on the neuronal membrane.

Extracellular glucose does not modulate either the response to activation of 5-HT3 receptors or the density of membrane-associated 5-HT3 receptors in SLN-labeled vagal afferent neurons

To examine whether the glucose-dependent modulation of the 5-HT-induced inward current was restricted to vagal afferent neurons innervating the GI tract, recordings were made from vagal afferent neurons labeled following application of DiI to the SLN, i.e., neurons presumed to innervate the larynx. In the presence of Krebs’ solution containing 5 mmol L−1 glucose, picospritz application of 5-HT induced an inward current of amplitude 890 ± 485 pA (= 5). In the presence of equisomolar Krebs’ solution containing 2.5 or 20 mmol L−1 glucose, however, the amplitude of the 5-HT-induced inward current was unaffected at 98 ± 2.5% and 101 ± 4.2% of control amplitude, respectively (Fig. 2A).

image

Figure 2.  Extracellular glucose concentration does not modulate the response to 5-HT or the density of membrane-associated 5-HT3 receptors in SLN-labeled vagal afferent neurons. (A) Electrophysiological traces illustrating the 5-HT-induced inward current in a single vagal afferent neuron labeled following application of DiI to the SLN. The neuron was voltage clamped at −50 mV prior to picospritz application of 5-HT. Superfusion with equiosmolar Krebs’ solution containing either 1.25 or 20 mmol L−1 glucose failed to modulate the amplitude of the 5-HT-induced inward current. (B) Photomicrographs of representative confocal images (1 μm thick) of acutely dissociated SLN-labeled vagal afferent neurons illustrating 5-HT3 receptor distribution (green). Neurons were exposed to 1.25, 5, or 20 mmol L−1 glucose prior to fixation. Note that there was no alteration in 5-HT3 receptor distribution in response to change in extracellular glucose concentration. Calibration bar = 20 μm. (C) Concentration–response curve illustrating the modulation of 5-HT3 receptor density in SLN-labeled vagal afferent neurons in response to extracellular glucose concentration. 5-HT3 receptor density associated with the neuronal membrane was expressed as a proportion of that neurons’ total 5-HT3 receptor density. Note that extracellular glucose concentration had no effect upon the proportion of membrane-associated 5-HT3 receptor density.

Download figure to PowerPoint

Immunohistochemical studies were also performed to assess whether the extracellular glucose concentration modulated the proportion of 5-HT3 receptors associated with the membrane of SLN-labeled vagal afferent neurons (Fig. 2B). In the presence of 1.25 mmol L−1 glucose, 25 ± 0.7% of the total 5-HT3 receptor density was membrane associated (= 21); this proportion was not changed by increasing the extracellular glucose concentration up to 20 mmol L−1 (25 ± 1.3%; = 14; > 0.05; Fig. 2C).

These results suggest that the glucose-dependent modulation of 5-HT3 mediated inward currents is not a ubiquitous property of vagal afferent neurons, but may be restricted to neurons innervating the GI tract.

Extracellular glucose modulates the ability of ondansetron to antagonize the 5-HT-induced inward current

To confirm that the extracellular glucose modulates 5-HT-induced inward current in GI vagal afferent neurons by modulating the number of available receptors, the ability of the 5-HT3 receptor selective antagonist, ondansetron (0.1 μmol L−1) to inhibit the response to 5-HT was tested at different extracellular glucose concentrations. In the presence of 5 mmol L−1 glucose, ondandsetron inhibited the 5-HT-induced inward current by 49 ± 10.9%, decreasing the current amplitude from 667 ± 179 to 348 ± 116 pA (= 7; < 0.05; Fig. 3A). In the presence of 2.5 mmol L−1 glucose, however, the same concentration of ondansetron inhibited the 5-HT response to a greater extent, decreasing the inward current from 704 ± 314 to 122 ± 68 pA, i.e., a 81 ± 10.9% decrease (= 5; < 0.05 vs control; < 0.05 vs inhibition in the presence of 5 mmol L−1 glucose). In contrast, in the presence of 10 or 20 mmol L−1 glucose, ondansetron was less effective and inhibited the 5-HT-induced inward current by only 45 ± 10.9 and 37 ± 10.7%, respectively, from 740 ± 261 to 497 ± 289 pA and from 342 ± 82 to 209 ± 67 pA, respectively (< 0.05; Fig. 3B). These results demonstrate that the ability of the receptor selective antagonist, ondansetron, to inhibit the 5-HT-mediated inward current was dependent upon the extracellular glucose concentration. As a single concentration of the antagonist would be expected to occupy the same number of 5-HT3 receptors, the decreased ability of ondansetron to block the response to 5-HT as glucose concentration increases suggests that elevating glucose levels may increase the number of active 5-HT3 receptors on the membrane of GI vagal afferent neurons.

image

Figure 3.  The ability of the 5-HT3 receptor selective antagonist, ondansetron, to inhibit the 5-HT-induced inward current in GI vagal afferent neurons is dependent upon extracellular glucose concentration. (A) Electrophysiological traces illustrating the effects of the 5-HT3 receptor selective antagonist, ondansetron (0.1 μmol L−1) to inhibit the 5-HT-induced inward current in a GI-projecting vagal afferent neuron. The neuron was voltage clamped at −50 mV prior to picospritz application of 5-HT. Ondansetron decreased the amplitude of the 5-HT-induced inward current. Superfusion of that same neuron with equiosmolar Krebs’ solution containing 20 mmol L−1 glucose increased the amplitude of the 5-HT-induced current; at this higher concentration of glucose, ondansetron was less effective at inhibiting the 5-HT-induced inward current. (B) Concentration–response curve illustrating the actions of extracellular glucose concentration to modulate the ability of ondansetron (0.1 μmol L−1) to inhibit the 5-HT-induced inward current in GI vagal afferent neurons. Note that as the extracellular glucose concentration increased, ondansetron became less effective at inhibiting the 5-HT-induced inward current, suggesting that as extracellular glucose levels increased, the number of membrane-associated 5-HT3 receptors also increased.

Download figure to PowerPoint

Extracellular glucose modulates the 5-HT-induced inward current in GI vagal afferent neurones via trafficking of 5-HT3 receptors

To determine whether increasing extracellular glucose concentration potentiates the responses of vagal afferent neurons to 5-HT by trafficking of 5-HT3 receptors to the cell membrane, the ability of glucose to modulate the 5-HT-induced inward current was assessed before and after exposure to the Golgi-disrupting agent, brefeldin-A, which we have demonstrated previously inhibits receptor trafficking within vagal neurocircuits.28 In GI vagal afferent neurons, picospritz application of 5-HT (100 μmol L−1) in the presence of 5 mmol L−1 extracellular glucose, induced an inward current of 397 ± 109 pA (= 5). In the presence of 20 mmol L−1 glucose, the 5-HT-induced inward current was increased by 56 ± 27% to 508 ± 40 pA (< 0.05). Following reperfusion with 5 mmol L−1 glucose and recovery of the 5-HT-induced inward current amplitude to control levels (389 ± 115 pA, i.e., 94 ± 8.5% of control amplitude, > 0.05), neurons were perfused for 5 min with brefeldin-A (5 μmol L−1) which itself did not alter the amplitude of the 5-HT-induced inward current (385 ± 117 pA, i.e., 97 ± 2.9% of control amplitude; > 0.05). In the continued presence of brefeldin-A, however, re-application of 20 mmol L−1 glucose no longer increased the amplitude of the 5-HT-induced inward current (338 ± 125 pA, i.e., 81 ± 7.0% of amplitude in the presence of 5 mmol L−1 glucose + brefeldin-A; < 0.05; Fig. 4A).

image

Figure 4.  Inhibition of protein trafficking prevents the glucose-induced modulation in response to 5-HT and the density of membrane-associated 5-HT3 receptors in GI-projecting vagal afferent neurons. (A) Electrophysiological traces illustrating the effects of the Golgi-disrupting agent, brefeldin-A (5 μmol L−1), to block the glucose-induced alteration in 5-HT-mediated inward current in a GI-projecting vagal afferent neuron. The neuron was voltage clamped at −50 mV prior to picospritz application of 5-HT (black trace). The magnitude of the 5-HT-induced inward current was increased following superfusion with equiosmolar Krebs’ solution containing 20 mmol L−1 glucose (red trace). Following washout and recovery, however, superfusion with brefeldin-A inhibited the ability of 20 mmol L−1 glucose to increase the response to 5-HT (blue trace), suggesting involvement of a protein-trafficking pathway. (B) Photomicrographs of representative confocal images (1 μm thick) of acutely dissociated GI-projecting vagal afferent neurons illustrating 5-HT3 receptor distribution (green). Neurons were exposed to either 5 or 20 mmol L−1 glucose in the absence or presence of the Golgi-disrupting agent brefeldin-A (5 μmol L−1) prior to fixation. Note that exposure to brefeldin-A inhibited the ability of 20 mmol L−1 glucose to increase the proportion of 5-HT3 receptors associated with the neuronal membrane, suggesting involvement of a protein-trafficking pathway. Calibration bar = 20 μm. (C) Graphical representation of the effects of brefeldin-A (5 μmol L−1) to disrupt the ability of extracellular glucose to modulate the proportion of 5-HT3 receptors associated with the membrane of GI vagal afferent neurons. In control conditions, varying the extracellular glucose concentration to 1.25 or 20 mmol L−1, respectively, altered the proportion of 5-HT3 receptors associated with the neuronal membrane (black bars). In the presence of brefeldin-A, however, altering extracellular glucose concentration had no effect upon membrane-associated 5-HT3 receptor density. *< 0.05 vs control (5 mmol L−1) glucose.

Download figure to PowerPoint

Immunohistochemical studies were also performed to assess the ability of brefeldin-A to disrupt the actions of extracellular glucose concentration to modulate the proportion of membrane-associated 5-HT3 receptors (Fig. 4B). In the continued presence of brefeldin-A (5 μmol L−1), 30 ± 0.7% of total 5-HT3 receptor density was associated with the neuronal membrane following exposure to 5 mmol L−1 glucose (= 27). Decreasing the concentration of glucose to 1.25 mmol L−1 did not increase receptor internalization, i.e., the proportion of membrane-associated receptor was unchanged at 28 ± 0.9% (= 26; > 0.05). Similarly, the externalization of receptors associated with increasing the extracellular glucose concentration to 20 mmol L−1 was blocked by brefeldin-A (28 ± 0.6%, = 41, > 0.05; Fig. 4C).

These data suggest that the actions of extracellular glucose to modulate the response to activation of 5-HT3 receptors, and the density of 5-HT3 receptors present on the neuronal membrane of GI vagal afferent neurons, are due to trafficking of receptors to and from the membrane.

d-glucose, but not l-glucose, modulates the 5-HT-induced inward current in GI vagal afferent neurons

We have demonstrated previously16,17 that d-glucose, but not l-glucose, increases glutamate release from the central terminals of vagal afferents via modulation of 5-HT3 receptors.17 To determine whether the observed effect of extracellular glucose on 5-HT-mediated inward currents in nodose ganglion neurons was similarly dependent upon d-glucose, recordings were made from GI vagal afferent neurons voltage clamped at −50 mV. In the presence of 5 mmol L−1d-glucose, picospritz application of 5-HT induced an inward current of amplitude 712 ± 326 pA. Following perfusion with equiosmolar Krebs’ solution containing 20 mmol L−1d-glucose, the magnitude of the 5-HT-induced inward current was increased by 43 ± 9.2% to 981 ± 474 pA (= 5; < 0.05). After re-exposure to 5 mmol L−1d-glucose and reversal of the increased 5-HT current (to 693 ± 289 pA, i.e., 104 ± 4.8% of control amplitude; > 0.05), neurons were perfused with Krebs’ solution containing 5 mmol L−1d-glucose + 15 mmol L−1l-glucose. The combination of d- and l-glucose had no effect on the amplitude of the 5-HT-induced inward current (738 ± 335 pA, i.e., 100 ± 5.1%, > 0.05; Fig. 5A).

image

Figure 5. d-glucose, but not l-glucose, modulates the response to 5-HT and the density of membrane-associated 5-HT3 receptors in GI-projecting vagal afferent neurons. (A) Electrophysiological traces illustrating the effects of d- and l-glucose on the response to 5-HT in a GI-projectin vagal afferent neuron. The neuron was voltage clamped at −50 mV prior to picospritz application of 5-HT (black trace). Superfusion with equiosmolar Krebs’ solution containing 20 mmol L−1d-glucose increased the magnitude of the 5-HT-induced inward current (red trace), whereas superfusion with 5 mmol L−1d-glucose + 15 mmol L−1l-glucose had no effect (blue trace). (B) Graphical representation of the effects of d- and l-glucose on the proportion of 5-HT3 receptors associated with the membrane of GI-projecting vagal afferent neurons. Neurons were exposed to 5 or 20 mmol L−1d-glucose or 5 mmol L−1d-glucose + 15 mmol L−1l-glucose prior to fixation. Elevating the extracellular d-glucose concentration increased the proportion of 5-HT3 receptors associated with the neuronal membrane, whereas elevating the extracellular l-glucose concentration had no effect. These results suggest that the effects of glucose are dependent upon glucose metabolism, rather than uptake, within the vagal afferent neurons. *< 0.05 vs control (5 mmol L−1) glucose.

Download figure to PowerPoint

Immunohistochemical studies were also carried out to investigate whether the ability of the extracellular glucose concentration to traffic 5-HT3 receptors to and from the membrane of GI vagal afferent neurons was dependent upon d-glucose. In the presence of 5 mmol L−1d-glucose, 26 ± 1.19% of 5-HT3 receptor density was membrane associated (= 64), increasing to 37 ± 2.6% in the presence of 20 mmol L−1d-glucose (= 38; < 0.05). In the presence of 5 mmol L−1d-glucose + 15 mmol L−1l-glucose, however, the proportion of membrane-associated 5-HT3 receptor density was unchanged at 26 ± 1.2% (= 31; > 0.05; Fig. 5B).

These data suggest that the glucose-dependent modulation of receptor density and response to activation of 5-HT3 receptors in GI vagal afferent neurons is dependent upon d- but not l-glucose, implying the involvement of metabolic events downstream to neuronal glucose uptake.

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgments
  9. Funding
  10. Disclosure
  11. Author Contribution
  12. References

Results of this study demonstrate that, in GI vagal afferent neurons, the extracellular concentration of glucose modulates: (i) the proportion of 5-HT3 receptors associated with the neuronal membrane, (ii) the magnitude of the 5-HT3 receptor mediated inward current, and (iii) the ability of the 5-HT3 receptor selective antagonist, ondansetron, to antagonize the 5-HT-induced inward current. This study also demonstrates that: (iv) the extracellular glucose concentration does not modulate either the 5-HT-induced inward current or the proportion of membrane-associated receptors in vagal afferent neurons innervating the SLN, suggesting that these responses may be selective for neurons innervating the GI tract, (v) these effects are blocked by the Golgi-disrupting agent Brefeldin-A, suggesting that glucose induces trafficking of premade receptors rather than de novo receptor synthesis, and (vi) that l-glucose does not mimic the actions of d-glucose, suggesting that a downstream metabolic pathway subsequent to neuronal glucose uptake is involved.

Since the initial studies demonstrating that glucose within the lumen of the intestine increases the firing rate of vagal afferent fibers,14,29,30 several studies have indicated that glucose acts, at least in part, via the release of 5-HT from mucosal enterochromaffin cell. In turn, this activates 5-HT3 receptors present on vagal afferent terminals within the GI tract, allowing the excitatory signal to be relayed to the brainstem.9–11,31 Electrophysiological studies on isolated afferent neurons have often been used as a convenient model for events at the peripheral terminal, based on the assumption that biophysical events are identical at both locations. While it remains unclear as to the exact extent to which this assumption holds true, previous studies have demonstrated that the response of GI vagal afferent neurons to neurohormones such as glucagon-like peptide 1 (GLP-1) do, indeed, mimic those of the afferent terminals.32 Intraluminal 5-HT itself increases the discharge rate of nodose ganglion neurons11 and, furthermore, studies in which dye was injected into the internal occipital artery have shown relatively uniform dye penetration into the nodose and jugular ganglia, suggesting that circulating factors have free access to vagal afferent neurons.33 Vagal afferent neurons themselves display a wide variety of receptors for neurotransmitters and neuromodulators hence, presumably, are capable of responding directly to diverse circulating factors. In this regard, following ingestion of a meal, the levels of platelet-free 5-HT rise from approximately 25 to 88 nmol L−1,34 as vagal afferent neurons from several species, including the rat, display both 5-HT2 and 5-HT3 receptors, one may reasonable assume that vagal afferent neurons may be additionally excited by the increase in circulating 5-HT levels following ingestion of a meal.21–24

In addition, therefore, to inducing the release of 5-HT from enterochromaffin cells and activating 5-HT3 receptors on vagal afferent terminals within the GI tract, glucose may also increase the ability of vagal sensory neurons and nerve fibers to respond to the released 5-HT, providing a means by which the vagal afferent signal can be amplified and/or prolonged in a temporally distinct manner. Trafficking of 5-HT3 receptors via regulated exocytosis has been shown previously to occur in expression systems,35,36 but, to our knowledge, this is the first functional and physiological demonstration of 5-HT3 receptor trafficking within peripheral neurons.

Studies in both animals and humans have shown that 5-HT3 receptor selective agonists induce relaxation of the proximal stomach,12,37 whereas 5-HT3 receptor selective antagonists accelerate gastric transit.10,38 Indeed, the 5-HT3 receptor selective antagonist ondansetron is effective in treatment of several GI motility disorders.39–41 Significantly, studies have also demonstrated that a larger dose of a 5-HT3 antagonist is required to induce the same increase in gastric motility and intestinal transit in diabetic animal models,38,42 suggesting that hyperglycemia may affect the response to 5-HT, possibly via an increase in the number of 5-HT3 receptors present on vagal neurocircuits.

We have demonstrated previously that glucose modulates excitatory synaptic transmission from the central terminals of vagal afferent fibers16 due to alteration in the number of presynaptic 5-HT3 receptors.17 In this study, we extend these earlier findings by demonstrating that glucose also regulates the trafficking of 5-HT3 receptors on GI-projecting vagal afferent neurons and, by consequence, modulates the responsiveness of these neurons to 5-HT. The results of this study suggest that circulating glucose levels modulate vagally mediated GI responses via alteration in the number of 5-HT3 receptors associated with the neuronal membrane of GI vagal afferent neurons. The effect of ondansetron to inhibit the 5-HT-induced inward current is more apparent at glucose levels below 5 mmol L−1 suggesting that higher concentrations of ondansetron may be required to inhibit the 5-HT-induced response in the diabetic state (see also38). Although it remains to be determined, it is possible that this glucose-dependent trafficking of 5-HT3 receptors may be a generalized phenomenon within vagal neurocircuits and, as such, may also occur on the peripheral nerve terminals within the intestine.

An elevation in extracellular glucose levels induced a rapid, and reversible, increase in the number of functional 5-HT3 receptors associated with the membrane of GI-projecting vagal sensory neurons, hence increased the magnitude of the response to 5-HT. This suggests that even a relatively small change in receptor distribution is able to produce a marked difference in the response to 5-HT. This phenomenon appears to be restricted to vagal afferent neurons innervating the GI tract, however, as glucose altered neither the response to 5-HT nor the proportion of 5-HT3 receptors associated with the neuronal membrane of SLN-labeled neurons (i.e., neurons presumed to innervate the larynx). As such, these data suggest that the vagal afferent neurons involved in vagally mediated GI reflexes such as oropharyngeal and esophageal peristalsis and relaxation of the lower esophageal sphincter43 are not susceptible to modulation by glucose.

The exact mechanism by which glucose regulates the density of membrane-associated 5-HT3 receptors on GI vagal afferent neurons remains to be elucidated. The relatively rapid speed with which glucose can increase the density of membrane-associated 5-HT3 receptors (within minutes) argues against de novo synthesis of receptors but, rather, implies that glucose traffics receptors to and from the membrane surface, as observed previously.36 We have demonstrated previously that G-protein coupled28,44–46 and ionotropic17 receptors on the nerve terminals of central vagal neurocircuits are exceptionally susceptible to trafficking to and from the neuronal membrane and, thus, are able to modulate the excitability and responsiveness of vagally mediated responses in accordance within physiological conditions in an on-demand fashion. While several reports have described changes in neurotransmitter receptor distribution patterns within GI vagal sensory neurons in response to alterations in feeding conditions,18–20 this study is the first to report that such receptor trafficking may occur rapidly enough to alter ongoing physiological responses on a minute-to-minute basis. Cumulatively, these studies all suggest that vagally mediated homeostatic reflexes are exceptionally ‘plastic’ and open to modulation by a variety of circulating factors.

Glucose is also able to alter the excitability of GI vagal afferent neurons directly, via closure of an ATP-sensitive potassium channel, in a manner similar to the canonical model described in pancreatic β-cells.15 In this study, however, increasing extracellular glucose concentration was rarely observed to induce an inward current in GI vagal afferent neurons, possibly due to methodological differences in electrophysiological recordings. In this study, which used conventional whole-cell patch clamp recording techniques, after entering whole-cell mode, an equilibration period of at least 5 min was allowed prior to examining the neuronal responses. As ATP was routinely included within the patch pipette solution, it is reasonable to assume that ATP-sensitive potassium channels were already closed prior to alteration in extracellular glucose concentration, rendering it difficult, if not impossible, to observe further glucose-dependent channel closure. In this study, however, the actions of extracellular glucose to modulate 5-HT3 receptor distribution and 5-HT-mediated responses were dependent upon d-glucose, as l-glucose was without effect (see also 17). These results suggest that metabolic events downstream of neuronal glucose uptake are required to observe the modulation of 5-HT3 receptor mediated responses. The nature of these downstream metabolic events are still open to investigation, but previous studies have demonstrated that 5-HT3 receptors can be trafficked to and from the membrane of receptor expression systems by protein kinase C (PKC).35,36 Furthermore, both acute and chronic hyperglycemia activate and upregulate PKC activity,47,48 suggesting that this second messenger system may be a likely candidate in the trafficking of 5-HT3 receptors within GI vagal afferent neurons.

Conclusions

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgments
  9. Funding
  10. Disclosure
  11. Author Contribution
  12. References

This study suggests that circulating glucose levels may affect GI functions via actions at multiple levels, both peripheral and central, confirming and extending the emerging concept that vagally mediated neurocircuits regulating autonomic functions are highly integrative and capable of a remarkable degree of plasticity and modulation.49 By regulating the number of functional 5-HT3 receptors present on the cell membrane of GI vagal afferent neurons, glucose is able to modulate its own ‘perception’, hence amplify or prolong vagal afferent signaling. The rapid timescale of the observed electrophysiological effects suggest that glucose may alter ongoing physiological responses of vagal afferent neurons on a minute-to-minute basis, although further studies to investigate these effects within physiological glucose ranges and timescales are required. Whether such receptor trafficking actions of glucose are restricted 5-HT3 receptors, or whether this is a more generalized phenomenon and can occur with other receptors involved in GI homeostatic reflexes, remains to be elucidated.

Acknowledgments

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgments
  9. Funding
  10. Disclosure
  11. Author Contribution
  12. References

The authors thank F Holly Coleman for performing the labeling surgeries, Dr R. Alberto Travagli for comments on previous versions of this manuscript, and W. Nairn Browning for support and encouragement.

Author Contribution

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgments
  9. Funding
  10. Disclosure
  11. Author Contribution
  12. References

TB, AET, SR, and KNB performed the research and analyzed the data; TB and KNB designed the research study; TB and KNB wrote the article.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgments
  9. Funding
  10. Disclosure
  11. Author Contribution
  12. References