Ghrelin is involved in the paracrine communication between neurons and glial cells

Authors


Address for Correspondence

Prof Inge Depoortere, PhD, Targid, Gut Peptide Research Lab, Gasthuisberg O&N1, box 701, Leuven 3000, Belgium.

Tel: +32-16-330675; fax: +32-16-330723;

e-mail: inge.depoortere@med.kuleuven.be

Abstract

Background

Ghrelin is the only known peripherally active orexigenic hormone produced by the stomach that activates vagal afferents to stimulate food intake and to accelerate gastric emptying. Vagal sensory neurons within the nodose ganglia are surrounded by glial cells, which are able to receive and transmit chemical signals. We aimed to investigate whether ghrelin activates or influences the interaction between both types of cells. The effect of ghrelin was compared with that of leptin and cholecystokinin (CCK).

Methods

Cultures of rat nodose ganglia were characterized by immunohistochemistry and the functional effects of peptides, neurotransmitters, and pharmacological blockers were measured by Ca2+ imaging using Fluo-4-AM as an indicator.

Key Results

Neurons responded to KCl and were immunoreactive for PGP-9.5 whereas glial cells responded to lysophosphatidic acid and had the typical SOX-10-positive nuclear staining. Neurons were only responsive to CCK (31 ± 5%) whereas glial cells responded equally to the applied stimuli: ghrelin (27 ± 2%), leptin (21 ± 2%), and CCK (30 ± 2%). In contrast, neurons stained more intensively for the ghrelin receptor than glial cells. ATP induced [Ca2+]i rises in 90% of the neurons whereas ACh and the NO donor, SIN-1, mainly induced [Ca2+]i changes in glial cells (41 and 51%, respectively). The percentage of ghrelin-responsive glial cells was not affected by pretreatment with suramin, atropine, hexamethonium or 1400 W, but was reduced by l-NAME and by tetrodotoxin. Neurons were shown to be immunoreactive for neuronal NO-synthase (nNOS).

Conclusions & Inferences

Our data show that ghrelin induces Ca2+ signaling in glial cells of the nodose ganglion via the release of NO originating from the neurons.

The gastrointestinal tract exerts complex functions from motility, digestion, secretion, and absorption to food sensing and regulation of food intake. Information about the status of these processes is conveyed to the central nervous system via the vagus nerve. Vagal afferent nerve endings express receptors for both orexigenic and anorexigenic hormones. The receptors are synthesized in the nodose ganglion, in which the afferent cell bodies reside, and transported to the nerve terminals through axonal transport.[1, 2]

The majority of the hormones that are released from the gastrointestinal tract are anorexigenic, such as cholecystokinin (CCK), glucagon-like peptide 1 (GLP-1), and peptide YY3-36 (PYY3-36), which inhibit gastric emptying. Ghrelin, which is secreted from the stomach, is the only circulating orexigenic hormone and stimulates not only food intake but also gastric emptying and intestinal motility.[3, 4] As vagotomy or treatment with capsaicin blocks ghrelin's effect on food intake and gastric motility in rodents, it is considered that ghrelin stimulates the vagus nerve to mediate its effects.[5-8] In humans, ghrelin also loses its ability to stimulate food intake in vagotomized patients.[9] In contrast, one study reported that the acute eating-stimulatory effect of intraperitoneal ghrelin does not require vagal afferent signaling.[10] Thus, it cannot be excluded that circulating ghrelin can also access the hypothalamus across an incomplete blood–brain barrier at the level of the arcuate nucleus.

Immunohistochemistry and in situ hybridization studies demonstrated that the ghrelin receptor (GRLN-R) is expressed in neurons of both rat and human nodose ganglia.[2, 7, 11] The ghrelin receptor is synthesized in vagal afferent cell bodies and transported to the afferent terminals to bind ghrelin released by the ghrelin-producing cells.[7] GRLN-R mRNA expression shows a circadian rhythm and is affected by starvation, vagotomy, and hormones (CCK, gastrin).[12] About 75% of GRLN-R containing neurons express other orexigenic receptors such as cannabinoid (CB)-1 and melanin-concentrating hormone (MCH)-1 receptors.[11] Moreover, GRLN-R neurons also express receptors for anorexigenic hormones such as CCK which in turn are also immunoreactive for the leptin receptor.[11, 13, 14] Whether the ghrelin hormone itself is expressed in the nodose ganglion is not clear. Burdyga et al.[11] found no ghrelin expression in the rat nodose ganglion, whereas Page et al.[15] found a low expression of ghrelin in the mouse nodose ganglion.

Previous studies have shown that vagal afferent neurons isolated from the nodose ganglion retain their physiological characteristics in short-term culture.[16, 17] The aim of the current study was to determine whether ghrelin has functional effects in cultures of the rat nodose ganglion by measuring ghrelin's effect on intracellular Ca2+ changes. Sensory neurons within sensory ganglia, such as the nodose ganglia, are surrounded by satellite glial cells (SGC). These cells do not merely serve a structural role, but are thought to have different functions such as influencing the microenvironment and modulating neurotransmission.[18] We investigated whether ghrelin activates or influences the interaction between both cell types. The ghrelin-induced responses were characterized pharmacologically and interactions with the anorexigenic peptide, CCK, were studied as well.

Material and Methods

Primary cell culture of nodose ganglia

Primary cell cultures were prepared from the nodose ganglia of 8-week-old Wistar-Han rats (Janvier, France). Nodose ganglia were removed from anaesthetized rats under a stereomicroscope and washed with Hank's Balanced Salt Solution (HBSS) supplemented with 1% penicillin/streptomycin. To allow enzymatic digestion, the tissue was incubated during 90 min with Earle's Balanced Salt Solution containing collagenase IA (0.5 mg mL−1; Sigma-Aldrich, St. Louis, MO, USA) and dispase (0.5 mg mL−1; Roche Applied Science, Basel, Switzerland). The cells were dispersed by gentle trituration through a siliconized Pasteur pipette and washed four times with HEPES buffered Dulbecco's Modified Eagle Medium supplemented with 10% fetal bovine serum (FBS) and 1% penicillin/streptomycin. The cells were seeded on glass coverslips and grown in an incubator (37 °C, 5% CO2) in neurobasal A medium supplemented with 10% FBS, 1% penicillin/streptomycin, 200 mmol L−1 glutamine, 50 nmol L−1 glutamate, 50 ng mL−1 mouse nerve growth factor 7S, 50 ng mL−1 human recombinant fibroblast growth factor, 1% N2 supplement, and 2% B27 supplement. Experiments were performed on day 2 of the cell culture.

All procedures were approved by the Ethical Committee for Animal Experiments from the University of Leuven.

Measurement of intracellular Ca2+ changes

Nodose ganglion cell cultures were loaded with 10 μmol L−1 Fluo-4-AM (Molecular Probes, Life Technologies, Paisley, UK) at room temperature in HEPES-Krebs buffer (pH 7.4) containing 148 mmol L−1 NaCl, 5 mmol L−1 KCl, 10 mmol L−1 glucose, 10 mmol L−1 HEPES, 1 mmol L−1 MgCl2, and 2 mmol L−1 CaCl2 during 40 min. After loading, the cells were rinsed for 15 min in buffer without dye and transferred to a cover glass chamber mounted on an inverted Zeiss Axiovert 200M microscope (Carl Zeiss, Oberkochen, Germany), with TILL Poly V light source (TILL Photonics, Gräfelfing, Germany) and a cooled CCD camera (PCO Sensicam-QE, Kelheim, Germany). Changes in intracellular Ca2+ concentration [Ca2+]i are reflected in Fluo-4-AM fluorescence intensity and recorded at 525/50 nm. Neurons were identified 5 min prior to drug administration by a brief (5 s) KCl (75 mmol L−1) depolarization. Cells of the nodose ganglion culture were stimulated via a local perfusion system (1 mL min−1) for 20 s with 10 μmol L−1 lysophosphatidic acid (LPA), the NO donor SIN-1 (0.5 mmol L−1), ATP (10 μmol L−1), ACh (0.1 mmol L−1), 1 μmol L−1 ghrelin, 1 μmol L−1 CCK, 0.1 μmol L−1 leptin or a combination of 1 μmol L−1 ghrelin, and 1 μmol L−1 CCK. Three coverslips with cultured cells were obtained per rat examined, and the responses of neurons and glia to three different agonists, superfused in a random order, were examined on two spots in each coverslip. In each spot, 20 neurons or glial cells were selected at random for data analysis. Data retrieved were pooled and considered as one experiment, which was repeated numerous times (n = # rats). In a different set of experiments, the effect of pharmacological blockers on the ghrelin-induced responses was tested. After a first stimulation with 1 μmol L−1 ghrelin, cells were superfused during 8.5 min with l-NAME (0.3 mmol L−1), suramin (0.1 mmol L−1), atropine (5 μmol L−1), hexamethonium (0.1 mmol L−1), 1400 W (1 μmol L−1) or tetrodotoxin (1 μmol L−1) before the cells were again superfused with 1 μmol L−1 ghrelin during 20 s in the presence of these inhibitors. The effect of the inhibitors on the percentage of ghrelin-responsive cells was compared with the percentage of responsive cells after preincubation with buffer. Regions of interest were drawn and average fluorescence signals were normalized to their initial value, all using Igor Pro (Wavemetrics, Lake Oswego, OR, USA).

Immunofluorescence staining

Cultured cells were fixed with 4% paraformaldehyde, washed and non-specific binding sites were blocked with 4% donkey serum and/or 4% goat serum, and permeabilized with 0.5% Triton-X-100. Cells were incubated with the primary antibodies during 24 h at 4 °C. The following primary antibodies were used: mouse anti-PGP-9.5 (1/500; Sanbio, Uden, The Netherlands), rabbit anti-GRLN-R (1/100; Phoenix Pharmaceuticals, Burlingame, CA, USA), guinea pig anti-SOX-10 (1/500)[19] (kind gift from Dr. Wegner, Erlangen, Germany), and rabbit anti-neuronal NO-synthase (nNOS; 1/400; Santa Cruz Biotechnology, Dallas, TX, USA). Cells were washed three times with PBS and incubated with the appropriate fluorochrome-labeled secondary antibodies: Alexa488-donkey anti-mouse (1/1000; Invitrogen, Life Technologies, Paisley, UK), Alexa594 donkey anti-rabbit (1/1000; Invitrogen), Alexa594 goat anti-guinea pig (1/1000; Invitrogen) or AMCA donkey anti-rabbit (1/1000; Jackson ImmunoResearch Laboratories, West Grove, PA, USA) during 2 h at room temperature. Finally, cells were rinsed three times in PBS and embedded in Citifluor (Citifluor Ltd, London, UK). Fluorescence was visualized under a fluorescence microscope (Olympus BX4, Olympus, Tokyo, Japan).

Statistics

All results are presented as mean ± SEM. The effect of pharmacological blockers on ghrelin-induced Ca2+ changes was analyzed by one-way anova analysis followed by Dunnett's post hoc testing. The proportion of delayed Ca2+ release events (delayed = event later than end of superfusion, non-delayed = event before end of superfusion) in glial cells and neurons was compared using a Pearson chi-squared test.

Results

Characterization of the cell culture

Cells in culture were divided into two groups based on their morphology on differential interference contrast (DIC) images: 18% were large cells with fibers and 82% were smaller with a spindle-like shape (Fig. 1A). Immunohistochemical studies revealed that the large cells stained for the neural marker PGP-9.5, while the nucleus of the spindle-like cells was immunoreactive for the glial cell marker SOX-10 (Fig. 1A).[20]

Figure 1.

Characterization of the cell cultures derived from rat nodose ganglia. (A) Differential interference contrast (DIC) image of the cell culture and immunohistochemical staining of the cell culture with SOX-10 (red), PGP-9.5 (green), and the merged image (yellow); bar = 100 μm. (B) Representative tracing of the increase in [Ca2+]i after KCl (75 mmol L−1) depolarization in a neuron and glial cell marked in the DIC picture (left) and percentage of cells responding (n = 23 rats). (C) Representative tracing of the [Ca2+]i increase after lysophosphatidic acid (LPA) (10 μmol L−1) addition in a neuron and glial cell marked in the DIC picture (left) and percentage of cells responding (n = 23 rats).

Further characterization of the cell culture was performed by measuring Ca2+ rises in response to depolarization with KCl (Fig. 1B) and stimulation with the glial cell activator LPA (Fig. 1C).[21] The majority (81 ± 5%) of the large, PGP-9.5-positive cells showed a fast [Ca2+]i rise in response to KCl (75 mmol L−1) depolarization, while only 6 ± 2% of the spindle-like, SOX-10-positive cells, responded. Stimulation of the cell culture with LPA (10 μmol L−1) induced a [Ca2+]i increase in 69 ± 9% of the cells with a SOX-10-positive nucleus but only in 21 ± 12% of the PGP-9.5-immunoreactive cells. The large, PGP-9.5-positive cells will be named neurons hereafter, whereas the spindle-like, SOX-10-positive cells will be designated as glial cells.

Responses of the cell culture to anorexigenic (CCK, leptin) and orexigenic (ghrelin) peptides

To examine whether nodose ganglion cells respond directly to peripherally released appetite-regulating hormones, we determined the effect of the anorexigenic peptides, CCK, and leptin, and of the orexigenic peptide, ghrelin, on [Ca2+]i rises in cultured cells. The neurons responded to CCK (31 ± 4%) but only a minority of the neurons reacted to either leptin (2 ± 1%) or ghrelin (3 ± 1%; Fig. 2A). However, of the K+-responsive neuronal fibers, 19 ± 6% responded to CCK, 16 ± 7% to leptin, and 15 ± 4% to ghrelin (Fig. 2B). The glial cells responded equally to the applied stimuli (CCK: 30 ± 2%; leptin: 21 ± 2%; ghrelin: 27 ± 2%; Fig. 2C).

Figure 2.

Percentage of neurons (A), neuronal fibers (B), and glial cells (C) showing a rise in [Ca2+]i after stimulation with 1 μmol L−1 cholecystokinin (CCK) (n = 13 rats), 1 μmol L−1 ghrelin (n = 15 rats) or 0.1 μmol L−1 leptin (n = 6 rats).

To determine whether functional subsets of glial cells exist in the nodose ganglion, the number of glial cells with overlapping responses to the neuropeptides was determined. About one third of the ghrelin-responsive glial cells (31 ± 3%) also showed CCK responses and vice versa (35 ± 5%; Table 1). A similar pattern was observed for the coresponses to ghrelin and leptin and for the coresponses to leptin and CCK.

Table 1. Characterization of subpopulations of glial cells. Number of glial cells with responses to ghrelin, CCK or both (n = 9 rats); ghrelin, leptin or both (n = 4 rats); CCK, leptin or both (n = 3 rats)
 Total cells examined# Ghrelin responsive cells# CCK responsive cells# Leptin responsive cells# cells with overlapping response
Coresponses ghrelin & CCK354143137 74
Coresponses ghrelin & Leptin12748 5227
Coresponses CCK & Leptin76 202630

Effect of coadministration of ghrelin and CCK

To identify possible interactions between ghrelin and CCK, we determined the effect of coadministration of ghrelin and CCK on [Ca2+]i changes in neurons and glial cells. Neurons responded to CCK, but hardly to ghrelin when applied separately. As CCK and ghrelin have opposite effects on neural firing of afferent nerves, we studied a possible inhibitory effect of ghrelin on [Ca2+]i changes in the presence of CCK. The amplitude of the response to CCK (relative rise 1.38 ± 0.10) did not differ from the amplitude observed after coadministration of ghrelin and CCK (1.24 ± 0.04) in neurons.

Glial cells responded to both ghrelin and CCK. In a subset of glial cells (5%), we found an additive effect between ghrelin and CCK (Fig. 3A). In another subset of glial cells (12%), no response to either ghrelin or CCK was observed, but coadministration of both peptides resulted in a synergistic response (Fig. 3B).

Figure 3.

Representative tracings of the effect of ghrelin, cholecystokinin (CCK), and coadministration of ghrelin and CCK on [Ca2+]i in glial cells. (A) additive effect between 1 μmol L−1 ghrelin and 1 μmol L−1 CCK (B) synergistic effect between 1 μmol L−1 ghrelin and 1 μmol L−1 CCK.

Characterization of ghrelin-induced responses

All PGP-9.5-positive neurons that showed a [Ca2+]i rise during KCl depolarization were immunopositive for the ghrelin receptor (GRLN-R; Fig. 4A). Nevertheless these neurons did not show a [Ca2+]i increase upon stimulation with ghrelin itself (Fig. 4B). In contrast, glial cells that only stained faintly for the GRLN-R, responded with a clear [Ca2+]i rise upon ghrelin stimulation (Fig. 4C). Ghrelin responses in glial cells were delayed since they appeared in 83% of the cases after stopping of the superfusion (20 s). This suggests that responses in glial cells might be secondary to a primary neuronal response. For comparison, for CCK the number of delayed Ca2+ events was significantly (P < 0.0001) lower in neurons (41%) than in glial cells (72%).

Figure 4.

Response of neurons and glial cells to ghrelin. (A) Differential interference contrast (DIC) image of neurons and glial cells costained for PGP-9.5 (green) and for the ghrelin receptor (red) and the merged image (yellow); bar = 100 μm. (B) fluorescence tracings of the increase in [Ca2+]i of a neuron and glial cell, marked in the DIC picture, after KCl (75 mmol L−1) depolarization and (C) after stimulation with ghrelin (1 μmol L−1).

The effect of candidate neurotransmitters, ACh, ATP, and NO, involved in the signaling between neurons and glial cells was investigated.[18] About 90 ± 10% of the neurons responded to 10 μmol L−1 ATP, while the response to 0.1 mmol L−1 ACh (24 ± 5%) and the NO donor, SIN-1 (0.5 mmol L−1; 15 ± 6%), was much lower (Fig. 5A). In contrast, almost none of the glial cells responded to ATP (7 ± 4%) while 41 ± 3% responded to ACh and 51 ± 4% to SIN-1.

Figure 5.

Neurotransmitters involved in the paracrine communication between neurons and glial cells. (A) Percentage of neurons and glial cells responding to ATP (10 μmol L−1), ACh (10 μmol L−1), and SIN-1 (0.5 mmol L−1; n = 4–5 rats). (B) Percentage of ghrelin-responsive glial cells responding after preincubation with buffer (n = 5), 5 μmol L−1 atropine (n = 4), 0.1 mmol L−1 hexamethonium (n = 6), 0.3 mmol L−1 l-NAME (n = 6), 0.1 mmol L−1 suramin (n = 4), 1 μmol L−1 1400 W (n = 5) or 1 μmol L−1 TTX (n = 4). (C) Immunofluorescence staining of a rat nodose ganglion culture for PGP-9.5 (green) and neuronal NO-synthase (nNOS; blue), and the merged image; bar = 100 μm. *P < 0.05, vs buffer-stimulated glial cells, #P < 0.05, ##P < 0.005 neurons vs glial cells.

The role of these neurotransmitters in the effect of ghrelin on [Ca2+]i changes in glial cells was studied as well. Cultures were first stimulated with ghrelin, consequently incubated with buffer or antagonist and then stimulated for a second time with ghrelin in the presence of buffer or antagonist. Only 44 ± 5% (P < 0.01, n = 4 rats) of the glial cells that responded to ghrelin during the first stimulation responded during the second stimulation after preincubation with buffer, indicating that desensitization of the GRLN-R occurred. Treatment with tetrodotoxin (TTX) reduced the number of glial cells responding to ghrelin to 17 ± 5% (P < 0.05), indicating the requirement of neuronal activation for the ghrelin-induced Ca2+ changes in glial cells (Fig. 5B). Neither suramin (P2 purinergic receptor antagonist), nor atropine (muscarinic receptor antagonist) or hexamethonium (nicotinic receptor antagonist) affected the percentage of ghrelin-responsive glial cells. The NO-synthase inhibitor, l-NAME, significantly (P < 0.05) reduced the number of glial cells responding to ghrelin to 17 ± 6%. These findings suggest that NO is the neurotransmitter responsible for activation of the glial cells. An immunofluorescence staining for nNOS confirmed the presence of nNOS in the neurons of the rat nodose ganglion (Fig. 5C). Treatment with the inducible NO-synthase (iNOS) blocker, 1400 W, did not affect the response of the glial cells to ghrelin.

Discussion

In this study, the effect of ghrelin on intracellular Ca2+ changes in a short-term culture of rat nodose ganglia was investigated. The culture consisted of neurons with fibers and spindle-like glial cells. The cell bodies of the neurons were only responsive to CCK, whereas the fibers were responsive to CCK, ghrelin, and leptin, albeit to a low extent. The glial cells on the other hand showed overlapping [Ca2+]i changes in response to CCK, ghrelin, and leptin. In addition, ghrelin-induced responses in glial cell were delayed and pharmacological interventions revealed that NO, originating from the neurons, is involved. Immunohistochemistry studies confirmed the presence of nNOS in neurons and addition of a NO donor to the culture medium mainly activated glial cells. These findings therefore suggest that ghrelin is involved in the paracrine communication between neurons and glial cells. Sensory ganglia, such as the nodose ganglia, contain cell bodies of primary afferent neurons that transmit sensory information via the vagus nerves from internal organs such as the heart, respiratory tract and stomach to the central nervous system. Satellite glial cells surround sensory neurons within sensory ganglia. Ultrastructural studies showed that SGCs are a distinct type of glial cells that are not fully similar to either astrocytes or oligodendrocytes.[22] In our short-term cell culture from rat nodose ganglia, we have used PGP-9.5 and SOX-10 as a marker for neurons and glial cells, respectively. SOX-10 belongs to the group of HMG (high mobility group) domain–containing transcription factors.[23, 24] During vertebrate development, it is highly expressed in the emerging neural crest and later in the developing peripheral nervous system (PNS) and central nervous system, where its occurrence is restricted to glial cells.[20] Staining of the nodose ganglion cell culture with this marker showed a clear nuclear staining in PGP-9.5-negative cells, suggesting that these cells are glia cells. These cells also responded to lysophosphatidic acid which elicits [Ca2+]i transients in enteric glia cells.[25] None of these cells showed responses to depolarization (KCl) which was an exclusive property of the neuronal cells.

As the vagus nerve is the most important gateway between the gastrointestinal system and the central nervous system, it is not a surprise that its cell body, the nodose ganglion, is sensitive to orexigenic and anorexigenic hormones. The presence of the receptor for CCK[26], leptin[14, 27], and ghrelin[2, 7, 11] has already been shown in tissue sections of the nodose ganglion. In the current study, we provide evidence that only CCK induced Ca2+ rises in cell bodies of cultured nodose ganglion neurons, whereas ghrelin and leptin induced responses only in neuronal fibers. On the other hand, all three neuropeptides (leptin, CCK and ghrelin) investigated induced functional responses in glial cells. Most of the immunohistochemistry studies in sections of the nodose ganglia focused on the sensory neurons rather than on the SGCs. In case a neuron is immunoreactive for a receptor, it is difficult to determine whether the thin sheath of SGCs surrounding it, is stained positive or not. In our short-term cell culture, SGCs migrated away from the neurons making it possible to distinguish between both types of cells. We found that neurons stained more intensively for the GRLN-R than glial cells although ghrelin selectively induced [Ca2+]i rises in glial cells, suggesting communication between neurons and glial cells in response to ghrelin.

Orexigenic and anorexigenic hormones do not only have a different effect on food intake, but also have opposing effects on vagal nerve activity. Administration of CCK increases the discharge of both gastric[28] and jejunal[29] vagal afferent nerve fibers via the CCK1 receptor. In addition, in vitro experiments showed that CCK and leptin act synergistically on the activation of gastric vagal afferents.[30, 31] The same results were obtained with cultures of neurons from the rat nodose ganglion.[32] In contrast to the stimulatory effects of CCK, ghrelin suppresses gastric vagal afferent discharge[5, 13] and selectively inhibits subpopulations of mechanically sensitive gastroesophageal vagal afferents.[15] The stimulatory effect of CCK on vagal afferent activity could be abolished by pretreatment with ghrelin and vice versa.[13] To unmask a possible inhibitory effect of ghrelin on [Ca2+]i changes in neurons, we studied the effect of coadministration of ghrelin and CCK on Ca2+ rises in neurons. An interaction between ghrelin and leptin was not studied because similar to ghrelin, leptin was ineffective in neurons. Ghrelin did not affect the amplitude of the CCK-induced [Ca2+]i changes. In glial cells, it was found that in subpopulations that responded to both CCK and ghrelin, coadministration of both peptides resulted in additive effects. Other subpopulations that did not respond to ghrelin or CCK showed a synergistic effect when the neuropeptides were coadministered. The neurotransmitters involved in the interneuron-glial cell communication were investigated. At present, the only evidence for chemical signaling from sensory neurons to SGCs concerns NO. Studies showed that about 5% of the neurons in dorsal root ganglia (DRG) of adult rats contained nNOS, whereas many SGCs contained the enzyme guanylate cyclase, activated by NO to catalyze cyclic GMP formation.[33-38] Aoki et al.[39] noted that SGCs accumulated the NO precursor, arginine, and proposed that DRG neurons send signals to SGCs via NO, and SGCs can talk to neurons by supplying them arginine to produce more NO. Magnusson et al.[35] also reported in cultured nodose ganglia a correlation between the number of cGMP-positive satellite cells and the number of NOS-positive neurons, as reflected by NADPH-diaphorase histochemistry. Our findings in the current study demonstrated the presence of nNOS in the neurons of rat nodose ganglia and showed that the NO donor, SIN-1, induced Ca2+ increases in 51% of the glial cells and only in 15% of the neurons. Furthermore, the percentage of ghrelin-responsive glial cells was significantly reduced by pretreatment with l-NAME or TTX, thereby confirming that NO, released from neurons after stimulation of the ghrelin receptor on neuronal fibers, is responsible for the observed Ca2+ rise in glial cells. The involvement of iNOS from surrounding glia was excluded as pretreatment of the cells with iNOS-blocker, 1400 W, was without effect. A possible direct effect of ghrelin on the glial cells cannot be excluded as the ghrelin-induced Ca2+ changes in the glial cells were not completely abolished by l-NAME or TTX.

Another candidate neurotransmitter is ATP. Sensory neurons respond to ATP largely via P2X3 receptors[40, 41] and it has been suggested that these receptors are important in the transmission of pain signals. The majority (90 ± 10%) of the neurons in our short-term culture responded to ATP, whereas almost none (7 ± 4%) of the glial cells responded. Results from experiments with suramin confirmed that ATP is not involved in neurotransmission of the ghrelin-induced responses between neurons and glial cells. This is in contrast to studies in mouse SGCs in trigeminal sensory ganglia and enteric glial cells showing an increase in intracellular Ca2+ in response to ATP.[25, 42, 43]

ACh is the main neurotransmitter in the PNS, but its role in sensory ganglia is not well defined. Sensory DRG neurons express nicotinic[44] and muscarinic[45] ACh receptors and display functional responses to ACh. We observed a response to ACh in 24 ± 5% of the neurons and in 41 ± 3% of the glial cells. Therefore, it is conceivable that neurons and glial cells can mutually communicate by cholinergic signaling. Our pharmacological studies showed that neither nicotinic receptors nor muscarinic receptors are involved in the ghrelin-induced Ca2+ signaling in glial cells in the rat nodose ganglion.

The data from this study suggest that glia in sensory nodose ganglia do not merely give mechanical support to neurons but have many essential functions from controlling the microenvironment to being able to receive and transmit chemical signals.

The NO-cGMP pathway has emerged as a neuroprotective signaling system involved in the communication between neurons and glia in DRG to prevent neuronal apoptosis.[46] Caloric restriction is associated with high plasma ghrelin levels which can be sensed by ghrelin receptors in the nodose ganglion and transmitted to the SGC via NO. Therefore, ghrelin might underpin many of the enhanced neuronal functions that are observed in response to caloric restriction, including neuroprotection which involves reduced apoptosis and increased mitochondrial function.[47]

It has been shown that other orexigenic peptides such as orexin and neuropeptide Y have a more or less well-defined inhibitory role in pain modulation.[48, 49] Ghrelin also has been shown to exert antinociceptive effects. Activation of spinal cord type 1a GRLN-Rs enhances inhibitory neurotransmission in a subset of deep dorsal horn neurons, where it blocks the response of peptidergic nociceptive afferents to painful stimuli.[50] The increased ghrelin levels associated with pathological conditions such as inflammation[51-53], cancer,[54, 55] and congestive heart failure[56] may trigger the release of inhibitory neurotransmitters in the nodose ganglia to counteract abnormal pain signaling conditions related to the pathological condition.

In conclusion, our study provides the first evidence that ghrelin induces NO-dependent Ca2+ responses in glial cells in the rat nodose ganglion and further adds evidence to the idea that glial cells actively participate in normal and pathological processes. It also suggests that ghrelin is a multifunctional hormone with not only important effects on food intake but perhaps also in neuroprotection and pain signaling.

Funding

This work is supported by grants from the Flemish Foundation for Scientific Research (contract FWO G.0693.08) and by a Methusalem grant from the University of Leuven for research on “The Brain-Gut Axis in Health and Disease: from Mucosal Integrity to Cortical Processing”.

Disclosure

No competing interests declared.

Author Contribution

BA, BD, AG, and TT performed the research and analyzed the data; PV wrote the imaging analysis software; BS, BA, and ID wrote the manuscript; JT, PV, and ID designed the research study.