Neurotransmitters involved in fast excitatory neurotransmission directly activate enteric glial cells


  • W. Boesmans,

    1. Laboratory of Enteric NeuroScience (LENS), KU Leuven, Leuven, Belgium
    2. Translational Research center for Gastrointestinal Disorders (TARGID), KU Leuven, Leuven, Belgium
    Search for more papers by this author
  • C. Cirillo,

    1. Laboratory of Enteric NeuroScience (LENS), KU Leuven, Leuven, Belgium
    2. Translational Research center for Gastrointestinal Disorders (TARGID), KU Leuven, Leuven, Belgium
    Search for more papers by this author
  • V. Van den Abbeel,

    1. Laboratory of Enteric NeuroScience (LENS), KU Leuven, Leuven, Belgium
    2. Translational Research center for Gastrointestinal Disorders (TARGID), KU Leuven, Leuven, Belgium
    Search for more papers by this author
  • C. Van den Haute,

    1. Laboratory for Neurobiology and Gene Therapy, KU Leuven, Leuven, Belgium
    Search for more papers by this author
  • I. Depoortere,

    1. Translational Research center for Gastrointestinal Disorders (TARGID), KU Leuven, Leuven, Belgium
    Search for more papers by this author
  • J. Tack,

    1. Translational Research center for Gastrointestinal Disorders (TARGID), KU Leuven, Leuven, Belgium
    Search for more papers by this author
  • P. Vanden Berghe

    1. Laboratory of Enteric NeuroScience (LENS), KU Leuven, Leuven, Belgium
    2. Translational Research center for Gastrointestinal Disorders (TARGID), KU Leuven, Leuven, Belgium
    Search for more papers by this author

Address for Correspondence
Pieter Vanden Berghe, Laboratory for Enteric NeuroScience (LENS), Translational Research Center for Gastrointestinal Disorders (TARGID), O&N1 #701, KU Leuven, Herestraat 49, B-3000 Leuven, Belgium.
Tel: +32 16 330153; fax: +32 16 345939;


Background  The intimate association between glial cells and neurons within the enteric nervous system has confounded careful examination of the direct responsiveness of enteric glia to different neuroligands. Therefore, we aimed to investigate whether neurotransmitters known to elicit fast excitatory potentials in enteric nerves also activate enteric glia directly.

Methods  We studied the effect of acetylcholine (ACh), serotonin (5-HT), and adenosine triphosphate (ATP) on intracellular Ca2+ signaling using aequorin-expressing and Fluo-4 AM-loaded CRL-2690 rat and human enteric glial cell cultures devoid of neurons. The influence of these neurotransmitters on the proliferation of glia was measured and their effect on the expression of c-Fos as well as glial fibrillary acidic protein (GFAP), Sox10, and S100 was examined by immunohistochemistry and quantitative RT-PCR.

Key Results  Apart from ATP, also ACh and 5-HT induced a dose-dependent increase in intracellular Ca2+ concentration in CRL-2690 cells. Similarly, these neurotransmitters also evoked Ca2+ transients in human primary enteric glial cells obtained from mucosal biopsies. In contrast with ATP, stimulation with ACh and 5-HT induced early gene expression in CRL-2690 cells. The proliferation of enteric glia and their expression of GFAP, Sox10, and S100 were not affected following stimulation with these neurotransmitters.

Conclusions & Inferences  We provide evidence that enteric glial cells respond to fast excitatory neurotransmitters by changes in intracellular Ca2+. On the basis of our experimental in vitro setting, we show that enteric glia are not only directly responsive to purinergic but also to serotonergic and cholinergic signaling mechanisms.


The enteric nervous system (ENS), a ganglionated plexus within the layers of the intestinal wall, is responsible for the control of several gastrointestinal (GI) functions such as motility, secretion, and absorption. This nerve network functions quite autonomously from the central nervous system (CNS) and is able to exert reflex-like activity through a series of neurons connected in a polysynaptic network.1 Apart from neurons, the ENS contains a large amount of glial cells that not only populate the myenteric and submucosal ganglia but also have cell bodies in the extra-ganglionic layers of the gut wall.2,3 Enteric glia were long believed to merely form a matrix that kept neurons in place, but in the past decade, many studies have expanded this physical support role to encompass several aspects of GI physiology.4

Besides their role in controlling epithelial barrier function,5–8 accumulating evidence indicates that enteric glial cells mediate inflammatory responses and serve as a link between the nervous and immune system in the GI tract.9–14 Enteric glia were furthermore shown to protect enteric neurons from oxidative stress and are involved in controlling enteric neuron phenotype and ENS motor function.15–17 Interestingly, recent work indicates that glial cells within the gut of adult animals have neurogenic potential and are capable of generating enteric neurons in response to injury.18,19

Similar to CNS astrocytes which are considered active partners in synaptic communication in the brain,20 enteric glia are believed to modulate enteric neurotransmission. It has been shown that enteric glia contain neurotransmitter precursors, have the machinery for uptake and degradation of neuroligands, express neurotransmitter receptors 3,21–24, and become activated during colonic migrating motor complexes.25 Moreover, neuronal activity induces purinergic signals to elicit Ca2+ transients in neighboring enteric glial cells,26–28 indicative of neuron-to-glia communication. Apart from purines, also other transmitters can activate enteric glia; these include serotonin (5-HT), adrenergic, cholinergic, and protease-activated receptor agonists, endothelins, and lysophosphatidic acid (LPA).29–37

So far, functional studies have been carried out in primary cell cultures or in intact isolated tissues. However, the intimate relationship between enteric neurons and glia, both physically and functionally, has hampered detailed examination of pure glial responses to ENS transmitters and complicated careful interpretation of several conclusions. Therefore, the aim of our study was to investigate the effect of neurotransmitters that are known to elicit fast excitatory potentials in enteric nerves [acetylcholine (ACh), 5-HT, and adenosine triphosphate (ATP)] in a rat enteric glia cell line and in human primary enteric glia both of which were void of neurons.

Materials and methods

CRL-2690 glia cell line

A stable enteric glia cell line, CRL-2690, established from rat small intestine,38 was purchased from American Type Culture Collection (Manassas, VA, USA). Cells were thawed and grown to confluence in DMEM medium containing 10% fetal bovine serum (FBS), 0.5% penicillin-streptomycin, and 1% G5 supplement (all from Invitrogen, Merelbeke, Belgium). Cells were trypsinized (using trypsin 0.25%; Invitrogen) and passaged every 5–6 days.

Generation of mitaequorin CRL-2690 cells

The CRL-2690 cell line was used to generate a stable line expressing the mitochondrial-targeted Ca2+ sensitive protein aequorin (mitAeq). To this end, naïve CRL-2690 cells were transduced using a lentiviral vector coding for apo-aequorin and a hygromycin-resistance gene.39 After 72 h, the cell culture medium was replaced by medium containing 200 μg mL−1 hygromycin (Invitrogen) to select for transduced cells until a pure mitAeq-CRL-2690 cell line was obtained.

Human enteric glia cultures

Four duodenal biopsies were taken from 13 subjects (six men, mean age 54 ± 11 and seven women, mean age 53 ± 14) who were referred to our gastroenterology unit for clinical and endoscopic evaluation. The subjects included in the study were undergoing endoscopy for presumed functional disorders. No GI lesions, whether inflammatory or neoplastic, were observed during the course of the endoscopy. The study protocol (ML7400) was approved by the Ethics Committee of the Leuven University Hospital, Belgium. Written informed consent was obtained from each subject. After removal, the biopsies were immediately immersed in sterile ice-cold (4 °C) Krebs solution (in mmol L−1: 120.9 NaCl, 5.9 KCl, 1.2 MgCl2, 2.5 CaCl2, 11.5 glucose, 14.4 NaHCO3, and 1.2 NaH2PO4) previously oxygenated (95% O2–5% CO2) and equilibrated at pH 7.4, and transferred to the laboratory. The submucosal plexus was isolated following a previously described protocol.40 After removal, the submucosal plexus was digested for 30 min at 37 °C in an enzymatic solution containing protease (1 mg mL−1; Sigma, Bornem, Belgium) and collagenase (1.25 mg mL−1; Sigma).11 The suspension was spun down, the ganglia were picked up, plated into 12-well plates, and topped-up with DMEM-F12 (Invitrogen) supplemented with 10% heat-inactivated FBS and 1% antibiotic-antimycotic solution (Sigma) and kept in an incubator at 37 °C continuously gassed with 95% O2–5% CO2.

Luminescence measurements

mitAeq-CRL-2690 cells were suspended using trypsin and charged with coelenterazine-AM (5 μmol L−1, 2 h; Molecular Probes, Invitrogen), a necessary co-factor for the luminescence reaction to occur. A 96-well plate was prepared containing agonists in various concentrations and inserted into a Microlumat Plus luminometer (Berthold technologies, Bad Wildbad, Germany), 100 μL of coelenterazine-loaded mitAeq-CRL cells (∼5.104 cells) were injected into each of the wells while luminescence at 490 nm was monitored during 20 s. Control solution and Triton X-100 (0.9%; Sigma) were used to determine the baseline (Lbckg) and maximal signal intensity (LtritonX), respectively. Results were expressed in arbitrary units of relative luminescence (rL) based on a background subtraction and normalization to the signals elicited by Triton-X (0.9%): rL = 100(L – Lbckg)(LtritonX– Lbckg)−1. A minimum of eight cultures were used per condition. Analysis was performed using Microsoft Excel 2007 (Microsoft, Redmond, WA, USA) and Igor Pro (Wavemetrics, Lake Oswego, OR, USA).

Microscopic imaging

Live fluorescence imaging was performed on an inverted Zeiss Axiovert 200M microscope (Carl Zeiss, Oberkochen, Germany), with TILL Poly V light source (TILL Photonics, Gräfelfing, Germany) and cooled CCD camera (PCO Sensicam-QE, Kelheim, Germany) using TillVisION (TILL Photonics).

Fluorescence of fixed samples was visualized under an epifluorescence microscope (BX41 Olympus; Olympus, Aartselaar, Belgium) with UMNUAUV and U-MWIY2 filter-cubes for visualizing blue and red probes, respectively. Images were recorded using Cell^F software on an XM10 (Olympus) camera. All image analysis was performed with custom-written routines in Igor Pro.

Fluo-4 - [Ca2+]i imaging

Confluent CRL-2690 cells and human enteric glial cells were suspended and seeded onto glass coverslips, and processed for Ca2+ imaging experiments as previously described.41,42 Cells were loaded with 10 μmol L−1 Fluo-4 AM (30 min; Molecular Probes, Invitrogen), rinsed, and transferred to a cover glass chamber mounted on the microscope stage. The cells were constantly perfused with control or drug containing HEPES-buffered solutions that were gravity fed (1mL min−1) via a multi-barrel perfusion system. All recordings were made at room temperature. Changes in intracellular Ca2+ concentration ([Ca2+]i) are reflected in Fluo-4 fluorescence intensity and were recorded at 525/50 nm.

Proliferation – cell viability assay

CRL-2690 cells were cultured in a 96-well plate (104 cells per well) in the presence of increasing concentrations of either of three neurotransmitters or arabinofuranosyl cytidine (ARA-C; Sigma), used as a control. A 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay was used to determine cell proliferation. In brief, MTT (0.5 mg mL−1) was added to the culture medium 24 h after exposure to the agonists for 2 h. The MTT is converted by chemical reduction in living cells, to a H2O insoluble blue formazan. The crystallized formazan was dissolved in DMSO and determined spectrophotometrically in a microplate reader (iEMS Reader MF; Labsystems, Helsinki, Finland) by measuring the absorbance at 540 nm (and background at 690 nm) as a measure of the total amount of healthy cells.


Cultured CRL-2690 cells and human enteric glia cultures were fixed in 4% paraformaldehyde-containing HEPES-buffered solution for 30 min (RT), washed, and permeabilized (0.1–0.5% Triton X-100) in the presence of blocking serum (4% donkey/goat serum in PBS). The cells were incubated overnight at 4 °C with primary antibodies rabbit anti-S100 (1 : 500; Dako, Glostrup, Denmark), guinea pig anti-Sox10 (1 : 250; Dr. Wegner, Erlangen, Germany),43 and rabbit anti-c-Fos (1 : 300; SantaCruz Biotechnologies, Santa Cruz, CA, USA) diluted in blocking buffer. After cells were rinsed with PBS for 3 × 10 min, secondary donkey antisera were applied: anti-mouse Alexa594 (1 : 1000; Molecular Probes, Invitrogen) and anti-rabbit AMCA (1 : 250; Jackson Immuno Research Labs, West Grove, PA, USA). Control experiments were performed by omitting the primary antibodies.

Analysis of c-Fos stainings

Dual color images of the c-Fos and DAPI stainings were taken on an Olympus BX41 microscope (see Microscopic imaging) and analyzed using Igor Pro. Firstly, the DAPI positive nuclei were automatically detected and used to generate regions of interest, which were overlaid onto the matched c-Fos image to calculate the intensity of the red signal. Based on the distribution of the c-Fos signals in unstimulated CRL-2690 cells, a cut-off value was set to include 90% of all control signals. Above this value c-Fos staining was assumed positive for all conditions tested.


Total RNA was extracted using Trizol (Invitrogen) and 2 μg was reverse transcribed using 200 units Superscript II (Invitrogen) to obtain cDNA. The latter served as a template for the PCR reaction [45 amplification cycles: 95 °C (10 s), 60 °C (15 s), 72 °C (15 s)]. Specific rat primers (Invitrogen) were designed (GFAP – forward: TCCTggAACAgCAAAACAAg – reverse: CAgCCTCAggTTggTTTCAT, S100: – forward: AgAgAggACTCCggCggCAA – reverse: AgggCAACCATggCCTTCTCCA, Sox10 – forward: AggACggCgAggCAgACgAT – reverse: CCgCTgAgCACCTggCTgAC) and the Light-Cycler 480 SYBR Green I Master mix (Invitrogen) was used to run the qPCR reaction on a Lightcycler 480 system (Roche Diagnostics, Mannheim, Germany). An inter-run calibrator was used and standard curves were created to obtain PCR efficiencies. Relative expression levels were calculated with the LightCycler 480 software, expressed relative to glyceraldehyde 3-phosphate dehydrogenase (GAPDH), and corrected for inter-run variability.

Data presentation and statistics

Unless stated otherwise, all data are presented as means (± SEM). The ‘n’ values are the numbers of cells analyzed for each group. ‘N’ values represent the number of fields-of-view examined per group. Data from different conditions were compared with independent Student’s t-test or analysis of variance (anova) with a Bonferonni post hoc test. Differences were considered significant if P < 0.05. Statistical analysis was performed with Microsoft Excel 2007 and GraphPad Prism 5.0 (GraphPad, San Diego, CA, USA).


Aequorin Ca2+ signaling in rat enteric glial cells

To rapidly and efficiently test the responsiveness of enteric glia upon different concentrations of neurotransmitters, we developed an enteric glial cell line that stably expresses the Ca2+ sensitive protein aequorin. By specifically targeting the aequorin to mitochondria (mitAeq) using the signaling sequence of human cytochrome C (subunit 8), a good signal to noise ratio was obtained, due to the high Ca2+ concentrations reached in the mitochondrial matrix.44 To test whether the mitAeq-CRL-2690 cells could be used as a functional reporter system, they were exposed either to control medium or Triton X-100 containing medium to perforate the plasma membrane, allowing Ca2+ ions to enter the cell, thus obtaining a maximal signal. Background luminescence was on average 72 ± 7.8 kcounts in 20 s, while maximum signals of 225 ± 23 kcounts were obtained with Triton X-100. Next, the mitAeq-CRL-2690 cells were exposed to different concentrations of either ACh, 5-HT, or ATP to investigate whether these neurotransmitters were able to induce [Ca2+]i changes in enteric glia. We found that all three neurotransmitters induced dose-dependent luminescence rises (Fig. 1A–C) with an EC50 value of 1.02 × 10−7, 1 × 10−6 and 4.5 × 10−5 mol L−1 for ACh, 5-HT, and ATP, respectively. These data indicate that cultured enteric glial cells are sensitive to neurotransmitters involved in fast excitatory neurotransmission in the ENS. As an internal control, mitAeq-CRL-2690 cells were also challenged with the known enteric glia activator LPA (Fig. 1D). We found that different concentrations of LPA activate mitAeq-CRL-2690 cells with an EC50 of 4.2 × 10−6 mol L−1.

Figure 1.

 Acetylcholine, serotonin, and ATP dose-dependently activate mitAeq-CRL-2690 cells. (A–D) Relative Ca2+ responses of mitAeq-CRL-2690 cells upon exposure to different concentrations [Log M for (A,B) and μmol L−1 for (C,D)] and acetylcholine (ACh, A), serotonin (5-HT, B), adenosine triphosphate (ATP, C), and lisophosphatidic acid (LPA, D). A sigmoid fit is represented by the gray dotted line.

Ca2+ waves in sheets of enteric glial cells

To further explore the effect of these neurotransmitters on enteric glia and to establish our findings in a more conventional in vitro setting, we tested the potential of ACh, 5-HT, and ATP to induce responses in sheets of naïve CRL-2690 cells grown on glass coverslips and loaded with Fluo-4 (Fig. 2). Acetylcholine (10 μmol L−1) induced Ca2+ transients (1.14 ± 0.06, = 12) that were present in 15% of the cells and were mediated by muscarinic receptors because responses were blocked by atropine (10 μmol L−1, 1.07 ± 0.01, = 6), but not by hexamethonium (100 μmol L−1, 1.12 ± 0.04, = 12). Thirty percent of the CRL-2690 cells were activated by 5-HT stimulation (10 μmol L−1, 1.18 ± 0.02, = 10). These Ca2+ responses were abolished by ketanserin (10 μmol L−1, 1.02 ± 0.018, = 10), but unaltered in the presence of ondansetron (10 μmol L−1, 1.22 ± 0.07, = 7), suggesting that serotonergic transmission to enteric glia is mediated via 5-HT2 receptors. Adenosine triphosphate (10 μmol L−1) induced a transient Ca2+ rise in almost all cells (1.85 ± 0.08, = 34). Responses were significantly reduced by the P2 blocker suramin (10 μmol L−1, 1.13 ± 0.03, = 38), but not by the specific P2Y1 blocker MRS2179 (10 μmol L−1, 1.75 ± 0.07, = 38). Electrical stimulation and high K+ depolarization did not induce any responses in CRL-2690 cells. Altogether, these data show that ACh, 5-HT, and ATP are able to induce [Ca2+]i changes in sheets of cultured rat enteric glial cells.

Figure 2.

 Fast excitatory neurotransmitters induce Ca2+ transients in cultured rat enteric glial cells. (A–C, upper panels) Representative gray scale images of CRL-2690 cells loaded with Fluo-4 at rest (left) and upon activation [right, activated cells in (A,B) are depicted by black arrowheads] with acetylcholine (ACh, 10 μmol L−1, A), serotonin (5-HT, 10 μmol L−1, B), and adenosine triphosphate (ATP, 10 μmol L−1, C). Scale bars: 50 μm. (A–C, lower left graphs) Representative traces of neurotransmitter-induced [Ca2+]i transients for three individual cells each. (A–C, lower right graph bars) Average amplitude of neurotransmitter-induced [Ca2+]i rises in CRL-2690 cells and the effect of cholinergic, serotonergic, and purinergic antagonists. Atropine (10 μmol L−1, A), ketanserin (10 μmol L−1, B), and suramin (10 μmol L−1, C) significantly reduced the amplitude of ACh, 5-HT, and ATP-induced [Ca2+]i transients, respectively (*< 0.05 vs control). Hexamethonium (100 μmol L−1, A), ondansetron (10 μmol L−1, B), and MRS2179 (10 μmol L−1, C) had no effect.

The effect of acetylcholine, serotonin, and ATP on human enteric glial cells

Next, we tested whether these three neurotransmitters could also generate Ca2+ transients in human enteric glial cells. Therefore, enteric glia obtained from human duodenal mucosal biopsies were grown in primary culture and prepared for Fluo-4 imaging (Fig. 3). The distinctive morphology of enteric glia was used to distinguish them from other cells (e.g. fibroblasts) present in the primary cultures. In addition, after live imaging, the identity of enteric glial cells was confirmed by immunohistochemical labeling for S100 and Sox10. Similar to their effect on CRL-2690 cells, we found that ACh, 5-HT, and ATP were able to induce Ca2+ transients in cultured human enteric glia. When challenged with ATP (10 μmol L−1), about half of the glial cells (53%) responded with a rise in [Ca2+]i (1.08 ± 0.01, = 45). Stimulation with 5-HT (10 μmol L−1) induced a Ca2+ transient of similar magnitude, but activated a smaller proportion (10%) of glial cells present in the primary cultures (1.08 ± 0.03, = 26 cells). Exposure to ACh (10 μmol L−1) caused a strong [Ca2+]i rise in 24% of glia (1.16 ± 0.01, = 60). The majority of human enteric glia also responded to LPA (1.05 ± 0.01, = 38 cells). None of the cells displayed a Ca2 + transient upon high K+ depolarization, indicating that neurons were not present in this primary culture (data not shown).

Figure 3.

 Acetylcholine, serotonin, and ATP induce Ca2+ transients in cultured human enteric glial cells. (A) Representative gray scale image of cultured human enteric glia loaded with Fluo-4. (B–D) Ca2+ responses of three cells (color-coded numbers in A) to acetylcholine (Ach, 10 μmol L−1, B), serotonin (5-HT, 10 μmol L−1, C), and adenosine triphosphate (ATP, 10 μmol L−1, D). The identity of cultured human enteric glia responding to the various agonists was confirmed by immunohistochemistry for S100 (E) and Sox10 (F). Scale bar: 25 μm. (G) Average amplitude of agonist-induced [Ca2+]i rises in cultured human enteric glial cells. (H) The proportion of identified human enteric glial cells that responded to ACh, 5-HT, ATP, and lisophosphatidic acid (LPA, 10 μmol L−1).

How do enteric glial cells interpret neurotransmitter signaling?

Although our data indicate that enteric glial cells are sensitive to neurotransmitters, little is known about how they interpret and process these signals. As a first approach, we therefore investigated whether stimulation by ACh, 5-HT, or ATP induces gene transcription in CRL-2690 cells (Fig. 4). Compared with the strong activation in terms of Ca2+ signaling, we found no evidence for activation of the early gene c-Fos by ATP (50 μmol L−1, 15.75 ± 0.79%, = 15, vs control: 13.24 ± 1.09%, = 28), even after 24 h incubation. In contrast, both ACh (1 μmol L−1, 22.69 ± 2.16%, = 15) and 5-HT (1 μmol L−1, 19.30 ± 1.45%, = 15) significantly increased the number of c-Fos-positive enteric glial cells compared with control. Stimulation by LPA (10 μmol L−1) resulted in expression of c-Fos in almost all cells (98.15 ± 0.28%, = 15) that were identified by counterstaining for DAPI.

Figure 4.

 Neurotransmitter-induced c-Fos expression in cultured rat enteric glia. (A–E) Representative fluorescence images of CRL-2690 cells that were incubated with control (A), acetylcholine (Ach, 1 μmol L−1, B), serotonin (5-HT, 1 μmol L−1, C), adenosine triphosphate (ATP, 50 μmol L−1, D), or lisophosphatidic acid (LPA, 10 μmol L−1, D) for 24 h and immunostained for c-Fos (red) and counterstained with DAPI (blue). Examples of c-Fos-positive cells are depicted with white arrowheads. Scale bar: 50 μm. (F) Proportion of c-Fos-positive cells (% of cells identified with DAPI) after incubation with the various agonists (anova: < 0.0001, Bonferonni post hoc: *< 0.05 vs control).

To examine whether induction of early gene transcription is accompanied by changed expression of glial cell markers, we isolated mRNA from CRL-2690 cells that were incubated with either ACh, 5-HT, ATP, or LPA, and performed quantitative RT-PCR using primers for GFAP, S100, and Sox10 (Fig. 5A–C). Incubation for 24 h with one of the three neurotransmitters did not result in altered GFAP, S100, or Sox10 expression. Similarly, LPA incubation did not change the expression of these typical enteric glia cell markers. Consistent with our findings at the transcriptional level, we found no differences in the expression of these glial cell markers by immunohistochemistry (data not shown). Accordingly, based on an MTT assay, these neurotransmitters did not affect the growth of CRL-2690 cells (Fig. 5D–F). The mitotic blocker ARA-C that dose-dependently abolished the proliferation of CRL-2690 cells was used as a positive control (Fig. 5G).

Figure 5.

 Neurotransmitters that evoke fast excitatory potentials in enteric neurons do not influence glial marker expression or proliferation of cultured rat enteric glial cells. (A–C) Relative glial fibrillary acidic protein (GFAP, A), Sox10 (B), and S100 (C) mRNA expression of CRL-2690 cells after 24 h incubation with control, acetylcholine (Ach, 1 μmol L−1), serotonin (5-HT, 1 μmol L−1), adenosine triphosphate (ATP, 50 μmol L−1), or lisophosphatidic acid (LPA, 10 μmol L−1). (D–G) CRL-2690 cell proliferation (expressed by relative formazan absorbance) after 24 h incubation with increasing concentrations (Log M) of ACh (D), 5-HT (E), ATP (F) or the mitotic blocker arabinofuranosyl cytidine (ARA-C, G). ARA-C dose-dependently reduced CRL-2690 cell proliferation (anova: = 0.0002, Bonferonni post hoc: *< 0.05 vs control).


To understand the precise involvement of glia in enteric neurotransmission and in conditions of disturbed ENS signaling, it is important to find out whether enteric glial cells can independently respond to particular neurotransmitters and ligands. Several studies have described that enteric glia are receptive to signaling molecules involved in enteric neurotransmission, but the close association between enteric neurons and glia complicates conclusive analysis. By using enteric glial cell cultures devoid of neurons together with a combination of Ca2+ imaging techniques, we show here that both rat and human enteric glia are responsive to ACh, 5-HT, and ATP. Thus, we ascertain that enteric glial cells can be activated directly by neurotransmitters involved in fast excitatory neurotransmission in the ENS.

Enteric neurons and glia are commonly derived from enteric neural crest cells45 and although it is not clear whether segregated precursor lineages exist for both during ENS development, enteric glia remain a source of neurons in the mature ENS.19 Apart from the possible temporary existence of glia-neuron intermediates, also the close spatial association between enteric glia, neurons, and their processes makes alienation of neuronal and glial signals extremely challenging. Indeed, enteric neurons and glia are densely packed into ganglia and glial cells often closely embrace nerve fibers,2,46 hence leaving conventional optical imaging techniques short on resolution. In the in vitro approach that we have adopted here, the absence of neurons obviously eliminates such problems. It also excludes the possibility that glial cells are activated secondary to neuronal activation, as is the case in purinergic neuron-to-glia communication described by our group and others.26,28 By analogy with these studies, our present observations confirm that ATP is a potent activator of enteric glia and that purinergic responses are mediated by P2 receptors. We previously demonstrated that P2Y1 receptors mediate the ATP-dependent paracrine communication between enteric neurons and glia in murine ENS cultures.26 Others have reported that enteric glia in the myenteric plexus of guinea pig28 and mouse27 respond through P2Y4 receptors. Remarkably, P2Y1 was recently reinstated as the key receptor for purinergic activation of enteric glia.47 Notwithstanding this apparent consensus, and adding to these differing data, we show here that MRS2179 had no effect on the ATP-induced Ca2+ transients displayed by CRL-2690 cells indicating that at least in this cell line, the P2Y1 receptor was not involved. The question whether the differences in species or experimental setting underlie these discrepancies will need to be addressed in future experiments.

While Kimball et al. only found a marginal proportion of enteric glia to be responsive to 5-HT (10−4mol L−1),29 30% of cultured CRL-2690 cells responded to serotonergic stimulation. An effect that is probably mediated by 5-HT2 receptors as ketanserin blocked the 5-HT-induced [Ca2+]i rises. Apart from 5-HT and ATP, we show that also ACh is able to directly activate enteric glia in vitro. Compared with recent reports where cholinergic-induced responses in enteric glial cells were assigned to nicotinergic receptors,25,35 hexamethonium did not affect Ca2+ transients evoked by ACh in CRL-2690 cells. In contrast, the muscarinic receptor blocker atropine reduced [Ca2+]i rises with about 50%. Future research is required to unravel the origin of these opposing results, although the mere presence of nerves could be a critical explanatory factor. Similar to previous studies, the enteric glia activator LPA also induced robust Ca2+ rises in CRL-2690 cells.26,34

Contrary to the benefits of the transformed enteric glial cell line to distinguish pure glial from neuronal signals, there are clearly a number of drawbacks that this in vitro approach brings about. CRL-2690 cells were shown to have similar physiological properties as primary cultured enteric glia, but the immortalization might have influenced the expression of particular genes.38 Morphologically, CRL-2690 cells also differ from enteric glia in vivo. As neuro-glial communication most likely occurs at the level of processes, this could have important implications for their intracellular Ca2+ signaling. Another disadvantage of cell lines is that they usually represent only one cellular subtype and that heterotypic interactions are lost. It is therefore puzzling that only a proportion of the cells in culture responded to the neurotransmitters we applied. Although CRL-2690 cells were consistently grown to confluence before experiments, it may be that some of the cells were in various states of differentiation, with different receptor expression or altered glia-to-glia communication as a consequence. To overcome the possible cell line-related issues, we tested whether ACh, 5-HT, and ATP could also evoke Ca2+ transients in human glial cells which were isolated from the submucosal plexus of duodenal biopsies. Similar to CRL-2690, human enteric glia also displayed [Ca2+]i rises upon exposure to these neurotransmitters and, although to a lesser extent, to the glial activator LPA. These findings again confirm that neuronal transmitters can directly activate glia; furthermore they prove for the first time that primary glial cells isolated from human biopsies can be used for functional measurements, which opens the possibility to also investigate glial cells from diseased conditions.

Several studies have described that glia-derived substances, such as glutathione, nitric oxide, and S100β control different aspects of GI homeostasis.6,11,35,48 However, the physiological aspects of information processing by enteric glial cells linked to their presumed activation by neurotransmitters are poorly understood. We observed that both ACh and 5-HT increase the number of c-Fos cells in culture. Although this effect greatly contrasts with the massive upregulation after exposure to LPA, it suggests that gene transcription is initiated by signaling cascades following cholinergic and serotonergic receptor stimulation. Interestingly, ATP did not induce early gene transcription in CRL-2690 cells, indicating that different mechanisms act downstream to purinergic vs cholinergic and serotonergic receptor stimulation in these cells. We found no evidence for alterations in glial cell viability or for changes in the expression of GFAP, S100, and Sox10, implicating that, at least such basic cell biological processes, are not controlled by these neurotransmitters. Thus, to fully comprehend the role of enteric glia in ENS physiology, future investigations that aim at elucidating the mechanisms by which they interpret and compute neuronal signals are highly warranted.

In conclusion, our study shows that immortalized rat enteric glia and human enteric glial cells kept in primary culture respond to fast excitatory neurotransmitters with changes in [Ca2+]i. On the basis of our experimental in vitro setting, we thereby provide evidence that enteric glia are not only directly responsive to purinergic but also to serotonergic and cholinergic signaling mechanisms and that this may be the mechanism by which glial cells become activated in case neurotransmitter is released from a varicose release site or leaks from the synaptic cleft of a dedicated synapse.


We thank the members of LENS for discussions and expert technical assistance. We kindly thank Dr. Wegner (Erlangen, Germany) for providing us with the antibody for Sox10.


W.B. and C.C. are postdoctoral fellows of the Fonds voor Wetenschappelijk Onderzoek (FWO, Belgium). This work was funded by Methusalem (BOF, KU Leuven, J.T.) and FWO (KN; G.0501.10, P.V.B.).


The authors have no financial or other conflicts to disclose.

Author Contributions

WB and PVB studied concept and design, interpreted the data, analyzed statistically, drafted and edited the manuscript; CC performed and analyzed experiments on human enteric glia and was involved in manuscript writing; VvdA performed and analyzed aequorin experiments; CVdH generated the mitAeq CRL-2690 cell line; ID provided scientific advice for the aequorin measurements and critically revised the manuscript; JT was responsible for mucosal biopsy collection; PVB wrote analysis software; and JT and PVB obtained funding.