Modulation of lipopolysaccharide-induced neuronal response by activation of the enteric nervous system

Rat ENS primary cultures (rENSpc) were performed as previously described using the small intestines of E15 Sprague-Dawley rat embryos (Janvier Laboratories SA, Le Genest-St-Isle, France) [15]. These procedures were approved by the local Animal Care and Use Committee (agreement E. 44011; INSERM, Nantes, France). Briefly, the small intestines of rat embryos were removed, diced in Hank’s Buffered Salt Solution (Sigma-Aldrich, Saint-Louis, Missouri, United States) and collected in 5 mL of Dulbecco’s modified Eagle’s medium (DMEM)-F12 (Gibco®, Life Technologies, Carlsbad, California, United States) (1:1) for digestion at 37°C for 15 minutes in 0.1% trypsin (Sigma-Aldrich). The trypsin reaction was stopped by adding medium containing 10% fetal calf serum and then treated by DNase I 0.01% (Sigma-Aldrich) for 10 minutes at 37°C. After triturating with a 10 mL pipette, cells were centrifuged at 750 rpm for 10 minutes. Cells were counted and then seeded at a density of 2.4 × 10⁵ cells/cm⁻² on 24-well plates previously coated with a solution of gelatin (0.5%; Sigma-Aldrich) in sterile phosphate buffered saline. After 24 hours, the medium was replaced with a serum-free medium DMEM-F12 (1:1) containing 1% of N-2 supplement (Life Technologies. Cultures were maintained for 14 days. Using this method, rENSpc is composed of well-organized ganglia interconnected by interganglionic fiber strands forming a network that lays on a smooth muscle cells monolayer (smooth muscle actin or SMA positive cells). Fluorescence-activated cell sorting (FACS) analysis, aimed to reveal the cellular composition of the rENSpc, revealed that enteric neurons and glial cells represent 5 to 6% and 20 to 40%, respectively, of cells present in culture (personal observation). Half of the medium was replaced every other day and the day before the experiment.

The rENSpc were treated with lipopolysaccharides (Escherichia coli and Salmonella typhosa; 1:1, Sigma-Aldrich) at 0.1 μg/ml for the indicated time, except for Figure 1A, where different concentrations were tested. For the purpose of establishing which pathways and receptors are implicated in TNF-α and TLR2 regulation, U0126 (mitogen-activated protein kinase kinase 1/2 or MEK1/2 inhibitor; 10 μM), compound C (5’-adenosine monophosphate-activated protein kinase (AMPK) inhibitor; 10 μM) (Calbiochem, Merk Millipore, Billerica, Massachusetts, United States) and anti-rat TNF-α (1 and 10 μg/ml; eBiosciences, San Diego, California, United States) were added 30 minutes prior to the addition of LPS or ENS stimulation. Pam3CSK4 (TLR1/2 agonist; 100 ng/ml; Invivogen, San Diego, California, United States), A438079 (selective P2X₇ antagonist; 30 μM; Tocris Bioscience, Bristol, United Kingdom), adenosine-5'-triphosphate (ATP) (100 μM) and 2’(3’)-O-(4-benzoylbenzoyl) adenosine-5'-triphosphate triethylammonium salt (BzATP) (selective P2X7 agonist; 100 μM; Sigma-Aldrich) were also used to treat ENS plus or minus LPS.

To study the effect of neuronal activity on cytokine secretion, rENSpc were electrically stimulated in 24-well plates fitted with a pair of platinum electrodes connected to an electrical stimulator (DualImpedance Research Stimulator, Harvard Apparatus Ltd, Edenbridge, United Kingdom). The electrical field stimulation (EFS) parameters used were trains of constant current pulses (pulse duration: 20 μs; amplitude: 8 V; frequency: 15 Hz) applied for seven hours, with reversal of electrode polarity every 30 minutes, and supernatants and lysates were collected after 24 hours for ELISA TNF-α measurements and quantitative PCR (qPCR). Neuronal activation was verified by analysis of c-Fos expression after seven hours of EFS (Additional file 1). Putative neuronal damage induced by EFS was also verified. Following seven hours of EFS, no change in neuron-specific enolase (NSE) in the culture medium or in protein gene product (PGP) 9.5 expression was observed as compared to control (non-stimulated condition), suggesting that EFS did not affect neuronal viability. Neuronal activity during the seven hours was also measured using an essential neuromediator, ATP.

Primary cultures of human and adult rat EGCs were obtained according to the procedure described by Soret et al. [16]. Briefly, human EGC were originated from colonic resections of patients with colorectal cancer (according to the guidelines of the French Ethics Committee for Research on Human and registered under the number DC-2008-402). Adult rat EGC were obtained from entire small intestines of Sprague-Dawley rats. After dissection of rat or human segments, the remaining longitudinal muscle-myenteric plexus (LMMP) was placed in GentleMACS tubes C in a GentleMACS dissociator (MiltenyiBiotec, Bergisch Gladbach, Germany) for enzymatic digestion (250 μl protease, 5 g/l⁻¹; 250 μl collagenase, 20 g/l⁻¹; 400 μl bovine serum albumin, 50 g/l⁻¹; Sigma-Aldrich) and mechanical dissociation in DMEM/F12 medium (supplemented with 10% heat-inactivated fetal calf serum, 100 IU/ml penicillin, 100 μg/ml streptomycin, 1.1 μg/ml amphotericin B, 20 μg/ml gentamicin, 6 mM glutamine and 2.1 g/l NaHCO₃; Life Technologies). Following gentle agitation at 37°C, pellets were washed and resuspended in DMEM/F12 complete medium. The suspension was placed under a microscope and ganglia were picked up using a pipette with a sterile tip and seeded in 24-well plate coated with gelatin. After 48 hours, the medium was entirely replaced and only EGC proliferated to reach confluency after two to four weeks. The primary cultures of EGC from rat embryos (JUG2 cell line) was obtained as described by Bach-Ngohou et al. [17] and derived from ENS primary culture after trypsinization and differential centrifugation. EGC were treated with LPS at a concentration of 0.1 μg/ml for seven hours before collecting supernatants.

Colonic segments were pinned in a dissection dish containing ice-cold, sterile oxygenated Krebs solution (Sigma-Aldrich). Mucosae, submucosal and circular muscles were carefully removed under a dissection microscope and the remaining LMMP was placed in a 24-well plate containing DMEM/F12 complete medium (supplemented as in EGC primary culture procedure). The following day, the medium was replaced with 1:1 DMEM-F12 containing 1% N-2 supplement, and specimens were electrically stimulated with the same equipment and parameters as for rENSpc stimulation. After seven hours, supernatants and lysates were collected and tissues were weighted to normalize the ELISA analysis.

After treatments, rENSpc were fixed in PBS containing 4% paraformaldehyde at room temperature for 30 minutes. Following pre-incubation in PBS containing 5% donkey serum and 0.5% Triton™ X-100 (Sigma-Aldrich) for 30 minutes, cells were incubated with the following primary antibodies overnight at 4°C: rabbit anti-phospho-p44/p42 (P-ERK), rabbit anti-phospho-acetyl coenzyme A (CoA) carboxylase (P-ACC; 1:500; Cell Signaling Technology, Beverly, Massachusetts, United States), goat anti-TLR2 (1:50; Santa Cruz Biotechnology, Dallas, Texas, United States), and goat anti-TNF-α (1:200; Santa Cruz Biotechnology). Anti-rabbit and anti-goat conjugated to Cyanine3 (1:500; Jackson Immunoresearch, West Grove, Pennsylvania, United States) were used as secondary antibodies for one hour. In the following step, the cells were labelled with primary antibodies for neuronal or glial markers: mouse anti-Hu (1:200; Molecular Probes®, Life Technologies), rabbit anti-PGP9.5 (1:1,000; Ultraclone Limited, Tebu-bio, Le Perray-en-Yvelines, France), mouse anti-Sox10 (1:500; R&D systems, Minneapolis, Minnesota, United States), rabbit anti-S100b (1:1,000; DakoCytomation, Glostrup, Denmark), mouse anti-S100b (1:1,000, Abcam, Cambridge, United Kingdom), and mouse anti-β-tubulin III (1:1,000; Sigma-Aldrich). After secondary antibodies (FluoProbes™488; 1:200; Interchim or Cyanine5; 1:200; Jackson Immunoresearch), images were acquired with an Olympus IX 50 fluorescence microscope coupled to a digital camera (model DP71, Olympus), analyzed with Cell B software (Soft Imaging System, Olympus), and treated with Image J software (National Institute of Health, Bethesda, Maryland, United States).

ENS cells cultured for different times (as indicated), with or without stimuli, LPS or agonists, were lysed in RA1 (Nucleospin RNAII, Macherey-Nagel, Düren, Germany) to separate RNA and proteins. Proteins were precipitated and pellets were resuspended in Phosphate-buffered saline/tris (2-carboxyethyl)phosphine (PSB/TCEP) (Macherey-Nagel). Samples were processed for electrophoresis using NuPAGE™ MES SDS buffer kit (Life Technologies) and separated on 4 to 12% Bis-Tris gels (NuPAGE, Life Technologies). Proteins were transferred to nitrocellulose membranes with the iBlot™ system (Life Technologies). After blocking with Tris-buffered saline (TBS), 0.1% Tween 20 and 5% nonfat dry milk for one hour, blots were incubated overnight at 4°C with primary antibodies diluted in TBS and 5% nonfat dry milk for rabbit anti-phospho-Thr172-AMPK (P-AMPK), rabbit anti-AMPK, rabbit anti-phospho-p44/p42 (P-ERK), rabbit anti-p44/p42 and rabbit anti-phospho-ACC (1:500; Cell Signaling Technology), and mouse anti-beta-actin (1:5000; Sigma-Aldrich). Immunoblots were probed with the appropriate horseradish peroxidase-conjugated secondary antibodies (Life Technologies) and visualized by chemiluminescence (ECL Crescendo or Forte, Merk Millipore). Quantitative analysis was performed using Image J software. The value of phosphorylated and total protein immunoreactivity was normalized to the amount of β-actin immunoreactivity and expressed as the fold increase relative to the control, taken as 1.

According to the manufacturer’s recommendations, total RNA extraction from cells was performed with Nucleospin RNAII kit (Macherey-Nagel) and 1 μg purified RNA was denatured and processed for reverse transcription using Superscript II reverse transcriptase (Life Technologies). PCR amplifications were performed using the Absolute Blue SYBR green fluorescein kit (Roche, Basel, Schweiz) run on a Rotor-Gene (Qiagen, Venlo, The Netherlands). The following primers were used:

S6 number: [GenBank:NM_001010]. Forward primer: 5’-CCAAGCTTATTCAGCGTCTTGTTACTCC-3’. Reverse primer: 5’-CCCTCGAGTCCTTCATTCTCTTGGC-3’. TNF-α number: [GenBank:NM_012675.3]. Forward primer: 5’-CTCATTCCTGCTCGTGGCGGG-3’. Reverse primer: 5’-TACGACGTGGGCTACGGGCTT-3’.TLR2 number: [GenBank:NM_198769.2]. Forward primer: 5’-CAGATGGCCAGAGGACTCA-3’. Reverse primer: 5’-AATGGCCTTCCCTTGAGAG-3’.

Relative quantification of gene expression was determined using the standard curve method and endogenous control ribosomal protein S6 mRNA. The ratios, TNF-α/S6 or TLR2/S6, were compared with control experimental conditions normalized to 1.

The concentrations of TNF-α and IL-6 were determined in rENSpc and human LMMP specimen culture supernatants by ELISA (BD OptEIA™ ELISA Set, BD Biosciences, Franklin Lakes, New Jersey, United States) according to the manufacturer's instructions. Absorbance measurements were performed at 450 nm on a spectrophotometric enzyme-linked immunosorbent sandwich assay (ELISA) plate reader (Varioskan™, Thermo Scientific, Waltham, Massachusetts, United States) using the SkanIt software (Thermo Scientific). For human LMMP specimens, the TNF-α concentration was expressed related to the weight of the specimen.

Data resulted from independent experiments performed in duplicate or triplicate. Statistical significance was evaluated using GraphPad Prism software (GraphPad Software Inc., La Jolla, California, United States). For time- and dose-dependent experiments, a two-way ANOVA test followed by a Bonferroni post-hoc test was used. The time-dependent response to LPS in Western blots was analyzed with a Kruskal-Wallis nonparametric ANOVA test, followed by Dunn’s post-hoc test. The differences between groups were calculated by a two-tailed Student’s t-test for nonparametric and unpaired data, or by Mann-Whitney U test.

To determine whether the ENS could respond to LPS by producing proinflammatory cytokines, we measured TNF-α concentrations in the supernatants of rENSpc treated with different concentrations of LPS for two to 24 hours. We have verified that neurons of the rENSpc expressed the TLR4 and were thereby able to respond to LPS (Additional file 2). The TNF-α concentration was significantly increased after eight hours of 1 μg/ml LPS stimulation, and between 12 to 24 hours for LPS concentrations ranging from 0.01 to 1 μg/ml (Figure 1A). We chose the intermediate LPS concentration (0.1 μg/ml) for the rest of the study. Using immunocytochemistry, we further aimed to identify the ENS cell type (neuronal or glial) expressing TNF-α. In rENSpc, enteric neurons were organized in ganglia connected to each other by interganglionic fiber strands as demonstrated by immunostaining using βIII-tubulin (Tuj) antibody (Figure 1B). EGCs, identified by S100β immunostaining, were also present in enteric ganglia (Figure 1B). We showed that in rENSpc treated with LPS for 24 hours, TNF-α immunoreactivity was detected in the cytoplasm of Tuj-positive cells, but not in S100β-positive cells (Figure 1B). To confirm that EGC do not produce TNF-α, we measured TNF-α production from human, adult and embryonic rat EGC in response to treatment with LPS. None of these cells responded to LPS treatment with an increase in TNF-α production (Figure 1C). Our results show that enteric neurons are the cells of the rENSpc producing TNF-α in response to LPS stimulation.

To assess if ENS activity could regulate TNF-α production, we measured LPS-induced TNF-α production in the supernatants of rENSpc after EFS and in those not treated with EFS. EFS had no significant effect on basal TNF-α production. However, EFS significantly inhibited TNF-α production induced by LPS in rENSpc (Figure 2A). qPCR analysis indicated that EFS inhibited the increase in TNF-α mRNA level observed after LPS (Figure 2B). These results show that LPS induce an increase in TNF-α transcript and protein levels, and that ENS activation reduces TNF-α production.

Various studies have demonstrated the anti-inflammatory effects of adenosine receptor activation [18],[19] and ATP is a common intercellular messenger produced by enteric neurons [20]. To determine their putative regulatory role in LPS-induced TNF-α production by the ENS, we studied the potential role of ATP and purinergic receptor activation. An increased extracellular ATP concentration remarkably decreased the LPS-induced production of TNF-α. The application of BzATP, a P2X7 agonist, induced a similar reduction in TNF-α production, as did ATP. The ATP effects were partially reversed by the selective P2X7 antagonist, A438079 (Figure 2C). These results show that LPS-induced TNF-α production is reduced by ATP stimulation of P2X7 receptors.

To decipher the signaling pathways responsible for LPS-induced TNF-α synthesis by the ENS, and its modulation by EFS, we analyzed the possible implication of the canonical extracellular signal-regulated kinase (ERK) pathway, and also of the 5'-adenosine monophosphate-activated protein kinase (AMPK) pathway. By Western blot analysis using phospho-specific antibodies recognizing the active form of the ERKs, we showed that phosphorylation of ERKs was increased after two hours of LPS stimulation, with no change in ERK expression (Figure 3A). Activation of the AMPK pathway can be measured by the phosphorylation of the Thr172 of the AMPK itself, or by the phosphorylation of acetyl-CoA carboxylase (ACC), an AMPK target, on its Ser19. Neither AMPK nor ACC phosphorylation changed significantly over time, with or without LPS treatment (data not shown). Immunocytochemical analyses revealed that the immunoreactivity for phospho-ERK and phospho-ACC concurred with Tuj and Hu (neuronal markers), but not with S100β or Sox10 (glial markers) immunoreactivity (Figure 3B). These data show that, while the AMPK pathway was constitutively activated in neuronal cells, ERKs were activated in neurons after LPS treatment. Pretreatment of rENSpc with U0126 or compound C inhibitors of the ERK or AMPK pathways, respectively, significantly reduced TNF-α mRNA expression (Figure 3C) and TNF-α production (Figure 3D). Interestingly, ERK phosphorylation was not affected by compound C (Figure 3E), and ACC-phosphorylation was not modified by U0126 (Figure 3F), again suggesting the independence of the two pathways. These results demonstrate that the AMPK and ERK pathways participate independently in LPS induction of TNF-α production by the ENS.

To determine if EFS could inhibit TNF-α production by modulating the ERK and AMPK pathways, we analyzed ERK and AMPK activation in rENSpc treated, or not, with LPS, in the presence or absence of EFS. EFS had no significant effect on basal pathway activity but, combined with LPS treatment, it significantly reduced the phosphorylation of ERK and AMPK compared to the LPS conditions (Figure 3G). This suggests that ENS activation limits TNF-α production by inhibition of LPS-induced ERK activation and decrease of AMPK activity.

To assess if the induction of TNF-α production by LPS and its regulation by EFS also occurs in human samples, we measured LPS-induced TNF-α production in the supernatants of human LMMP after, or not, EFS. EFS had no significant effect on basal TNF-α production. However, EFS significantly inhibited TNF-α production induced by LPS in the LMMP (Figure 4A). The application of BzATP, a P2X7 agonist, induced a similar reduction in TNF-α production (Figure 4B). Pretreatment of LMMP with U0126 or compound C inhibitors of the ERK or AMPK pathways, respectively, significantly reduced TNF-α production (Figure 4C). These results show that in human LMMP as in rENSpc, LPS induces an increase in TNF-α production that depends on AMPK and ERK pathways, and that ENS activation reduces this production by stimulation of the P2X7 receptors.

To further investigate the ENS response to LPS, we measured, using qPCR, TLR2 mRNA expression in rENSpc. TLR2 mRNA was detected in control conditions and LPS treatment induced a strong increase in its expression (Figure 5C). To determine if this regulation also occurred at the protein level, and to identify the cells expressing TLR2, we compared immunoreactivity for TLR2 with immunoreactivity for the neuronal marker Tuj or for the glial marker S100β on rENSpc stimulated, or not, with LPS. TLR2 immunoreactivity was detected slightly in our control conditions but clearly appeared in Tuj-positive cells, and rarely in S100β-positive cells, when rENSpc were treated with LPS (Figure 5A). To confirm that TLR2 expression is mainly neuronal, we have analyzed TLR2 immunoreactivity in human EGC. The glial marker S100β staining of cultures of human EGC always present highly positive cells and low positive cells. Only the highly S100β-positive EGC also expressed the TLR2 (Figure 5B), but no changes were observed after LPS stimulation. All together these data show that mainly enteric neurons express the TLR2 and respond to LPS through TLR2 upregulation.

To determine if TNF-α production induced by LPS could be involved in the TLR2 upregulation observed in enteric neurons in response to LPS, we measured TLR2 mRNA in rENSpc stimulated by LPS in the absence or presence of TNF-α-neutralizing antibodies. Anti-TNF-α antibodies significantly reduced the expression of the TLR2 mRNA induced by LPS (Figure 5D). In the same manner, EFS stimulation of the rENSpc significantly reduced the TLR2 mRNA expression induced by LPS (Figure 5E). These data show that TNF-α production induced by LPS upregulated the TLR2 receptor expression and depends on ENS activity.

TLR2, as well as TLR4, activation in different cell types can induce TNF-α production [21]. To assess the involvement of TLR2 in the ENS response to LPS, we measured TNF-α production in rENSpc stimulated with the TLR1/2 agonist, Pam3CSK4. Pam3CSK4 induced a significant increase in TNF-α production that could be further increased after the addition of LPS (Figure 5F). These results suggest that TLR2 activation can induce TNF-α production by enteric neurons.

An additional functional consequence of TNF-α regulation by ENS activity could be the regulation of other cytokines, such as interleukin-6 (IL-6). To address this issue, we measured IL-6 production by rENSpc treated with LPS in the absence or presence of TNF-α-neutralizing antibodies. Like TNF-α, IL-6 production was increased by LPS, but this was delayed in time compared to the TNF-α response (data not shown). TNF-α-neutralizing antibodies had no effect on basal IL-6 production but, surprisingly, significantly increased that induced by LPS (Figure 6A). EFS stimulation of the ENS increased the IL-6 production (Figure 6B). This suggests that ENS activation, through TNF-α downregulation, can potentiate IL-6 production.