Background
The enteric nervous system (ENS) is the largest component of the peripheral nervous system and is constituted by two cell types, neurons and enteric glial cells (EGC). It is organized in two major ganglionated plexuses, namely the Meissner’s submucosal (SMP) and the Auerbach’s myenteric (LMMP) plexuses, which control gastrointestinal motility, secretion and blood flow, participate in maintenance of the epithelial barrier and modulate various processes of the local immune response [
1,
2]. Although a number of studies have demonstrated that the ENS undergoes structural and phenotypic plastic changes during inflammatory responses [
3], growing evidence suggests that it is not only a bystander, but an active player during inflammation. Indeed, both enteric neurons and EGCs have been shown to contribute to the resulting inflammatory phenotype following chemically induced colitis [
4,
5], probably through the release of several immune mediators [
6‐
9]. Moreover, perineural inflammation is dense in biopsies from inflammatory bowel disease patients and is correlated to major histocompatibility complex class II expression in EGCs [
10], further illustrating the neuroimmune interactions existing in the intestine.
Toll-like receptors (TLR) are transmembrane receptors that recognise different highly conserved microorganism-associated molecular patterns (MAMP), as well as other molecules such as damage-associated molecular patterns [
11,
12]. Upon MAMP binding, the cytoplasmic domain recruits different adapter proteins to trigger a variety of signalling pathways that ultimately activate transcription factors such as nuclear factor-kB (NF-kB), activating protein-1 and interferon regulatory factors, which in turn promote the production of pro-inflammatory cytokines [
13,
14]. TLRs are expressed in most human tissues including the gastrointestinal tract [
15], where they have been proposed to mediate the cross-talk between host cells and commensal microflora due to the key role they play in the innate immune response and in shaping adaptive immunity [
16,
17].
In recent years, expression of some TLRs in the ENS has been described [
18,
19], and their roles in intestinal motility, apoptosis and normal development have progressively been unravelled [
19,
20]. In addition, their immune functions have been also addressed by some authors [
7,
21], but there are still several questions to be answered, such as whether differential recognition of MAMPs by the ENS translates into microbial-selective responses, or whether TLR-mediated signalling is involved in expansion of the inflammatory response. In this context, we aimed to characterise expression and functionality of TLR2/4/9 in the ENS, focusing on their responses in terms of cytokine, chemokine and chemoattraction induction after single or combined ligand challenges. Our results show that neurons are the main TLR-expressing cells in enteric plexuses and integrate TLR signals promoting diverse responses depending on the stimulus received, bringing some new evidence regarding their roles in defence against pathogens and inflammation enhancement.
Methods
Reagents and antibodies
All culture media, foetal bovine serum, antibiotics, N-2 supplement and 4′,6-diamidino-2-phenylindole (DAPI) were from Life Technologies (El Prat de Llobregat, Spain). Trypsin, DNase I, gelatin and Bay 11-7082 were from Sigma (Madrid, Spain). The synthetic diacylated lipopeptide Pam2CSK4, a TLR2/6-specific agonist, was purchased from InvivoGen (San Diego, USA). Lipopolysaccharide (LPS) stimulation of TLR4 was performed with a mixture 1:1 of LPS from Escherichia coli O55:B5 and Salmonella typhosa, both purchased from Sigma. Phosphorothioate-modified class B CpG oligonucleotides (ODN) 1826 5′-TCCATGACGTTCCTGACGTT-3′ and 1826 control (cODN) 5′-TCCATGAGCTTCCTGAGCTT-3′ synthesised by Tib-Molbiol (Berlin, Germany) were used to stimulate TLR9. Primary antibodies used in immunofluorescence were rabbit monoclonal anti-TLR2 (1:500; Abcam, Cambridge, UK), rabbit polyclonal anti-TLR4 (1:100; Novus Biologicals, Cambridge, UK), mouse monoclonal anti-TLR9 (1:100; Novus Biologicals), chicken polyclonal anti-GFAP (1:500; Antibodies-online, Aachen, Germany), mouse monoclonal anti- S100β (1:1000; Abcam), mouse monoclonal anti-HuC/D (1:200; Life Technologies), chicken polyclonal anti-β-tubulin III (1:500; Abcam), rabbit monoclonal anti-NF-kB p65 (1:50; Cell Signaling Technology, Danvers, USA), goat polyclonal anti-IL-6 (1:250; Santa Cruz Biotechnology), mouse monoclonal anti-MCP-1 (1:100; Novus Biologicals) and rabbit polyclonal anti- ionized calcium-binding adapter molecule (IBA)-1 (1:500; Wako Chemicals). Secondary antibodies used were Alexa Fluor 488 and 568 donkey and goat anti-rabbit IgG, respectively; Alexa Fluor 568 goat anti-mouse IgG and Alexa Fluor 647 donkey anti-mouse IgG (1:500; all from Life Technologies), CF488A donkey anti-chicken IgY (1:2000) and CF488A donkey anti-mouse IgG (1:500; both from Biotium, Hayward, USA). For Western blotting, rabbit monoclonal anti-phospho-IkBα (1:1000; Cell Signaling Technology), mouse monoclonal anti-β-actin (1:5000; Sigma), horseradish peroxidase (HRP)-linked goat anti-rabbit IgG (1:10,000; Cell Signaling Technology) and HRP-linked sheep anti-mouse IgG (1:100,000; GE Healthcare, Barcelona, Spain) antibodies were used. TLR2 neutralisation was performed with a mouse monoclonal anti-TLR2 antibody (clone T2.5, 10 μg/mL; Novus Biologicals).
Animals
For ex vivo experiments, 10-week-old male Sprague-Dawley rats were purchased from Charles River (Les Oncins, France) and housed in specific pathogen-free conditions, under a controlled temperature (20 ± 2 °C) and photoperiod (12 h/12 h light-dark cycle), with free access to food and water. Animals were euthanized by CO2 inhalation and distal ileum was removed and placed in ice-cold oxygenated Krebs solution for subsequent manipulation. Colons were flushed with Krebs solution, opened along the mesenteric border, pinned flat in a dissection dish and fixed in Lana’s fixative (4 % paraformaldehyde, 14 % picric acid in 0.4 M phosphate buffer).
For in vitro experiments, pregnant Sprague-Dawley rats purchased from Charles River were killed by CO2 inhalation followed by cardiac puncture exsanguination. Pregnant uteri were removed and kept in ice-cold phosphate-buffered saline (PBS) for further dissection.
All animal procedures performed were approved by the Ethical Committee of the Universitat Autònoma de Barcelona (code 2669).
Cell cultures
Isolation and culture of rat embryonic ENS was performed as described elsewhere [
22]. Briefly, intestines of rat embryos (E16) were removed and finely diced in PBS. Tissue fragments were digested with trypsin and DNase I, and cells obtained were counted and seeded at a density of 2.4 × 10
5 cells/cm
2 on 24- or 48-well plates, previously coated with a 0.5 % gelatin solution in sterile PBS. Stimulation was performed for 24 h after 15-day culture in serum-free medium (DMEM-F12 (1:1)) containing 1 % of N-2 supplement.
The murine macrophage cell line RAW 264.7 (ATCC® TIB-71) was purchased from the American Type Culture Collection and cultured in DMEM supplemented with 10 % heat-inactivated foetal calf serum.
Stimulation experiments were performed for 24 h with either 100 ng/mL Pam2CSK4, 100 ng/mL LPS, 1 μM ODN 1826, 1 μM cODN 1826 or combinations of these ligands. For NF-kB inhibition and TLR2 neutralisation experiments, cultures were pre-treated for 1 h with 15 μM Bay 11-7082 or 10 μg/mL anti-TLR2 antibody before MAMP stimulation.
In costimulation experiments, comparison between expected additive effects and measured effects of TLR ligand combinations was calculated according to the model of functional interaction, represented by the following equation as described in [
23]: E(ODN + LPS)expected = E(ODN)measured + E(LPS)measured – E(ODN)measured * E(LPS)measured.
Immunofluorescence
Fixed colon was dissected under a stereo microscope to obtain whole-mount preparations of the SMP and the LMMP. Adult tissues, as well as ENS cultures grown on cover-slips, were blocked for 1 h in PBS containing 4 % horse serum, 0.1 % Triton X-100 and 0.01 % NaN3. Samples were incubated overnight at 4 °C with combinations of TLR2, TLR4 or TLR9 with Hu C/D, β-tubulin III (neuronal markers), GFAP, S100β (glial markers), IL-6, MCP-1 or IBA-1 antibodies. Secondary antibodies to rabbit, mouse or goat IgG and chicken IgY were used to detect bound primary antibodies. All samples were mounted in Vectashield aqueous anti-fading mounting medium (Vector Laboratories) and analysed under a Zeiss LSM 700 confocal laser microscope (Carl Zeiss, Madrid, Spain). To avoid overlapping, control preparations were previously single stained for one marker, and their fluorescence was evaluated with the three different lasers used for final image acquisition. Band-pass filters were set up to avoid cross-talk between channels, and sequential acquisition was performed during experiments to excite fluorochromes one at a time.
Real-time RT-PCR analysis
Total RNA from ENS culture was extracted using the RNeasy Mini Kit (QIAGEN, Las Matas, Spain), quantified by optical densitometry and assessed for integrity by on-chip gel electrophoresis with the Experion™ System (Bio-Rad Laboratories, el Prat de Llobregat, Spain). Then, 100 ng of RNA were retro-transcribed by using the Transcriptor First-strand cDNA Synthesis Kit (Roche Applied Science, Mannheim, Germany) for reverse-transcriptase polymerase chain reaction (RT-PCR). Primer sequences listed in Table
1 were designed to span introns using the Universal ProbeLibrary Assay design Center (
https://lifescience.roche.com/webapp/wcs/stores/servlet/CategoryDisplay?tab=Assay+Design+Center&identifier=Universal+Probe+Library&langId=-1), and checked for specificity through BLAST search. PCR amplifications were performed using the LC480 SYBR Green I Mastermix (Roche Applied Science) according to manufacturer’s protocol, and run on a LightCycler 480 II instrument (Roche Applied Science). Absence of coamplification products was assured by generating a final melting curve for each reaction and by loading PCR products on a denaturing 2 % agarose gel, stained with SYBR safe (Life Technologies) and visualized under UV transillumination. Specificity of the primers was also determined by sequencing these amplification products. Messenger RNA (mRNA) level of expression of the genes of interest was corrected to that of the S6 housekeeping gene and calculated by the ΔΔCt method.
Table 1
List of primers used for real-time RT-PCR analysis
rS6
| CCAAGCTTATTCAGCGTCTTGTTACTCC | CCCTCGAGTCCTTCATTCTCTTGGC | NM_017160 |
rTLR2
| CAGATGGCCAGAGGACTCA | AATGGCCTTCCCTTGAGAG | ENSRNOT00000013025.3 |
rTLR4
| GGATGATGCCTCTCTTGCAT | TGATCCATGCATTGGTAGGTAA | NM_019178.1 |
rTLR9
| TCCGTGACAATCACCTCTCTT | GGTCCAGGTCTCGCAGATT | NM_198131.1 |
In order to compare mRNA expression levels of the receptors in basal conditions, absolute mRNA levels were estimated by determining the difference between the cycle threshold (Ct) of the target receptor and the Ct of the housekeeping gene, as described elsewhere [
21,
24]. According to their ΔCt to the S6 gene, genes were classified as high-expressed (ΔCt less than 5 cycles), intermediate-expressed (ΔCt from 5 to 15 cycles), low or rare-expressed (ΔCt superior to 15 cycles) and undetectable (ΔCt superior to 40 cycles).
Western blot
ENS cultures were harvested in RIPA lysis buffer (Millipore, Madrid, Spain) containing 2 mM sodium orthovanadate, phosphatase inhibitor cocktail 3 (Sigma) and a tablet of Complete™ protease inhibitors cocktail (Roche Applied Science). Protein samples (30 μg) were separated on a 10 % acrylamide gel containing 0.1 % sodium dodecyl sulfate and transferred to a nitrocellulose membrane with the iBlot™ Dry Blotting System (Life Technologies). Membranes were blocked for 1 h at room temperature with 5 % non-fat dry milk in Tris-buffered saline (100 mM NaCl, 10 mM Tris, pH 7.5) with 0.1 % Tween 20 (TBST), and incubated overnight at 4 °C with primary antibodies in a 5 % BSA solution in TBST. Bound antibodies were detected with HRP-conjugated anti-rabbit or anti-mouse antibodies, and visualized by enhanced chemiluminescent detection (ECL advance, GE Healthcare). Membranes were stripped for 15 min in Reblot buffer (Millipore), followed by extensive washing in TBST before reblocking with 5 % non-fat dry milk in TBST and reprobing for β-actin determination. Bands were imaged in a LAS-3000 Imager (Fujifilm, Tokyo, Japan) and quantified with Multigauge 3.0 software (Fujifilm). To allow comparison between different membranes, the density of the bands was referred to that of untreated controls and normalized to the amount of β-actin in the same sample.
TNF-α, IL-6 and MCP-1 ELISA
Culture supernatants were centrifuged, aliquoted and frozen, and further assayed with the corresponding BD OptEIA™ ELISA Sets (BD), following manufacturer’s recommended assay procedures. Final cytokine or chemokine values were related to the total protein amount of the sample, which was determined by using the BCA protein assay kit (Pierce, Rockford, USA).
Migration assays
Twenty-four hours after stimulation of ENS primary culture with MAMPs, conditioned supernatants were centrifuged, placed into 24-well plates and left to equilibrate for an hour with the transwell insert. Then, 105 RAW 264.7 macrophages were seeded in the upper chamber of the 8-μm-pore transwell inserts, and allowed to migrate for 4 h at 37 °C and 5 % CO2. After fixation in 4 % paraformaldehyde, cells on the upper surface of the transwell membrane were removed by rubbing with a sterile cotton swab, and cells on the lower surface were stained with DAPI. The average number of migrating cells was determined by counting eight fields per membrane at ×100 under a Carl Zeiss Axioskop 40 FL epifluorescence microscope equipped with a Zeiss AxioCam MRm camera (Carl Zeiss, Germany). Each experiment was performed in duplicate.
Statistical analysis
Results are presented as mean values ± S.E.M. of at least three independent experiments. All data were compared using one-way or two-way ANOVA, followed by Tukey’s post hoc test (unless otherwise stated). Where stated, randomised block design analysis was performed to minimise the variability due to differences between individual culture responses. Data analysis and plot were performed with GraphPad Prism 5.0 software (GraphPad Software Inc., La Jolla, USA). Randomised block design analyses were performed with Minitab 15 Statistical Software (Minitab Inc., Pennsylvania, USA). A P value <0.05 was considered to be significant.
Discussion
Since their initial characterisation in human in the late 90s [
13], TLRs have been widely described in several tissues and cells, including central and peripheral nervous systems [
15,
30]. In recent years, the study of their role in the ENS has gained attention; however, there are still only a few works addressing their participation in the immune response to microbes. In this study, we report the expression and functionality of TLR2/4/9 in embryonic ENS primary cultures, emphasising in their interactions in terms of cytokine and chemokine production. Our results indicate that enteric neurons respond to MAMPs through secretion of inflammatory mediators, integrating their signals to trigger a tailored response to each challenge. Indeed, activation of TLR2 or TLR9 upon LPS stimulation enhanced macrophage chemoattraction, while potentiating different pro-inflammatory responses. Taken together, these findings suggest that TLRs confer the ENS the ability to discriminate microbial signals and expand inflammation by promoting particular pro-inflammatory microenvironments and interacting with resident immunocytes to chemoattract immune cells. These facts further support the idea that the ENS is an immunologically active tissue.
Expression of TLRs within the ENS has been previously assessed by other groups in full-thickness tissue sections of the small intestine [
18,
25]. In this regard, colocalisation studies in SMP and LMMP whole-mounts from adult rat colon offer an improved perspective to discern immunoreactive cell populations. Our findings in these tissues confirm previous data about TLR2 expression, which had been described in both HuC/D
+ and GFAP
+ cells [
19]. Conversely, we could only observe TLR4 staining in some subsets of enteric neurons [
20,
25], even though expression in EGCs has been already reported [
18,
21]. In this regard, our results agree with parallel studies in central nervous system cultures pointing out that neurons have more prominent TLR4 expression than astrocytes [
29]. Finally, the TLR9 expression patterns we report here match with previous work that localise this receptor in neuronal somas and β-tubulin III
+ microtubules from dorsal root ganglia cultures [
31] and various brain sections [
32]. Interestingly, TLR distribution in embryonic ENS primary cultures was very similar to that of adult tissues, which might indicate that few modifications affect their localisation during development. Indeed, minor TLR2 and TLR4 expression variations have been observed in the mouse developing brain, whereas TLR9 increases in adulthood [
32].
Functional TLRs activate the NF-kB pathway in several cell types, including neurons [
33] and EGCs [
21]. This transcription factor has been shown to convey signals from lipopeptides, LPS and class B CpG ODNs, leading to production of cytokines and chemokines [
13,
14,
34]. Our results indicate that NF-kB is also involved in ENS-mediated recognition of MAMPs, but additionally suggest that other signalling cascades may participate in cytokine and chemokine production. Indeed, inhibition with Bay 11-7082 resulted in significant decrease in TNF-α, IL-6 and MCP-1 release, but not complete abrogation. Therefore, it is likely that other pathways, such as the MAPK and the 5-adenosine monophosphate-activated protein kinase, are responsible for the observed mediator production after NF-kB blockade, as they have been previously shown to play important roles in production of TNF-α in embryonic ENS culture [
7] and central nervous system neurons [
29].
Release of cytokines and other immunomodulatory molecules is a common feature following TLR recognition of MAMPs [
13]. However, a number of adaptions have developed in the gastrointestinal tract to prevent TLR-driven immune responses to commensal microbes [
35]. Some of these adaptions might also occur in the ENS, as we observed no significant pro-inflammatory or chemoattractive responses in embryonic ENS culture upon Pam2CSK4 or ODN 1826 stimulation, despite the fact that they induced NF-kB activation. Similar observations have been reported in human EGCs, in which translocation into the nucleus of the NF-kB p50 subunit takes place after exposure to enteropathogenic bacteria, but not to probiotic strains [
21]. These findings might underlie selective mechanisms to signal microbe presence while minimising responses to those non-pathogenic.
We have reported an active participation of resident macrophages in IL-6 and MCP-1 production, which poses an important limitation to the use of these embryonic ENS cultures for immunologic research purposes. Indeed, the neuronal or macrophage origin of the released cytokines could not be quantified, making it difficult to define the actual contribution of each cell type to the inflammatory microenvironment and chemoattractive effects. We propose a major role for enteric neurons because (1) expression of TLR2/4/9 was preferentially seen in these cells; (2) TNF-α expression, which has been localised by our collaborators in enteric neurons following LPS exposure [
7], was altered upon TLR2/4 and TLR4/9 interactions; (3) MCP-1 reactivity following LPS stimulation accumulated majorly in neurons and (4) TLR2 up-regulation upon LPS challenge occurred in neurons [
7]. Even though the involvement of resident immunocytes may be important, embryonic macrophages display important phenotypic differences with adult macrophages, such as no major histocompatibility complex (MHC) class II expression or poor cytokine production [
36,
37]. Similarly, embryonic EGCs did not release IL-6 upon MAMP challenge, which is in contrast with previous observations in adult cells [
8,
21].
Consistent up-regulation of TLR2 following MAMP or cytokine stimulation has been described in astrocytes, neurons and other cell types [
7,
24,
38,
39]. Induction of TLR2 mRNA is dependent on reactive oxygen species [
39] and ultimately on NF-kB, which has different binding sites on the TLR2 promoter region [
40]. Overexpression of this receptor is necessary to reach maximal NF-kB activation after MAMP recognition [
38], possibly through signalling the formation of lipoproteins bearing lipid oxidation end-products [
41,
42]. Our experiments using a TLR2-blocking antibody agree with these explanations, showing that in our setup, TLR2 induction is crucial to obtain maximal TNF-α, IL-6 and MCP-1 production following LPS stimulation. Furthermore, increased chemoattraction after combined challenge of TLR2/4 suggests that these receptors may interact in ways other than those we determined to enhance the release of cytokines or chemokines. These additional interactions might implicate activation of additional signalling pathways, since we report reduced responsiveness to NF-kB inhibitors in ENS-mediated release of IL-6 and MCP-1 after Pam2CSK4 + LPS stimulation.
In addition to TLR2/4 cross-responsiveness phenomena [
43,
44], other TLR interactions after combined MAMP challenge have been documented [
23,
45,
46]. Here, we demonstrate strong synergic responses in TNF-α release after costimulation with TLR4/9 ligands. Furthermore, IL-6 production and chemoattraction of macrophages by supernatants from ODN 1826 + LPS-treated embryonic cultures were also significantly enhanced, suggesting that additional chemokines might be induced after combined TLR4/9 challenge. Similar interactions between both receptors have been characterised in microglia, dendritic cells and bone marrow-derived macrophages [
23,
46,
47]. The molecular mechanisms which may account for such effects are TLR9 up-regulation following LPS stimulation, as we and others have described [
46], as well as enhanced signalling and duration through alternative pathways, such as the MAPK [
47]. From a functional point of view, the ability to integrate different stimuli from the same microorganism contributes to discrimination of commensals and pathogens, inducing robust pro-inflammatory responses against the latter [
48]. Recognition of Gram-negative motifs from both membrane (LPS) and nucleus (CpG DNA) is hence integrated in a synergic way, triggering strong pro-inflammatory responses and promoting chemoattraction of immune cells. In contrast, detection of Gram-positive and Gram-negative molecular patterns (Pam2CSK4 and LPS) lead to similar chemotactic responses but to a not so harsh pro-inflammatory environment, which might result into a milder priming of the subsequent effector response.
Different studies have addressed the involvement of central and peripheral nervous system neurons in defence against pathogens. These studies demonstrate that LPS stimulation activates NF-kB and MAPK pathways in these cells to induce production of cytokines like TNF-α and IL-6, and chemokines such as RANTES and KC [
29,
33]. Participation of enteric neurons in inflammation has been also demonstrated. Transgenic mice displaying increased and reduced numbers of neurons develop, respectively, more and less severe experimental colitis than wild-type littermates, suggesting that neurons play pro-inflammatory roles [
4]. Furthermore, different signalling pathways are activated in these cells upon IL-1β or LPS challenge [
7,
49], leading to production of IL-8 and TNF-α [
6,
7]. Our findings show that enteric neurons express TLR2/4/9,signal through the NF-kB pathway and release IL-6 and MCP-1 in response to LPS challenge. These observations add further evidence to the pro-inflammatory roles of neurons, and set them as active players in neuroimmune interactions in the gastrointestinal tract, participating in microbial recognition and priming subsequent immune responses.
Abbreviations
DAPI, 4',6-diamidino-2-phenylindole; EGC, enteric glial cell; ENS, enteric nervous system; GFAP, glial fibrillary acidic protein; HRP, horseradish peroxidase; IBA-1, ionized calcium-binding adapter molecule; IkB, inhibitor of kB; IL-6, interleukin-6; LMMP, longitudinal muscle myenteric plexus; LPS, lipopolysaccharide; MAMP, microorganism-associated molecular pattern; MAPK, mitogen-activated protein kinase; MCP-1, monocyte chemoattractant protein-1; NF-kB, nuclear factor-kB; ODN, CpG oligonucleotide; PBS, phosphate-buffered saline; PCR, polymerase chain reaction; SMP, submucosal plexus; TLR, toll-like receptor; TNF-α, tumour necrosis factor-α
Acknowledgements
Work in EF’s group is also supported by Generalitat de Catalunya (SGR-2009-997). The authors would like to thank J. Chevalier, M. Biraud and M. Ribeiro-Neunlist (INSERM U913) for training and technical support on ENS PC preparation; N. Barba (INc from the Universitat Autònoma de Barcelona) for excellent technical assistance in confocal microscopy imaging and Drs. L. Grasa (Universidad de Zaragoza), MT. Abreu (University of Miami) and R. Herzog (University of Florida) for providing TLR2−/−, TLR4−
/
− and TLR9−
/
− tissues for confirmation of antibody specificity.