|Home | About | Journals | Submit | Contact Us | Français|
Vesicular glutamate transporters (VGLUTs) have been extensively studied in various neuronal systems, but their expression in visceral sensory and autonomic neurons remains to be analyzed in detail. Here we studied VGLUTs type 1 and 2 (VGLUT1 and VGLUT2, respectively) in neurons innervating the mouse colorectum. Lumbosacral and thoracolumbar dorsal root ganglion (DRG), lumbar sympathetic chain (LSC), and major pelvic ganglion (MPG) neurons innervating the colorectum of BALB/C mice were retrogradely traced with Fast Blue, dissected, and processed for immunohistochemistry. Tissue from additional naïve mice was included. Previously characterized antibodies against VGLUT1, VGLUT2, and calcitonin gene-related peptide (CGRP) were used. Riboprobe in situ hybridization, using probes against VGLUT1 and VGLUT2, was also performed. Most colorectal DRG neurons expressed VGLUT2 and often colocalized with CGRP. A smaller percentage of neurons expressed VGLUT1. VGLUT2-immunoreactive (IR) neurons in the MPG were rare. Abundant VGLUT2-IR nerves were detected in all layers of the colorectum; VGLUT1-IR nerves were sparse. A subpopulation of myenteric plexus neurons expressed VGLUT2 protein and mRNA, but VGLUT1 mRNA was undetectable. In conclusion, we show 1) that most colorectal DRG neurons express VGLUT2, and to a lesser extent, VGLUT1; 2) abundance of VGLUT2-IR fibers innervating colorectum; and 3) a subpopulation of myenteric plexus neurons expressing VGLUT2. Altogether, our data suggests a role for VGLUT2 in colorectal glutamatergic neurotransmission, potentially influencing colorectal sensitivity and motility.
Glutamate, the principal excitatory neurotransmitter in neurons throughout the nervous system (Fonnum, 1984; Hökfelt et al., 1984; Watkins, 2000), is incorporated into synaptic vesicles by vesicular glutamate transporters (VGLUTs) (Liguz-Lecznar and Skangiel-Kramska, 2007). To date, three subtypes (VGLUT1, VGLUT2, and VGLUT3) have been identified (Ni et al., 1994; Bellocchio et al., 2000; Aihara et al., 2000; Takamori et al., 2001; Fremeau et al., 2002).
Initially, glutamatergic neurons were morphologically studied under the optical or electron microscope by immunohistochemical detection of the neurotransmitter itself (Storm-Mathisen et al., 1983; Ottersen, 1987; De Biasi and Rustioni, 1988). However, the presence of glutamate in a neuron does not directly imply its use as a neurotransmitter since the amino acid also serves as a precursor for the synthesis of other molecules, including GABA (Broman et al., 2000). Alternatively, detection of indirect markers such as glutaminase, the enzyme synthesizing glutamate (Kaneko et al., 1990), or glutamate transporters located in neuronal and glial plasma membranes (see Kanner and Schuldiner, 1987; Danbolt, 2001) has also been used. However, because of the high specificity of VGLUTs to a glutamatergic phenotype, the use of selective antibodies or in situ hybridization probes for their detection has become the gold standard for the identification of glutamatergic neurons (Fremeau et al., 2004; Moriyama and Yamamoto, 2004) and glutamate in glia (Montana et al., 2004; Anlauf and Derouiche, 2005).
The expression of VGLUT1 and VGLUT2 transcripts and proteins in dorsal root ganglion (DRG) sensory neurons innervating the skin has been studied in detail in the rat (Oliveira et al., 2003; Hwang et al., 2004; Landry et al., 2004; Morris et al., 2005), mouse, and guinea pig (Morris et al., 2005; Brumovsky et al., 2007). More recently, the presence of chimeric enhanced green fluorescent protein (EGFP) under control of the VGLUT3 promoter was shown in a subpopulation of L4-5 DRG neurons (Seal et al., 2009). In the rodent spinal cord all three transporters are present, although with different patterns of distribution, including dorsal vs. ventral or lateral horns, neuropil vs. cell body expression (Oliveira et al., 2003; Todd et al., 2003; Li et al., 2003a, b; Landry et al., 2004; Alvarez et al., 2004; Morris et al., 2005; Brumovsky et al., 2007; Seal et al., 2009). In contrast to nonvisceral neurons, the expression of VGLUTs in neurons innervating visceral organs has been studied mostly in relation to the upper gut.
The gut is a visceral structure subject to control by extrinsic and intrinsic innervations (see Robinson and Gebhart, 2008). The extrinsic innervation includes sensory nerves contributed by nodose and DRG neurons (Blackshaw et al., 2007) and autonomic (sympathetic and parasympathetic) nerves derived from neurons located in the lumbar sympathetic chain and prevertebral ganglia (Furness, 2006; Burnstock, 2009a). The presence of VGLUT2 has been shown in nodose ganglia and DRG neurons innervating the stomach or ileum of rat (Tong et al., 2001; Corbett et al., 2005), as well as in specialized axonal terminations of vagal afferent fibers in esophageal myenteric ganglia (Berthoud et al., 1997) called intraganglionic laminar endings (IGLEs) in the rat, mouse (Raab and Neuhuber, 2003, 2004, 2005), and guinea pig (Zagorodnyuk et al., 2003). In addition, esophageal IGLEs in the guinea pig (Zagorodnyuk et al., 2003) and rat (Ewald et al., 2006), but not the mouse (Kraus et al., 2007), express VGLUT1. A few studies in the lower gut show that VGLUTs are also present in axonal terminations in the guinea pig (Olsson et al., 2004) and mouse (Spencer et al., 2008) rectum.
Sensory and motor neurons located in the myenteric and submucosal plexuses, and collectively referred to as enteric neurons, form the intrinsic innervation of the gut (Furness et al., 2004; Burnstock, 2009a). Enteric neurons containing glutamate have been described in rat and guinea pig ileum (Liu et al., 1997) as well as in rat stomach (Tsai et al., 1994) and VGLUT1 and/or VGLUT2 was reported present in enteric neurons in guinea pig, rat, mouse (Tong et al., 2001) and human ileum (Linke et al., 2008) and rat and mouse esophagus (Ewald et al., 2006; Kraus et al., 2007). However, in guinea pig colon evidence suggests that VGLUT1 and VGLUT2 are not expressed by myenteric plexus neurons (Olsson et al., 2004).
Despite extensive study of VGLUTs, a comprehensive account of their expression in rodent visceral sensory neurons innervating the distal gut has not been reported. Therefore, in the present study we analyzed the expression of VGLUT1 and VGLUT2 in mouse DRG, lumbar sympathetic chain, and major pelvic ganglion neurons retrogradely labeled from the colorectum using Fast Blue. Identification of VGLUT1- and VGLUT2-expressing colorectal neurons was made by immunohistochemistry using two well-characterized antibodies (Kawamura et al., 2006; Brumovsky et al., 2007). In addition, we explored the potential glutamatergic phenotype of colorectal myenteric plexus neurons by immunohistochemistry and riboprobe in situ hybridization. Finally, we studied coexpression of these transporters with CGRP in DRG and the colorectum. Portions of these data have been reported in abstract form (Gebhart et al., 2008).
Male BALB/c (Taconic, Germantown, NJ; 7–8 weeks old) were used in all experiments. All research protocols adhered to the United States Public Health Service policies regarding the care and use of animals in research and were reviewed and approved by the Institutional Animal Care Use Committee of the University of Pittsburgh.
The colon of six BALB/c male mice, 6 weeks old, was injected with the fluorescent retrograde neuronal tracer Fast Blue (FB, 2% in saline; EMS-Chemie, Gross Umstadt, Germany) (Robinson et al., 2004). Under aseptic conditions, animals were anesthetized with isoflurane (Hospira, Lake Forest, IL) and injected with buprenorphine (0.1 mg/kg; Bedford Labs, Bedford, OH) for postoperative analgesia. Following laparotomy, three to five injections (≈5 μl, using a Hamilton syringe with a 25G needle) of FB were made at different sites into the wall of the descending colon. Care was taken not to allow the dye to spread to areas other than the injection site, both by containing any spillage using a cotton tip and by rinsing the peritoneal cavity with sterile saline before suturing muscle and skin. After treating the skin incision with dibucaine ointment (1%; Perrigo, Allegan, MI), mice were allowed to recover in a warm environment under close observation.
Twelve days after injection of FB, and including seven naïve BALB/C mice, the animals were deeply anesthetized using sodium pentobarbital (60 mg/kg, intraperitoneally [i.p.]; Ovation Pharmaceuticals, Deerfield, IL) and perfused via the ascending aorta with 20 ml of Tyrode’s buffer (37°C), followed by 20 ml of a mixture of 4% paraformaldehyde and 0.2% picric acid dissolved in 0.16 M phosphate buffer (pH 6.9) (37°C) and 50 ml of the same mixture at 4°C, the latter for ≈5–7 minutes. Thoracolumbar (T8-L1; TL) and lumbosacral (L6-S2; LS) DRGs, the major pelvic ganglion (MPG), the lumbar sympathetic chain (LSC), and the colorectum were dissected out and postfixed for 90 minutes at 4°C in the same fixative and immersed in 10–20% sucrose in phosphate-buffered saline (PBS) (pH 7.4) containing 0.01% sodium azide and 0.02% bacitracin (both from Sigma, St. Louis, MO) (4°C) for 48 hours. After embedding in Tissue-Tek O.C.T. compound (Sakura, Torrance, CA) and deep freezing on dry ice, tissue was sectioned in a cryostat (Leica, Heidelberg, Germany) at 12-μm (DRG, LSC) or 20-μm thicknesses (MPG, colon).
Table 1 lists the antibodies used in this study. The VGLUT1 and VGLUT2 antibodies were raised against the C-terminal sequence of rat VGLUT1 (531–560 amino acid residues, GenBank accession number U07609; Miyazaki et al., 2003) or rat VGLUT2 (519–582 amino acid residues, GenBank accession number AF271235; Miyazaki et al., 2003). Antibodies against C-terminus polypeptide-glutathione-S transferase (GST) fusion proteins were raised and affinity purified by affinity media coupled with GST-free polypeptides (Miyazaki et al., 2003; Kawamura et al., 2006; Miura et al., 2006). Antibody specificity was tested by immunoblot of mouse brain extracts, with detection of a single band at 60 kDa for both VGLUT1 and VGLUT2 (Miyazaki et al., 2003; Kawamura et al., 2006; Miura et al., 2006), and also by immunostaining patterns in the adult mouse brain, which were identical to previous reports by immunohistochemistry and in situ hybridization. Specifically for DRGs, the VGLUT1 and VGLUT2 antibodies produced immunohistochemical patterns identical to those previously described in an earlier publication (Brumovsky et al., 2007). In addition, combined in situ hybridization–immunohistochemistry studies show colocalization of transcript and protein of these two VGLUTs in the same DRG neurons (see Results).
Finally, the rabbit polyclonal antibody raised against the rat CGRP (Sigma) is a delipidized whole antiserum that recognizes CGRP in both rats and mice (due to the conserved sequence of CGRP between the two rodents). In dotblot experiments there is negligible cross-immunoreactivity with calcitonin, somatostatin, and amylin (Zhang et al., 2001). Moreover, the CGRP immune signal is completely abolished in DRGs and the dorsal horn of the spinal cord of mice lacking CGRP (Zhang et al., 2001).
Sections were washed twice in PBS and incubated for 24 hours at 4°C with rabbit anti-VGLUT1 (1:4,000; #VGluT1-Rb-Af500; Frontier Science, Hokkaido, Japan) (Kawamura et al., 2006) or guinea pig anti-VGLUT2 (1:8,000; #VGluT2-GP-Af670-1; Frontier Science) (Miyazaki et al., 2003; Brumovsky et al., 2007) antibodies, diluted in 0.01 M PBS containing 0.3% Triton X-100 and 0.5% bovine serum albumin (BSA).
To visualize immunoreactivity, sections were processed using a commercial kit based on tyramide signal amplification (Adams, 1992) (TSA Plus, NEN Life Science Products, Boston, MA). Briefly, the sections were washed in TNT buffer (0.1 M Tris-HCl, pH 7.5; 0.15 M NaCl; 0.05% Tween 20) for 10 minutes, incubated with TNB buffer kit (0.1 M Tris-HCl, pH 7.5; 0.15 M NaCl; 0.5% Dupont Blocking Reagent; NEN) for 30 minutes at room temperature (RT) and incubated for 60 minutes with either a donkey antirabbit or antiguinea pig/horseradish peroxidase (HRP) conjugate (Jackson ImmunoResearch, West Grove, PA) diluted 1:200 in TNB buffer. The sections were washed twice in TNT buffer and incubated in a biotinyl tyramide-fluoride thiocyanate (BT-FITC) conjugate kit (NEN) diluted 1:700 in amplification diluent kit (NEN) for 30 minutes at RT.
After incubation with VGLUT1 or VGLUT2, antiserum, and followed by the TSA plus technique for visualization of each primary antibody (as above), some sections were washed in PBS, incubated with a rabbit anti-CGRP antibody (1:8,000; #C8198; Sigma), and processed according to Coons (1958). Sections were incubated for 24 hours at 4°C with the anti-CGRP antiserum, washed twice in PBS, and further incubated using a tetramethyl rhodamine isothiocyanate (TRITC)-conjugated donkey antirabbit antibody (1:400; Jackson ImmunoResearch).
To test for nonspecific staining by the secondary antibody, the primary antibody was omitted from a few control sections. Also, single-stained sections were processed for comparison with double-stained sections. The rabbit VGLUT1 antiserum was studied after preadsorption with the C-terminal 531–560 amino acid peptide (10−5 and 10−6M), which gave no visible fluorescent signal (data not shown). CGRP and VGLUT2 antisera have been thoroughly characterized in previous publications (Robinson et al., 2004; Brumovsky et al., 2007). Specificity of the VGLUT antibodies was further assessed through combined in situ hybridization–immunohistochemistry studies (see below).
Four male BALB/c mice, 7–8 weeks old, were transcardially perfused with fixative as above and their colorectum removed. Immediately after, and under a dissecting microscope, the colorectum was opened along the longitudinal axis, pinned flat, and denuded of the mucosal, submucosal, and circular muscle layers using fine dissection forceps. The remaining tissue (longitudinal muscle layer and myenteric plexus) was incubated in the same fixative for 90 minutes and transferred into wells containing PBS, after which free-floating immunohistochemistry was initiated.
Whole-mount preparations were placed in 24-well cell culture chambers (Corning, Corning, NY), rinsed three times in PBS (15 minutes each), and incubated for 24 or 48 hours at 4°C with rabbit anti-VGLUT1 (1:4,000) or guinea pig anti-VGLUT2 (1:8,000) antiserum diluted in 0.01 M PBS containing 0.3% Triton X-100 and 0.5% BSA. For all solutions 300–500 μl per well was used. To visualize immunoreactivity, tissue was processed using a modified TSA protocol. Briefly, the whole mounts were washed twice in TNT buffer for 15 minutes, incubated with TNB buffer for 1 hour at RT, and then incubated for 24 hours at 4°C with a donkey antirabbit or antiguinea pig/HRP conjugate (Jackson ImmunoResearch) diluted 1:400 in TNB buffer. The preparations were washed twice in TNT buffer (15 minutes each) and incubated in a BT-FITC conjugate diluted 1:1,000 in amplification diluent for 2 hours at RT, then washed twice in TNT (15 minutes each) and twice in PBS. Some whole mounts were additionally incubated with propidium iodide (PI; Sigma) diluted at 0.0001% in TNT, followed by three washes in TNT buffer.
Male, 7-week-old, BALB/c mice (n = 4) were used for in situ hybridization analysis. Mice were briefly sedated using CO2 before rapid decapitation. The colorectum, brainstem, and L4-5 DRGs were quickly removed, embedded in O.C.T. (Tissue-Tek), and frozen over dry ice. Sections of colorectum and brainstem (20 μm) were cut in a cryostat, thaw-mounted onto Superfrost Plus (Fisher Scientific, Waltham, MA) glass slides, and stored at −20°C until hybridization. Adjacent sections through the different tissues were processed for in situ hybridization localization of VGLUT1 and VGLUT2 mRNAs using 35S-labeled cRNA probes as described previously (Seroogy and Herman, 1997; Numan et al., 2005; Dickerson et al., 2009). Briefly, the slide-mounted sections were brought to RT and then placed in 4% paraformaldehyde for 10 minutes. This was followed by washes in 0.1 M phosphate buffer (PB), 0.1 M PB/0.2% glycine, and 0.25% acetic anhydride in 0.1 M triethanolamine. The sections were then dehydrated with increasing concentrations of ethanol, delipidated in chloroform, and air-dried. Sections were hybridized for 18–24 hours at 60°C in hybridization cocktail (0.15 mg/ml yeast tRNA, 10% dextran sulfate, 50% formamide, 1× Denhardt’s solution, 1 mM EDTA, 20 mM Tris-HCl, 40 mM dithiothreitol, 0.33 mg/ml denatured salmon sperm DNA) and the 35S-labeled cRNA probe at a concentration of 1.0 × 106 cpm/50 μl per slide.
Sense and antisense probes complementary to the coding region of mouse VGLUT1 (nucleotides 855–1788; GenBank accession number XM_133432.2) and VGLUT2 (nucleotides 848–2044; GenBank accession number NM_080853.2) were generously provided by Dr. Akiya Watakabe, National Institute for Basic Biology, Okazaki, Japan (Nakamura et al., 2007). The sense and antisense cRNA probes were prepared by in vitro transcription using appropriate linearized DNA constructs in the presence of the corresponding RNA polymerase (T3 or T7) and 35S-UTP (New England Nuclear, Perkin Elmer, Waltham, MA). For posthybridization treatment, sections were washed several times in 4× SSC (1× SSC = 0.015 M sodium citrate, 0.15 M sodium chloride at pH 7.0) containing 10 mM sodium thiosulfate at 37°C. The sections were then incubated in ribonuclease A (0.05 mg/mL) for 30 minutes at 45°C. This was followed by several washes in decreasing concentrations of SSC (2×, 0.5×, and 0.1×) at 37°C. All but the final wash also contained 10 mM sodium thio-sulfate. The sections were briefly rinsed in distilled water (dH2O), dipped in 95% ethanol, and finally air-dried.
Sections were dipped in NTB2 nuclear track emulsion (Kodak, Rochester, NY; 1:1 in dH2O), air-dried, and exposed in sealed slide boxes at 4°C for 3–7 days and 6–12 weeks, for DRGs and colon, respectively. The emulsion was developed in D19 (Kodak) and fixed with Rapidfix (Kodak). The slides were counterstained with cresyl violet (Sigma) and coverslipped with DPX mounting solution (Fluka, Buchs, Switzerland). As controls for specificity, some sections were pretreated with ribonuclease A (0.05 mg/ml) for 30 minutes at 45°C before hybridization with the 35S-labeled cRNA probes. Some sections were also hybridized with sense-strand 35S-labeled riboprobes against each VGLUT. Finally, several tissue sections were incubated in hybridization cocktail that lacked the radioactive probe as a chemography control. No specific labeling was observed under any of these conditions.
In order to establish if protein and transcript for VGLUT1 and VGLUT2 do coexist in the same DRG neurons, and to define the reliability of the antibodies and riboprobes utilized in the present study, we carried out combined ribop-robe in situ hybridization–immunohistochemistry experiments (Seroogy et al., 1994; Seroogy and Herman, 1997). Briefly, naïve male 7-week-old BALB/c mice (n = 2) were deeply anesthetized, perfused, and the L4-5 DRGs quickly removed and treated as described above for immunohistochemistry. Tissue sections were cut in a cryostat, thaw-mounted onto Superfrost Plus glass slides, and processed for in situ hybridization localization of VGLUT1 and VGLUT2 mRNAs using the appropriate riboprobes (see above). Following hybridization and posthybridization rinses in descending concentrations of SSC buffer, slides were rinsed with sterile PB and incubated overnight at 4°C with VGLUT1 or VGLUT2 antibodies (see Table 1) at a concentration of 1:3,000 and 1:6,000, respectively. This was followed by immunohistochemical processing using the avidin-biotin technique (Vectastain kit; Vector Laboratories, Burlingame, CA) and with diaminobenzidine as the chromagen. Slides were then rinsed in 50 mM Tris, dehydrated in a series of ethanol at increasing concentrations, finishing in Xylene, and air-dried. After 24 hours slides were dipped in Kodak NTB autoradiography emulsion as described above.
All sections and whole-mount preparations were cover-slipped using 2.5% DABCO in glycerol (Sigma) and examined with a Nikon Eclipse E600 fluorescence microscope (Nikon, Tokyo, Japan) with appropriate filters and a Retiga 2000 R Fast CCD camera (Q-Imaging, Surrey, British Columbia, Canada) using IPLab software (Scanalytics, Vancouver, Canada). For colocalization analysis of the colon, MPG, and the LSC, a Fluoview FV 1000 confocal laser scanning biological microscope equipped with 10× (0.45 N.A.), 20× (0.75 N.A.), and 60× oil (1.40 N.A.) objectives was used (Olympus, Tokyo, Japan). The FITC labeling was excited using a 547–514 nm argon multiline laser. For the detection of TRITC and TMR, a 543 nm HeNe laser was used.
Resolution, brightness, and contrast of the images were optimized using the Adobe Photoshop CS3 software (Adobe Systems, San Jose, CA). Because confocal imaging of FB was not possible due to lack of appropriate filters, images were composed by merging separate optical (FB) and confocal (other markers of interest) photomicrographs.
Retrogradely traced colorectal neurons were quantified in T8-T13 and L6-S2 DRGs. Every fifth DRG section was used to quantify the number of FB-positive (colonic) neuron profiles (NPs), as well as the number of colonic VGLUT1-, and VGLUT2-immunoreactive (IR) NPs present in each section. In total, 5–8 sections per ganglion were used for quantification. The cell body diameters of a representative sample of each type of quantified neurons (more than 20 and up to 69 neurons per type) were measured using the public domain NIH program ImageJ (developed at the U.S. National Institutes of Health and available on the Internet at http://rsb.info.nih.gov/ij/download.html), and these data used to correct the raw counts, based on the Abercrombie correction factor (Abercrombie, 1946; Guillery, 2002). Percentages of VGLUT-expressing colorectal neurons were obtained by counting, within the total number of FB-positive colorectal neurons, those expressing or not VGLUT1 or VGLUT2.
Due to the very low background obtained with the TSA plus technique, we were able to detect a variety of immunofluorescent signal intensities for both VGLUT1 and VGLUT2, ranging from intensely to lightly stained NPs. Moreover, negative neurons were seen as virtually completely black, and thus most neurons showing VGLUT-signal were easily identified. However, to avoid occasional false-positives we counted those NPs exhibiting an intensity of immunostaining higher than two orders of magnitude, and when intracellular structures such as the Golgi apparatus or the endoplasmic reticulum could be observed. In the case of the combined riboprobe in situ hybridization–immunohistochemistry, it should be noted that the immunohistochemical VGLUT1 and VGLUT2 signals obtained when using the avidinbiotin/diaminobenzidine technique produced a background staining that appeared to also be influenced by the antibody used. In the present study we discriminated between VGLUT1, VGLUT2, and background by only including those NPs that showed a signal that was at least three orders of magnitude above the negative signal obtained from, for example, the nerve bundles in the DRGs (not shown). Riboprobe in situ hybridization signals were analyzed in such a way that only neurons where the density of accumulation of silver grains over a single NP was more than 5 times that of background were counted as labeled.
Data are presented as mean ± SEM and were statistically analyzed using Student’s t-test for independent samples; P < 0.05 was considered significant.
In the present study we studied the expression of VGLUT1 and VGLUT2 in colorectal DRG neurons and neurons in the myenteric plexus, following two different approaches: 1) by immunohistochemical detection of VGLUT1 and VGLUT2 using two previously characterized antibodies (Table 1; DRG and myenteric neurons); and 2) by evaluating the coexpression of VGLUTs transcript and protein by means of combined riboprobe in situ hybridization–immunohistochemistry, utilizing previously characterized VGLUT riboprobes (Nakamura et al., 2007).
Both the VGLUT1 and VGLUT2 antibodies produced distinct immunohistochemical signals (Fig. 1A,B, respectively) that also matched previously described patterns of immunoreactivity (see below and Morris et al., 2005; Brumovsky et al., 2007). However, to certify the specificity of the probes we carried out combined riboprobe in situ hybridization–immunohistochemistry studies and found that both VGLUT1 (Fig. 1C) or VGLUT2 (Fig. 1D) transcript and protein in fact coexpress, as shown by the colocalization of their respective in situ hybridization–immunohistochemical signals.
The rodent colorectum is innervated by lumbar splanchnic (LSN) and pelvic (PN) nerves with cell bodies located in T8-L1 and L6-S2 DRGs, respectively (Robinson and Gebhart, 2008). In the mouse, colorectal DRG neurons can be divided into small (≈300 μm2), medium (350–600 μm2), and large (650–>1,100 μm2) (Robinson et al., 2004; Tan et al., 2008). Retrograde tracing from the colorectum revealed a discrete number of FB-positive NPs of different sizes per section in these DRGs that were easily differentiated from FB-negative NPs (Figs. 2, ,44).
Both VGLUT1 and VGLUT2 were expressed in mouse DRG neurons projecting to the colorectum (Table 2; Fig. 2). VGLUT1 was present in 15 ± 1.8% of TL and 8 ± 1.1% of LS colorectal DRG NPs (Table 2; Fig. 2A–C), the majority of which appeared to be medium- to large-sized (Fig. 2A–C). VGLUT2, by far the most abundant VGLUT in DRG neurons, was detected in similar high proportions in TL (98 ± 1.2%) and LS (97 ± 1.4%) NPs of different sizes (Table 2; Fig. 2D–F). Colorectal DRG neurons lacking VGLUT2 were also found (Fig. 2G–I). In contrast, VGLUT1 was present in a significantly greater percentage of TL colorectal DRG NPs compared with its LS counterpart (Table 2; P = 0.014). Finally, in addition to colorectal neurons, abundant VGLUT2- and a considerable number of VGLUT1-IR, FB-negative neurons were detected in all DRGs (Figs. 2, ,44).
Colocalization analysis showed that 79.6 ± 2.1% of TL and 83.6 ± 2.3% of LS colorectal DRG NPs coexpressed VGLUT2 and CGRP (Figs. 3, 4A–D). Many FB-negative VGLUT2-IR NPs coexpressing CGRP were also detected (Fig. 4A–D). In contrast to VGLUT2, virtually no VGLUT1 colorectal DRG NPs coexpressed CGRP in any of the studied DRGs (Fig. 4E–H). Moreover, FB-negative VGLUT1-IR NPs also lacked CGRP (Fig. 4E–H). CGRP-only colorectal DRG neurons were also seen (Fig. 4E–H).
In the lumbar sympathetic chain (LSC), several neurons were retrogradely labeled from the colorectum by FB (Fig. 5B, E). None of these neurons showed the presence of VGLUT1 or VGLUT2 (Fig. 5A–C, 5D–F, respectively). However, a number of VGLUT2-IR fibers, as well as what seemed to be synaptic varicosities (Fig. 5D–F), were detected in LSCs.
In the MPG, many neurons were retrogradely labeled from the colorectum by FB (Fig. 5H, K, N, Q). None of these neurons exhibited VGLUT1-like immunoreactivity (LI) (Fig. 5G–I), and the latter was only detected in a few varicose profiles (Fig. 5G). VGLUT2-IR colorectal MPG neurons were rarely found (Fig. 5J–L). As observed in the LSC, many VGLUT2-IR fibers and varicosities were detected in the MPG (Fig. 5J, L, M, O, P, R). Moreover, some of these varicosities appeared to be arranged in baskets surrounding colorectal (Fig. 5J–L, 5M–O) as well as noncolorectal MPG (Fig. 5P–R) neurons.
VGLUT1-LI was restricted to a few nerve fibers found in the submucosal (Fig. 6A, B) and mucosal layers (Fig. 6C) of the colorectum. In contrast, a profuse VGLUT2-IR innervation was detected throughout (Fig. 6D), including the myenteric plexus (Figs. 6D, E) and submucosal and mucosal layers (Fig. 6D, F), even reaching the luminal aspects of the latter (Fig. 6F). Additional, nonspecific immunofluorescent cells were observed, usually in the mucosal layers (Fig. 6D, F). These cells were also seen after omission of the primary antibody (data not shown), strongly suggesting nonspecific binding of the secondary antibody.
VGLUT1-LI was not detected (24-hour, free-floating immunohistochemistry) in whole-mount preparations of the longitudinal muscle layer and myenteric plexus neurons and their projections (Fig. 7A). In contrast, a profuse VGLUT2-IR fiber network (Fig. 7B) was detected, often connecting different myenteric plexus ganglia (Fig. 7B) and forming numerous varicosities surrounding unstained myenteric plexus neurons (Fig. 7C). Prolonged free-floating immunohistochemistry of whole-mount preparations (48-hour) revealed the presence of a small population of VGLUT2-IR myenteric plexus neurons (Fig. 7D–F). These neurons were sparsely distributed among the many myenteric ganglia present in the whole-mount preparation, and neighboring other unstained myenteric plexus neurons (Fig. 7D–F).
In accordance with immunohistochemical results, in situ hybridization for detection of VGLUT mRNAs in the colorectum revealed the presence of VGLUT2 mRNA in a subpopulation of myenteric plexus neurons (Fig. 8A–D). VGLUT2 mRNA-positive neurons were present only at this location and absent in the submucosal plexus or elsewhere (Fig. 8A). In situ hybridization of sections tangential to the surface of the colorectum confirmed these observations. Thus, only a fraction of all neurons throughout the whole myenteric plexus were VGLUT2 mRNA-positive (Fig. 8B–D). In fact, not all myenteric ganglia contained VGLUT2 mRNA-positive neurons, and often only two or three neurons per ganglia expressed this VGLUT (Fig. 8C, D). Several VGLUT2 mRNA-positive NPs were detected in L4-5 DRGs (Fig. 8E) using the antisense riboprobe, and the signal was absent with the sense riboprobe in both DRGs (Fig. 8F) and colorectum (Fig. 8G). The lack of signal seen with the sense probe certifies the specificity of the in situ hybridization signal. In contrast, no VGLUT1-mRNA expression (Fig. 8H–K) was detected in any of the colorectal layers. The same antisense riboprobes exposed a population of DRG neurons expressing VGLUT1 (Fig. 8L); this signal was absent when incubating sections with the corresponding sense probe (Fig. 8M).
Colocalization analysis in the colorectum revealed that VGLUT2 highly coexpresses with CGRP in fibers present in the mucosal layer of the organ (Fig. 9A–C). In contrast, at the level of the myenteric plexus, VGLUT2 and CGRP appeared to be present in independent fiber populations (Fig. 9D–I).
In this study we characterized the glutamatergic innervation of the mouse colorectum using immunohistochemistry and in situ hybridization for the detection of VGLUT1 and VGLUT2. Our results reveal that VGLUT2 is the principal VGLUT in both the extrinsic and intrinsic innervations of the mouse colorectum, with a much lower percentage of colorectal DRG neurons expressing VGLUT1. Therefore, most extrinsic colorectal neurons in TL and LS DRGs appear to be glutamatergic. Significantly, the majority of these VGLUT2-expressing neurons are also peptidergic, suggesting the possibility of neurotransmitter interactions upon release from synaptic terminals. VGLUT2-expressing nerve fibers are present in all layers of the colorectal wall, including the myenteric plexus and the mucosa. Further analysis of the myenteric plexus revealed a discrete population of enteric neurons expressing VGLUT2 transcript and protein, suggesting local synthesis. In contrast, no VGLUT1-positive neurons were detected in the myenteric plexus. The presence of rare MPG neurons expressing VGLUT2 suggests an additional autonomic source of glutamatergic innervation of the colorectum.
Previous studies support the presence of VGLUTs in afferent sensory nerve terminals in the gut of the guinea pig, rat, and mouse (Raab and Neuhuber, 2003, 2004, 2005; Zagorodnyuk et al., 2003; Olsson et al., 2004; Ewald et al., 2006; Spencer et al., 2008). However, the proportion and phenotype of VGLUT-expressing vagal or DRG neurons giving rise to these endings has received limited attention. In the rat, Tong et al. (2001) and Corbett et al. (2005) reported expression of VGLUT2 (not quantified) in ganglia innervating the stomach (Tong et al., 2001; Corbett et al., 2005) or ileum (Tong et al., 2001). In addition, expression of VGLUT2 and VGLUT3 protein in rat nodose ganglion neurons supplying the aortic depressor nerve has also been described (Lin and Talman, 2006).
In the present study we show that more than 95% of mouse colorectal DRG neurons express VGLUT2. These neurons varied in size, in agreement with previous studies showing that colonic neurons in the mouse (Robinson et al., 2004; Sugiura et al., 2007; Robinson and Gebhart, 2008) and rat (Keast and De Groat, 1992; Traub et al., 1999) are of small-to-medium size, including some with larger profiles. The abundant expression of VGLUT2, which not only encompassed colorectal, but also nontraced DRG neurons, correlates well with our earlier study in mouse ((Brumovsky et al., 2007), where we observed that up to 65% of the total number of L4-5 (nonvisceral) DRG neurons express VGLUT2 (see Brumovsky et al., 2007). However, even though the hind-paw skin was the most likely destination, we did not trace this or other potential target tissues for the VGLUT2-expressing DRG neurons (Brumovsky et al., 2007). That most if not all colorectal neurons in the mouse express VGLUT2, compared to the 65% expression reported in nonvisceral neurons (Brumovsky et al., 2007), suggests differences in the glutamatergic nature of neurons at different DRG levels, as well as target organs. Accordingly, Tong et al. (2001) reported that the majority of rat thoracic DRG neurons express VGLUT2, and that a subpopulation of these neurons innervates the ileum. In contrast to the abundance of VGLUT2, VGLUT1 was detected in less than 20% of colorectal DRGs. These observations are in agreement with the more discrete expression of VGLUT1 in nonvisceral mouse DRG neurons (Morris et al., 2005; Brumovsky et al., 2007).
In the rat (Christianson et al., 2006; Robinson and Gebhart, 2008) and mouse (Robinson et al., 2004; Christianson et al., 2006; Robinson and Gebhart, 2008), the afferent innervation of the colorectum is provided by TL and LS DRG neurons. Interestingly, differences in the degree of expression of some neurochemical markers exist between mouse colorectal TL and LS DRG neurons (Brierley et al., 2005a; Christianson et al., 2006; Robinson and Gebhart, 2008). In support, we found that whereas the proportions of VGLUT2-expressing colorectal TL and LS DRG neurons are comparable, VGLUT1 is generally lower at LS levels. Even though further research is required to understand the significance of these differences, they may relate to recently reported electrophysiological differences in the properties of TL and LS colorectal afferent neurons (Brierley et al., 2004, 2005a, b).
Through coexpression analysis with CGRP, we found that around 80% of all VGLUT2-IR colorectal DRG neurons are peptidergic, in clear contrast to the virtual absence of the peptide in colorectal neurons expressing VGLUT1. It has long been appreciated that CGRP is contained in rat visceral sensory neurons (Keast and De Groat, 1992; Bennett et al., 1996; Wang et al., 1998), findings that have been confirmed by recent work in rat (Christianson et al., 2006) and mouse (Robinson et al., 2004; Christianson et al., 2006). Overall, most studies report that 50–80% of all colorectal DRG neurons express CGRP (Robinson and Gebhart, 2008), supporting our past (Brumovsky et al., 2007) and present observations of the peptidergic nature of VGLUT2-expressing neurons. Nonpeptidergic colorectal VGLUT2-expressing neurons (≈20%), as well as the few VGLUT1-expressing neurons, could belong to the population of DRG neurons (see Bennett et al., 1996; McMahon and Priestley, 2005). However, at least in nonvisceral mouse DRG neurons, VGLUT1 and IB4 are never coexpressed (Morris et al., 2005; Brumovsky et al., 2007). Further study is necessary to elucidate the full phenotype of these neurons.
Initial observations in the adult mammalian brain established that VGLUTs exhibit a complementary distribution (Fremeau et al., 2001; Schäfer et al., 2002; Kaneko and Fujiyama, 2002; Boulland et al., 2004). However, recent studies suggest that some neurons in the central (Boulland et al., 2004) and peripheral (Li et al., 2003; Landry et al., 2004; Brumovsky et al., 2007) nervous systems express more than one type of VGLUT. In the present work the overwhelming presence of VGLUT2 in virtually all colorectal DRG neurons suggests that some may also express VGLUT1. We previously described a moderate VGLUT1 and VGLUT2 coexpression in nonvisceral DRG neurons (Brumovsky et al., 2007). Moreover, colocalization between these two VGLUTs has been shown in rat (Ewald et al., 2006) and mouse (Kraus et al., 2007), but not in guinea pig (Zagorodnyuk et al., 2003) esophageal IGLEs. Also in nonvisceral systems, Todd et al. (2003), Alvarez et al. (2004), and Persson et al. (2006) demonstrated frequent colocalization of VGLUT2 and VGLUT1 in primary afferent varicosities extending into the deep dorsal horn. Interestingly, afferent fibers terminating deeply in the dorsal horn originate from different types of low-threshold mechanoreceptors, and these may include visceral mechanoreceptors (Willis and Coggeshal, 2004). Thus, it could be speculated that medium- to large-sized colorectal DRG neurons expressing VGLUT1 may subserve a mechanoreceptive function in the colorectum.
Colorectal DRG neurons also express a number of neurochemical markers associated with nociception. For example, the transient receptor potential cation channel, subfamily V, member 1 (TRPV1), a nonselective cation channel activated by pH, heat, and capsaicin (Caterina et al., 1997; Tominaga et al., 1998), is reportedly expressed in a large proportion of colorectal DRG neurons in both rat (Christianson et al., 2006; Chaban, 2008) and mouse (Robinson et al., 2004; Christianson et al., 2006). P2X purinoceptor 3 (P2X3), a purinoceptor that has also been implicated in nociception (Wirkner et al., 2007), particularly in visceral organs (Shinoda et al., 2009; Burnstock, 2009b), is also expressed by a proportion of colorectal DRG neurons (Robinson et al., 2004). Because the great majority of extrinsic colorectal neurons described here exhibited VGLUT2-LI, it is likely that they also coexpress TRPV1 and/or P2X3 (in addition to other receptors or channels not studied here), suggesting a relevant role for this VGLUT in the physiopathology of visceral disorders associated with discomfort and pain. Importantly, coexpression of VGLUT2 with TRPV1 has been previously reported in nerve fibers terminating in the mouse rectum (Spencer et al., 2008).
We could not assess the presence of VGLUT3 protein in DRG and autonomic neurons, and their colorectal projections. Commercially available antibodies regularly used to study VGLUT3-expressing nerve fibers in the central nervous system are largely ineffective in the study of DRG somata and their axonal projections (however, we are currently analyzing colorectal neurons in transgenic mice that express eGFP under control of VGLUT3 regulatory sequences, and exploring the expression of VGLUT3 mRNA in lumbosacral DRG neurons through riboprobe in situ hybridization).
Here we show that VGLUT2-LI is observed in the form of abundant varicosities around unstained colorectal myenteric plexus neurons, in agreement with a similar description in the guinea pig small intestine for glutamate-LI (Liu et al., 1997). In contrast, the expression of VGLUT1 in nerve fibers was very discrete. In support, Olsson et al. (2004) showed that ≈3% of anterogradely traced guinea pig rectal nerve varicosities terminating in the myenteric plexus contain VGLUT1, whereas ≈11% exhibit VGLUT2-LI, suggesting that VGLUT2 may be the most relevant VGLUT in mouse colorectal physiology. Similarly, unidentified L4-5 DRG neurons, as well as hairy and glabrous skin nerve terminals in the mouse (Brumovsky et al., 2007), express more VGLUT2 than VGLUT1. The LSC and MPG, both of which are important contributors of nerve fibers in the colorectum (Furness, 2006), could be additional sources of VGLUTs. However, analysis of these two ganglia in the present study only revealed the rare occurrence of VGLUT2-IR neurons in the MPG.
Basal as well as stimulated (electrical and chemical) release of glutamate from longitudinal muscle myenteric plexus has been shown in the guinea pig ileum, presumably from enteric neurons (Wiley et al., 1991; Sinský and Donnerer, 1998). The abundance of VGLUT2 in colorectal nerve fibers, most likely originating in VGLUT2-expressing DRG colorectal neurons shown here, suggests also an important contribution by extrinsic sources to the presence and action of glutamate in the gut. In fact, electrical stimulation of the sciatic nerve elicits the release of glutamate in the hindpaw skin of rat (deGroot et al., 2000).
In accordance with the peptidergic nature of the majority of colorectal DRG neurons, we found a high degree of coexpression of CGRP and VGLUT2 in nerve fibers innervating the colorectal mucosa. However, most VGLUT2 and CGRP-IR fibers located in areas of the myenteric plexus appear to belong to different nerve populations. This is intriguing and could imply that nonpeptidergic VGLUT2-IR colorectal DRG neurons selectively innervate the myenteric plexus. Supporting this idea, it has been reported that VGLUT2-containing varicosities in the esophageal myenteric plexus of rat are not CGRP-IR (Raab and Neuhuber, 2003). However, one further source of VGLUT2 could be myenteric plexus neurons themselves (see below).
The presence of glutamate in rodent enteric neurons has been previously described in small intestine myenteric (Liu et al., 1997; Kirchgessner et al., 1997) and sub-mucosal (Liu et al., 1997; Kirchgessner, 2001) plexuses from rat (Liu et al., 1997) and guinea pig (Liu et al., 1997; Kirchgessner et al., 1997), as well as in myenteric ganglia of the rat stomach (Tsai et al., 1994). Many of these neurons were shown to express the plasma membrane glutamate transporter EEAC1 (Rothstein et al., 1994; Liu et al., 1997; Kirchgessner et al., 1997), suggesting glutamate reuptake mechanisms. In humans the presence of glutamate has been reported in the submucosal and myenteric plexus as well as in nerve fibers innervating the circular muscle layer of the large intestine (Giaroni et al., 2003).
Evidence for the expression of VGLUTs in enteric neurons has been somewhat controversial. On the one hand, the expression of VGLUT2 has been suggested in ileal enteric neurons from different species (Tong et al., 2001) as well as in mouse esophagus (Raab and Neuhuber, 2005). VGLUT1 was found in cholinergic and nitrergic neurons in rat (Raab and Neuhuber, 2003, 2004; Ewald et al., 2006) and mouse (Kraus et al., 2007) esophageal myenteric plexus. Moreover, expression of all three VGLUTs was shown in human small intestine myenteric plexus neurons, interganglionic varicose fibers, and perisomatic puncta (suggesting early gestational regulation; Linke et al., 2008). On the other hand, studies in guinea pig (Raab and Neuhuber, 2003, 2004; Zagorodnyuk et al., 2003; Ewald et al., 2006) esophagus and guinea pig rectum (Olsson et al., 2004), failed to detect VGLUT1 or VGLUT2 in enteric neurons.
Here we found evidence for the synthesis of VGLUT2 in only a limited number of mouse colorectal myenteric plexus neurons. Immunoreactivity for this VGLUT was only observed in a small population of neurons scattered throughout the plexus. Because it is possible that the presence of VGLUT2-LI in the soma of myenteric plexus neurons was derived from the endocytic internalization of VGLUTs after their release from primary afferent fibers (Fremeau et al., 2001), we carried out in situ hybridization studies to investigate the potential synthesis of VGLUTs in these neurons. We found that some neurons exhibited VGLUT2 mRNA, but not VGLUT1 mRNA, supporting the idea of synthesis of VGLUTs in some intrinsic mouse colorectum neurons. The phenotype of the VGLUT2-containing mouse myenteric plexus neurons remains unknown. They could be glial instead of neuronal profiles, although in the mouse and rat esophagus VGLUT2 myenteric plexus neurons lack the classical glial marker GFAP (Raab and Neuhuber, 2004). Moreover, Liu et al. (1997) often noted glutamate coexpression in enteric neurons expressing substance P or choline acetyltransferase in the small intestine of guinea pig and rat.
There is limited information about the expression of glutamatergic receptors in colorectal extrinsic (afferent) neurons. However, expression of several subunits of the excitatory NMDA receptor have been identified in the majority of rat DRG neurons from T9 to L6 (Sato et al., 1993; McRoberts et al., 2001; Marvizon et al., 2002). In support, more than 70% of retrogradely labeled, cultured rat colorectal DRG neurons responded to NMDA with an increase in intracellular calcium content (McRoberts et al., 2001). Moreover, superfusion of rat colorectal tissue sections with NMDA and serine stimulates the outflow of SP- and CGRP-LIs, an effect abolished by the addition of capsaicin and presumably occurring in TRPV1-expressing primary afferents (McRoberts et al., 2001). Altogether, the extensive expression of VGLUT2 (and therefore possibly also glutamate) in colorectal DRG neurons in the mouse suggests that NMDA receptors expressed in colorectal afferent terminals may have a role as auto- and/or heteroreceptors.
Enteric neurons expressing NMDA receptors have also been shown in rat colorectum (Burns et al., 1994; McRoberts et al., 2001; Del Valle-Pinero et al., 2007) and stomach (Burns et al., 1994). In the guinea pig ileum, application of L-glutamate or NMDA agonists exerts an excitatory action on muscle contraction, an effect largely dependent on the activation of myenteric plexus neurons (Wiley et al., 1991; Sinský and Donnerer, 1998; Kirchgessner, 2001). In humans, NR1, the most ubiquitous NMDA receptor subunit, was localized in cholinergic neurons as well as to glia in the large intestine (Giaroni et al., 2003). In addition, other glutamatergic receptors, including mGluR4, were found in normal colonic crypt epithelium and myenteric plexus neurons in human (Chang et al., 2005). Finally, pharmacological and morphological evidence also supports the presence of group I metabotropic glutamate receptors in submucosal and myenteric plexus neurons of the guinea pig, rat, and mouse colorectum (Nasser et al., 2007).
Several lines of evidence support a role for glutamate and its receptors in visceral sensation and nociception, acting at both central and peripheral sites. Thus, intrathecal lumbar administration of NMDA facilitates the response to colorectal distension in the naïve rat (Kolhekar and Gebhart, 1994, 1996). Conversely, hypersensitivity induced by repeated colorectal distension (Gaudreau and Plourde, 2004) or intracolonic application of inflammogens such as zymosan (Coutinho et al., 1996, 1998) or turpentine (Ide et al., 1997) is significantly attenuated by spinal (Coutinho et al., 1996; Ide et al., 1997) or ventral medulla (Coutinho et al., 1998) microinjection of NMDA receptor antagonists. Moreover, systemic application of memantine, a noncompetitive, peripherally acting NMDA receptor antagonist, reduces the behavioral response to noxious colorectal distension in conscious rats (McRoberts et al., 2001).
The behavioral events described above are at least partially dependent on the modulation of primary afferent activity. NMDA-induced CGRP release from rat spinal cord slices receiving input from DRG neurons innervating the colon has been shown to significantly increase in colitis (Li et al., 2006). Moreover, trinitrobenzene sulfonic acid (TNBS)-induced colitis in rats results in an increased NMDA current density in colorectal and noncolorectal DRG neurons, most likely through the upregulation and phosphorylation of the NR2B subunit (Li et al., 2006). Peripheral activation of ionotropic receptors eliciting excitatory effects has also been shown in the peripheral endings of rat colorectum (McRoberts et al., 2001) and mouse esophagus (Page et al., 2005), further supporting the involvement of the glutamatergic system in colorectal hypersensitivity. Metabotropic glutamate receptors are also involved in visceral hypersensitivity. mGluR5 (group I) antagonists inhibit colorectal distension-evoked visceromotor and cardiovascular responses in conscious rats, at least partially through an effect on peripheral afferent endings (Lindström et al., 2008); however, central effects are also likely (Zhu et al., 2005).
Glutamate receptor activation in enteric neurons can modulate the physiology of the distal gut. Thus, the activation of ionotropic as well as metabotropic (Jankovic et al., 1999; Giaroni et al., 2000) receptors in the rat rectum (Jankovic et al., 1999) and guinea pig colon (Giaroni et al., 2000) induces strong concentration-dependent contractions. Also, selective upregulation (Zhou et al., 2006b) and phosphorylation (Zhou et al., 2006a) of NMDA receptors in myenteric plexus neurons after TNBS-induced colitis has been reported in rats, suggesting a contribution of the enteric nervous system to the physiopathology of colorectal motility and nociception. In support, a recent study shows that NMDA receptor antagonists reduce the motility and inflammatory response in TNBS-treated rats (Varga et al., 2010).
Functional gastrointestinal disorders are characterized by altered motility, discomfort, pain, and hypersensitivity in both the organ (e.g., to balloon distension) and its area of somatic referral, typically expressed as increased tenderness to palpation. These disorders are notoriously difficult to manage and there is thus a need to develop novel therapeutic approaches. The results presented here not only highlight the importance of the glutamatergic system in the mouse colorectum, but also suggest that VGLUT2, by far the most abundantly expressed VGLUT, may be a potentially interesting pharmacological target for the modulation of visceral pain and hypersensitivity. Because glutamate and its receptors have been implicated in inflammatory and neuropathic pain, several glutamatergic receptor antagonists have been tested in rodents and humans. However, despite encouraging antihyperalgesic effects, these compounds have been associated with undesired adverse effects (Sawynok, 2003; Carlton, 2009; Finch et al., 2009; Sigtermans et al., 2009).
A different strategy for central and peripheral modulation of glutamatergic signaling could be blockade of VGLUT activity. Recently synthesized VGLUT inhibitors have demonstrated high in vitro efficacy (Carrigan et al., 2002; Patel et al., 2007). Unfortunately, selective VGLUT inhibitors have yet to be reported. However, in support of their potential usefulness, a 50% reduction in VGLUT2 protein expression in heterozygous VGLUT2 null mice was associated with reduced/attenuated mechanical and cold hyperalgesia in the spared nerve injury model, while leaving unaffected other types of sensory processing, including acute nociception and inflammatory hyperalgesia (Moechars et al., 2006; Leo et al., 2009). This effect was not replicated in heterozygote VGLUT1 null mutants, suggesting the importance of VGLUT2 in glutamatergic signaling and neuropathic pain (Leo et al., 2009). More recently, the targeted deletion of VGLUT2 in TRPV1 (Lagerström et al., 2010) or the sodium channel Na(v)1.8-positive sensory neurons (Lagerström et al., 2011), revealed decreased responsiveness to thermal (Lager-ström et al., 2010), mechanical, and inflammatory non-visceral pain (Lagerström et al., 2011). Similar mechanisms could take place in visceral systems, as colorectal neurons express TRPV1 (Christianson et al., 2006; Chaban, 2008), and a role for Na(v)1.8 in visceral pain and hyperalgesia was previously reported (Laird et al., 2002). It is unclear if the effects on neuropathy-induced hypersensitivity observed in VGLUT2 heterozygous mice depend on peripheral or central mechanisms or both. However, the abundance of VGLUT2 in nonvisceral (Brumovsky et al., 2007) and colorectal primary afferent neurons (demonstrated here) suggests that a reduction in its amount per vesicle could have profound effects on the size of glutamatergic quanta (Williams, 1997; Daniels et al., 2004; Moechars et al., 2006), and thus alter neurotransmission mechanisms, both at the target organ and spinal levels.
In conclusion, we present a systematic analysis of the expression of VGLUT1 and VGLUT2 in DRG neurons giving rise to colorectal afferents in the mouse. Contributions of the enteric and autonomic nervous systems were also addressed. The results reveal that VGLUT2, as compared to VGLUT1, is the most abundant VGLUT in extrinsic nerve terminals in the colorectum, suggesting the presence of glutamate and its potential release upon adequate stimuli. Currently ongoing studies in our laboratory will determine if VGLUT3 is also expressed by colorectal DRG neurons. Finally, the discrete number of myenteric neurons expressing VGLUT2 suggests that either very few enteric colorectal neurons in the mouse are glutamatergic or that VGLUT3 (or another unidentified VGLUT) participates in the incorporation of glutamate in enteric neurons in this organ.
Grant sponsor: National Institutes of Health (NIH); Grant number: NS19912 (to G.G.); Grant sponsor: IASP Early Career Research Award (to P.B.).
We thank Dr. Akiya Watakabe, National Institute for Basic Biology, Okazaki, Japan, for the generous donation of the VGLUT1 and VGLUT2 riboprobes. We also thank Mrs. Pamela Cornuet and Mr. Tim McMurray for excellent technical assistance, and Ms. Ann Hemmerle for assistance with some of the illustrations.