Corresponding author K. M. Sanders: Department of Physiology and Cell Biology, University of Nevada School of Medicine, Reno, NV 89557, USA. Email: email@example.com
Recent studies have demonstrated that intramuscular interstitial cells of Cajal (ICC) are preferential targets for neurotransmission in the stomach. Terminals of enteric motor neurones also form tight, synaptic-like contacts with ICC in the small intestine and colon, but little is known about the role of these cells in neurotransmission. ICC at the deep muscular plexus (ICC-DMP) of the small intestine express neurokinin 1 receptors (NK1R) and internalize these receptors in response to exogenous substance P. We used NK1R internalization as an assay of functional innervation of ICC-DMP in the murine small intestine. Under basal conditions NK1R-like immunoreactivity (NK1R-LI) was mainly observed in ICC-DMP (519 cells counted, 100% were positive) and myenteric neurones. ICC-DMP were closely apposed to substance P-containing nerve fibres. Of 338 ICC-DMP examined, 65% were closely associated with at least one substance P-positive nerve fibre, 32% were associated with at least two, 2% were associated with more than two nerve fibres and 1% with none. After electrical field stimulation (EFS, 10 Hz; 1 min) NK1R-LI was internalized in more than 80% of ICC-DMP, as compared to 10% of cells before EFS. Internalization of NK1R was not observed in myenteric ICC or smooth muscle cells in response to nerve stimulation. Internalization of NK1R-LI was blocked by the specific NK1 receptor antagonist WIN 62577 (1 μm) and by tetrodotoxin (0.3 μm), suggesting that internalization resulted from stimulation of receptors with neurally released neurokinins. These data suggest that ICC-DMP are primary targets for neurokinins released from enteric motor neurones in the intestine.
Intramuscular interstitial cells of Cajal (ICC-IM) in gastrointestinal muscles lie in close association with varicose nerve terminals of enteric motor neurones (e.g. Daniel & Posey-Daniel, 1984; Wang et al. 1999, 2000). These morphological observations have led many investigators to suggest that ICC-IM may be involved in mediating neural inputs to the gastrointestinal tract. ICC-IM express receptors for a number of neurotransmitter substances (cf. Epperson et al. 2000) supporting a potential role for these cells as mediators of enteric motor neurotransmission. Physiological experiments have demonstrated the role for ICC-IM in enteric neurotransmission by showing that gastrointestinal muscles lacking ICC-IM have greatly reduced postjunctional responses to nerve stimulation in the lower oesophageal sphincter and stomach (Burns et al. 1996; Ward et al. 1998, 2000; Beckett et al. 2002, 2003; Suzuki et al. 2003).
In the small intestine, intramuscular ICC are concentrated in the region of the deep muscular plexus and are commonly referred to as ICC-DMP. Immunohistochemical and ultrastructural studies have shown that varicosities of excitatory and inhibitory enteric motor neurones form close, synaptic-like associations with ICC-DMP (Wang et al. 1999, 2003). For example, varicosities containing substance P-like immunoreactivity (a chemical indicator of excitatory motor nerve terminals) form synaptic contacts with ICC-DMP, and ICC-DMP express neurokinin 1 receptors (NK1R; cf. Sternini et al. 1995; Portbury et al. 1996; Lavin et al. 1998; Vannucchi & Faussone-Pellegrini, 2000). Exposure of small intestinal muscles to exogenous substance P causes NK1R internalization in ICC-DMP (Lavin et al. 1998). Taken together, these data suggest that ICC-DMP could be an important site of excitatory innervation in the small intestine, but direct tests demonstrating functional innervation of ICC-DMP by excitatory neurones have not been performed. The white-spotting (W/WV) and Steel-dickie (Sl/Sld) mutations have disruptions in the Kit signalling pathway that lead to a loss of ICC in the myenteric region of the small intestine (Ward et al. 1994), but ICC-DMP are not lost in these animals (Ward et al. 1994, 1995) and no animal model has been identified in which ICC-DMP are selectively lesioned. Therefore, other techniques are needed to evaluate the role of ICC-DMP in enteric motor neurotransmission. In the present study, we have examined the expression of NK1R in the murine small intestine and utilized the internalization of NK1R as an assay for functional innervation of ICC-DMP by enteric excitatory motor neurones.
BALB/c mice between the ages of 30 and 60 days postpartum were obtained from Jackson Laboratory (Bar Harbour, ME, USA). Animals were anaesthetized by isoflurane (Baxter, Deerfield, IL, USA) inhalation and exsanguinated after cervical dislocation. The abdomens were immediately opened and the entire gastrointestinal tract from 0.5 cm above the lower oesophageal sphincter to 1 cm above the internal sphincter was removed and placed in modified (low Ca2+, high Mg2+) Krebs Ringer Buffer (mKRB; see below) for 30 min at room temperature. The use and treatment of animals was approved by the Institutional Animal Use and Care Committee at the University of Nevada.
Functional immunohistochemical studies
After incubating in mKRB for 30 min a 4 cm segment of ileum starting 2 cm from the ileo-caecal sphincter was isolated for further dissection. The ileum was opened along the mesenteric border and the luminal contents were washed away with mKRB. The mucosa was removed by sharp dissection and the remaining strips of tunica muscularis were pinned to the Sylgard elastomer (Dow Corning Corp., Midland, MI, USA) base of a dissecting dish with the mucosal side of the circular muscle layer facing upward. Tissues were cut into small strips (5 × 10 mm) and pinned to Sylgard elastomer panels (1 × 10 × 15 mm at 110% of resting length and width). For electrical field stimulation of tissues, parallel platinum electrodes were placed on either side of the muscle strips and were connected to an electric field stimulator. Tissues were subsequently immersed in an organ bath and allowed to equilibrate in oxygenated KRB (97% O2–3% CO2) at 37.5 ± 0.5°C for 60 min before experiments were initiated. Before experiments were performed, L-NA (0.1 mm) and monensin (5 μm) was added to the KRB. Neural responses were elicited by square wave pulses of electrical field stimulation (EFS; 0.5 ms duration, 10 Hz, train duration of 60 s, 15 V) using a Grass SD9 stimulator (Quincy, MA, USA).
In some experiments intestines were exposed to substance P (1 μm in KRB) for 1 h at 4°C and then introduced to KRB (without substance P) for 20 min at 37°C.
After stimulation with EFS or exogenous substance P the Sylgard elastomer panels with attached strips of tunica muscularis were rapidly transferred to Zamboni fixative (2% paraformaldehyde made up in a 1.5% saturated picric acid solution, 0.1 m phosphate buffer, pH 7.3 at 4°C).
For examination of whole mount preparations the tunica muscularis were pinned to the Sylgard floor of a dissecting dish and stretched to 110% of their resting length before being fixed with Zamboni fixative for 1 h at room temperature. The muscle strips were removed from the Sylgard dish and washed with 0.01 m phosphate-buffered saline (PBS, pH 7.4) with 0.3% Triton X-100 overnight with several changes of the solution. Tissues on which functional immunohistochemical studies were performed were also washed after fixation as described above. Non-specific antibody binding was reduced by incubation of the tissues in bovine serum albumin (BSA, 1% in PBS, Sigma, St Louis, MO, USA) for 1 h at room temperature. Tissues were incubated with antibody to neurokinnin 1 receptor (NK1R, rabbit polyclonal antiserum, Sigma S8305, 1 : 2000) or mixture of NK1R and substance P (guinea-pig polyclonal antiserum, Chemicon AB5892, 1 : 1000) diluted with PBS containing Triton X-100 (0.3%) for 48 h at 4°C.
To demonstrate colocalization of NK1R and Kit immunoreactivities, the tunica muscularis was fixed in acetone (10 min at 4°C), washed with PBS and incubated with BSA (made up in PBS) for 1 h at room temperature. Tissues were incubated with primary antibodies to both NK1R and Kit (ACK2, rat monoclonal antibody, 5 μg ml−1, Gibco BRL, Gaithersburg, MD, USA) for 24 h at 4°C consecutively.
For cryostat studies using NK1R and substance P antibodies, small intestines were flushed with KRB before being fixed with Zamboni fixative for 4 h at room temperature. Following fixation, tissues were washed with PBS, immersed in 20% sucrose containing PBS and embedded in Tissue-Tek (Miles, Elkhart, IN, USA) before being quickly frozen in liquid nitrogen. Cryostat sections were cut at 10 μm thickness using a Leica CM3050 cryostat and collected on Vectabond-coated (Vector Laboratories, Burlingame, CA, USA) dry glass slides. Sections were preincubated with BSA (1% made up in PBS) for 1 h before being incubated with antibodies against NK1R (1 : 2000) and substance P (1 : 1000) at room temperature overnight. For examination of Kit labelling on cryostat sections intact small intestines were flushed with KRB and embedded in Tissue-Tek. After cutting, sections were immediately fixed in acetone (4°C for 10 min), washed with PBS, and preincubated with BSA (1% made up in PBS) for 1 h before being incubated with antibodies with NK1R and Kit at room temperature overnight.
For secondary antibodies, Alexa Fluor 488-coupled goat antirabbit IgG (for NK1R), Alexa Fluor 594-coupled goat anti-rat IgG (for ACK2) and Alexa Fluor 594-conjugated goat anti-guinea-pig IgG (for substance P) were used, respectively. All secondary antibodies were obtained from Molecular Probes (Eugene, OR, USA) and diluted to 1 : 200 in PBS. After tissues were incubated with primary antibodies, tissues were washed with PBS for at least 1 h, before incubation in secondary antibodies for 1 h at room temperature. After washing with PBS, specimens were mounted with Vectashield (Vector Laboratories, Burlingame, CA, USA). Control tissues were prepared by either omitting primary or secondary antibodies from the incubation solutions.
Tissues were examined with a Bio-Rad MRC 600 confocal microscope (Hercules, CA, USA) or a Zeiss LSM 510 META with excitation wavelengths of 488 nm 568 nm (Bio-Rad) or 543 nm (Zeiss). Some confocal micrographs shown are digital composites of Z-series scans of several optical sections (1–15 × 0.5–1.3 μm) through a depth of full or partial thickness of musculature. Final images were constructed with Bio-Rad Comos software or Zeiss LSM 510 META software.
Solutions and drugs
Muscles were maintained in KRB (37.5 ± 0.5°C; pH 7.3–7.4) containing (mm): Na+, 137.4; K+, 5.9; Ca2+, 2.5; Mg2+, 1.2; Cl−, 134; HCO3−, 15.5; H2PO4−, 1.2; dextrose, 11.5 and bubbled with 97% O2–3% CO2. Alternatively they were maintained in mKRB containing (mm): Na+, 137.1; K+, 5.9; Ca2+, 0.5; Mg2+, 15; Cl−, 157.3; HCO3−, 15.5; H2PO4−, 1.2; dextrose, 11.5 and bubbled with 97% O2–3% CO2. Solutions of tetrodotoxin (TTX; Sigma), Nω-nitro-l-arginine (L-NA; Sigma), 17-β-hydroxy-17-α-ethynyl-δ-4-androstano[3,2-b]pyrimido[1,2-a]benzimidazole (WIN 62577; Sigma), monensin sodium salt (Sigma) and substance P acetate salt (Sigma) were dissolved in distilled water or ethanol at 0.1–10 mm and diluted in KRB to the stated final concentrations.
To examine the expression of neurokinin-1 receptor-like immunoreactivity (NK1R-LI) in the murine small intestine, standard immunohistochemical procedures were performed on whole mount preparations and cryostat sections cut parallel and transverse to the circular and longitudinal muscle layers. Whole mount preparations revealed a dense network of cells with NK1R-LI running parallel to the long axis of circular smooth muscle cells (Fig. 1A). The cells with NK1R-LI were spindle shaped, had a prominent nuclear region, and possessed numerous spiny projections that extended from the main axis to contact the cell bodies and processes of adjacent cells with NK1R-LI. Cryostat sections revealed that the NK1R cells were located along the inner aspect of the circular muscle layer in the region of the deep muscular plexus (DMP, Fig. 1D). Double labelling with NK1R-LI and Kit-like immunoreactivity (Kit-LI) in whole mounts and cryostat cross-sections showed that the cells with NK1R-LI located in the region of the DMP also contained Kit-LI (Fig. 1A–F). Co-localization of NK1R-LI and Kit-LI in the cells near the inner aspect of the circular muscle layer demonstrated that the cells with NK1R-LI were interstitial cells of Cajal of the deep muscular plexus (ICC-DMP; see Torihashi et al. 1995). Observations of cryostat cross-sections also revealed that NK1R-LI and Kit-LI were located around the periphery of ICC-DMP (Fig. 1D–F), suggesting distribution of NK1R in the plasma membrane. Double labelling experiments using whole mounts and cryostat cross-sections revealed that NK1R-LI could not be resolved in ICC at the level of the myenteric plexus (ICC-MY) or in circular and longitudinal smooth muscle cells (Figs 1A–C and 2A–C). To determine if all, or only a subpopulation of, ICC-DMP were immunopositive for NK1R-LI, cells with NK1R-LI were counted in intestinal tissues from three animals double labelled with antibodies against NK1R and Kit. Of 519 ICC-DMP (i.e. cells labelled with Kit antibody), 100% were immunoreactive for NK1R. There were no ICC-DMP that were not immunoreactive for NK1R and no NK1R immunopositive cells in the DMP that did not express Kit-LI. NK1R-LI was also expressed in a subpopulation of myenteric neurones (Figs 2A, C). These neurones displayed a typical morphology of Dogiel type II, as previously demonstrated (Sternini et al. 1995; Grady et al. 1996a; Sayegh & Ritter, 2003).
For functional innervation of ICC-DMP to occur via NK1R, ICC-DMP must be associated with enteric nerve fibres that contain and release neurokinins. Double labelling experiments were performed on whole mount preparations (using single optical sections of 1.3 μm) and on cryostat sections using antibodies against substance P and NK1R. Varicose enteric nerve fibres with substance P-LI were closely associated with ICC-DMP with NK1R-LI (Fig. 3). The association between nerve fibres with substance P-LI and ICC-DMP occurred over distances greater than 200 μm (Fig. 3A–C). For example, with 338 ICC-DMP examined in three animals with substance P and NK1R double labelling, at least 64.8% of cells were associated with one substance P-LI positive nerve fibre, 32.3% were associated with two substance P-LI nerve fibres, 2.1% were associated with more than two nerve fibres and 0.8% were associated with none. We were unable to find obvious morphological associations between substance P nerve fibres and ICC-DMP in less than 1% of the ICC-DMP examined. Cryostat sections also confirmed the close association between substance P-LI nerve fibres and ICC-DMP. In cryostat sections nerve fibres with substance P-LI were not observed in close association with ICC-MY suggesting that although substance P is present in enteric neurones at the level of the myenteric plexus functional innervation does not occur through NK1R in these cells in the small intestine (Fig. 3D–F).
NK1R expressed by a variety of cells are internalized upon stimulation with substance P (cf. Larsen et al. 1989; Bowden et al. 1994; Grady et al. 1996b; Mann et al. 1999), and a previous study showed that stimulation with exogenous substance P induced internalization of NK1R in ICC-DMP of the guinea-pig small intestine (Lavin et al. 1998). We tested whether NK1R of the murine intestine could be internalized by substance P stimulation. Muscles were exposed to solutions containing substance P (1 μm) for 1 h at 4°C, and then warmed to 37°C for 20 min to allow receptor internalization. After this period the muscles were fixed and examined for NK1R internalization. Under control conditions (muscles not exposed to substance P) NK1R was diffusely distributed in ICC-DMP and could not be resolved in smooth muscle cells (Fig. 4A). Exposure to substance P resulted in particulate NK1R-LI in ICC-DMP and particulate labelling of circular smooth muscle cells and myenteric neurones (Fig. 4B–D). These data show that receptors in ICC-DMP and smooth muscle cells are available to exogenous substance P.
We exploited the phenomenon of NK1R internalization to determine whether ICC-DMP are functionally innervated by excitatory motor nerve terminals in the DMP. Examination of unstimulated whole mount preparations showed, as above, diffuse distribution of NK1R-LI on the cell surface in the perinuclear region and processes of ICC-DMP (Figs 5A, B). Following electrical field stimulation (EFS; 10 Hz for 1 min, 0.5 ms pulse duration), NK1R-LI translocated from a diffuse distribution in the plasma membrane to an endosomal-like distribution within the cell cytoplasm (Figs 6A, B). A summary of the number of ICC-DMP in which NK1R internalization was observed relative to total number of NK1R-LI cells in each preparation revealed that 80.1 ± 1.3% of ICC-DMP showed internalization of NK1R-LI after EFS compared to 10.4 ± 1.0% of ICC-DMP prior to EFS (P < 0.01; n= 10; each experiment on tissues of separate animals, Fig. 7). We were unable to resolve NK1R-LI in smooth muscle cells following EFS.
The specificity of involvement of neurokinins in EFS-induced internalization of NK1R-LI was tested by treating tissues with the specific NK1R antagonist WIN 62577 (1 μm) for 20 min prior to EFS. In these experiments NK1R internalization was 11.3 ± 1.4% in control conditions and 15.3 ± 1.1% in the presence of WIN 62577 (Fig. 8, P= 0.08; n= 3). WIN 62577 blocked internalization of NK1R-LI in response to EFS, suggesting that neurokinin stimulation of NK1R mediated receptor internalization in ICC-DMP (Fig. 8). Internalization of NK1R-LI in response to EFS was likely to be due to release of neurokinins from enteric nerve terminals since incubation of tissues with tetrodotoxin (TTX; 0.3 μm) for 20 min prior to EFS blocked receptor internalization. In these experiments NK1R internalization was 12.2 ± 1.1% in control conditions and 14.2 ± 1.5% in the presence of TTX (Fig. 9; P= 0.34; n= 3).
Studies in animals lacking intramuscular ICC have shown that ICC mediate a significant portion of postjunctional responses to enteric excitatory and inhibitory neurotransmission (Burns et al. 1996; Ward et al. 2000; Beckett et al. 2002, 2003; Suzuki et al. 2003). The role of ICC in neurotransmission appears to depend upon the close anatomical relationship between varicose nerve terminals and ICC (see Daniel & Posey-Daniel, 1984; Burns et al. 1996), because when this relationship breaks down, such as in models of type 1 diabetes, neuromuscular transmission is compromised (Ordog et al. 2000). Close, synaptic-like contacts between enteric neurones and ICC also occur in the colon (Wang et al. 2000) and in the deep muscular plexus of the small intestine (Wang et al. 1999, 2003). In this study we utilized functional immunohistochemical techniques to demonstrate that ICC-DMP are innervated by enteric excitatory motor neurones.
ICC-DMP of several species have been shown to express NK1R (Sternini et al. 1995; Portbury et al. 1996; Lavin et al. 1998; Lecci et al. 1999; Vannucchi et al. 1999; Vannucchi & Faussone-Pellegrini, 2000). One study demonstrated that exposure of guinea-pig small intestinal muscles to exogenous substance P caused receptor internalization and aggregation of receptor in endosome-like structures in the cytoplasm of ICC-DMP (Lavin et al. 1998). Internalization of NK1R-LI was also observed in ICC of the myenteric plexus of the guinea-pig ileum, where NK1R-LI was not observed until the tissues were exposed to substance P. In the present study, and others, unstimulated smooth muscle cells of the small intestine were found to be either immunonegative for NK1R or to have weak immunoreactivity (e.g. Southwell & Furness, 2001). Exposure to exogenous substance P increased resolution of NK1R-LI (this study and see Lavin et al. 1998). Resolution of NK1R-LI in smooth muscle cells after stimulation demonstrates the presence of NK1R by smooth muscle cells, which is well documented from pharmacological and physiological studies. The fact that resolution of NK1R-LI on smooth muscle cells is poor under basal conditions suggests that expression of these receptors is much less in smooth muscle cells than in ICC-DMP. It is possible that the receptors are concentrated as they are internalized into endosomes, and this might improve the ability to resolve immunoreactivity.
The increase in NK1R-LI in smooth muscle cells after stimulation with exogenous substance P, but not after EFS, suggests that NK1R of smooth muscle cells are not significantly stimulated by endogenous neurokinins under either basal conditions or during stimulation of enteric motor neurones. This observation supports the hypothesis that primary innervation by neurokinin-containing motor neurones (i.e. excitatory neurones) occurs via ICC-DMP. In our experiments electrical field stimulation (10 Hz for 60 s) caused TTX-sensitive receptor internalization in ICC-DMP, but this relatively robust level of stimulation caused no resolvable NK1R-LI in smooth muscle cells. This finding suggests that, in comparison to ICC-DMP, smooth muscle cells are not exposed to concentrations of neurokinins from neural release that are high enough to produce detectable receptor internalization. The data also suggest that different cell populations are exposed to neurokinins released from nerve terminals than are stimulated by exogenous substance P in ‘organ bath’ pharmacological experiments, which is a well-developed concept in autonomic neurotransmission (Hirst et al. 1992).
NK1Rs are also internalized in myenteric neurones stimulated with substance P (Mann et al. 1999). It is interesting to note that treatment of tissues with monensin to trap internalized receptors resulted in a high level of intracellular NK1R-LI without addition of substance P. This was attributed to stimulation of receptors (and internalization) by endogenous release of substance P since the effect was blocked by solutions with low Ca2+–high Mg2+ (to block transmitter exocytosis) and by SR140333, a specific antagonist of NK1R. We noted a low background of internalized NK1R in unstimulated tissues in the presence of monensin. This observation suggests there is a low basal level of substance P release from motor nerve terminals that impinge upon ICC-DMP and may indicate low basal firing rates for excitatory motor neurones at frequencies that would support neurokinin release.
Whole mount images showed that receptor internalization occurred throughout ICC-DMP, suggesting multiple points of exposure to neurokinins during nerve stimulation. This is consistent with double labelling experiments showing enteric motor neurones coursing along the lengths of ICC-DMP and many points of contact between substance P-containing varicosities and ICC-DMP. We also found that all ICC-DMP (as identified with Kit antibody) were immunopositive for NK1R, and virtually every ICC–DMP had close associations with substance P-containing excitatory nerve terminals. Thus, there are not specialized populations of ICC-DMP designed to mediate excitatory or inhibitory neurotransmission. This idea is reinforced by studies of inhibitory pathways because ICC-DMP were widely immunopositive for soluble guanylyl cyclase and type 1 cGMP-dependent protein kinase, the receptor and signalling pathway activated by nitric oxide (Salmhofer et al. 2001). These findings suggest that motility disorders resulting from loss of ICC should not show preferential loss of excitatory or inhibitory neural regulation.
In summary, endogenous neurokinins induce endocytosis of NK1R in ICC-DMP of the murine intestine. In many cells ligand-induced endocytosis and recycling of neurokinin receptors have correlated with loss and recovery of functional binding sites (cf. Bowden et al. 1994) suggesting that this mechanism participates in regulation of peptidergic neurotransmission. Exogenous substance P also increased NK1R internalization in ICC-DMP and in smooth muscle cells. ICC-DMP are densely innervated by substance P-containing neurones and stimulation of intrinsic neurones resulted in NK1R internalization in ICC-DMP, but not in smooth muscle cells. Our data support the hypothesis that ICC-DMP are the primary sites of functional innervation by neurokinin-containing (excitatory) motor neurones. Thus, ICC-DMP are likely to participate in : (i) the postjunctional response of the intestine to nerve stimulation, and (ii) regulation of the overall responsiveness of the gut to motor nerve stimulation.
This project was supported by a program project from the NIH : P01 DK41315 and by DK 57236. Morphological studies were supported by a Core laboratory facility supported by this program (i.e. Core C) and an equipment grant from the NCRR for the Zeiss LSM510 confocal microscope (1 S10 RR16871).