Loss of interstitial cells of Cajal and development of electrical dysfunction in murine small bowel obstruction

Authors


  • Author's present address I.-Y. Chang: Department of Anatomy, College of Medicine, Chosun University, Kwangju, Korea.

Abstract

  • 1Partial obstruction of the murine ileum led to changes in the gross morphology and ultrastructure of the tunica muscularis. Populations of interstitial cells of Cajal (ICC) decreased oral, but not aboral, to the site of obstruction. Since ICC generate and propagate electrical slow waves in gastrointestinal muscles, we investigated whether the loss of ICC leads to loss of function in partial bowel obstruction.
  • 2Changes in ICC networks and electrical activity were monitored in the obstructed murine intestine using immunohistochemistry, electron microscopy and intracellular electrophysiological techniques.
  • 3Two weeks following the onset of a partial obstruction, the bowel increased in diameter and hypertrophy of the tunica muscularis was observed oral to the obstruction site. ICC networks were disrupted oral to the obstruction, and this disruption was accompanied by the loss of electrical slow waves and responses to enteric nerve stimulation. These defects were not observed aboral to the obstruction.
  • 4Ultrastructural analysis revealed no evidence of cell death in regions where the lesion in ICC networks was developing. Cells with a morphology intermediate between smooth muscle cells and fibroblasts were found in locations that are typically populated by ICC. These cells may have been the redifferentiated remnants of ICC networks.
  • 5Removal of the obstruction led to the redevelopment of ICC networks and recovery of slow wave activity within 30 days. Neural responses were partially restored in 30 days.
  • 6These data describe the plasticity of ICC networks in response to partial obstruction. After obstruction the ICC phenotype was lost, but these cells regenerated when the obstruction was removed. This model may be an important tool for evaluating the cellular/molecular factors responsible for the regulation and maintenance of the ICC phenotype.

The hypertrophy of smooth muscle is a physiological response to the increased functional demands placed on an organ. This phenomenon is well documented in both vascular and visceral smooth muscle tissues, and has been studied extensively in the bladder, where obstruction of the urethra can lead to hypertrophy of the organ (Steers & DeGroat, 1988). Hypertrophy of the gastrointestinal (GI) muscle wall, associated with a partial obstruction, has also been documented in both clinical situations and animal models. With partial obstructions there are dramatic changes in the gross morphology and ultrastructure of smooth muscle cells within the tunica muscularis (Gabella, 1979, 1987). Changes in the muscle wall are not just limited to smooth muscle cells; there is also a decrease in the density of innervation of the hypertrophic intestine and the levels of expression of neuromessengers are also altered (Ekblad et al. 1998). To date, however, there have been few studies of the functional changes that occur in the GI tract as a result of mechanical obstructions. Mechanical obstructions of the GI tract that result from either congenital and/or acquired defects in both sphincteric and non-sphincteric regions are associated with pathologies that lead to a variety of motility disorders. Malformation or dysfunction of the sphincters, such as in infantile hypertrophic pyloric stenosis (Oue & Puri, 1999), idiopathic megarectum and megacolon (Gattuso et al. 1998) or stenosis of the recto-anal region of the GI tract, as occurs in Hirschsprung's disease, often lead to hypertrophy of the tunica muscularis of the GI tract oral to the site of occlusion (Lane, 1966; Webster, 1973). The genetic factors that have been implicated in the aetiology of diseases such as Hirschsprung's disease may be similar in humans and animal models (Robertson et al. 1997). Hypertrophy of the tunica muscularis can also be a consequence of acquired obstructions, as in achalasia (Friensen et al. 1983) or Chagas’ disease (Smith, 1982), growths of both benign and malignant tumours into the lumen or growths that compress the muscularis of the gut wall (Zollinger, 1988).

We studied the effects of bowel obstruction on the networks and electrical behaviour of the interstitial cells of Cajal (ICC). Immunohistochemical analyses and electron microscopy were performed on tissues both oral and aboral to a site of mechanical obstruction. The results of these studies demonstrate that a significant loss of ICC networks is induced oral, but not aboral, to a site of partial obstruction. We also tested whether defects in ICC were associated with changes in electrical slow wave activity and responses to enteric nerve stimulation. Our results suggest that ICC were not lost by cell death, but rather the phenotype of the ICC changed from functional cells that express Kit (a common marker for ICC) to a non-functional cell type. When obstructions were relieved, ICC networks were restored and slow wave activity returned, suggesting that ICC can be regenerated in adult tissues from which they have disappeared due to a pathophysiological process.

METHODS

Animal treatment

Male Balb-C mice (n= 34) aged between 40 and 60 days, and with a body weight of 23–26 g, were used for the studies of intestinal obstruction. These animals were obtained from breeder pairs purchased from Harlan Sprague-Dawley (Indianapolis, IN, USA). The mice were anaesthetized with a single i.p. dose of pentobarbital (Nembutal 3 μl g−1; Abbott Laboratories, Chicago, IL, USA). When the absence of a hind-limb pinch-withdrawal reflex was verified, a paramedial (left-hand side) laparotomy was performed. A loop of intestine was exposed and a polyethylene clip (6 mm in length, 5 mm external diameter, 4 mm internal diameter) was placed over a piece of intestine 30–50 mm oral to the ileocaecal sphincter. The clip size was designed to avoid obstruction of the resting diameter, but to reduce the ability of a bolus of material to pass across the site. The clip was also placed with a cut side between lateral extensions of the mesenteric vascular bed to avoid vascular injury. Following insertion of the clip, the intestine was replaced within the abdomen, and the abdomen was sutured closed. The animals were allowed to recover post-operatively on a heated blanket. Throughout the surgery and recovery the animals were intermittently oxygenated with a 97 % O2-3 % CO2 mixture. In animals with partial occlusions no differences were observed in eating or defaecation habits. A group of sham-operated animals was also prepared as controls. These animals underwent the same surgical preparations and had clips installed in the same anatomical positions, but the clips used on these animals had an internal diameter of 5 mm and produced no occlusion of the small intestine. Experimental animals, non-operated controls and sham-operated controls were killed for experiments at various times after the initial surgery (24 h to 74 days) by CO2 inhalation followed by decapitation and exsanguination.

In some experiments, after the occlusion clip had been in place for 14 days, the animals were re-anaesthetized and a second laparotomy was performed. After confirming intestinal hypertrophy (the major gross pathological change noted after partial occlusion), the clips were removed to relieve the occlusion and the site was marked with a piece of surgical thread. These animals were subsequently allowed to recover for a period of 1–2 months before being killed for electrophysiological and morphological studies. The animals were maintained and the experiments performed in accordance with the National Institutes of Health Guide for the care and use of laboratory animals, and all procedures used were approved by the institutional animal use and care committee at the University of Nevada.

Electrophysiological experiments

After animals were killed, the entire GI tract from 5 mm oral to the lower oesophageal sphincter to 10 mm oral to the internal anal sphincter was removed and placed in Krebs-Ringer buffer (KRB; see below). A 116 mm-long region of small intestine was isolated. This included the region of small intestine 10 mm aboral to the occlusion clip, 6 mm beneath the clip and 100 mm oral to the clip. The bowel was opened along the mesenteric border and the lumenal contents were washed away with KRB. Segments of the bowel were pinned to the base of a Sylgard silicone elastomer dish (Dow Corning, Midland, MI, USA) and the mucosa was removed by sharp dissection. Strips of muscle (25 mm × 5 mm) were cut parallel to the longitudinal muscle layer from specific sites both aboral and oral to the site of the occlusion clips. The muscles were placed in a recording chamber with the submucosal aspect of the muscle facing upwards. Cells were impaled with glass microelectrodes with resistances of 50–90 MΩ. Transmembrane potentials were recorded with a standard electrometer (Intra 767; World Precision Instruments, Sarasota, FL, USA). Data were recorded on digital tape (Vetter, Robersburg, PA, USA) and hard copies were made by replaying the tapes through a polygraph (Gould RS 3200, Cleveland, OH, USA). Experiments were performed in the presence of nifedipine (1 μm; Sigma, St Louis, MO, USA) in the perfusion solution to reduce contractions and facilitate extended periods of cell impalement. We have shown previously that slow waves in the murine intestine are not affected by nifedipine (Ward et al. 1994). In some experiments, parallel platinum electrodes were placed on either side of the muscle strips and neural responses were elicited by square-wave pulses of electrical field stimulation (EFS; 0.5 ms duration, 1–20 Hz, train duration of 1 s, supramaximal voltage) using a Grass S48 stimulator (Quincy, MA, USA).

Solutions and drugs

The bath chamber was perfused constantly with oxygenated KRB of the following composition (mm): NaCl 118.5; KCl 4.5; MgCl2 1.2; NaHCO3 23.8; KH2PO4 1.2; dextrose 11.0; CaCl2 2.4. The pH of the KRB was 7.3–7.4 when bubbled with 97 % O2-3 % CO2 at 37 ± 0.5 °C. After pinning, the muscles were left to equilibrate for at least 1 h before experiments were begun.

Analysis of electrophysiological data

Data are expressed as means ±s.e.m. Differences in the data were evaluated by Student's t test. The level of statistical significance was set at P < 0.05. The n values reported in the text refer to the number of muscle strips from which recordings were performed. Each muscle strip used was taken from a separate animal. Several electrical parameters were analysed: (1) resting membrane potential (RMP); (2) slow wave amplitude; (3) duration (time to 90 % repolarization from the initial upstroke); and (4) frequency. The figures displayed were made from digitized data using Adobe Photoshop 4.0.1 (Adobe, Mountain View, CA, USA) and Corel Draw 7.0 (Corel, Ontario, Canada).

Immunohistochemistry

Immunohistochemical analysis was performed on pieces of muscle at the same distances from the site of occlusion and directly adjacent to the regions of tissue as used for electrophysiology. For whole-mount preparations, the mucosa was removed by sharp dissection and the remaining strips of tunica muscularis (2 cm × 1 cm) were pinned to the base of a dish filled with Sylgard elastomer (Dow Corning), with the mucosal side of the circular muscle layer facing upwards. Tissues were fixed in acetone (4 °C; 10 min). Following fixation, preparations were washed for 60 min in phosphate-buffered saline (PBS; 0.01 m, pH 7.4). Tissues were then incubated in 10 % goat serum containing 0.3 % Triton X-100 for 1 h at room temperature to reduce non-specific antibody binding. For examination of ICC, tissues were incubated overnight at 4 °C with a rat monoclonal antibody raised against Kit protein (ACK2; 5 μg ml−1 in PBS; Gibco-BRL, Gaithersburg, MD, USA). Immunoreactivity was detected using fluorescein isothiocyanate (FITC)-conjugated secondary antibody (FITC-anti-rat; Vector Laboratories, Burlingame, CA, USA; 1:100 in PBS, 1 h, room temperature). Control tissues were prepared in a similar manner, omitting ACK2 from the incubation solution.

For cryostat sections, tissues from different regions along the intestine (oral and aboral to the clip site) were fixed in acetone as described above prior to being dehydrated in graded sucrose solutions, embedded in Tissue Tek (Miles, IL, USA), and frozen in liquid nitrogen. Cryostat sections were cut at 10 μm thickness and pre-incubated in goat non-immune serum for 1 h (10 % in PBS) before being incubated with ACK2 (5 μg ml−1 in PBS) at 4 °C overnight. Immunoreactivity was detected with FITC-conjugated secondary antibody (goat anti-rat 1:100, 1 h, room temperature; Vector Laboratories). Control tissues were prepared in a similar manner, but ACK2 was omitted from the incubation solution.

Tissues were examined with a BioRad MRC 600 (Hercules, CA, USA) confocal microscope with an excitation wavelength appropriate for FITC (488 nm). Confocal micrographs are digital composites of z-series scans of 10–30 optical sections through a depth of 10–30 μm. Final images were constructed with Comos software (BioRad).

Light and electron microscopy

Small intestines from 10 animals (5 control and 5 treated) were fixed with 2 % paraformaldehyde, 2.5 % glutaraldehyde, 3 % sucrose and 1.25 mm CaCl2 in 0.05 m cacodylate buffer (pH 7.4) at 4 °C for 2 h. Tissues were subsequently washed in cacodylate buffer (4 × 15 min) prior to post-fixation with 1 % osmium tetroxide for 1 h. After fixation, tissues were dehydrated through a graded series of alcohols, followed by immersion in propylene oxide for 10 min, before being infiltrated and finally embedded in Eponate 12 resin (Ted Pella, Redding, CA, USA). Semi-thin and ultra-thin sections were cut parallel and transverse to the long axis of the circular muscle layer at different sites oral and aboral to the clip site. Semi-thin sections were examined with a Leitz Diaplan microscope using phase contrast microscopy. Ultra-thin sections were stained with uranyl acetate (10 min) and counterstained with lead citrate (5 min) before being viewed with a Philips CM10 transmission electron microscope. Using light microscopy, five fields of view were analysed for each region of interest from five control and five treated animals. The n values reported in the text refer to the number of animals analysed.

RESULTS

Morphological changes

Polyethylene clips (6 mm in length, 5 mm exterior diameter, 4 mm interior diameter) were placed around the ileum for periods of 24 h to 14 days. These clips resulted in a partial occlusion of the intestine. Partial occlusion for 48–72 h produced no obvious changes in bowel morphology, but after more extended periods (e.g. 4 days), we observed significant distension of the bowel oral to the site of the clips. For up to 50 mm oral to the site of the clip, the diameter of the intestine was increased by approximately 250 % (i.e. from a diameter of 2.5 ± 0.25 mm to 6.75 ± 0.4 mm; n= 8 animals; P < 0.05). Distension of the small intestine extended 80–110 mm oral to the clip (Fig. 1A and B). Aboral to the clip (i.e. 5–10 mm from the bottom edge of the clip) there was no significant difference in the diameter of the small intestine (as compared 6 control and 5 sham-operated animals, i.e. 2.5 ± 0.2 mm).

Figure 1.

A, gross distension of the murine small intestine 14 days after insertion of a clip to create a partial occlusion. Distension of the intestine occurred proximal to the site of the clip. B, the gastrointestinal tract after its removal from the animal. Scale bars are as indicated in each panel.

Others have reported that the bowel walls of guinea-pigs and rats undergo extensive hypertrophy proximal to a region of partial obstruction (Gabella, 1990). We examined the hypertrophic and/or hyperplastic changes in the tunica muscularis of the murine intestine oral and aboral to the site of occlusion. At a distance 1 mm oral to the site of obstruction, the smooth muscle layers increased in thickness from an average of 26.1 ± 2.3 μm in control animals (n= 5) to 279 ± 9 μm (n= 5) 14 days after administration of the occlusion clip. The increase in thickness of the circular muscle layer was attributed to a change in the number and diameter of smooth muscle cells. The circular and longitudinal muscle layers of control animals had an average cross-sectional number of 7.4 ± 0.9 (n= 5) and 4.5 ± 0.7 (n= 5) cells, respectively. Fourteen days after clip insertion the number of cells 1 mm oral to the partial occlusion in both the circular and longitudinal muscle layers had increased to 21.3 ± 1.8 (n= 5) and 14.3 ± 0.5 (n= 5) per cross-section, respectively. These data are summarized in Table 1. With conventional electron microscopy we found that the ultrastructure of smooth muscle cells also changed dramatically. The diameter of the smooth muscle cells increased from a control value of 2.2 ± 0.3 μm (n= 300 cells from 5 animals) to 8.9 ± 1.1 μm (n= 300 cells from 5 animals) and the cells of both muscle layers had a more irregular profile (cell diameters were measured in transverse sections at the level of the nuclei from cells within 10 image fields and at 8 sites oral and aboral to the site of partial occlusion). These observations confirm the morphological changes in smooth muscle presented previously in detailed studies of other animal models of partial occlusion (Gabella, 1979, 1987). Therefore, we did not further characterize the specific changes in smooth muscle cells that occurred in the murine ileum.

Table 1. Tunica muscularis parameters oral and aboral to the site of partial obstruction
Distance from partial obstruction clipThickness of circular muscle layer (μm)No. of circular muscle cellsThickness of longitudinal muscle layer (μm)No. of longitudinal muscle cells
  1. Distances are given oral (+) and aboral (−) to the site of the partial obstruction clip. n= 5 control and 5 treated animals.

Control 16.6 ± 0.1 7.4 ± 0.9 9.5 ± 0.1 4.5 ± 0.7
−5 mm 83.5 ± 5.515.6 ± 0.961.5 ± 3.213.4 ± 1.3
+1 mm186.3 ± 1421.3 ± 1.892.5 ± 4.514.3 ± 0.5
+25 mm 98.5 ± 9.115.3 ± 1.163.9 ± 8.211.1 ± 0.6
+75 mm 16.0 ± 1.1  8 ± 0.411.9 ± 1.1 4.9 ± 0.3
+100 mm15.96 ± 0.6 7.3 ± 0.210.2 ± 0.5 4.4 ± 0.2

We examined how partial obstruction might affect ICC networks by performing immunohistochemistry for Kit, a commonly used marker for ICC in the GI tract (Ward et al. 1994; Huizinga et al. 1995; Torihashi et al. 1995). Whole-mounts and cryostat sections of the tunica muscularis were labelled with an anti-Kit antibody (ACK2). In control and sham-operated animals Kit-like immunoreactivity (Kit-LI) was localized in two distinct populations of ICC within the tunica muscularis. As reported previously, cells displaying Kit-LI in the small intestine formed an extensive network between the circular and longitudinal muscle layers at the level of the myenteric plexus (IC-MY), and a second population of ICC was found at the level of the deep muscular plexus in the circular muscle layer (IC-DMP).

IC-MY and IC-DMP were not noticeably affected in animals 48–72 h after administration of the occlusion clips. However, 14 days after the initiation of partial occlusions, IC-MY and IC-DMP could not be resolved by Kit immunohistochemistry for up to 10 mm oral to the occlusion site (Fig. 2). There was a gradient in the labelling of cells with Kit-LI in tissues between 25 and 100 mm oral to the clip. The intensity of Kit-LI was indistinguishable from normal levels at 100 mm from the obstruction (Fig. 2). Normal ICC networks were observed in tissue sampled at 5–10 mm aboral to the occlusion site.

Figure 2.

Kit-like immunoreactivity (Kit-LI) in ICC networks (arrows) shown in confocal images of whole-mounts at various sites oral and aboral to the site of partial occlusion lasting 14 days

A, ICC networks of normal appearance at a distance 5 mm aboral to the occlusion clip. ICC were also observed at the site of the occlusion clip (B). Immediately proximal, and for distances up to 50 mm oral to the occlusion clip, ICC networks were disrupted (C and D). ICC were absent at 1 mm oral to the clip, and only weak Kit-LI was observed at 25 mm. At 75–100 mm oral to the clip (E and F) ICC networks were essentially normal. G–I, cryostat sections through regions aboral and oral to the occlusion clip. These sections demonstrate that ICC networks between the circular and longitudinal muscle layers at the level of the myenteric plexus (ic-my) and those at the level of the deep muscular plexus in the circular muscle layer (ic-dmp) were present aboral to (G) and at 100 mm oral to (I) the site of partial obstruction. However, at 25 mm oral to the obstruction (H), IC-MY and IC-DMP were greatly reduced in number. The arrow and arrowhead in H represent the levels of the myenteric plexus and the deep muscular plexus across the section, respectively. Scale bars are as indicated in each panel. A, B, E and F–I are reconstructions from 10 × 1 μm optical sections. C and D are reconstructions from 30 × 1 μm optical sections.

Although loss of Kit labelling in GI muscles is typically a good indicator of loss of the ICC phenotype, we also examined tissues with transmission electron microscopy to determine whether the reduction in Kit expression was associated with an actual reduction in ICC. With electron microscopy, IC-MY were easily distinguishable from adjacent smooth muscle cells in control animals, as described previously (Torihashi et al. 1995). They had an electron-dense cytoplasm and a well-developed rough endoplasmic reticulum (ER). IC-MY also possessed abundant mitochondria, numerous caveolae along the plasma membrane, and a continuous basal lamina. IC-DMP displayed the same ultrastructural features as described previously (Torihashi et al. 1995); they possessed numerous mitochondria and were closely associated with nerve processes.

At distances 5–10 mm aboral to the partial occlusions, cells with typical IC-MY and IC-DMP ultrastructures were observed (Fig. 3A–B), confirming the presence of Kit-LI in this region. In contrast, typical ICC were not apparent 1–10 mm oral to the occlusion clips in either the myenteric plexus or deep muscular plexus (Fig. 3C–D). We observed cells in the spaces typically populated with ICC that had substantially different morphologies from normal ICC. These cells had ultrastructural features that were intermediate between fibroblasts and ICC, for example: (1) the cytoplasm was not as electron dense as in ICC; (2) fewer mitochondria were observed; (3) few caveoli were present; (4) an abundance of rough and smooth ER were observed; and (5) sparse basal lamina was found (see Fig. 3C–D). Therefore, we have termed these cells ‘intermediate’ cells. More proximal to the site of the clip (i.e. 50–100 mm), cells with ultrastructures similar to IC-MY were readily observed between the circular and longitudinal muscle layers (Fig. 3E–H). These cells possessed an electron-dense cytoplasm, numerous mitochondria and a well-developed ER. Caveolae were observed within the plasma membrane, and a continuous basal lamina was present. IC-DMP were also observed in close association with nerve bundles in tissues taken 50–100 mm oral to the occlusion clips. At 50 mm from the occlusion clip, ‘intermediate’ cells with the features described above were intermixed with ICC of normal appearance, suggesting that the lesion in ICC networks was still developing at this level. No evidence of cell death or necrosis (e.g. shrinking, condensed or fragmented nuclei, indicative of apoptosis, or cell swelling or bursting) was observed at any level. Furthermore, there was an absence of macrophage accumulation.

Figure 3.

Electron microscopy confirmed the loss of ICC oral, but not aboral to the site of the occlusion clip applied for 14 days

A, a typical IC-MY interposed between the circular muscle layer (cm) and a myenteric ganglion (mg). This section was from a tissue sample taken 5 mm aboral to the site of the occlusion clip. As typically observed, the ICC has an abundance of mitochondria (m). Intermediate cells (arrowhead) with less-dense cytoplasm and an abundance of rough ER were also occasionally observed in this region. At the same distance aboral to the clips, ICC were found in the deep muscular plexus (IC-DMP) within the circular muscle layer. These cells were in close association with nerve bundles (nb; B). Immediately oral to the site of the occlusion clips, cells with ultrastructural features of IC-MY and IC-DMP were absent. Intermediate cells (see text for description of features) were found in this region (arrowhead; C) in both the myenteric plexus and deep muscular plexus (C and D, respectively). E and F show IC-MY and IC-DMP, respectively, that were occasionally observed 50 mm oral to the site of partial occlusion. IC-MY and IC-DMP at this level were intermixed with intermediate cells (arrowheads; E). At 100 mm oral to the occlusion clip, IC-MY and IC-DMP with ultrastructural features typical of ICC were observed in all animals. Scale bar in H represents 1 μm and is the same for all panels.

Electrophysiological studies

Loss of ICC has been associated with specific phenotypes in mice: (1) absence of slow wave activity (Ward et al. 1994); and (2) reduced neurotransmission (Burns et al. 1996; Ward et al. 1998, 2000). Electrophysiological studies were therefore performed on circular muscles at several sites both oral and aboral to the partial obstructions and at different time points after administration of the clip. Control circular muscle cells of the ileum (i.e. from animals that had not undergone surgical procedures; n= 10) had resting membrane potentials (RMP) averaging −64.0 ± 3.5 mV, and slow waves of 29.7 ± 1.0 mV in amplitude and 2.3 ± 0.3 s in duration occurred at a frequency of 28.5 ± 1.5 cycles min−1. This activity was not statistically different from values reported previously (Ward et al. 1994). The electrical activities of ileal circular muscle cells were not significantly different in animals 48–72 h after partial occlusion clips were inserted (data not shown). However, 14 days after insertion of the clip, the electrical activity of the circular layer had changed considerably. At sites just oral to the clip (i.e. 1 mm oral), circular muscle cells were depolarized to −53.0 ± 2.3 mV and slow waves were absent in four out of the eight muscles studied. In the remaining four muscles, slow waves were greatly reduced in amplitude, duration and frequency (e.g. to an average amplitude of 3.3 ± 2.2 mV, duration of 0.9 ± 0.4 s and frequency of 11.5 ± 4.9 cycles min−1; Fig. 4; P < 0.001 compared to control tissues). Cells were also depolarized at sites more oral to the clip, and at 10 mm RMP averaged −41.0 ± 3.2 mV. Slow wave amplitude and frequency were reduced to 2.8 ± 0.6 mV and 20.9 ± 3.9 cycles min−1, respectively. One of the seven tissue samples taken from this position was electrically quiescent. In general, at greater than 10 mm the magnitude of the effects of partial occlusion decreased as a function of distance from the clip. RMP and slow wave frequency were not distinguishable from control values at 50 mm from the sites of the clips. Slow wave amplitude was significantly reduced for distances up to 60 mm from the site of the clip (Fig. 4), but was essentially normal at 100 mm compared to controls.

Figure 4.

Electrical slow waves recorded oral and aboral to the site of a 14 day partial occlusion

Slow waves were present 5 mm aboral (A) and at the site of the occlusion (B). Oral to the occlusion (1–5 mm), slow waves were either absent or greatly reduced in amplitude (C and D). From 25–50 mm slow wave activity was attenuated (E and F). Between 75 and 100 mm oral to the occlusion slow waves were normal (G and H).

In contrast to the dramatic changes in electrical activity observed oral to the clips, slow wave activity was relatively normal aboral to the clips. For example, at 5 mm aboral to the clip site, the RMP of the circular muscle cells averaged −60.2 ± 3.6 mV and slow waves were 22.2 ± 2.6 mV in amplitude, 2.2 ± 0.12 s in duration and occurred at a frequency of 26.8 ± 1.2 cycles min−1 (Fig. 4). Figure 5 summarizes data from many regions oral and aboral to the site of occlusion.

Figure 5.

Summary of electrical parameters from cells oral and aboral to the 14 day partial occlusion

A, resting membrane potential (RMP) decreased immediately oral to the site of the occlusion, and returned to normal approximately 75–100 mm oral to the occlusion clip. B, slow wave amplitudes were greatly depressed immediately oral to the occlusion site. Slow waves were essentially normal 75–100 mm oral to the occlusion site. C, the effect of obstruction on slow wave frequency. D, the effect of obstruction on slow wave duration. Both frequency and duration were decreased immediately oral to the site of obstruction, but were normal 25–50 mm oral to the site of occlusion. – indicates sites aboral and + indicates sites oral to the obstruction clip. Levels of significance compared to controls are indicated by asterisks (*P < 0.05; **P < 0.001).

Responses to the stimulation of intrinsic nerves with EFS (0.5 ms in duration delivered at a frequency of 1, 5, 10 and 20 Hz for 1 s; i.e. 1–20 s−1) were also compared in control and partially occluded ileal tissues. Normal responses to nerve stimulation in the ileum consisted predominantly of hyperpolarization of the membrane potential, averaging 4.5 ± 1.5 mV, and a reduction in slow wave amplitude (by 4 ± 1 mV) immediately following stimulation (n= 8). Increasing the stimulus frequency from 1 pulse to 10 Hz increased the amplitude of the hyperpolarization response (to 9.0 ± 2.5 mV) and further decreased the amplitudes of slow waves during or immediately following stimulation by 10.0 ± 2 mV. These responses persisted for one to several slow wave cycles after cessation of the stimulus. Inhibition of the slow waves was often followed by a period in which membrane potential was depolarized for several slow waves before returning to control levels (Fig. 6A). These responses were similar to previously reported neural responses in this region of the murine small intestine (Ward et al. 1994). Similar responses were also recorded in sham-operated animals (data not shown). In contrast, little or no response to EFS (1–20 Hz) was observed in 23 cells from six muscles, taken 1–10 mm oral to the sites of partial obstruction (Fig. 6B). In the six cells impaled in one particular muscle sample, small excitatory responses consisting of transient membrane depolarizations averaging 2.8 ± 0.7 mV in amplitude and 2.3 ± 0.6 s in duration were observed instead of the typical inhibitory junction potentials observed with a single pulse. In the cells of the remaining muscles there was no response to EFS. With a stimulation of 10 Hz, the response of three muscles was an excitatory junction potential averaging 4.0 ± 0.7 mV in amplitude and 3.4 ± 0.7 s in duration. More proximal to the clip sites (i.e. 50–100 mm oral), EFS responses similar to control responses were observed (not shown).

Figure 6.

Post-junctional neural responses were attenuated in tissues oral to the site of partial occlusion

A, neural responses to electrical field stimulation (EFS: 1–20 Hz for 1 s, 0.5 ms pulse duration, indicated by the arrow and bars) recorded from the circular muscle layer of a control ileum. Responses typically consisted of a frequency-dependent hyperpolarization in membrane potential and inhibition of slow wave electrical activity during or immediately following stimulation, which was followed by a slower developing membrane depolarization. B, response to EFS in a muscle taken 5 mm oral to a partial occlusion clip (14 days after insertion of the clip). Neural responses were greatly attenuated at all frequencies tested (1–20 Hz, for 1 s, 0.5 ms pulse duration; indicated by the arrow and bars). C, recovery of neural responses 30 days after removal of the partial obstruction clip. Inhibitory and excitatory neural responses were recorded 5 mm oral to the site where the partial obstruction clip had been placed.

The electrical activities of tissues from control or sham-operated animals were also investigated. Two control experiments were performed in which the small intestine was occluded for a relatively short period (e.g. 48 h). The small bowel of these animals appeared to be normal and electrical activity recorded from tissues oral to the site of occlusion was comparable to control values. For example, the RMP of circular muscle cells 1–10 mm oral to the clip averaged −68 ± 2.5 mV, and slow waves were 30 ± 3 mV in amplitude and 2.2 ± 0.2 s in duration. Slow waves occurred at a frequency of 28 ± 2 cycles min−1. A further series of experiments was performed on sham-operated animals in which a polyethylene clip (5 mm internal diameter) was inserted 14 days before electrical activity was recorded. Tissues from these animals had ICC networks of normal appearance. The RMP of circular muscle cells 1–10 mm oral to the clip positions averaged −62.6 ± 3 mV, and slow waves were 28 ± 3 mV in amplitude, 2.3 ± 0.25 s in duration and occurred at a frequency of 27 ± 2 cycles min−1 (n= 8). These values were not statistically different from activity recorded from control tissues that had not been surgically prepared with non-occluding clips. These data suggest that the changes in electrical activity recorded in tissues from animals with an occluding bowel clip were not a consequence of non-specific surgical trauma, irritation from the occlusion clip, or manipulation of the ileum.

Recovery after removal of the obstruction

In an additional series of experiments we examined the ability of tissues to regenerate the ICC phenotype and normal electrical responses after surgical treatment to relieve the mechanical obstruction. An occlusion clip was inserted in each of six animals and left for 14 days. Observation of the bowels of these animals after 14 days revealed distention and evidence of hypertrophy, as described above. With this group of animals, however, the clips were removed, and the animals were allowed to recover for 30–60 days before electrical activity and histological examinations were performed. Gross examination after the 30–60 day recovery periods revealed that the small bowel had returned to a more normal appearance. After the recovery period, the diameter of the bowel within 50 mm of the site of the occlusion clip averaged 3.0 ± 0.3 mm (Fig. 7).

Figure 7.

Restoration of normal anatomy after removal of the partial occlusion

A partial occlusion clip was placed on the intestine for 14 days and then removed. A period of 30 days was allowed before the animal was killed. The distension of the small intestine caused by 14 days of partial occlusion was greatly reduced during the recovery period. Scale bar, 5 mm.

Examination of Kit-LI in tissues 30 days after clip removal demonstrated recovery of Kit expression and re-establishment of the ICC phenotype in regions 1–25 mm oral to the site where the occlusion clip had been (Fig. 8). Kit-LI was restricted to cells located at the levels of the myenteric and deep muscular plexuses. In order to confirm the return of the ICC phenotype in regions oral to the site of the occlusion clip, electron microscopy was performed on tissues after 30 days of recovery. Cells with ultrastructures typical of IC-MY and IC-DMP were observed in the myenteric plexus region and within the deep muscular plexus, respectively (Fig. 9). Occasionally, intermediate cells were still observed after 30 days of recovery.

Figure 8.

ICC networks recovered after removal of the partial obstruction

A, Kit-LI 5 mm aboral to the site of the clip after a 30 day recovery period. B, ICC networks (arrows) at the site of the partial obstruction. C, Kit-LI 5 mm oral to the partial obstruction site. D, Kit-LI 25 mm oral to the partial obstruction. Scale bar, 50 μm in all panels. A–D are 1 μm z-steps through a tissue thickness of 10 μm.

Figure 9.

Electron microscopy confirmed recovery of ICC 30 days after removal of the partial obstruction clips

A and B, typical IC-MY adjacent to myenteric ganglia (mg) in tissues within 5 mm of the site of the obstruction clip 30 days after removal of the clip. IC-MY had a characteristic electron-dense cytoplasm, abundance of mitochondria, caveolae and gap junctions with adjacent IC-MY (A; inset). After this period of recovery, occasional intermediate cells (arrowhead) were still present (B). This cell was found in the myenteric plexus region and lies adjacent to a circular muscle cell (cm). The cytoplasm of the intermediate cell contained mitochondria and an abundance of rough ER. C and D, restoration of IC-DMP oral to the site of partial obstruction 30 days after clip removal. The cytoplasm of IC-DMP was electron dense and contained numerous mitochondria. Caveolae were observed along the plasma membrane. IC-DMP also formed gap junctions (arrow) with circular smooth muscle cells (C and inset) and were in close association with nerve bundles (nb: D). Circular smooth muscle cells also returned to more normal morphology after removal of the partial obstruction clip (A–D). Scale bars are as indicated in each panel.

There was also partial recovery of the electrical activity during the recovery period. Circular muscle cells 1 mm oral to the site of the clip continued to be depolarized (e.g. RMP averaged −52 ± 1.9 mV), but slow waves had increased to 10.0 ± 2.1 mV in amplitude, 3.1 ± 0.4 s in duration, and occurred at a frequency of 19.2 ± 1.9 cycles min−1 (n= 5; Fig. 10). Electrical activity was closer to control levels at more proximal sites. At 10 mm oral to the clip sites, RMP averaged −51 ± 1.5 mV and slow wave amplitude and frequency increased to 12.8 ± 2.4 mV and 22.6 ± 1.2 cycles min−1. RMP and slow wave frequency were at control levels 20 mm oral to the site of clip insertion, and slow wave amplitude was normal at 25–30 mm from the clip sites (Fig. 10). Figure 11 summarizes electrical parameters after 30 days of recovery from partial bowel obstruction.

Figure 10.

Recovery of slow waves after removal of partial obstruction clips

A, electrical slow waves 5 mm aboral to the site of obstruction 30 days after removal of the clip. B, slow waves at the site of obstruction. C–E, partial recovery of slow waves 1–25 mm oral to the site of occlusion. F–H, normal slow wave activity 50–100 mm oral to the occlusion.

Figure 11.

Summary of electrical parameters 30 days after removal of the partial occlusion clips

A, the RMP was partially restored oral to the occlusion site during the recovery period. Slow wave activity was also partially restored (B–D) during the recovery period. – indicates sites aboral and + indicates sites oral to the obstruction site. Levels of significance compared to controls are indicated by asterisks (*P < 0.05; **P < 0.001).

Responses to EFS also partially recovered after removal of the occlusion clips. After 30 days of recovery, small inhibitory junction potentials of 2.0 ± 1.0 mV were observed in muscles 1–10 mm oral to the site of occlusion in response to a single pulse of EFS (0.5 ms in duration; 12 cells from 3 animals; P < 0.05 when compared to neural responses 14 days after partial obstruction). Increasing the stimulus frequency from 1 pulse to 10 Hz increased the amplitude of inhibitory junction potentials to 6.0 ± 1.4 mV and reduced the amplitude of slow waves during or immediately following stimulation by 5.0 ± 1.5 mV (P < 0.01 when compared to neural responses 14 days after partial obstruction; Fig. 6C).

DISCUSSION

The results of several experiments have suggested that ICC generate pacemaker activity in phasic GI muscles, actively propagate slow waves, and mediate neural inputs from enteric motor neurones (see Sanders et al. 1999). The data in this study add support to this hypothesis: when ICC were lost from the small bowel oral to a partial obstruction, slow waves and neural responses were also lost. Recovery of ICC after removal of the obstruction caused restoration of slow wave activity. Neural responses were more resistant to recovery; however, distinct responses were observed in one animal 60 days after removal of the obstruction. Loss of ICC would be expected to result in dramatic motor dysfunction, and numerous studies have documented loss of ICC in GI motor disorders (Faussone-Pellegrini & Cortesini, 1985; Rumessen, 1996; Vanderwinden et al. 1996a,b; Isozaki et al. 1997; Kenny et al. 1998; Hagger et al. 1998; He et al. 2000; Ordog et al. 2000). The fact that ICC losses have been noted in such a variety of GI motor disorders is likely to be very important, however, and it is possible that common molecular signals generated during a variety of pathophysiological processes might converge on factors regulating the ICC phenotype. Animal models of GI motor disorders are extremely important, therefore, as a means of determining the molecular/ genetic messages that maintain or regulate the ICC phenotype. The possible involvement of ICC as a primary factor in motor dysfunction is an exciting new hypothesis in neurogastroenterology research.

In the present study we have developed a murine model of mechanical obstruction of the bowel that allows both a temporal and a spatial analysis of ICC phenotype and the integrity of ICC networks. We found that ICC are lost from the small intestine in the region immediately oral to a partial mechanical obstruction, confirming a previous report that Kit-LI was lost at the level of deep muscular and myenteric plexuses immediately oral to the site of partial obstruction in a rat model of hypertrophic ileum (Ekblad et al. 1998). A major finding in our study was that loss of ICC from tissues oral to a partial obstruction was associated with functional deficits in slow wave generation and propagation, and in motor neurotransmission. Of course, it is possible that changes in smooth muscle cells and the density of innervation that also occur oral to the regions of obstruction could also contribute to specific functional defects in these tissues. We found a gradient in the severity of damage to ICC networks, and in regions approximately 100 mm oral to the obstruction, ICC networks and function were essentially normal. It is also interesting to note that the region immediately aboral to the obstruction was not severely affected. The defects in ICC function developed over an extended period. A partial obstruction of 1–2 days was without discernible effect, but within 2 weeks there was a substantial loss of ICC. Thus, the gradient in the lesion over many millimetres oral to the obstruction and the time-dependent development of the lesion offers future experimental opportunities for investigation of the time course and spatial grading of the molecular signals that cause the loss of the ICC phenotype.

An important observation in the present study was that loss of ICC in areas oral to the occlusion clips was not associated with ultrastructural evidence of cell death. We have shown previously that when GI tissues are treated with neutralizing Kit antibodies, the ICC phenotype is lost as a result of a change in ICC towards a smooth muscle phenotype (Torihashi et al. 1999). At the present time we do not know the fate of ICC in the partial occlusion model, but we suggest that in the absence of cell death, ICC must change into another phenotype. In ileal tissues from animals with a partial occlusion, we noted an unusual cell type, referred to as ‘intermediate’ cells, in the anatomical locations typically occupied by ICC. These cells had features of both ICC and fibroblasts. Whether intermediate cells are the remnants of ICC networks (i.e. the product of a re-differentiation process that ICC undergo) or another cell type that is also affected by the mechanical obstructions has not yet been determined. Additional studies with markers for smooth muscle and other cell types will be needed to determine whether the fate of ICC in the occlusion model is similar to ICC in tissues treated with Kit antibodies.

Having demonstrated that loss of ICC in tissues oral to the occlusion site is not a result of cell death, we hypothesized that conditions might exist in which the ICC phenotype could be re-established. We found that simply removing the occlusion and allowing the animals to recover for at least 30 days accomplished this: Kit-positive ICC returned, and we noted a diminution of the intermediate cells. After the recovery period, networks of ICC of normal appearance were observed and there was partial recovery of electrical slow wave activity and responses to enteric nerve stimulation. These observations suggest that adult tissues retain the ability to regenerate functional ICC, and to re-establish these cells in the appropriate anatomical locations. The discovery of the molecular signals responsible for the restoration of functional ICC is an important goal of future investigations.

Previous studies have demonstrated that synaptic-like contacts occur between intramuscular ICC and varicosities of enteric neurones (Ward et al. 2000; Wang et al. 2000), and similar contacts exist between enteric neurones and IC-DMP in the small intestine (Wang et al. 1999). The close associations between ICC and enteric neurones appear to be important for enteric motor neurotransmission (Burns et al. 1996; Ward et al. 2000). Thus, the recovery of function in the small bowel after removing an occlusion clip is likely to include the re-establishment of close (synaptic) contacts between ICC and nerve varicosities. This suggests that the targeting molecules responsible for establishing contacts between nerves and ICC are expressed beyond the early developmental period when the initial innervation of IC-DMP occurs. In addition, Ekblad et al. (1998) demonstrated that several populations of enteric neurones underwent changes in neurotransmitter expression in the obstructed gut. It is possible that the loss of ICC-enteric neurone interactions might exert a trophic influence on the expression of neurotransmitters.

The disappearance and subsequent recovery of the ICC phenotype and function in an animal model is potentially an extremely important tool for future research because this model provides experimental access to molecular/ genetic signals that regulate the ICC phenotype. The spatial gradient in the degree of loss of ICC and the time-dependent development of the lesion in ICC networks make it possible to design experiments to investigate the onset of changes in the ICC phenotype and the reasons for the differences in severity of the lesion. By performing analyses of molecular/genetic signals that regulate the ICC phenotype in mice, human homologues might be ascertainable, and it may be possible to advance our knowledge of the factors that control the ICC phenotype in human patients. It is possible that a variety of pathophysiological scenarios that ultimately result in motility disorders converge on the same pathways that produced the dramatic changes in ICC networks observed in the present study. Determining the factors that restore ICC networks from re-differentiated, non-functional cells may provide new therapeutic opportunities for the treatment of motility disorders in humans.

Acknowledgements

These experiments were supported by grants DK40569 and DK 57236 from the National Institutes of Health. The morphological studies were supported by the core laboratory facility of a Program Project Grant (DK 41315).

Ancillary