Slow wave propagation and plasticity of ... - Wiley Online Library

1039

Exp Physiol 96.10 pp 1039–1048

Research Paper

Slow wave propagation and plasticity of interstitial cells of Cajal in the small intestine of diabetic rats Wim J. E. P. Lammers1 , H. M. Al-Bloushi2 , S. A. Al-Eisaei2 , F. A. Al-Dhaheri2 , B. Stephen1 , R. John2 , S. Dhanasekaran3 and S. M. Karam2

Experimental Physiology

From the Departments of 1 Physiology, 2 Anatomy and 3 Pharmacology, Faculty of Medicine and Health Sciences, United Arab Emirates University, Al Ain, United Arab Emirates

The number of myenteric interstitial cells of Cajal (ICC-MY), responsible for the generation and propagation of the slow wave in the small intestine, has been shown to decrease in diabetes, suggesting impairment of slow-wave (SW) propagation and related motility. To date, however, this expected decrease in SW propagation has neither been recorded nor analysed. Eleven rats were treated with streptozotocin and housed in pairs with 11 age-matched control animals. After 3 or 7 months, segments of duodenum, jejunum and ileum were isolated and divided into two parts. One part was processed for immediate freezing, cryosectioning and immunoprobing using anti-c-Kit antibody to quantify ICC-MY. The second part was superfused in a tissue bath, and SW propagation was recorded with 121 extracellular electrodes. In addition, a cellular automaton was developed to study the effects of increasing the number of inactive cells on overall propagation. The number of ICC-MY was significantly reduced after 3 months of diabetes, but rebounded to control levels after 7 months of diabetes. Slow-wave frequencies, velocities and extracellular amplitudes were unchanged at any stage of diabetes. The cellular automaton showed that SW velocity was not linearly related to the number of inactive cells. The depletion of ICC-MY is not as severe as is often assumed and in fact may rebound after some time. In addition, at least in the streptozotocin model, the initial reduction in ICC-MY is not enough to affect SW propagation. Diabetic intestinal dysfunction may therefore be more affected by impairments of other systems, such as the enteric system or the muscle cells. (Received 12 April 2011; accepted after revision 6 July 2011; first published online 8 July 2011) Corresponding author W. J. E. P. Lammers: Department of Physiology, Faculty of Medicine and Health Sciences, PO Box 17666, Al Ain, United Arab Emirates. Email: [email protected]

The pathophysiology of gastrointestinal (GI) dysfunction in diabetes mellitus is multifactorial. Contributing causes could be damage to the vagal (Steinberg et al. 2006) or sympathetic nerves (Camilleri & Malagelada, 1984), damage or dysfunction of the enteric nervous system (Nowak et al. 1986; Belai & Burnstock, 1990; Xue & Suzuki, 1997), hyperglycaemia (Soffer et al. 1999) and damage or destruction of interstitial cells of Cajal (ICCs; Sanders et al. 1999; Huizinga et al. 2009). These effects are not equally distributed throughout the GI system, and some tissues or systems may be more affected in some areas than in others (Sanders et al. 2002; Wang et al. 2009). On top of that, compensatory mechanisms and plasticity of the relevant tissues (Ordög et al. 2000; Ordög, 2008; Huizinga et al. 2009) will certainly affect the C 2011 The Authors. Journal compilation C 2011 The Physiological Society

ensuing pathophysiology, as does the fact that slow-wave dysrhythmias may occur in diabetic patients (He et al. 2001; Sanders et al. 2002; Ordög, 2008). It is therefore not clear which factor plays what role in diabetic GI dysfunction. Recently, attention has focused on the loss of myenteric (MY) ICCs, and its role in the generation and propagation of slow waves in the diabetic small intestine (He et al. 2001; Yamamoto et al. 2008). However, to the best of our knowledge, macroscopic propagation of the slow wave has not yet been recorded and analysed in diabetes. Here we analysed the propagation of slow waves in a streptozotocin model and took the opportunity to delineate the expected loss of slow-wave propagation and to test whether it correlated with the expected loss in ICC-MY. DOI: 10.1113/expphysiol.2011.058941

1040

W. J. E. P. Lammers and others

Methods The experimental protocol was approved by the Animal Research Ethics Committee, Faculty of Medicine & Health Sciences, United Arab Emirates University. Twenty-two young adult male Wistar rats (body weight 318 ± 23 g) were used in this study. Eleven were treated with an intraperitoneal injection of streptozotocin at 60 mg kg−1 and housed in pairs together with 11 agematched control rats. Blood glucose levels and body weights were monitored (Table 1). After 3 months, five control and five diabetic rats were deprived of food overnight and then decapitated. The small intestines were quickly removed and immersed in cold oxygenated Tyrode solution. The remaining rats (six control and six diabetic animals) were decapitated 7 months after streptozotocin injection and their intestines treated in a similar manner. From each preparation, 10-cm-long segments from the duodenum, jejunum and ileum were isolated. The 3 cm proximal part of each segment was immediately frozen using OCT medium (EM Sciences, Hatfield, PA, USA) at the temperature of liquid nitrogen and then processed for cryosectioning. Frozen sections (15 μm thick) were fixed in 4% paraformaldehyde in PBS and processed for immunohistochemistry using the avidin–biotin complex method as previously described (Karam et al. 1997). Briefly, peroxidase activity was inhibited by incubating tissue sections in PBS containing 3% hydrogen peroxide for 1 h. The primary antibody used was a rabbit polyclonal anti-c-Kit antibody (Dako, Glostrup, Denmark; 1:400 dilution). Following incubations in biotinylated antirabbit immunoglobulin G (1:500 dilution) and avidin peroxidase (1:1000 dilution), diaminobenzidine was used to visualize areas of antigen–antibody binding sites. Finally, tissue sections were counterstained with Haematoxylin to label all nuclei, including those of immunoprobed cells. For quantification of c-Kitimmunolabelled ICC-MY, at least three tissue sections of each control and diabetic rat were examined. Labelled ICCs located in between the two smooth muscle layers of muscularis externa were quantified per high power field and averaged for each group of rats.

Table 1. Body weight and blood glucose levels 3 months

n Body weight (g) Blood glucose (mg dl−1 )

7 months

Control

Diabetic

Control

Diabetic

5 348 ± 18 108 ± 9

5 264 ± 28∗ 342 ± 51∗

6 365 ± 23 119 ± 10

6 245 ± 32∗ 375 ± 54∗

∗

Significant difference from control P < 0.001.


The three remaining 7 cm distal segments were positioned in an organ bath for electrophysiological recordings. The preparations were opened along the mesenteric border and pinned at the corners with the serosal side facing upwards. The preparations were superfused at a rate of 100 ml min−1 with a modified Tyrode solution (mM: 130 NaCl, 4.5 KCl, 2.2 CaCl2 , 0.6 MgCl2 , 24.2 NaHCO3 , 1.2 NaH2 PO4 and 11 glucose), saturated with carbogen (95% O2 –5% CO2 ) and kept at a constant pH (7.35 ± 0.05) and temperature (37 ± 0.5◦ C). Extracellular recordings were performed with an 11 × 11 extracellular electrode array (Fig. 1). The array consisted of Teflon-coated silver wires (0.3 mm diameter) that extended 3–5 mm below the array base to allow Tyrode solution to reach the serosal surface of the tissue (Lammers et al. 1993). The interelectrode distance was 1 mm. All electrodes were connected through shielded wires to 121 AC amplifiers where the signals were amplified (×4000), filtered (bandwidth 2–400 Hz), digitized (8 bits, 1 kHz sampling rate per channel) and stored on a PC hard disk (Lammers et al. 1993, 2005). Electrical recordings from all 121 electrodes were made in a unipolar configuration with a large silver plate in the tissue bath acting as the indifferent pole. After the experiment, representative samples of spontaneous activities were chosen and the signals (duration of 64 s) transferred to a personal computer for further analysis. The signals were digitally filtered (20-point moving average) to remove 50 Hz noise and displayed on screen in sets of 11 channels at a time (Fig. 1D). The local activation time of a slow wave was identified by the moment of maximal negative slope and marked with a cursor (Lammers et al. 1993, 2005), whereby the reference time was determined by the timing of the first detected slow wave (t = 0.0 s at electrode no. 1). After all recorded slow waves had been analysed, their activation times were displayed in the format of a grid of the original recording array of electrodes (Fig. 1E). In the case of a low-quality signal or if no deflections were visible, those sites on the maps were left empty (open circles). In the maps, isochrones were drawn manually around areas activated in time steps of 500 ms. From the local activation times, the direction and the magnitude of the slow-wave propagation was calculated (Lammers et al. 2005) and averaged (Fig. 1F, arrow). The amplitudes of the local waveforms were also measured and averaged. The slow-wave frequency was measured by averaging the sequential intervals from one representative electrogram (Fig. 1D, electrogram no. 11). From each 64 s recording, five or six slow-wave cycles were analysed and averaged. For the simulation studies, a cellular automaton model (Lammers et al. 1987; Mase et al. 2005) was developed and written in REALsoftware (Fig. 2). The model consisted of an isotropic array of 100 × 20 cells. C 2011 The Authors. Journal compilation C 2011 The Physiological Society


Slow waves and interstitial cells of Cajal in diabetic small intestine

Pacing was performed from a linear array located at one end of the array. Propagation from one cell to its horizontal or vertically oriented neighbouring cell took place after a delay √ of five time steps, and in the oblique direction in 5 2 time steps. Propagation rate in the horizontal and vertical directions was kept the same to ensure isotropic propagation conditions (Lammers et al. 2002). Three different propagation models were tested; an excited cell could activate four, eight or 24 neighbouring

cells (Fig. 2A). For each test run, propagation times were measured from the pacing site to the opposite side of the sheet (normal at 500 time steps; Fig. 2C). Several test runs were performed with all cells active (Fig. 2C) and when randomly inactivating an increasing number of cells (Fig. 2D and E). All data are presented as means and standard deviations. Student’s unpaired t test was used in this study to determine statistical significance.

Figure 1. Method of recording and analysing isolated segments from the small intestine A and B show the opening of the tubular preparation and its positioning in an organ bath, serosal side facing upwards, and the lowering of the 121-electrode array onto the tissue. C shows that occasionally the array was larger than the tissue width. From the central column of recording electrodes (1–11), electrograms are displayed in D. From the first slow-wave cycle, the rapid downstroke was time marked (times indicated in milliseconds) and these values plotted, as measured at all recording electrodes, in the map in E. In E, isochrones are drawn around areas activated in 500 ms time steps and false coloured as shown in F. From these data, the velocity of the slow wave and its frequency (D) were calculated. C 2011 The Authors. Journal compilation C 2011 The Physiological Society

1041

1042


Results During the resection of the small intestinal segments, it was noted that the small intestines of all diabetic rats (3 and 7 months) were full and expanded, suggesting slow transit, while the control intestines were largely empty.

Histology results

Immunolocalization of c-Kit in the muscle layer of the three intestinal segments of control and the 3-month-


diabetic rats revealed a decrease in the amount of immunolabelled ICC-MY in the diabetic group (Fig. 3). Quantification showed that the difference was statistically significant in all three segments of the small intestine (Fig. 4). However, when the intestinal segments of the 7-month-diabetic rats and their control group were immunoprobed with anti-c-Kit antibody, the labelling pattern was reversed. There was a tendency to find more labelled ICC-MY in the diabetic than in the control rats (Figs 3 and 4), although this did not reach statistical significance.

Figure 2. Simulation of propagation A, three different models of activating neighbouring cells were tested. B, size and orientation of the simulation sheet used, with simultaneous pacing of all the 20 cells in the first left column. C, result of propagation through a homogeneous sheet, reaching the last column (right) in 500 time steps. D, 10% of the cells are randomly selected to be inactive, as indicated by the filled squares. E, propagation through this sheet increased the propagation time by 5%. See main text for further details. C 2011 The Authors. Journal compilation C 2011 The Physiological Society



1043

Figure 3. Immunoprobing of c-Kit in the muscle layer of 3- and 7-month-diabetic rats and their agematched controls Immunolabelled ICC-MY are indicated by the arrows. The bar in the lower left panel indicates 40 μm, and this scale is applicable to all panels.

Slow-wave mapping

Positioning the 121-electrode array immediately revealed a propagating slow wave in all control and streptozotocintreated segments from all three parts of the small intestine. Figure 5 shows representative electrograms and propagation maps obtained from both control and 3-month-diabetic rats. In all cases, slow waves were

recorded and propagated sequentially from one location to the next. From these analyses, the frequencies, propagation velocities and amplitudes of the extracellular slow wave were measured and are presented in Fig. 6. We did not find significant differences in all three characteristics between the two control groups and those obtained after 3 and 7 months diabetes, nor did we find any significant correlations when pairing the individual data with ICCMY counts. Simulation study

Figure 4. Counts of immunolabelled ICC-MY per high-power field in 3- and 7-month-diabetic rats and their age-matched controls ∗ P < 0.05; n.s., not significant. C 2011 The Authors. Journal compilation C 2011 The Physiological Society

With the cellular automaton model described in the Methods, experiments were performed in which the time of propagation for the impulse to conduct from one end of a preparation to the other was measured. In each experiment, the number of inactive cells was increased in steps of 5% until the stage was reached when the impulse was no longer able to reach the other end of the strip (total propagation block). As the inactive cells were randomly distributed, each experiment was repeated five times. The simulations were repeated for three neighbouring models; the four-, eight- and 24-neighbouring models. Examples of these experiments are shown in Fig. 7. The left column shows the propagation with the fourneighbour propagation model with 20, 35, 50, 65 and 80% inactive cells. The time below each map indicates the

1044



Figure 5. Electrograms and propagation maps from segments obtained from the duodenum (left panels), the jejunum (central panels) and the ileum (right panels) from a control and from a rat 3 months after streptozotocin treatment As shown in the electrograms from the control and the STZ-treated rat, slow waves are visible and propagation occurred in all segments of the small intestine.

C 2011 The Authors. Journal compilation C 2011 The Physiological Society



increase in propagation time achieved in each experiment, whereby t = 500 is the normal unimpeded propagation time (Fig. 2). The left column (four-neighbour model) shows that the propagation time increased with increasing numbers of inactive cells and that total block was achieved at values higher than 35%. In the central and in the right columns, the eight- and 24-neighbouring models were tested. Use of these propagation models resulted in a marked drop in propagation time, and total block was only achieved at a much higher percentage of inactive cells. The results of all simulations, plotted in Fig. 8, show the effect of increasing numbers of inactive cells in the three neighbouring propagation models. They show the following results: (a) propagation is not linearly related to the number of randomly distributed inactive cells; (b) total propagation blocks require large numbers of inactive cells; and (c) increasing communication between cells inhibits the depression in propagation.

1045

(2006) used imaging of Ca2+ transients as a marker for propagation and showed very nicely (their Fig. 3) the variability in activation timing of neighbouring ICCs, similar to the variations we evoked in our simulation with moderate (10–15%) inhomogeneities. They also showed that, at least in the small intestine, Ca2+ transients propagate isotropically, similar to the extracellularly recorded slow-wave propagation shown some time ago (Lammers et al. 2002) and which formed the basis for

Discussion In this study, we have documented the reduction in ICCMY at 3 months of diabetes, as also shown by others (Yamamoto et al. 2008), and the repopulation of ICCMY at 7 months, which has not been shown before in the small intestine, but seems to be similar to the plasticity in ICC-MY shown in the stomach (Wang et al. 2009). To our initial surprise, we did not see the expected reduction in slow-wave frequency, velocity and/or amplitude suggested by the transient reduction in ICC-MY at 3 months of diabetes. As the ICCs in the ICC-MY are assumed to function in a network, the question arises of how many cells could be removed before overall propagation is affected. This question arose many years ago in a study of propagation in cardiac muscle (Lammers et al. 1987), and we adapted that crude cellular automaton to this particular issue. Even at the most basic level of four connections, we did not find a linear relationship between the number of cells affected and the slow-wave velocity. In the present simulation, we extended the original study by increasing the potential number of connections between neighbouring cells, which showed that slowwave propagation becomes even more robust with an increase in interconnectivity. Qualitatively similar results were recently described, again in cardiac tissue, in a much more elegant model (Steinberg et al. 2006) than illustrated in our cellular automaton. There is not much information regarding the architecture of the ICC-MY in the small intestine, especially related to the type and number of connections between individual ICCs. Rumessen and coworkers (Rumessen & Thuneberg, 1991; Rumessen & Vanderwinden, 2003) showed that ICC-MY were mostly arranged in bundles of three to seven cells. Park et al. C 2011 The Authors. Journal compilation C 2011 The Physiological Society

Figure 6. Bar graphs of the frequency, conduction velocity and amplitude of the slow waves recorded from the 3- and 7-month control animals (white) and from the 3- and 7-month-diabetic (black) segments from the duodenum, jejunum and ileum There were no statistical differences between any groups in all three segments.

1046



Figure 7. Simulated propagations with various numbers of inactive cells in the three propagation models See main text for further details.

Figure 8. Increase in propagation time with increasing number of randomly distributed inactive cells With the four-neighbour propagation model, propagation time rapidly increased with increasing numbers of inactive cells until propagation block was achieved at 38% inactive cells. The increase in propagation time was much less with the eight- and especially the 24-neighbouring propagation model. The encircled data points are illustrated in Fig. 7. See main text for further details.

imposing an isotropic structure on our present simulation. Lee et al. (2007) also described the complex architecture of the ICC-MY in the human jejunum, which consists of a mixed population of bipolar and multipolar cells and bundles of ICC-MY running both parallel, longitudinal and obliquely, something we tried to emulate in our primitive simulation by varying the number of connections between neighbouring cells. In addition, there is no information on the type of abnormality that may occur in the diabetic small intestine. Loss of ICC-MY could occur in a random fashion at an individual level, or in variable patches that may reach several square millimetres, similar to what is described in the diabetic stomach (Ordög et al. 2000), or possibly even in larger sections, possibly accumulating to a complete circumferential lesion. Defining this level of detail will be necessary to determine the effects of ICC-MY loss on slow-wave propagation and propagation block. Several studies have shown that the effect of diabetes, both in patients (Forster et al. 2005) and in the streptozotocin model, induces differential effects in the C 2011 The Authors. Journal compilation C 2011 The Physiological Society



stomach (Ordög et al. 2000; Wang et al. 2009). In the stomach, the deleterious effects seem to be more pronounced in the antrum than in the corpus with, probably as compensation, activities emerging in the fundic area (Tougas et al. 1992; Ordög et al. 2000). We have found no such spatial variations along the rat small intestine. Both the decrease in ICC-MY at 3 months and the repopulation at 7 months were similar at all three monitored areas, while the slow-wave propagation was not affected in any of these areas. The results in this study indicate that there is not a linear relationship between the number of ICC-MY and slowwave frequency, velocity or amplitude. Furthermore, even though the animals were severely diabetic and did show an initial but transient reduction in ICC-MY, propagation of the slow wave was unaffected. This suggests that, at least in the small intestine, motility based on slow-wave activities may not be as affected in diabetes as some have suggested and seems to shift the onus to other parts of the GI tract within the enteric system and/or the muscle layers. The limitations of the present study should be clear. The immunohistochemical analysis was restricted to the frozen tissue sections without any wholemount preparations. Electrophysiological recordings were performed in isolated segments of the whole intestine that were superfused with Tyrode solution. Finally, the simulation model used is a relatively simple cellular automaton model, and no attempts were made to develop a more realistic model, as have recently appeared in the literature (Buist et al. 2010; Du et al. 2010). Within these constraints, however, we propose that the loss of ICC-MY seen in diabetes has the following characteristics: (a) it is not as large as often is assumed; (b) it is transient; and (c) it is does not have such a great effect on slow-wave generation and propagation as hitherto assumed. These data could be helpful in elucidating a small part of the complex GI dysfunctions in diabetes mellitus. References Belai A & Burnstock G (1990). Changes in adrenergic and peptidergic nerves in the submucous plexus of streptozocin-diabetic rat ileum. Gastroenterology 98, 1427–1436. Buist ML, Corrias A & Poh YC (2010). A model of slow wave propagation and entrainment along the stomach. Ann Biomed Eng 38, 3022–3030. Camilleri M & Malagelada JR (1984). Abnormal intestinal motility in diabetics with the gastroparesis syndrome. Eur J Clin Invest 14, 420–427. Du P, O’Grady G, Gibbons SJ, Yassi R, Lees-Green R, Farrugia G, Cheng LK & Pullan AJ (2010). Tissue-specific mathematical models of slow wave entrainment in wild-type and 5-HT2B knockout mice with altered interstitial cells of Cajal networks. Biophys J 98, 1772–1781. C 2011 The Authors. Journal compilation C 2011 The Physiological Society

1047

Forster J, Damjanov I, Lin Z, Sarosiek I, Wetzel P & McCallum RW (2005). Absence of the interstitial cells of Cajal in patients with gastroparesis and correlation with clinical findings. J Gastrointest Surg 9, 102–108. He CL, Soffer EE, Ferris CD, Walsh M, Szurszewski JH & Farrugia G (2001). Loss of interstitial cells of Cajal and inhibitory innervation in insulin-dependent diabetes. Gastroenterology 121, 427–434. Huizinga JD, Zarate N & Farrugia G (2009). Physiology, injury, and recovery of interstitial cells of Cajal: basic and clinical science. Gastroenterology 137, 1548–1556. Karam SM, Yao X & Forte JG (1997). Functional heterogeneity of parietal cells along the pit-gland axis. Am J Physiol Gastointest Liver Physiol 272, G161–G171. Lammers WJEP, Al-Kais A, Singh S, Arafat K & El-Sharkawy TY (1993). Multielectrode mapping of slow wave activity in the isolated rabbit duodenum. J Appl Physiol 74, 1454–1461. Lammers WJEP, Stephen B, Slack JR & Dhanasekaran S (2002). Anisotropic propagation in the small intestine. Neurogastroenterol Motil 14, 357–364. Lammers WJEP, Ver Donck L, Schuurkes JAJ & Stephen B (2005). Peripheral pacemakers and patterns of slow wave propagation in the canine small intestine in vivo. Can J Physiol Pharmacol 83, 1031–1043. Lammers WJEP, Wit AL & Allessie MA (1987). Effects of anisotropy on functional reentrant circuits: preliminary results of computer simulation studies. In Activation, Metabolism and Perfusion of the Heart, ed. Sideman S & Beyar R, pp. 133–151. Martinus Nijhoff Publishers, Doordrecht. Lee HT, Hennig GW, Fleming NW, Keef KD, Spencer NJ, Ward SM, Sanders KM & Smith TK (2007). The mechanism and spread of pacemaker activity through myenteric interstitial cells of Cajal in human small intestine. Gastroenterology 132, 1852–1865. Mase M, Faes L, Antolini R, Scaglione M & Ravelli F (2005). Quantification of synchronization during atrial fibrillation by Shannon entropy: validation in patients and computer model of atrial arrhythmias. Physiol Meas 26, 911–923. Nowak TV, Harrington B, Kalbfleisch & Amatruda JM (1986). Evidence for abnormal cholinergic neuromuscular transmission in diabetic rat small intestine. Gastroenterology 91, 124–132. Ordög T (2008). Interstitial cells of Cajal in diabetic gastroenteropathy. Neurogastroenterol Motil 20, 8–18. Ordög T, Takayama I, Cheung WKT, Ward SM & Sanders KM (2000). Remodeling of networks of interstitial cells of Cajal in a murine model of diabetic gastroparesis. Diabetes 49, 1731–1739. Park KJ, Hennig GW, Lee HT, Spencer NJ, Ward SM, Smith TK & Sanders KM (2006). Spatial and temporal mapping of pacemaker activity in interstitial cells of Cajal in mouse ileum in situ. Am J Physiol Cell Physiol 290, C1411–C1427. Rumessen JJ & Thuneberg L (1991). Interstitial cells of Cajal in human small intestine. Ultrastructural identification and organization between the main smooth muscle layers. Gastroenterology 100, 1417–1431.

1048


Rumessen JJ & Vanderwinden JM (2003). Interstitial cells in the musculature of the gastrointestinal tract: Cajal and beyond. Int Review Cytol 229, 115–208. Sanders KM, Ordög T, Koh SD, Torihashi S & Ward SM (1999). Development and plasticity of interstitial cells of Cajal. Neurogastroenterol Motil 11, 311–338. Sanders KM, Ordög T, Koh SD & Ward SM (2002). Physiology and pathophysiology of the interstitial cells of Cajal: from bench to bedside. IV. Genetic and animal models of GI motility disorders caused by loss of interstitial cells of Cajal. Am J Physiol Gastrointest Liver Physiol 282, G747–G756. Soffer EE, Thongsawat S, Hoogwerf BJ & Shah A (1999). Effect of acute hyperglycemia on jejunal compliance and peristaltic reflex in healthy humans. Dig Dis Sci 44, 50–55. Steinberg BE, Glass L, Shrier A & Bub G (2006). The role of heterogeneities and intercellular coupling in wave propagation in cardiac tissue. Philos Transact A Math Phys Eng Sci 364, 1299–1311. Tougas G, Hunt RH, Fitzpatrick D & Upton AR (1992). Evidence of impaired afferent vagal function in patients with diabetes gastroparesis. Pacing Clin Electrophysiol 15, 1597–1602.


Wang XY, Huizinga JD, Diamond J & Liu LW (2009). Loss of intramuscular and submuscular interstitial cells of Cajal and associated enteric nerves is related to decreased gastric emptying in streptozotocin-induced diabetes. Neurogastroenterol Motil 21, 1095–e92. Xue L & Suzuki H (1997). Electrical responses of gastric smooth muscles in streptozotocin-induced diabetic rats. Am J Physiol Gastrointest Liver Physiol 272, G77–G83. Yamamoto T, Watabe K, Nakahara M, Ogiyama H, Kiyohara T, Tsutsui S, Tamura S, Shinomura Y & Hayashi N (2008). Disturbed gastrointestinal motility and decreased interstitial cells of Cajal in diabetic db/db mice. J Gastroenterol Hepatol 23, 660–667. Acknowledgements The authors would like to thank Dr M. A. H. Al Sultan and Mr M. Shafiullah for taking care of the animals and Professor John Morrison, who kindly give us the opportunity to sample the small intestine from the animals in his project. The cellular automaton software can be downloaded from www.smoothmap.org software.

C 2011 The Authors. Journal compilation C 2011 The Physiological Society