Release of 5-Hydroxytryptamine From the Mucosa Is Not Required for the Generation or Propagation of Colonic Migrating Motor Complexes

GASTROENTEROLOGY 2010;138:659 – 670

Release of 5-Hydroxytryptamine From the Mucosa Is Not Required for the Generation or Propagation of Colonic Migrating Motor Complexes DAMIEN J. KEATING and NICK J. SPENCER

BACKGROUND & AIMS: The pacemaker mechanism that underlies the cyclic generation of colonic migrating motor complexes (CMMCs) is unknown, although studies have suggested that release of 5-hydroxytryptamine (5-HT) from enterochromaffin cells in the mucosa is essential. However, no recordings of 5-HT release from the colon have been made to support these suggestions. METHODS: We used real-time amperometry to record 5-HT release directly from the mucosa in mouse isolated colon to determine whether 5-HT release from enterochromaffin cells was required for CMMC generation. RESULTS: We found that 5-HT was released from mucosal enterochromaffin cells during many, but not all, CMMC contractions. However, spontaneous CMMCs still were recorded even after removal of the mucosa, and submucosa and submucosal plexus when all release of 5-HT had been abolished. CMMC pacemaker frequency was slower in the absence of the mucosa, an effect reversed by focal application of exogenous 5-HT onto the myenteric plexus. Despite the absence of the mucosa and all detectable release of 5-HT, ondansetron significantly reduced CMMC frequency, suggesting that 5-HT3 receptor blockade slows the CMMC pacemaker via a mechanism independent of 5-HT release from enterochromaffin cells. CONCLUSIONS: Our results show that 5-HT can be released dynamically during CMMCs. However, the intrinsic pacemaker and pattern generator underlying CMMC generation lies within the myenteric plexus and/or muscularis externa and does not require any release of 5-HT from enterochromaffin cells. Endogenous release of 5-HT from enterochromaffin cells plays a modulatory role, not an essential role, in CMMC generation. View this article’s video abstract at www.gastrojournal. org.

C

onsiderable evidence has been presented to suggest that 5-hydroxytryptamine (5-HT) plays an important role in the control of a variety of different patterns of gastrointestinal motility.1– 6 Enterochromaffin cells in the gastrointestinal mucosa synthesize and store the largest quantity (⬎90%) of the serotonin in the body, yet their role in the generation of cyclic motor activities of the gastrointestinal tract remains unclear. There has been

considerable interest in recent years in the functional role of enterochromaffin cells because a number of studies have shown that in a variety of intestinal diseases such as Crohn’s disease and ulcerative colitis, an increased number of enterochromaffin cells populate the intestine7 and altered serotonin signaling has been shown to modify gastrointestinal motility in human patients and laboratory animals.8 –11 Colonic migrating motor complexes (CMMCs) have been recorded from the colon in a variety of mammals,12–14 including human beings,15 but the pacemaker controlling CMMC cycling frequency has eluded researchers. It is clear that the CMMC pacemaker must lie within the colon wall itself because CMMCs occur in an isolated colon. It also is known that the CMMC pacemaker is modified by various pharmacologic agents, in particular 5-HT3 antagonists.16 –18 Recent studies suggest that 5-HT release from enterochromaffin cells is essential for CMMC generation because removal of the mucosa abolished spontaneous CMMCs.18 These conclusions were based, however, on mechanical recordings from smooth muscle, which reveal no information about 5-HT release. Because no real-time recordings have been made of 5-HT release from enterochromaffin cells in the colon, the role of enterochromaffin cells and 5-HT release in CMMC generation remains contentious. Identification of the CMMC pacemaker and an understanding of the role of endogenous 5-HT release in CMMC generation is of supreme clinical importance. 5-HT3 antagonists have been used extensively in human beings as leading therapies to relieve the symptoms of irritable bowel syndrome,19 despite their site of action remaining elusive. In this study, we have used amperometry to record the dynamic release of 5-HT from enterochromaffin cells in the isolated mouse colon to determine whether 5-HT release from enterochromaffin cells is required for the generation and propagation of CMMCs. We show that 5-HT is released from enterochromaffin cells during many, but not all, CMMC contractions. However, when the mucosa and submucosal plexus are removed, Abbreviations used in this paper: CMMC, colonic migrating motor complexes; 5-HT, 5-hydroxytryptamine. © 2010 by the AGA Institute 0016-5085/10/$36.00 doi:10.1053/j.gastro.2009.09.020

BASIC– ALIMENTARY TRACT

Department of Human Physiology, Flinders University, School of Medicine, South Australia, Australia

660

KEATING AND SPENCER

CMMCs remain, albeit at reduced frequency. Thus, the intrinsic pacemaker that entrains CMMC rhythmicity and pattern generation is located within the myenteric plexus and/or muscularis externa and does not require any release of 5-HT from enterochromaffin cells, or the submucosal plexus, as originally thought.

Materials and Methods Preparation of Tissues C57BL/6 mice (20 –90 days old) of either sex were euthanized humanely by inhalation of anesthetic (pentobarbital) followed by cervical dislocation, approved by the animal welfare committee at Flinders University. The entire colon was removed and placed in either ice-cold Krebs solution, or room temperature Krebs solution (see later), both of which were bubbled constantly with carbogen gas (95% O2/5% CO2). A midline incision was made along the mesenteric border and the entire colon was pinned mucosal-side uppermost in a Sylgard-lined (Dow Corning, Midland, Michigan) Petri dish containing oxygenated Krebs solution. The Krebs solution used contained the following: 118 mmol/L NaCl, 4.7 mmol/L KCl, 1.0 mmol/L NaHPO4, 25 mmol/L NaHCO3, 1.2 mmol/L MgCl, 11 mmol/L D-glucose, and 2.5 mmol/L CaCl2. BASIC– ALIMENTARY TRACT

Amperometric Measurements of Serotonin Release From Enterochromaffin Cells Amperometric recordings from the mouse colon were made using the method we have described previously for amperometric recordings made on chromaffin cells in the adrenal medulla,20 the only major difference was that we adjusted the holding potential to ⫹375 mV to selectively detect the oxidation current that can be attributed to release 5-HT only from enterochromaffin cells. Currents caused by the oxidation of 5-HT were recorded using an EPC-7 amplifier (List Medical, Darmstadt, Germany) and Pulse software (HEKA Electronic, Lambrecht/ Pfalz, Germany), sampled at 10 KHz, using a low pass filter of 1 KHz, using an ITC-18 A-D interface (Instrutech Corporation, Great Neck, NY). Carbon fiber electrodes used were mounted on an electronic micromanipulator (Sutter MP-285; Sutter Instruments, Novato, CA), so that the carbon fiber electrode could be lowered repeatedly to within a few microns of a desired depth within the epithelial layer. The whole colon was placed mucosa uppermost in a Sylgard-lined organ bath and perfused continuously with oxygenated Krebs solution that was temperature controlled at 35°C–37°C by using an automatic temperature controller (TC-344B; Warner Instrument Corporation, Hamden, CT). A carbon fiber electrode (5-␮m diameter, ProCFE; Dagan Corporation, Minneapolis, MN) was placed into the epithelial layer and ⫹375 mV was applied to the electrode under voltage clamp conditions. To avoid contact with the mucosa or tissue surface during colonic contractions, carbon fiber electrodes were placed 100 ␮m above the tissue surface. This ensured no 5-HT

GASTROENTEROLOGY Vol. 138, No. 2

release or signal artifacts were evoked by the electrode compressing the tissue.

Simultaneous Mechanical Recordings From the Circular Muscle Combined With Amperometry To correlate direct release of 5-HT from enterochromaffin cells with cyclic CMMCs, it was necessary to record the mechanical activity of the circular muscle in at least 2 independent sites along the whole isolated colon. To do this, we used 2 independent Grass (FT-03C) isometric force transducers (Grass, Quincy, MA) connected via fine suture thread to 2 spring stainless steel claws that anchored to the colon (Figure 1A). A low level of resting tension (5–10 mN) was applied to all preparations. Mechanical recordings were made under near-isotonic conditions, whereby the CM layer could shorten during contraction (Figure 1). The isometric force transducers were connected to 2 custom-made preamplifiers (Biomedical Engineering, Flinders University) and then to a Powerlab (model 4/30; AD Instruments, Bella Vista, New South Wales, Australia).

Measurements and Statistics Data in the results section are presented as means ⫾ SEM. The use of “n” in the Results section refers to the number of animals on which observations were made. Measurements of CMMC half-duration were made between the half-amplitude contraction on the rising phase of a single CMMC, with the half-amplitude point on the recovery phase of the same contraction. Measurements of CMMC propagation velocity were made selectively from only those CMMC contractions that propagated from the oral to anal recordings sites, or from the anal to oral recording sites. Data sets were considered statistically significant if P values less than .05 were reached.

Results Effects of Removal of the Mucosa, the Submucosa, and the Submucosal Plexus on Spontaneous CMMCs Release of 5-HT from enterochromaffin cells has been proposed to be essential for CMMC generation in mouse colon.18 It was therefore of particular interest to us to determine whether CMMCs still would occur when the mucosa was removed from the full length of mouse colon. To test this, we first sharp dissected and removed the mucosa, submucosa, and submucosal plexus from the entire length of mouse colon (Figure 1F). In these dissected preparations (n ⫽ 21), spontaneous CMMCs consistently were preserved, despite the absence of the mucosa, submucosa, and submucosal plexus (Figure 1E). This strongly suggested that the intrinsic pacemaker and pattern generator underlying CMMC generation is not dependent on the mucosa, or release of any substances from enterochromaffin cells. To confirm that in these

5–HT AND CMMCS

661

Figure 1. Removal of the mucosa, submucosa, and submucosal plexus does not prevent spontaneous CMMC generation or propagation. (A) Control preparation with intact mucosa. Simultaneous mechanical recordings were made from the proximal, mid-, and distal colon. (B) Normal CMMC activity. (C) H&E staining shows an intact mucosa, lamina propria, and muscularis externa. (D) Diagrammatic representation of the same recording set up as in panel A, but mechanical recordings were made from preparations devoid of mucosa, submucosa, and submucosal plexus. (E) Spontaneous CMMCs still occur despite removal of these structures. CMMCs occurred at slower frequencies. (F) H&E staining from the same preparation used in panel E, no mucosa, submucosal ganglia, or submucosa are present.

dissected preparations the mucosa and submucosal plexus was removed completely, we stained 6 preparations for H&E. In all preparations, it was clear that the entire mucosa, lamina propria, and submucosal plexus had been removed (Figure 1F) and only consisted of circular and longitudinal muscle with myenteric plexus. In contrast, in intact preparations, the mucosa, lamina propria, and submucosa were preserved and readily identified (Figure 1C). The properties of CMMCs were different after removal of the mucosa and submucosal plexus. The mean interval between spontaneous CMMCs was 1.4 ⫾ 0.1 minutes (range, 0.95–2.8 min; n ⫽ 17; Figure 2A, horizontal line) in intact preparations, compared with the significantly less frequent and more variable interval of 3.6 ⫾ 0.3 minutes (range, 2.0 – 6.6 min; P ⬍ .001; analysis of variance [ANOVA] with Dunn’s method; Figure 2B) in dissected preparations with their mucosa, submucosa, and submu-

cosa removed. A comparative histogram reveals this increased variability in the intervals between CMMCs between the 2 preparations (Figure 2C). We examined whether the propagation velocity or the direction of propagation of CMMCs changed when the mucosa and submucosal plexus were removed. In preparations with mucosa present, the propagation velocity of anally propagating CMMCs (3.8 ⫾ 0.3 mm/sec) was not significantly different from the propagation velocity of anally propagating CMMCs in dissected preparations devoid of mucosa, submucosa, and submucosal plexus (4.8 ⫾ 0.4 mm/sec; P ⫽ .21; 51 CMMCs, n ⫽ 11; ANOVA with Dunn’s method). Similarly, the propagation velocity of orally migrating CMMCs in preparations with the mucosa present was 3.5 ⫾ 0.4 mm/sec, and without the mucosa, submucosa, and submucosal plexus was 5.0 ⫾ 1.0 mm/sec (P ⫽ .08; 9 CMMCs, n ⫽ 5; ANOVA with Dunn’s method). There were significant changes between

BASIC– ALIMENTARY TRACT

February 2010

662

KEATING AND SPENCER

GASTROENTEROLOGY Vol. 138, No. 2

Figure 2. Effects of removal of the mucosa, submucosa, and submucosal plexus on CMMC intervals. (A and B) X-Y scatter plot of the individual intervals between successive CMMC contractions in (A) intact preparations and in (B) preparations with mucosa, submucosa, and submucosal plexus removed. The black horizontal line represents the mean. (C) Bar chart histogram comparison of CMMC intervals obtained from intact preparations (mucosa present) and superimposed with data from preparations with the mucosa, submucosa, and submucosal plexus removed.

BASIC– ALIMENTARY TRACT

the direction of propagation of CMMCs in intact preparations (Figure 3A), compared with dissected preparations with their mucosa, submucosa, and submucosal plexus removed (Figure 3B). The majority of CMMCs in intact preparations originated in the midcolon (68 of 200; 34%) and propagated both orally and anally from their site of origin. This was significantly more than in mucosa-free preparations (28 of 138; 20%; P ⬍ .01, chisquare test). The proportion of CMMCs originating at

the oral end of the colon and propagating through to the anal end was lower in the intact preparations (23%) than in mucosa-free preparations (36%; P ⬍ .01). No difference was seen in the proportion of CMMCs occurring simultaneously, anal-mid-oral, or in an undetermined manner. Interestingly, the amplitudes of CMMCs in the proximal colon were of consistently greater amplitude in mucosa-free dissected preparations (42.1 ⫾ 5.4 mN; n ⫽ 21) compared with preparations with the mucosa and sub-

Figure 3. Differences in characteristics of spontaneous CMMCs after removal of the mucosa, submucosa, and submucosal plexus. (A and B) Direction of propagation of CMMCs. In intact preparations, most CMMCs were initiated in the midcolon, then propagate both orally and anally (34%). (B) In dissected mucosal-free preparations, most CMMCs were initiated in the proximal colon, then propagated anally (36%), whereas 20% were initiated in the midcolon, then propagated orally and anally. (C) Changes in CMMC amplitudes and (D) durations between the 3 regions of colon in intact and mucosa-free preparations.

February 2010

mucosa present (29.4 ⫾ 3.3 mN; n ⫽ 17; P ⫽ .06; Figure 3C). However, in the mid- and distal colon, spontaneous CMMCs were consistently smaller in amplitude in mucosa-free preparations (P ⬍ .001; ANOVA with Tukey test; n ⫽ 21; Figure 3C). The durations of CMMCs in the proximal, midcolon, and distal colon were not significantly different between intact preparations or those devoid of mucosa and submucosal plexus (P ⬎ .05; n ⫽ 17; Figure 3D).

Mucosal Compression Evokes Release of 5-HT From Enterochromaffin Cells

663

ditions, at a holding potential of ⫹375 mV, a resting steady-state concentration of 5-HT was recorded (mean, 22.0 ⫾ 5.8 ␮mol/L; n ⫽ 7), when the carbon fiber electrode was positioned immediately above the mucosal surface, suggesting that high concentrations of 5-HT were being secreted, at least in vitro, from the mucosa. Then, a motorized micromanipulator was used to lower the carbon fiber electrode into the mucosa, to a depth of 100 ␮m, to determine whether enterochromaffin cells released increasing concentrations of 5-HT in response to mechanical stimuli. We found that as soon as the electrode penetrated the mucosa, a transient 5-HT oxidation current was detected that equated to a peak concentration of release of 119 ⫾ 26 ␮mol/L (n ⫽ 7). This transient peak in release was followed by a sustained release of 5-HT (mean concentration, 45.6 ⫾ 5.6 ␮mol/L; n ⫽ 7).

BASIC– ALIMENTARY TRACT

Next, we sought to determine whether enterochromaffin cells released 5-HT under resting in vitro conditions and whether this release may be necessary for the control of CMMC activity. Under voltage clamp con-

5–HT AND CMMCS

Figure 4. Amperometric recordings of 5-HT release from enterochromaffin cells during mucosal compression. (A) Increasing concentrations of exogenous 5-HT generate increasing oxidation currents at the carbon fiber electrode. (B) At a holding potential of ⫹375 mV, mucosal compression (100 ␮mol/L) into the epithelial layer generated a peak oxidation current. During maintained compression of the electrode, switching holding potentials back to 0 mV abolished the current. The current was restored when the holding potential was stepped back to ⫹375 mV. (C) Upper panel, the transient and sustained oxidation current generated by initial contact of the electrode into the epithelium. Lower panel, at a holding potential of 0 mV, the same mucosal compression stimulus failed to evoke any current. (D) Results showing that at ⫹375 mV, the peak transient current was 119 ␮mol/L and the sustained current was 45 ␮mol/L, whereas at a holding potential of 0 mV no current was activated during compression.

664

KEATING AND SPENCER

GASTROENTEROLOGY Vol. 138, No. 2

BASIC– ALIMENTARY TRACT

Figure 5. Amperometric recordings reveal cyclic release of 5-HT from enterochromaffin cells during spontaneous CMMCs. (A) Diagrammatic representation of the recording set up for simultaneous mechanical recordings and amperometric recordings. (B) Spontaneous anally propagating CMMCs are shown where the CMMC contraction starts in the proximal colon and propagates to the distal colon. As the CMMC propagates past the midcolon, the carbon fiber electrode detects a release of 5-HT.

These peaks in evoked release were assumed to represent the summed release of 5-HT from a local population of enterochromaffin cells that were distorted mechanically and inhabited the vicinity of the carbon fiber electrode tip. To test that the amperometric signal generated by mucosal compression was not caused by deformation of the electrode as it entered the epithelium, we applied the same compression stimulus at a different site along the mucosa, but this time at a holding potential of 0 mV (Figure 4). At this potential 5-HT is not oxidized.1 It was found that by applying the same mucosal compression stimulus, but at a holding potential of 0 mV, no oxidation current was detected, confirming that the current generated was not caused by distortion of the electrode itself (Figure 4), but rather by oxidation of 5-HT because it was released from enterochromaffin cells.

Amperometric Recordings of 5-HT Release From Enterochromaffin Cells During Spontaneous CMMCs Because 5-HT release from enterochromaffin cells has been inferred as being essential for CMMC generation,18 we wished to investigate further the role of 5-HT

in CMMCs and further confirm that our dissection had indeed ablated 5-HT release from these preparations. To test this, intact whole segments of mouse colon were pinned as sheet preparations with the mucosa facing uppermost (Figure 5A). In this recording configuration, mechanical recordings were made under near-isotonic conditions, whereby the CM layer could shorten during contraction. In all preparations studied, where CMMC contractions could be seen clearly to have propagated between the proximal and distal recording sites, we placed the carbon fiber electrode on the mucosal surface. We found that in 93% of CMMC contractions that propagated between the proximal and distal recording sites, the contractions could be correlated temporally with a release of 5-HT from the midcolon (Figure 5B). The maximum peak release of 5-HT measured during CMMC contractions was 1.53 ⫾ 0.08 mmol/L (range, 0.4 –3.1 mmol/L; 66 CMMCs; n ⫽ 7; Supplementary Figure 1), with a mean duration of release of 17.9 ⫾ 1.4 seconds (range, 1.1– 47.9 s; 66 CMMCs; n ⫽ 7). Each release event likely represented the summed activity of multiple release sites from many neighboring enterochromaffin cells be-

cause no discrete unitary currents were resolved. Hexamethonium (300 ␮mol/L) immediately abolished CMMCs and the associated cyclic release of 5-HT (data not shown).

Amperometric Recordings From the Mouse Colon After Removal of the Mucosa, Submucosa, and Submucosal Plexus Our results suggest that the mucosa is not required for CMMC generation or propagation. However, we sought to ensure, at the same time as when CMMCs still were recorded, whether all detectable release of 5-HT had been prevented in mucosa-free preparations. We therefore sharp-dissected off the mucosa, submucosa, and submucosal plexus from the entire full length of colon, and recorded mechanical activity from the proximal and distal colon, while simultaneously recording amperometry from the circular muscle layer of the midcolon (Figure 6A). From these preparations, we could not detect any resting or background concentration of 5-HT in the organ bath (0.0 ⫾ 0.0 pA; n ⫽ 7; Figure 6B).

Effects of Focal Application of 5-HT on CMMC Pacemaker Activity Because CMMCs occurred at slower frequencies when the mucosa was removed from the colon (Figure 2), we sought to determine if CMMC frequency would in-

5–HT AND CMMCS

665

crease in mucosa-free preparations if 5-HT was spritzed onto the myenteric plexus and circular muscle. We found that spritzing 5-HT (50 ␮L of 100 ␮mol/L; Figure 7A) onto the circular muscle caused an instantaneous increase in CMMC frequency to that seen in intact preparations (Figure 7B; n ⫽ 4; P ⬍ .05). In intact preparations (mucosa present), when the same stimulus was applied directly to the mucosa, there was no change in CMMC frequency (Figure 7C; n ⫽ 3; P ⬎ .05), suggesting that exogenous 5-HT does not readily cross the mucosal border in intact preparations and has little access to the CMMC pacemaker in the myenteric plexus.

Effects of 5-HT3 Receptor Blockade on Spontaneous CMMCs in Preparations Devoid of Mucosa, Submucosa, and Submucosal Plexus In previous studies on intact isolated whole mouse colon, the 5-HT3 antagonist ondansetron, at a concentration of 1 ␮mol/L, has been shown to potently decrease the pacemaker frequency of CMMCs.16,17 The assumption drawn from such studies has been that ondansetron acts by blocking 5-HT3 receptors on enterochromaffin cells, thereby decreasing the release of 5-HT, leading to an inhibition of CMMC frequency.16 –18 However, in light of our findings that the removal of the mucosa did not prevent CMMC generation, we were

Figure 6. Effects of removal of the mucosa, submucosal plexus, and submucosa on spontaneous CMMC characteristics. (A) Diagrammatic representation of the recording technique. The carbon fiber electrode was lowered into the circular muscle layer of the midcolon, while mechanical recordings were made from the proximal and distal colon. (B) After removal of the mucosa, submucosa, and submucosal plexus, spontaneous CMMCs still were recorded, even though no background or cyclic release of 5-HT could be detected.

BASIC– ALIMENTARY TRACT

February 2010

666

KEATING AND SPENCER

GASTROENTEROLOGY Vol. 138, No. 2

BASIC– ALIMENTARY TRACT

Figure 7. Exogenous 5-HT spritzed onto the myenteric plexus and circular muscle potently increases CMMC pacemaker frequency. (A) Diagrammatic representation of the preparation. A spritz pipette was used to apply exogenous 5-HT onto the circular muscle and myenteric plexus, in a preparation with the mucosa, submucosa, and submucosal plexus removed. (B) Left-hand side of trace shows spontaneous CMMCs. When 50 ␮L of 5-HT (100 ␮mol/L) was puffed onto the circular muscle layer, an immediate and sustained increase in CMMC pacemaker activity occurs. (C) When 5-HT was spritzed onto the mucosa of intact preparations there was no change in CMMC interval. In mucosa-free preparations, 5-HT spritzed onto the circular muscle caused a significant decrease in CMMC interval.

interested in whether ondansetron still would decrease the pacemaker frequency of CMMCs in mucosa-free preparations. In such preparations where no basal or cyclic release of 5-HT was recorded using amperometry, ondansetron (1 ␮mol/L) significantly decreased the frequency of CMMCs (Figure 8B) by 196% from a mean control interval of 2.8 ⫾ 0.3 to 5.5 ⫾ 0.6 minutes (n ⫽ 7; P ⫽ .003; Figure 8A), and decreased CMMC amplitudes in the proximal colon from 71.8 ⫾ 17.0 to 39.8 ⫾ 7.8 mN (n ⫽ 7; P ⫽ .04). In intact preparations, with the mucosa present, ondansetron (1 ␮mol/L) decreased CMMC frequencies by 280% from 1.5 ⫾ 0.1 to 4.3 ⫾ 0.7 minutes (Figure 8A; n ⫽ 6; P ⫽ .006), where CMMC amplitudes were reduced by 21% from 38.1 ⫾ 4.0 to 30.0 ⫾ 9.9 mN (n ⫽ 6; P ⫽ .45). At 3 ␮mol/L, ondansetron abolished CMMCs (n ⫽ 3).

Discussion In this study, we sought to identify the location of the intrinsic pacemaker and neurogenic pattern generator that underlies the cyclic generation of CMMCs in the mouse colon; and whether the release of 5-HT from enterochromaffin cells is essential for CMMC generation, as has been suggested. We identified that the intrinsic

pacemaker and pattern generator underlying CMMC generation is located within the myenteric plexus and/or muscularis externa, and does not require any release of 5-HT from enterochromaffin cells, or the presence of the mucosa or submucosal plexus. The major properties of spontaneous CMMCs were remarkably unaltered after the removal of these structures. Also, we show that the mechanism by which 5-HT3 antagonists slow the CMMC pacemaker frequency does not require enterochromaffin cells, or suppression of 5-HT release from enterochromaffin cells, as has been assumed.

5-HT Can Be Released From Enterochromaffin Cells During Spontaneous CMMCs and by Mucosal Compression We were particularly interested in whether 5-HT was released from enterochromaffin cells during spontaneously occurring CMMCs and, if so, what was the functional role of this release. We have several reasons to conclude that our amperometric recordings were detecting Ca2⫹-dependent 5-HT exocytosis from enterochromaffin cells. First, at the holding potential of ⫹375 mV, 5-HT is oxidized whereas other mucosally released compounds such as melatonin and tyramine are not.21 Sec-

5–HT AND CMMCS

667

Figure 8. Effects of ondansetron on CMMC activity recorded from a colonic preparation devoid of mucosa, submucosa, and submucosal plexus. (A) Ondansetron (1 ␮mol/L) significantly increased the intervals between CMMCs. (B) Typical recording from a dissected preparation (devoid of mucosa, submucosa, and submucosal plexus), showing that initially ondansetron (1 ␮mol/L) disrupted CMMCs, then potently increased the interval between CMMCs, even though the mucosa and submucosal ganglia were absent. No release of 5-HT could be detected in these preparations.

ond, in similar conditions, the addition of the serotonin transporter antagonist fluoxetine significantly increases the oxidizable current.22,23 Finally, removal of extracellular Ca2⫹ reduces this current in line with the occurrence of Ca2⫹-dependent 5-HT exocytosis.23 When simultaneous amperometric and mechanical recordings were made, we found that most CMMCs (93%) could be correlated temporally with a release of 5-HT, which reached surprisingly high concentrations, up to 1.5 mmol/L. Interestingly, Bertrand1 used amperometry to record the dynamic release of 5-HT from enterochromaffin cells during fluid-filled peristalsis in the guinea-pig small intestine. He found that during peristalsis, 5-HT concentrations peaked at approximately 5 ␮mol/L and the peak 5-HT release evoked by mucosal compression was approximately 15 ␮mol/L.1 In the mouse colon we detected considerably higher peak concentrations of 5-HT released during mucosal compression, in the order of approximately 120 ␮mol/L. The higher concentrations of 5-HT recorded from the mouse colon may be a reflection of many factors, including greater density of enterochro-

maffin cells in the colon, greater mechanosensitivity of the release mechanisms for 5-HT in the colon, and species differences. These results imply that considerably greater quantities of 5-HT are released from enterochromaffin cells in the mouse colon compared with the guinea-pig ileum.

What Is the Role of 5-HT Release From Enterochromaffin Cells During CMMCs? The functional role of the high concentration of 5-HT that can be released from enterochromaffin cells during CMMC contractions remains unclear. An initial impression was that release of 5-HT from enterochromaffin cells might act to stimulate mucosally projecting Dogiel type II neurons, which have been shown to project into the mucosa of the mouse colon.24 These then might activate, or amplify, the underlying intrinsic neural circuitry that generates CMMCs. This hypothesis seemed plausible because at least in guinea-pigs a population of mucosally projecting Dogiel type II neurons behave functionally as intrinsic sensory neurons25,26 that respond to

BASIC– ALIMENTARY TRACT

February 2010

668

KEATING AND SPENCER

BASIC– ALIMENTARY TRACT

luminally applied acid and exogenous 5-HT26 and have nerve endings that lie in close apposition to enterochromaffin cells.27 However, our data in mouse colon suggests that nerve endings of mucosally projecting neurons are not required for CMMC generation, nor is the release of any substances from enterochromaffin cells in the mucosa. This strongly suggests that the intrinsic neural circuitry underlying CMMC generation and propagation does not require activation of the nerve endings of mucosally projecting Dogiel type II neurons (Supplementary Figure 2). This study illustrates that one role of 5-HT release from enterochromaffin cells is to modulate the frequency at which CMMCs occur, even though this release is clearly not a requirement for CMMC generation or propagation. Given the excitatory effect of spritzed 5-HT onto the myenteric plexus, and the inhibitory effect of ondansetron on CMMCs in mucosa-free preparations, further attention now needs to be paid to the role of the recently identified 5-HT– containing interneurons in the mouse colon28 because these neurons may be potential intrinsic pacemakers of CMMC activity. Our results show that 5-HT is released by mucosal distortion and during some, but not all, spontaneous CMMC contractions. Thus, it might be expected that a moving fecal pellet would induce 5-HT release in vivo, and this 5-HT would be transported out of the lumen by the serotonin transporter,22,23 where it might modulate the firing of the intrinsic pacemaker, which must lie in the myenteric plexus.

How Do the Properties of CMMCs Change When 5-HT Release From the Mucosa Has Been Prevented? One of the differences between intact preparations and dissected preparations that had the mucosa, submucosa, and submucosal plexus removed was that the frequency of CMMCs was slower in preparations devoid of mucosa. The combination of our dissection, 5-HT spritzing, and ondansetron results strongly imply that 5-HT from the mucosa is not essential for CMMC generation, but rather acts as a modulator of CMMC frequency. Interestingly, CMMCs in this study occurred considerably faster (mean interval, ⬃1.1 min) compared with all other previous reports of CMMC intervals measured from intact tube-type preparations, which occurred at mean intervals of approximately 3.6 minutes.14,29,30 This is likely owing to the cut sheet preparations we used in this study, which allow the myenteric ganglia greater access to the 5-HT in the organ bath compared with tube preparations. In dissected, mucosa-free preparations, the amplitudes of CMMCs were larger in the proximal colon compared with control, and smaller in the mid- and distal colon. The reason for this is unclear. It may be owing to dissection damage, although if this is the case it is not clear why CMMCs in the proximal colon were consistently larger in mucosa-free preparations. One rea-

GASTROENTEROLOGY Vol. 138, No. 2

son may be the intriguing coincidence we observe between the most common site of CMMC origin in both preparations—midcolon in intact preparations and proximal colon in dissected preparations, and the greatest CMMC amplitude. The propagation velocities of orally and anally propagating CMMCs were no different after removal of the mucosa, or submucosal plexus, suggesting that myenteric neural circuitry controls the rate at which CMMCs migrate along the colon, not 5-HT release from enterochromaffin cells.

Effects of 5-HT Application to the Mucosa and Myenteric Plexus A major observation was that spritzing exogenous 5-HT onto the myenteric plexus and circular muscle increased the frequency of CMMCs (Figure 7). However, when the same stimulus was applied to the mucosa, there was no change in CMMCs. This suggests that the control of the CMMC pacemaker frequency is modulated primarily within the myenteric plexus and not at the level of the mucosa as previously thought and that exogenous 5-HT applied to the mucosa probably has little access to penetrate the mucosa and reach the myenteric plexus. It has been shown that 5-HT applied to the mucosa activates the nerve endings of mucosally projecting Dogiel type II neurons.31 Therefore, our observation that 5-HT had no effect when applied to the mucosa is consistent with our conclusion that activation of mucosally projecting neurons is not essential for CMMC generation.

Where Do 5-HT3 Antagonists Act to Slow the Pacemaker Frequency of CMMCs ? A number of studies have shown that 5-HT3– receptor antagonists potently slow CMMC frequency in rodents16 –18 and similarly retard colonic transit in human beings, making them highly effective in treating diarrhea-predominant irritable bowel syndrome. However, the site of action of 5-HT3 antagonists in slowing CMMC frequencies and decreasing colonic transit has remained unclear. Because the vast majority of 5-HT in the body is synthesized and stored in enterochromaffin cells within the mucosa and because enterochromaffin cells express 5-HT3 receptors, 5-HT3 antagonists were thought to act by inhibiting release of 5-HT from enterochromaffin cells. Interestingly, in preparations devoid of mucosa and where no release of 5-HT was detected, ondansetron still reduced the frequency of CMMCs (Figure 8), suggesting that the inhibitory effect of 5-HT3 antagonists on CMMC pacemaker frequency is not mediated via the suppression of 5-HT release from enterochromaffin cells. More likely ondansetron slows CMMCs by suppressing fast synaptic transmission at 5-HT3 receptors within the myenteric plexus.32

February 2010

We found that the myenteric plexus in the mouse colon was exquisitely sensitive to dissection damage. In the process of removing the mucosa, submucosa, and submucosal plexus, even the slightest damage to the myenteric plexus prevented the spontaneous generation of CMMCs. We noted that if the mucosa and submucosa were physically peeled off the underlying myenteric plexus, spontaneous CMMCs usually failed to occur, or, at least they failed to propagate past the site of the dissection damage. However, we found that we consistently could preserve the pacemaker and pattern generator underlying spontaneous CMMCs if the mucosa, submucosa, and submucosal plexus were delicately cut off of the myenteric plexus, with fine microscissors to avoid any tearing or pulling of the underlying myenteric ganglia. A recent study18 reported that spontaneous CMMCs were abolished in the mouse colon when the mucosa was removed from the colon. It was not stated how these investigators removed the mucosa from their preparations, and no amperometric or histologic measurements were made. There is evidence of differences between different strains of mice,28 however, this could not have accounted for the opposite results obtained between our study and that of Heredia et al18 because the same age and strain of mice were used and the same method of euthanasia.

Conclusions The findings of the current study show that in the mouse colon, 5-HT can be released from enterochromaffin cells during spontaneously occurring CMMCs. However, this release of 5-HT is not required for CMMC generation or propagation because CMMCs persisted after removal of the mucosa and blockade of all release of 5-HT from enterochromaffin cells. Rather, 5-HT release from enterochromaffin cells appears to modulate the regularity of timing of CMMCs. Taken together, these results lead to the inescapable conclusion that the intrinsic pacemaker and pattern generator underlying CMMC generation and propagation is located within the myenteric plexus and/or muscularis externa (Supplementary Figure 2). We suggest that the cyclic generation of CMMCs requires cyclic neurogenic activity within the myenteric plexus, and not basal or cyclic release of 5-HT from enterochromaffin cells.

Supplementary Materials Note: To access the supplementary material accompanying this article, visit the online version of Gastroenterology at www.gastrojournal.org, and at doi: 10.1053/j.gastro.2009.09.020.

669

References 1. Bertrand PP. Real-time measurement of serotonin release and motility in guinea pig ileum. J Physiol 2006;577:689 –704. 2. Michel K, Zeller F, Langer R, et al. Serotonin excites neurons in the human submucous plexus via 5-HT3 receptors. Gastroenterology 2005;128:1317–1326. 3. Mawe GM, Coates MD, Moses PL. Review article: intestinal serotonin signalling in irritable bowel syndrome. Aliment Pharmacol Ther 2006;23:1067–1076. 4. Furness JB. The enteric nervous system. Oxford, UK: Blackwell Publishing, 2006. 5. Galligan JJ. Actions of sumatriptan on myenteric neurones: relief from an old headache in the enteric nervous system? Neurogastroenterol Motil 2007;19:1–3. 6. Grundy D, Schemann M. Enteric nervous system. Curr Opin Gastroenterol 2007;23:121–126. 7. Bishop AE, Pietroletti R, Taat CW, et al. Increased populations of endocrine cells in Crohn’s ileitis. Virchows Arch A Pathol Anat Histopathol 1987;410:391–396. 8. Lomax AE, Linden DR, Mawe GM, et al. Effects of gastrointestinal inflammation on enteroendocrine cells and enteric neural reflex circuits. Auton Neurosci 2006;126 –127:250 –257. 9. O’Hara JR, Lomax AE, Mawe GM, et al. Ileitis alters neuronal and enteroendocrine signalling in guinea pig distal colon. Gut 2007; 56:186 –194. 10. Krauter EM, Strong DS, Brooks EM, et al. Changes in colonic motility and the electrophysiological properties of myenteric neurons persist following recovery from trinitrobenzene sulfonic acid colitis in the guinea pig. Neurogastroenterol Motil 2007;19:990 – 1000. 11. Banerjee S, Akbar N, Moorhead J, et al. Increased presence of serotonin-producing cells in colons with diverticular disease may indicate involvement in the pathophysiology of the condition. Int J Colorectal Dis 2007;22:643– 649. 12. Spencer NJ. Control of migrating motor activity in the colon. Curr Opin Pharmacol 2001;1:604 – 610. 13. Ro S, Hwang SJ, Muto M, et al. Anatomic modifications in the enteric nervous system of piebald mice and physiological consequences to colonic motor activity. Am J Physiol Gastrointest Liver Physiol 2006;290:G710 –G718. 14. Roberts RR, Bornstein JC, Bergner AJ, et al. Disturbances of colonic motility in mouse models of Hirschsprung’s disease. Am J Physiol Gastrointest Liver Physiol 2008;294:G996 –G1008. 15. Hagger R, Kumar D, Benson M, et al. Periodic colonic motor activity identified by 24-h pancolonic ambulatory manometry in humans. Neurogastroenterol Motil 2002;14:271–278. 16. Fida R, Bywater RA, Lyster DJ, et al. Chronotropic action of 5-hydroxytryptamine (5-HT) on colonic migrating motor complexes (CMMCs) in the isolated mouse colon. J Auton Nerv Syst 2000; 80:52– 63. 17. Bush TG, Spencer NJ, Watters N, et al. Effects of alosetron on spontaneous migrating motor complexes in murine small and large bowel in vitro. Am J Physiol Gastrointest Liver Physiol 2001; 281:G974 –G983. 18. Heredia DJ, Dickson EJ, Bayguinov PO, et al. Localized release of serotonin (5-Hydroxytryptamine) by a fecal pellet regulates migrating motor complexes in murine colon. Gastroenterology 2009;136:1328 –1338. 19. Viramontes BE, Camilleri M, McKinzie S, et al. Gender-related differences in slowing colonic transit by a 5-HT3 antagonist in subjects with diarrhea-predominant irritable bowel syndrome. Am J Gastroenterol 2001;96:2671–2676. 20. Keating DJ, Chen C. Activin A stimulates catecholamine secretion from rat adrenal chromaffin cells: a new physiological mechanism. J Endocrinol 2005;186:R1–R5.

BASIC– ALIMENTARY TRACT

Importance of the Dissection Process in Preserving Spontaneous CMMCs

5–HT AND CMMCS

670

KEATING AND SPENCER

BASIC– ALIMENTARY TRACT

21. Bertrand PP. Real-time detection of serotonin release from enterochromaffin cells of the guinea-pig ileum. Neurogastroenterol Motil 2004;16:511–514. 22. Bertrand PP, Hu X, Mach J, et al. Serotonin (5-HT) release and uptake measured by real-time electrochemical techniques in the rat ileum. Am J Physiol Gastrointest Liver Physiol 2008;295: G1228 –G1236. 23. Bian X, Patel B, Dai X, et al. High mucosal serotonin availability in neonatal guinea pig ileum is associated with low serotonin transporter expression. Gastroenterology 2007;132:2438 –2447. 24. Furness JB, Robbins HL, Xiao J, et al. Projections and chemistry of Dogiel type II neurons in the mouse colon. Cell Tissue Res 2004;317:1–12. 25. Bornstein JC, Furness JB, Kunze WA. Electrophysiological characterization of myenteric neurons: how do classification schemes relate? J Auton Nerv Syst 1994;48:1–15. 26. Kunze WA, Bornstein JC, Furness JB. Identification of sensory nerve cells in a peripheral organ (the intestine) of a mammal. Neuroscience 1995;66:1– 4. 27. Furness JB, Jones C, Nurgali K, et al. Intrinsic primary afferent neurons and nerve circuits within the intestine. Prog Neurobiol 2004;72:143–164. 28. Neal KB, Parry LJ, Bornstein JC. Strain-specific genetics, anatomy and function of enteric neural serotonergic pathways in inbred mice. J Physiol 2009;587:567–586. 29. Fida R, Lyster DJ, Bywater RA, et al. Colonic migrating motor complexes (CMMCs) in the isolated mouse colon. Neurogastroenterol Motil 1997;9:99 –107. 30. Spencer NJ, Bywater RA. Enteric nerve stimulation evokes a premature colonic migrating motor complex in mouse. Neurogastroenterol Motil 2002;14:657– 665. 31. Bertrand PP, Kunze WA, Furness JB, et al. The terminals of myenteric intrinsic primary afferent neurons of the guineapig ileum are excited by 5-hydroxytryptamine acting at 5-hydroxytryptamine-3 receptors. Neuroscience 2000;101:459 – 469.

GASTROENTEROLOGY Vol. 138, No. 2

32. Galligan JJ, Parkman H. Recent advances in understanding the role of serotonin in gastrointestinal motility and functional bowel disorders. Neurogastroenterol Motil 2007;19(Suppl 2):1– 4. 33. Spencer NJ, Hennig GW, Dickson E, et al. Synchronization of enteric neuronal firing during the murine colonic MMC. J Physiol 2005;564:829 – 847. 34. Spencer NJ, Bayguinov P, Hennig GW, et al. Activation of neural circuitry and Ca2⫹ waves in longitudinal and circular muscle during CMMCs and the consequences of rectal aganglionosis in mice. Am J Physiol Gastrointest Liver Physiol 2007;292:G546 – G555.

Received July 14, 2009. Accepted September 16, 2009. Reprint requests Address requests for reprints to: Nick Spencer, PhD, Senior Lecturer, Department of Human Physiology, School of Medicine, Flinders University, South Australia, Australia. e-mail: nicholas.spencer@flinders.edu.au; fax: (61) 8-8204-5768. Acknowledgments The authors would like to acknowledge the outstanding technical assistance provided by Melinda Kyloh. Damien J. Keating and Nick J. Spencer contributed equally in experimental design, analysis, and writing the manuscript. Conflicts of interest The authors disclose no conflicts. Funding This work was funded by a grant from the National Health and Medical Research Council in Australia (grants 535034 and 535033 to N.S., and grant 441112 to D.J.K.). D.J.K. is also supported by a BioInnovation SA Fellowship.

February 2010

5–HT AND CMMCS

670.e1

Supplementary Figure 1. Characteristics of 5-HT release events associated with propagating CMMC contractions. (A) A histogram showing the peak 5-HT concentration reached during CMMCs. (B) Histogram shows the range of durations of 5-HT release associated with individual CMMC contractions. (C) Histogram showing the amount of charge (in picocoloumbs, pC) detected at the electrode during CMMC-induced 5-HT release.

670.e2

KEATING AND SPENCER

GASTROENTEROLOGY Vol. 138, No. 2

Supplementary Figure 2. Diagrammatic representation of the neural circuitry underlying CMMC generation. The intrinsic neural circuitry underlying CMMC generation has been shown to consist of the synchronous firing of ascending and descending interneurons,33 leading to simultaneous activation of orally projecting excitatory motor neurons and anally-projecting inhibitory motor neurons to the circular muscle.34 5-HT is released from enterochromaffin cells during CMMCs, but is not required for CMMC generation. Submucosal ganglia or activation of nerve endings in the mucosa is not required for spontaneous CMMC generation.