Effect of a novel 5-HT3 receptor agonist MKC-733 on upper gastrointestinal motility in humans
Professor R. C. Spiller, Division of Gastroenterology, University Hospital, Nottingham, NG7 2UH, UK.
Background : Although 5-HT3 antagonists have been used to treat chemotherapy-induced emesis and diarrhoea-predominant irritable bowel syndrome, the effects of 5-HT3 agonists in humans are unknown.
Aim : To determine the effect of MKC-733, a selective 5-HT3 receptor agonist, on upper gastrointestinal motility.
Methods : Oral MKC-733 (0.2, 1 and 4 mg) was compared with placebo in three randomized, double-blind, cross-over studies in healthy males. Antroduodenal manometry was recorded for 8 h during fasting and 3 h post-prandially (n = 12). Gastric emptying and small intestinal transit were determined by gamma-scintigraphy (n = 16). Gastric emptying, accommodation and antral motility were determined by echoplanar magnetic resonance imaging (n = 12).
Results : MKC-733 (4 mg) increased the number of migrating motor complexes recorded in the antrum and duodenum (P < 0.001), but had no effect on post-prandial motility. MKC-733 delayed scintigraphically assessed liquid gastric emptying (P = 0.005) and accelerated small intestinal transit (P = 0.038). Echoplanar magnetic resonance imaging confirmed the delayed gastric emptying (P < 0.001) and demonstrated a significant increase in cross-sectional area of the proximal stomach (P < 0.01).
Conclusions : MKC-733 delays liquid gastric emptying in association with relaxation of the proximal stomach, stimulates fasting antroduodenal migrating motor complex activity and accelerates small intestinal transit.
Serotonin (5-hydroxytryptamine, 5-HT) has potent effects on gastrointestinal motor and secretory function that are mediated through a large family of receptor subtypes. 5-HT3 receptors are present on many enteric neurones,1 the mucosal terminals of extrinsic afferent neurones,2 and also neurones in the brain, spinal cord, heart, blood vessels and skin.3 5-HT3 antagonists inhibit small intestinal4 and pancreatic5 secretion, delay large bowel transit,6 inhibit the gastro-colonic response7 and improve pain and discomfort in diarrhoea-predominant irritable bowel syndrome.8 They have little effect on gastric emptying9 or accommodation,10 but have been shown to inhibit the gastric migrating motor complex (MMC).11 By contrast, little is known about the effects of 5-HT3 agonists, such as 2-methyl 5-hydroxytryptamine, on the human gastrointestinal tract. MKC-733 [(R)-N-(3-quinuclidinyl)-7-oxo-4,7-dihydrothieno[3,2b]pyridine-6-carboxamide hydrochloride] is an orally active, selective 5-HT3 agonist. In vitro studies have shown that MKC-733 has a high affinity for canine intestinal smooth muscle 5-HT3 receptors.12 MKC-733 dose dependently stimulates contraction13 and basal electrolyte secretion14 in the isolated guinea pig colon, effects that are antagonized by the selective 5-HT3 antagonist ondansetron. In some animal models, MKC-733 may also accelerate gastric emptying.15 These preliminary studies suggest that 5-HT3 agonists may accelerate gastrointestinal transit; however, there are important species differences in 5-HT receptor pharmacology, making it important to confirm these findings in humans. Our aim was to investigate, in humans, the effects of this 5-HT3 agonist on: (i) the gastric and small intestinal transit of a test meal; (ii) the fed and fasted antral motility; and (iii) the gastric accommodation; the safety and tolerability of MKC-733 were also assessed.
Forty healthy male subjects, aged between 18 and 45 years, were recruited by advertisement. Eligible subjects had no previous history of significant illness or gastrointestinal surgery and were taking no regular medication. Physical examination and resting electrocardiogram were normal and serum biochemistry, haematology and urinalysis were within the normal ranges for our laboratory. Additional inclusion criteria included a body mass index (BMI) of 19–29 kg/m2, alcohol consumption of no more than 28 units per week (one unit of alcohol equals 1/2 pint of beer or lager, one glass of wine or 1/6 gill of spirits) and cigarette consumption of less than five per week. Written consent was obtained from each subject. The study was approved by the Ethics Committee of the University of Nottingham and conducted according to the principles of the Declaration of Helsinki, 1996.
The study was performed in the Departments of Surgery, Medical Physics and Magnetic Resonance at the University of Nottingham between November 1999 and April 2000. The effects of MKC-733 on upper gastrointestinal motility were assessed in three separate, double-blind, randomized, placebo-controlled, cross-over studies using antroduodenal manometry (n = 12), gamma-scintigraphy (n = 16) and echoplanar magnetic resonance imaging (n = 12). No over-the-counter or prescribed medications were allowed for 7 days before the study. During the 48 h before each study day, subjects avoided strenuous physical activity, alcohol and caffeine-containing foods. Following an overnight fast for approximately 10 h, the subjects were treated in the morning at the start of the study. Each subject received placebo and three different doses of oral MKC-733 (0.2 mg, 1 mg and 4 mg), which were chosen on the basis of efficacy data demonstrating accelerated gastric emptying in dogs at doses of 0.01–0.1 mg/kg (equivalent to 0.7–7 mg in a 70-kg man) and phase I studies demonstrating the safety and tolerability of single doses up to 4 mg. The study medications were given under direct observation in a blind, randomized order on four separate study days approximately 1 week apart. All medications appeared identical and were packaged in numbered containers by personnel not involved in the study according to a computer-generated pseudo-randomization procedure (Hartington Statistics and Data Management Ltd, London, UK). All data processing was performed blind with regard to treatment before breaking the randomization code.
End-points and sample size
Manometric study. The primary end-point during fasting was the motility index of antral contractions during phase II of the MMC. Our previous measurements in healthy subjects suggested a motility index of 102 ± 50 mmHg.min (mean ± standard deviation). By using 12 subjects, it was possible to detect a doubling of the motility index with a power of 90%. The secondary end-points were the periodicity of the MMC, the number and amplitude of antral contractions during phase II and the mean durations of each phase. The primary end-point during the post-prandial phase was the motility index of antral contractions and the secondary end-points were the number and amplitude of antral contractions.
Scintigraphic study. The primary end-point was the half-time of gastric emptying (T50%) of the solid meal from the whole stomach. In previous studies, using a similar meal, the T50% value in healthy males was found to be 129 ± 22 min. By using 16 subjects, it was possible to detect a difference of emptying time of 40 min with a power of 90%. The secondary end-points were the T50% value of gastric emptying of the liquid meal from the whole stomach, the T50% values of gastric emptying of the solid and liquid meals from the proximal stomach and the small intestinal transit time of the solid and liquid meals.
Echoplanar magnetic resonance imaging study. The primary end-point was the T50% value of gastric emptying. In previous studies, the mean T50% value of a similar meal in healthy volunteers was 76 ± 21 min. By using 12 subjects, it was possible to detect a difference of emptying time of 40 min with a power of 90%. The secondary end-points were the speed and frequency of antral contractions and the cross-sectional area of the proximal stomach 30 min after the meal.
Study methods. Antroduodenal manometry was recorded using a six-channel, solid state, transducer catheter with a weighted tip (Gaeltec, Dunvegan, Isle of Skye, UK). The catheter was inserted on the day before the study and connected to a Flexilog 3000 ambulatory recorder (Oakfield Instruments Ltd, Eynsham, UK). The subjects were discharged home and the catheter was allowed to migrate into the small bowel overnight. Subjects returned to the manometry unit at 08.00 h on the following morning and the position of the catheter was adjusted so that the proximal four transducers were placed in the gastric antrum (20, 22.5, 25 and 27.5 cm from the tip) and the distal two transducers in the duodenum (5 and 10 cm from the tip). Fasting manometry was recorded for 8 h after drug/placebo administration, during which time the subjects relaxed in a sitting position. A second identical dose of drug or placebo was given 8 h after the first dose. Thirty minutes later, the subjects ate a meal consisting of chicken, rice, a bag of potato chips and water as required (energy value, 618 kcal). Post-prandial manometry was recorded for 3 h following the meal.
Analysis of manometric data. Analysis of the different phases of the MMC was performed by direct visual inspection by an experienced scientist (JW), blind with regard to treatment, with the aid of Flexilog 3000 analysis software. The phases of the MMC were defined according to the criteria used by Bjornsson and Abrahamsson.16 Data were calculated using the most distal antral and duodenal channels and only contractions exceeding a cut-off of 10 mmHg were included in the analysis. The periodicity of the MMC (number of cycles per hour) was calculated by dividing the total number of antroduodenal phase III activities by 8 h. The total number and amplitude of antral contractions and the motility index (number of contractions × median contraction amplitude × median contraction duration/2 × time) were calculated during phase II of the MMC and during the post-prandial phase. The durations of phase I, II and III of the MMC recorded from the duodenum were expressed as the mean duration of all the phases recorded (excluding those interrupted at the start and finish of the recording).
Test meal. Each subject received a mixed solid and liquid test meal which has previously been used in our department and has known reproducibility.17 Two pancakes were made from 160 g of a mixture containing 1 packet of pancake powder (Sainsbury's Supermarkets Ltd, Holborn, London, UK), 1 egg and 200 mL water; 4 MBq of 99mTc-labelled ion-exchange resin (Amberlite IRA 416; size, 0.3–1.2 mm) was added to the mixture before cooking. The pancakes were eaten with 8 g of sugar and lemon juice. A milkshake was made from 15 g of shake powder with 200 mL of full fat milk and labelled with 0.5 MBq of 111In (energy value of both meals, 490 kcal).
Study methods. Images were acquired using an IGE maxi-camera II gamma-camera (IGE Medical Systems, Slough, UK) equipped with a medium-energy (300 keV maximum), general purpose collimator. Images of the high-energy 111In photopeak (247 keV) and low-energy 99mTc photopeak (141 keV) were recorded using dual-energy windows set at 20%. Thirty minutes after ingesting MKC-733 or placebo, subjects drank the radio-labelled milkshake and immediately afterwards anterior and posterior gamma-camera images were taken in a standing position. The radio-labelled pancakes were then eaten and imaging was repeated. Both meals were eaten within a 10-min period (time, 0 h). Thereafter, images of the abdomen were taken every 20 min for the first 4 h. Subjects ate an unlabelled sandwich for lunch (time, 4 h). Following this, they were imaged every hour for a further 8 h (time, 4–12 h). A second dose of the study drug was given at time 7.5 h followed 30 min later by an unlabelled microwaved pasta dish with salad and a small cheesecake (energy value, 600 kcal).
Analysis of scintigraphic data. Analysis of the scintigraphic images was performed by an experienced scientist (EB) blind with regard to treatment. Scintigraphic images were acquired and displayed as a 128 × 128 pixel matrix (Hermes Software, Nuclear Diagnostics, Gravesend, UK). A template of the gastrointestinal tract was constructed for each volunteer using an external reference marker to align the individual images. Regions of interest were drawn around the whole stomach, body, small intestine and colon for each image. The raw data were presented as the total number of counts in the region of interest area. The geometric mean of the anterior and posterior images was calculated from the radioactive counts in each stomach region. These were corrected for background radiation, the radioactive half-life of each radionuclide and cross-talk between the two different energy windows. Data were normalized and presented as a graph of activity against time. The half-times of gastric emptying (T50%) from the whole stomach and the proximal stomach were calculated from the activity vs. time graph using a linear or exponential model according to best fit. The proximal stomach was defined as the portion of the stomach proximal to a line drawn at 45° downwards from the angulus. The small intestinal transit time was calculated from the time taken for 50% of the meal to arrive in the ascending colon minus the T50% value of gastric emptying.
Echoplanar magnetic resonance imaging study
Test meal. The test meal was prepared by adding 1% (w/w) locust bean gum powder (food grade Ceratonia Siliqua; Lucas Meyer Colloids Ltd, Chester, UK) to hot water. To this was added 50 mL of concentrated lipid (olive oil emulsified with sorbitan stearate, Crillet 3 VEG and Vykamol 83G; Croda Food Services, Lancashire, UK), 30 g of sucrose and 0.5% concentrated banana flavouring (energy value, 323 kcal). The solutions were blended using a food mixer and kept at 90 °C for 1 h before being allowed to cool slowly. This meal has been used extensively in our laboratory and provides excellent images for the analysis of gastric emptying.18
Study methods. Our methods have been described previously.19, 20 In brief, single shot echoplanar magnetic resonance images were acquired on a purpose-built, whole-body, 0.5-Tesla scanner equipped with actively shielded gradients and a 50-cm diameter birdcage coil. The in-plane resolution was 3.5 mm × 2.5 mm and the slice thickness was 1 cm. Each image was acquired in 130 ms using a 128 × 128 matrix with an effective echo time of 40 ms. Thirty minutes after ingesting MKC-733 or placebo, the subjects consumed the 500-mL test meal within 10 min. Directly after ingestion of the meal, transverse, multi-slice sets of images were acquired from the heart to the kidneys to determine the volume of the meal in the stomach. Repeat scans were performed every 15 min until the stomach was empty, up to a maximum time of 120 min. At 5 and 40 min after the test meal, a multi-slice set of images was acquired through the antrum to assess the antral contraction speed and frequency. During imaging, the subjects lay in a supine position, but relaxed in a sitting position outside the magnet between scans.
Analysis of echoplanar magnetic resonance imaging data. Analysis of the different echoplanar magnetic resonance images was performed by an experienced scientist (LM) blind with regard to treatment. A region of interest was drawn around the gastric contents on each image (Analyse software, Biomedical Imaging Resource, Mayo Foundation, Rochester, MN, USA). The volume of the gastric contents was calculated by summing the areas across the slices. The T50% value of gastric emptying was calculated from the volume–time graph using a linear or exponential model according to best fit. Relaxation of the proximal stomach was estimated by measuring the maximal cross-sectional area of a transverse slice through the proximal stomach 30 min after the test meal. The proximal stomach was identified at its widest point and surrounding anatomical landmarks, such as the liver and spine, were used to position each one of the four data sets at exactly the same level through the gastric lumen. The cross-sectional area was measured by drawing a region of interest manually using Analyse. The speed and frequency of antral contractions were derived from a motility plot using in-house software as described previously.20
Statistical analysis was performed on the ‘intention-to-treat’ population. The data were normally distributed and are presented as the mean ± standard deviation. Treatment differences were analysed using repeated measures analysis of variance (anova) with factors for subject, treatment and period. Tests for first-order treatment carry-over were performed, but were not significant. If the test for overall treatment effect was significant, pairwise comparisons were performed comparing each dose with placebo. A P value of less than 0.05 was considered to be significant. In the event that anova was not significant, but there was a clear dose–response effect, post hoc analysis of the data was performed using a test of linear trend. A further analysis of the effect of nausea was performed by comparing subjects with and without nausea (scintigraphic liquid emptying only) and excluding subjects who experienced nausea (gastric emptying, relaxation and MMC frequency) using non-paired and paired t-tests. Statistical analysis was performed by an external statistician blind with regard to the treatment given (Hartington Statistics and Data Management Ltd, London, UK) using SAS version 6.12 software (SAS Institute, Cary, NC, USA).
Forty-eight subjects were screened and, of these, 40 were eligible for randomization. Sixteen subjects took part in the scintigraphic study (mean age, 23.4 ± 4.0 years; BMI, 23.2 ± 2.7 kg/m2), one subject having failed screening due to a BMI above the normal range. Twelve subjects entered the manometric study (mean age, 27.2 ± 5.6 years; BMI, 23.8 ± 1.8 kg/m2), two subjects having failed screening due to abnormal laboratory values. Twelve subjects entered the echoplanar magnetic resonance imaging study (mean age, 22.0 ± 4.2 years; BMI, 22.4 ± 2.1 kg/m2), one having failed screening due to a BMI below the normal range, one due to withdrawal of consent and three due to abnormal laboratory values. All subjects completed the scintigraphic and echoplanar magnetic resonance imaging studies according to the protocol. A dosing error occurred during the manometric study and data for the 0.2-mg and 4-mg doses were excluded for the subject involved. Technical faults with the manometric equipment led to a failure of the datalogger to record any data during the placebo study in one subject and, in a further subject, no data were recorded from the duodenal transducers following 0.2 mg of MKC-733.
The most marked effect of MKC-733 was a dose-dependent increase in the number of MMC cycles recorded from the antrum and duodenum (P < 0.001 for 4 mg vs. placebo) (Table 1). This increase resulted from a significant reduction in the mean duration of each phase I cycle with no change in the mean durations of phase II and III. Following the 4-mg dose of MKC-733, one subject experienced nausea and one subject experienced both nausea and vomiting. When these patients were excluded from the analysis, the periodicity of the MMC following 4 mg remained significantly higher than for placebo (mean ± s.d.: 0.51 ± 0.13 vs. 0.30 ± 0.09 cycles per hour; paired t-test; P = 0.002).
Table 1. The effect of MKC-733 on the number of migrating motor complexes (MMCs) recorded from the antrum and duodenum per hour (periodicity) and the mean duration of each phase of the duodenal MMC
|Placebo (n = 11)||0.34 ± 0.11||60 ± 34||58 ± 44||6.0 ± 2.1|
|0.2 mg (n = 11)||0.40 ± 0.15||41 ± 27†||72 ± 47||6.3 ± 2.3|
|1 mg (n = 12)||0.42 ± 0.11||38 ± 23‡||73 ± 51||6.5 ± 2.7|
|4 mg (n = 11)||0.54 ± 0.14*||32 ± 18‡||58 ± 29||6.5 ± 2.4|
There was a trend to an increased number of antral contractions recorded during phase II of the MMC (anova, P = 0.097; linear trend, P = 0.036). The motility index and amplitude of the antral contractions during phase II were not significantly different from placebo for any of the doses. MKC-733 had no significant effect on the post-prandial motility index, the number of antral contractions or the amplitude of antral contractions during the 3 h following the meal.
MKC-733 significantly delayed the half-emptying time of the liquid meal from the whole stomach (P = 0.005 for 4 mg vs. placebo). After a 4-mg dose, this delay was significantly greater in subjects who complained of nausea (n = 7) compared with those who did not (n = 9) (69.7 ± 18.7 vs. 48.1 ± 16.5 min; unpaired t-test; P = 0.028). Emptying of the solid meal was also delayed, but this did not reach statistical significance (Table 2). There was also a delay in the emptying of both meals from the proximal stomach, which was significant for the solid meal (P = 0.016 for 4 mg vs. placebo), but not for the liquid meal (Table 2). MKC-733 dose dependently decreased the small intestinal transit time of the liquid meal (anova, P = 0.169; linear trend, P = 0.035) (Table 2). The small intestinal transit time of the solid meal was not determined as the meal did not reach the colon within the study period.
Table 2. The effect of MKC-733 on gastric emptying and small intestinal transit time measured by gamma-scintigraphy
|Placebo (n = 16)||44 ± 15||105 ± 21||40 ± 16||75 ± 32||276 ± 99|
|0.2 mg (n = 16)||44 ± 19||104 ± 27||41 ± 22||79 ± 29||265 ± 121|
|1 mg (n = 16)||43 ± 18||107 ± 26||46 ± 23||73 ± 27||247 ± 96|
|4 mg (n = 16)||58 ± 20*||113 ± 28||49 ± 22||89 ± 26†||223 ± 91‡|
Echoplanar magnetic resonance imaging study
The gastric half-emptying time of the test meal was significantly delayed (P < 0.001 for 4 mg vs. placebo) (Table 3) and MKC-733 significantly increased the cross-sectional area of the proximal stomach 30 min after the meal (P < 0.01 for 4 mg vs. placebo) (Table 3). The two subjects who experienced nausea after 4 mg of MKC-733 had the longest gastric half-emptying times of 133 and 99 min. When the data from these two patients were excluded, gastric emptying was still significantly slower following 4 mg compared with placebo (mean ± s.d.: 60 ± 21 vs. 40 ± 14 min; paired t-test; P = 0.023) and the proximal gastric area remained significantly larger (mean ± s.d.: 63 ± 6 vs. 51 ± 12 cm2; paired t-test; P = 0.044). There was no significant difference between treatment groups with regard to the number or speed of antral contractions following the meal (Table 3).
Table 3. The effect of MKC-733 on gastric motility measured by echoplanar magnetic resonance imaging
|Placebo (n = 12)||40 ± 13||1.6 ± 0.3||3.0 ± 0.3||50 ± 12|
|0.2 mg (n = 12)||39 ± 14||1.5 ± 0.2||3.0 ± 0.2||47 ± 15|
|1 mg (n = 12)||49 ± 20||1.6 ± 0.2||3.0 ± 0.2||56 ± 13|
|4 mg (n = 12)||70 ± 30*||1.5 ± 0.2||2.9 ± 0.3||64 ± 6†|
MKC-733 was well tolerated and no subjects withdrew from the studies as a result of adverse events. The most frequent symptoms were flushing, mild nausea and diarrhoea (Table 4). No abnormalities in blood pressure, pulse or resting electrocardiogram were observed during the study period.
Table 4. The number of patients experiencing symptoms following a single oral dose of MKC-733 or placebo in all three studies (n = 40)
|Abdominal pain||2||1||4|| 7|
This is the first study to describe the effects of a selective 5-HT3 agonist on upper gastrointestinal motility in humans. By combining manometry and scintigraphy with the novel echoplanar magnetic resonance imaging, we have been able to define the mechanisms involved in some detail. MKC-733 delays gastric emptying in association with relaxation of the proximal stomach, increases antroduodenal MMC activity and accelerates small intestinal transit.
We believe that these effects are mediated via 5-HT3 receptor stimulation, as MKC-733 demonstrates selective 5-HT3 receptor agonist activity in a number of animal models.12–15 In 5-HT3 receptor binding assays using membrane preparations of canine small intestinal muscle, MKC-733 strongly inhibited the specific binding of the radioactive ligand (3H)GR-65630 with a Ki value of 0.8 nm, compared with 0.51 µm for the 5-HT4 receptor and > 100 µm for 5-HT1, 5-HT2 and 5-HT7 receptors. There is weak affinity for M1 (Ki = 7.7 µm) and M2 (Ki = 3.3 µm) receptors, but no significant affinity for other receptors, including α1, α2, β2, nicotine, substance P, calcium channel (T,L), γ-aminobutyric acidA/B, dopamine-1, 2 and 3 and histamine-1 receptors. The intrinsic activity of MKC-733 in humans is unknown, but MKC-733 is a partial agonist in rats and guinea-pigs (intrinsic activities of 0.2 and 0.4, respectively) and a full agonist in the mouse (intrinsic activity of 1.0), suggesting important species differences in 5-HT3 receptor structure.14
The effects of MKC-733 demonstrated in this study are compatible with the predicted actions of a selective 5-HT3 agonist. Immunohistochemical studies in the rat have demonstrated a wide distribution of 5-HT3 receptors on cell bodies and fibres in the enteric plexuses, extrinsic afferent neurones and interstitial cells of Cajal,1 suggesting that selective 5-HT3 agonists should stimulate intrinsic and extrinsic motor and secretory reflexes and increase visceral afferent signalling. Activation of 5-HT3 receptors in the enteric and muscular layers of the intestine is normally prevented by the re-uptake of 5-HT into the crypt epithelial cells via a specific serotonin re-uptake transporter.21 However, it is likely that MKC-733 activates a much wider population of 5-HT3 receptors throughout the intestinal wall. In vitro pharmacological studies have suggested that 5-HT3 autoreceptors on enterochromaffin cells stimulate 5-HT release,22 which could potentiate the action of MKC-733 by stimulating other mucosal 5-HT receptor subtypes. However, 5-HT3 receptors have not yet been demonstrated on 5-HT-containing entero-endocrine cells by immunohistochemistry in humans.23
Selective 5-HT3 antagonists do not have a marked effect on gastric emptying9, 24 or proximal gastric motor function10, 25 in humans. In contrast, 4 mg of MKC-733 induced relaxation of the proximal stomach and delayed emptying of liquids. Similar gastric motor changes have been reported previously in association with nausea,26, 27 which occurred following a 4-mg dose in seven of the 16 volunteers during the scintigraphic study and two of the 12 volunteers during the echoplanar magnetic resonance imaging and manometric studies. Delayed gastric emptying in the scintigraphic study was observed only in those patients experiencing nausea, whereas delayed emptying and fundus relaxation occurred independently of nausea in the echoplanar magnetic resonance imaging study. Although there does not appear to be a causal link between the sensory and motor changes, it is likely that both effects result from the stimulation of 5-HT3 receptors on vagal afferents in the intestine or in the brainstem.28 The increased frequency of nausea observed following the mixed liquid/solid test meal compared with the liquid meal or during fasting also suggests an important effect of food on the subjective experience of nausea.
MKC-733 significantly increased the number of antroduodenal MMC cycles recorded during the 8-h fasting period. Abnormal intestinal phase III-like activity has previously been reported in patients with idiopathic nausea and vomiting29 and, in the volunteer who experienced vomiting after 4 mg of MKC-733, we observed three isolated bursts of phase III activity in the hour following emesis. The effect of MKC-733 was independent of nausea and vomiting, however, and all phase III activity included in our analysis was recorded as part of the normal fasting MMC cycle. The important role of 5-HT3 receptors in regulating the gastric MMC has previously been confirmed in studies with the selective 5-HT3 antagonist ondansetron.11
MKC-733 accelerated the small intestinal transit of the liquid meal. 5-HT3 antagonists do not affect small intestinal transit in humans,30 although transit is accelerated by serotonin re-uptake inhibitors.31 In humans, peristalsis induced by mucosal stimulation is mediated via 5-HT1P/4 receptors and unaffected by 5-HT3 receptor blockade.32 However, MKC-733 may stimulate intestinal contraction by acting directly on excitatory 5-HT3 receptors in the myenteric plexus. Accelerated small bowel transit could also be the result of 5-HT3-mediated small intestinal4 and pancreatic5 secretion.
The flushing and itching which occurred in half of the subjects following 4 mg is a predictable effect of a 5-HT3 agonist, as intradermal injection of 5-HT causes cutaneous erythema and burning pain that is inhibited by 5-HT3 receptor antagonists.33 Such symptoms suggest widespread systemic effects of the drug, although, in phase I studies, peak plasma levels were very low (mean ± standard deviation: 0.5 mg, 229 ± 98 pg/mL; 1 mg, 474 ± 100 pg/mL; 4 mg, 1513 ± 149 pg/mL). The total bioavailability of the drug in humans is unknown, but is likely to be low (approximately 3% in rats and 17% in dogs) and, following a 4-mg dose, 97.7% of the drug is excreted in the faeces. Skin symptoms were less intense and occurred less frequently following a second dose of the drug, suggesting tachyphylaxis. We observed no significant cardiovascular changes during this study, although, in animals, 5-HT3 receptor agonists can induce transient bradycardia and hypotension due to stimulation of cardiac vagal 5-HT3 receptors34 (von Bezold–Jarisch reflex). Diarrhoea (usually one or two loose stools later that day only), flatulence and colicky abdominal pain were all mild and well tolerated. 5-HT3 receptor antagonists are known to delay colonic transit6 and, in a recent animal study, a selective 5-HT3 agonist induced giant migrating contractions in the colon.35
In this study, we assessed upper gastrointestinal motility using three different techniques. Both gamma-scintigraphy and manometry are widely available and provide complementary information, but require separate visits. Gastric echoplanar magnetic resonance imaging is a relatively recent, non-invasive technique that allows gastric emptying and proximal and distal gastric motility to be assessed simultaneously without exposure to ionizing radiation. In the present study, we found good correlation between these different methods. A potential disadvantage of echoplanar magnetic resonance imaging is that it can only be performed in the supine position. A significant postural effect is unlikely as the ultrafast echoplanar magnetic resonance imaging technique allowed us to keep the subjects upright for most of the study and there was agreement with the scintigraphic measurements of liquid emptying performed in an upright position. Another potential disadvantage of echoplanar magnetic resonance imaging is that subjects are required to hold their breath in expiration during scanning to prevent movement artefacts, which may be difficult in some patients. Gastric accommodation is normally measured using a barostat balloon, although this procedure is time consuming, invasive and has low patient acceptability. In this study, we estimated gastric accommodation by measuring the cross-sectional area of a transverse slice through the proximal stomach in order to measure the degree of circumferential relaxation or contraction. A recent study has demonstrated that the volume change profile of the whole stomach measured by echoplanar magnetic resonance imaging correlates well with the gastric accommodation response measured by barostat,36 and our previous validation study showed volume measurements by echoplanar magnetic resonance imaging to be accurate to within ± 5%.37 Recently, single-photon emission computed tomography has been successfully used to measure gastric accommodation,38 although this technique also involves exposure to ionizing radiation. Our observation that MKC-733 had no measurable effect on the post-prandial antral motility assessed by echoplanar magnetic resonance imaging is in agreement with the findings from the manometric study.
We can only speculate on the potential role of 5-HT3 agonists in patients with digestive disorders, although systemic side-effects may limit their clinical use. MKC-733 accelerates gastric emptying in some animal models,15 and it is possible that the stimulatory effect of MKC-733 on the gastric MMC may improve gastric emptying in patients with severe gastroparesis and grossly delayed emptying of indigestible residues. More promisingly, the increased passage of loose stools and enhanced small intestinal transit suggest that 5-HT3 agonists may be useful in the treatment of constipation. Indeed, in a small preliminary study, MKC-733 improved colonic transit, increased stool frequency and relieved symptoms in subjects with constipation.39
This study was supported by Mitsubishi-Pharma Corporation. Dr N. S. Coleman was supported by an educational grant from Mitsubishi-Pharma Corporation.
For references to Tables 1 to 5 in the text, supplementary tables summarising cardiovascular responses are provided for those interested (http: http://www.blackwellpublishing.com/products/journals/suppmat/apt/apt1797/apt1797sm.htm).