5‐Hydroxytryptamine effects in the gut: the 3, 4, and 7 receptors

In the gut, serotonin (5-hydroxytryptamine: 5-HT) exerts a variety of physiological effects on enterocytes, smooth muscle cells, intrinsic enteric neurons, and extrinsic afferents, which are related to the expression of multiple 5-HT receptor types and subtypes regulating motility, vascular tone, secretion and perception. This, and the evidence that several factors involved in the control of 5-HT (production/release/re-uptake mechanisms) are changed in some syndromes, such as the carcinoid and the irritable bowel syndrome (IBS), have led to the idea that these conditions are partly based on a dysregulation of 5-HT function. Indeed, 5-HT3 antagonists are effective in relieving dysmotility and diarrhoea associated with carcinoid syndrome and a variety of symptoms (including abdominal pain) in diarrhoea-predominant IBS, whereas 5-HT4 agonists act as prokinetic agents in the lower gut (to treat constipated IBS patients) and in the upper gastrointestinal tract. In this issue of the journal, Crowell et al. observe some expected and unexpected effects of tegaserod, a partial 5-HT4 receptor agonist, on gastric emptying in a murine model of diabetes mellitus. This review will also briefly appraise 5-HT3 receptor activity, and outline the role on gut function of 5-HT7 receptors, the most recently characterized members of the 5-HT receptor family. In the C57BLKS/J db/db transgenic mouse, a model of diabetes mellitus that has been shown to have delayed gastric emptying, Crowell et al. found that tegaserod was able to accelerate gastric emptying at a low dose (0.1 mg kg−1, i.p.). The prokinetic effect was gradually reduced with increasing doses, up to the complete loss of the prokinetic effect with the use of higher doses (e.g. 2.0 mg kg−1, i.p.). The accelerating effect produced by 0.1 mg kg−1 was completely prevented by the selective 5-HT4 antagonist, GR113808. Obviously, the mechanisms responsible for the inverse dose–response relationship of tegaserod on gastric emptying are not clear, as acknowledged by the authors, and they are difficult to simply reconcile with a partial 5-HT4 agonistic effect of the drug. Although a direct activation of 5-HT4 receptors located in myenteric neurons is a well recognized mechanism for gastrointestinal prokinesis,1,2 a compound with the pharmacokinetic and pharmacodynamic profile of tegaserod may exert its prokinetic effect through an alternative mechanism. Tegaserod is a compound similar in structure to 5-HT, which also possesses partial agonist properties at the 5-HT4 receptor, as observed in electrically stimulated cholinergic twitch contractions in the guinea pig ileum.3 Under these conditions, the intrinsic activity of tegaserod was found equal to 0.2, which corresponds to one-fifth of that possessed by the full agonist 5-HT (intrinsic activity: 1.0).3 As the intrinsic activity of a drug is a measure of its efficacy in inducing a response in a given model system, this means that tegaserod is able to induce a potentiating effect on cholinergic twitch contractions, which is only 20% of that exerted by 5-HT. A partial agonist can also be considered pharmacologically as a partial antagonist, especially when its intrinsic activity is closer to zero (a condition that characterizes full antagonism) than to 1 (a condition that characterizes full agonism). Therefore, partial agonists are defined as drugs that produce submaximal tissue responses and competitively block the effects of agonists of higher intrinsic activities.4 Experimentally, the antagonist properties of a partial agonist can be demonstrated towards a full agonist (in this case, 5-HT), whose effects are expected to be antagonized in a dose-dependent manner. Most of 5-HT in the body is synthesized, stored in and released from the enterochromaffin cells of the gastrointestinal mucosa.5 There is abundant evidence that the mechanisms of 5-HT release can be modulated by a complex variety of heteroreceptors and by several 5-HT autoreceptors, including 5-HT3 and 5-HT4 receptors located in the surface membrane of enterochromaffin cells.6 In both animal and human intestinal preparations,7–9 activation of 5-HT3 receptors triggers a positive feedback mechanism leading to an increase in 5-HT release. In contrast, activation of 5-HT4 receptors is associated with a negative feedback mechanism leading to a reduction of 5-HT release. In human and pig isolated tissues, these effects were competitively antagonized by selective antagonists at 5-HT3 and 5-HT4 receptors, respectively.8 In vascularly perfused guinea-pig ileum preparations, 5-HT3 antagonists were found to reduce 5-HT release, whereas 5-HT4 antagonists enhanced 5-HT release, suggesting a tonic influence of endogenous 5-HT in modulating its own release from enterochromaffin cells.7 Based on this assumption, after i.p. administration, circulating tegaserod is distributed from the mucosal microcirculation to the mucosal cells where it may also function as a partial 5-HT4 antagonist, thus counteracting the effect of the full agonist 5-HT at the inhibitory 5-HT4 autoreceptor on enterochromaffin cells. By reducing the influence of the negative feedback mechanism caused by endogenous 5-HT, the net effect of tegaserod will result in an enhancement of both 5-HT release and 5-HT availability at neuronal 5-HT4 receptors, the sites in close proximity to the mucosa that are responsible for gastrointestinal prokinesis (Fig. 1). Therefore, tegaserod can be considered as a modulator of endogenous 5-HT release, which ultimately exerts its physiological effects on gastrointestinal motility. This is probably the mechanism through which the low i.p. dose of tegaserod exerted its prokinetic effect in the study from Crowell et al., via a mechanism that is exquisitely sensitive to the 5-HT4 antagonist, GR113808. Conversely, the inhibitory effect on motility caused by higher i.p. doses of tegaserod can be explained through a blockade of neuronal 5-HT4 receptors, because of the enhanced amounts of tegaserod released from the microvasculature, which cannot be overcome in a competitive manner by endogenous 5-HT (Fig. 1). The scheme illustrates the mode of action of i.p. administered tegaserod on gastric emptying. Left panel: at a low dose, tegaserod is released from the mucosal microcirculation in the vicinity of the mucosa, where it competitively antagonizes inhibitory 5-HT4 autoreceptors located on enterochromaffin cells (EC cell), leading to enhancement of 5-HT release (vertical arrow). 5-HT, in turn, stimulates neuronal 5-HT4 receptors on enteric pathways to promote gastric emptying, via a mechanism antagonized by the 5-HT4 antagonist, GR113808. Right panel: at high doses, increasing amounts of tegaserod are released from mucosal microcirculation. Under these conditions, tegaserod markedly blocks both inhibitory 5-HT4 autoreceptors and neuronal 5-HT4 receptors, via a mechanism that cannot be overcome by endogenous 5-HT. This leads to unaltered gastric emptying. Symbols ‘+’ and ‘−’ indicate positive and negative feedback mechanisms at 5-HT3 and 5-HT4 receptors, respectively. A similar mode of action can also be hypothesized for orally administered tegaserod (the route of administration in humans). Pharmacokinetic studies revealed that the absolute bioavailability of tegaserod is approximately 10%, and food is able to reduce by a further 50% the Cmax and area under the curve of tegaserod.10 This means that a very substantial amount of oral tegaserod is not absorbed; in fact more than two-thirds of the drug is excreted unchanged in stool.11 Theoretically, the unabsorbed component of tegaserod might exert its gastrointestinal prokinetic effect through a topical action in the intestinal lumen. In this respect, a low (virtually unabsorbed) oral dose of tegaserod blocks inhibitory 5-HT4 autoreceptors on enterochromaffin cells, causing an enhancement of 5-HT release and 5-HT availability at neuronal 5-HT4 receptors (Fig. 2). At high oral doses, the enhanced amounts of absorbed tegaserod may hinder prokinesis by blocking neuronal 5-HT4 receptors, an effect that cannot be counteracted by endogenous 5-HT (Fig. 2). The scheme illustrates the mode of action of orally administered tegaserod on gastrointestinal transit. Left panel: at a low dose, tegaserod is virtually unabsorbed and may competitively antagonize inhibitory 5-HT4 autoreceptors located on enterochromaffin cells (EC cell), leading to enhancement of 5-HT release (vertical arrow). 5-HT, in turn, stimulates neuronal 5-HT4 receptors on enteric pathways that promote gastrointestinal transit. Right panel: at high doses, increasing amounts of tegaserod are gradually absorbed. Under these conditions, tegaserod blocks both inhibitory 5-HT4 autoreceptors and neuronal 5-HT4 receptors, via a mechanism that cannot be overcome by endogenous 5-HT. This leads to unaltered gastrointestinal transit. Symbols ‘+’ and ‘−’ indicate positive and negative feedback mechanisms at 5-HT3 and 5-HT4 receptors, respectively. The evidence that in patients with constipation-predominant IBS tegaserod produced a bell-shaped effect on overall symptomatology, abdominal discomfort and constipation relief in the dose range 1–24 mg (the effect of the drug initially increased and then decreased with the use of the highest dose)12 is in favour of a mechanism related to a partial 5-HT4 antagonism of the drug. Therefore, if this represents a general mode of action of tegaserod that also operates in humans, physicians and patients should be made aware that an increase of drug dosage would not be associated with a more pronounced prokinetic effect, but rather with at least no effect or at worst, a deceleration of gastrointestinal transit. 5-HT4 receptors are G protein-coupled metabotropic receptors, which promote cyclic AMP formation in both enteric nerve and muscle, where they cause transmitter release (e.g. acetylcholine) and relaxation (e.g. in the human colon), respectively. They are distributed from the oesophagus to the large intestine. In humans, they have been recently identified on cholinergic nerves supplying both the longitudinal and circular layer of the colon.13,14 This wide distribution of 5-HT4 receptors may represent a problem when absorbable drugs endowed with ‘pure’ 5-HT4 agonism and directly activate neuronal 5-HT4 receptors located on intrinsic pathways, are used as prokinetics of both the upper and the lower gastrointestinal tract. Thus, a pure 5-HT4 agonist used to prevent gastric stasis might be expected to cause diarrhoea, an adverse event for diabetic or dyspeptic patients. The development of diarrhoea would not be perceived as an adverse event by patients taking the drug for constipation-predominant IBS or functional constipation. Originally, cisapride was proposed as a prokinetic of the upper gastrointestinal tract. This compound gave conflicting results on human colonic motility. Cisapride also exerts moderate antagonism at the 5-HT3 receptor, an effect associated with colonic (but not with gastric) inhibition of motility.15 On pharmacological grounds, renzapride is a mixed 5-HT4 agonist/5-HT3 antagonist with a pharmacological profile similar to that of cisapride, and therefore it would be expected to be a prokinetic agent of the upper gastrointestinal tract.16 Surprisingly, renzapride was recently found to cause a clinically significant acceleration of colonic transit (especially in the ascending colon) in female patients with constipation-predominant IBS.17 Cisapride gave conflicting results on colonic motility, and IBS, possibly because of its 5-HT3 antagonist properties.18 Further studies are required to assess whether mixed 5-HT4 agonists/5-HT3 antagonists are of value in the treatment of constipation-predominant IBS. 5-HT3 receptors are ligand-gated cation channels, which are found on neuronal tissue and on enterochromaffin cells in the gut. They mediate fast depolarization and transmitter release (including acetylcholine) in neuronal tissue. Blockade of 5-HT3 receptors by full antagonists, such as alosetron, blunts visceral perception, enhance colonic compliance, and reduce colonic motor response to a meal, fluid secretion, and colonic transit.19 They are used in the treatment of diarrhoea-predominant IBS. Their mode of action probably involves a simultaneous blockade of 5-HT3 receptors on enterochromaffin cells (which mediate a positive feedback mechanism on 5-HT release) and on extrinsic and intrinsic neuronal pathways mediating perception, motility and secretion. This widespread receptor blockade may account for their therapeutic effectiveness and adverse reactions, including constipation, which have led to a restricted use of alosetron in women with severe, diarrhoea-predominant IBS refractory to conventional therapy.20 The relationship between ischaemic colitis (which has been attributed to this class of compounds) and 5-HT3 antagonism is still unclear. To minimize the occurrence of adverse reactions, without losing the therapeutic efficacy it is important to ‘start slowly and go slowly’, and identify the minimal effective dose for each individual. About 30% of patients experienced constipation with 5-HT3 antagonists.21 The design of novel molecules with various degrees of partial 5-HT3 antagonism would allow a less effective blockade of endogenous 5-HT at the 5-HT3 receptor on both enterochromaffin cells and neuronal pathways, leading the ability of 5-HT or the partial antagonist to exert a residual ‘rescue’ intrinsic activity on colonic motility. 5-HT7 receptors are also G protein-coupled metabotropic receptors, which promote cyclic AMP formation. In the gut, they were originally identified in effector cells, causing muscle relaxation in preparations of animal and human stomach, small and large intestine.22–24 Recent electrophysiological, immunohistochemical and functional studies in the guinea-pig ileum have revealed the presence of 5-HT7 receptors in various subpopulations of enteric neurons. 5-HT7 receptors were found to mediate slow synaptic transmission in myenteric AH neurons.25 Immunohistochemical study demonstrated the co-localization of 5-HT7 receptors with calbindin and NeuN, two specific markers for AH neurons, in both the myenteric and the submucosal plexus, as well as with a dense subset of myenteric neurons containing nNOS or VIP.26,27 Experiments aimed at evaluating the role of endogenous 5-HT and 5-HT7 receptors on ileal peristalsis revealed that the selective 5-HT7 antagonist, SB-269970, significantly reduced (by about 30%) the accommodation of the circular muscle during the relaxatory phase of peristalsis,26,27 a process mediated by NO released from short descending inhibitory motor neurons.28 We hypothesized that 5-HT released by enterochromaffin cells following mucosal stimulation activated AH neurons via 5-HT7 receptors, and, in turn, activated descending (including serotonergic) interneurons,29 to provide synaptic input to 5-HT7 receptors located on nNOS inhibitory motor neurons, leading to circular muscle relaxation. In this respect, relaxation results from the cross-talk between paracrine and neuronal 5-HT, with participation of 5-HT7 receptors on various enteric neurons. Another class of receptors, the ‘5-HT1P receptor’ located on AH neurons, has been repeatedly proposed as the main target of 5-HT released by enterochromaffin cells following mucosal stimulation.30 This receptor is the only 5-HT receptor which has not been cloned yet. Moreover, its pharmacology, electrophysiology, and tissue distribution show an impressive similarity with those of the 5-HT7 receptor, as 5-HT1P receptors are G protein-coupled metabotropic receptors mediating slow synaptic transmission in AH neurons.30 Like 5-HT7 receptors in the guinea-pig ileum, 5-HT1P receptors are present in the rat stomach on nNOS containing neurons.31 Lastly, both receptors have a postjunctional location leading to direct smooth muscle relaxation, via cyclic AMP production.22,32 Ongoing studies in various laboratories will soon unveil whether the two receptors are identical or separate entities. A change in the number of EC cells and/or an impaired production/re-uptake of 5-HT have been associated with changes in bowel motility, secretion and visceral sensitivity in functional gastrointestinal disorders, such as IBS33. As endogenous 5-HT activates 5-HT7 receptors, it is likely that, under pathophysiological conditions, overstimulation of these receptors may cause an exaggerated relaxation of the circular muscle that could contribute to abdominal bloating, a symptom that accompanies many functional bowel disorders.34 5-HT7 antagonists might offer interesting opportunities to circumvent this problem. 5-HT7 receptors are also expressed by primary afferent nociceptors in the lumbar dorsal root ganglia,35 and may be a target for relief of abdominal pain. 5-HT7 agonists may also have potential with other ‘investigational’ fundus-relaxing drugs, such as nitrates, sumatriptan, buspirone and clonidine, to restore gastric accommodation for improving postprandial symptoms, such as early satiety.36 In conclusion, in the gut, 5-HT acts as a paracrine signalling molecule released by enterochromaffin cells and as a transmitter released by some descending serotonergic interneurons. 5-HT4 agonists and 5-HT3 antagonists have proved to be effective in the treatment of multiple symptoms of IBS and other functional bowel disorders. Because of its peculiar pharmacokinetic and pharmacodynamic profile, it is likely that the ‘partial 5-HT4 agonist’, tegaserod, exerts its prokinetic activity (at low doses) via partial antagonism at 5-HT4 autoreceptors on enterochromaffin cells. Novel and safer serotonergic drugs are needed, and agonists and antagonists at the 5-HT7 receptor appear to have a promising molecular target to affect gut functions. The author is supported by grants from the University of Pavia (FAR 2004) and the Italian Ministry of Education, University and Scientific Research (FIRB 2001; PRIN 2004).