Dr G. M. Mawe, Given Building D403A, Department of Anatomy and Neurobiology, 89 Beaumont Ave, University of Vermont, Burlington, VT 05405, USA. E-mail: email@example.com
Alterations in motility, secretion and visceral sensation are hallmarks of irritable bowel syndrome. As all of these aspects of gastrointestinal function involve serotonin signalling between enterochromaffin cells and sensory nerve fibres in the mucosal layer of the gut, potential alterations in mucosal serotonin signalling have been explored as a possible mechanism of altered function and sensation in irritable bowel syndrome. Literature related to intestinal serotonin signalling in normal and pathophysiological conditions has been searched and summarized.
Elements of serotonin signalling that are altered in irritable bowel syndrome include: enterochromaffin cell numbers, serotonin content, tryptophan hydroxylase message levels, 5-hydroxyindoleacedic acid levels, serum serotonin levels and expression of the serotonin-selective reuptake transporter. Both genetic and epigenetic factors could contribute to decreased serotonin-selective reuptake transporter in irritable bowel syndrome. A serotonin-selective reuptake transporter gene promoter polymorphism may cause a genetic predisposition, and inflammatory mediators can induce serotonin-selective reuptake transporter downregulation.
While a psychiatric co-morbidity exists with IBS, changes in mucosal serotonin handling support the concept that there is a gastrointestinal component to the aetiology of irritable bowel syndrome. Additional studies will be required to gain a more complete understanding of changes in serotonin signalling that are occurring, their cause and effect relationship, and which of these changes have pathophysiological consequences.
Irritable bowel syndrome (IBS) is a common gastrointestinal (GI) disorder associated with alterations in motility, secretion and visceral sensation. A range of clinical symptoms characterize this disorder, including altered stool frequency and form, abdominal pain and bloating. Impairment in quality of life is common for the significant proportion of the population who suffer from IBS.1–3 Specific clinical criteria differentiate IBS from other functional GI disorders as well as from inflammatory bowel disease (IBD) and other non-functional GI disorders. There are no specific biological, radiographic, endoscopic or physiological markers that have been identified in this disorder, and therefore, the diagnosis of IBS is based solely on these clinical criteria. Despite extensive interest and investigation, the pathogenesis of IBS remains unclear. Improved understanding of the enteric nervous system (ENS), and the neuroenteric pathways involved in GI function, may provide a conceptual framework to integrate previously described motility, sensory, and cognitive features of the disease and to identify new therapeutic targets. In this article, we will specifically review the evolving understanding of the important roles of serotonin (5-HT) as a signalling molecule in the gut and the pathophysiology of functional GI disorders like IBS. We have searched and summarized relevant literature related to 5-HT signalling in the mucosa of the intestines under normal conditions and changes in key elements of 5-HT signalling that have been reported in IBS and in response to inflammation.
The roles of serotonin as a signalling molecule in the gut
The ENS is by far the most vast and complex component of our peripheral nervous system.4 Distinguishing features of the ENS include the quantity of neurones, the variety of neurotransmitters and associated receptors, and the existence of intrinsic reflex circuitry. The number of neurones in the human small intestines alone (∼100 000 000) approximates the number of neurones in the entire spinal cord. Almost every known neurotransmitter can be found in the ENS, and most receptors associated with these neurotransmitters are also found in the ENS. The GI tract is the only part of the body with neural reflexes that are housed entirely within the organ. Unlike reflexes associated with neural control of other organ systems or smooth muscle, complete neural circuits that include sensory neurones, interneurones, and motor neurones are contained within the wall of the intestines, and these circuits are responsible for motility, secretion and vascular tone in the gut.
One of the signalling molecules with an unambiguous physiological role in the ENS is 5-HT.5 5-HT is often thought of as a neurotransmitter exclusive to the central nervous system (CNS), because of its well-defined roles in depression, arousal, pain pathways and other CNS functions. However, the major source of bioavailable 5-HT in the human body is located in the bowel, primarily in a subset of epithelial cells called enterochromaffin (EC) cells.6 EC cells express the enzymatic machinery, including the rate limiting enzyme, tryptophan hydroxylase (TpH), to synthesize 5-HT which is then stored in secretory granules. 5-HT is released from EC cells in response to luminal stimuli, including mechanical forces. Once released, 5-HT acts on receptors located on the processes of sensory neurones that pass into the lamina propria. These include branches of intrinsic sensory neurones whose cell bodies are located in submucosal and myenteric ganglia, as well as sensory neurones located in spinal (dorsal root) ganglia and vagal (nodose) ganglia (see Figure 1). As a result, release of 5-HT from EC cells initiates activation of motor, secretory and vasodilatory reflexes, as well as stimulation of afferent signals to the brain and spinal cord.
An important facet of efficient intercellular signalling is the termination of the signal. In the case of 5-HT signalling in the brain, nerve terminals that release 5-HT express a serotonin-selective reuptake transporter (SERT), so they are uniquely capable of terminating the signals that they initiate.7, 8 5-HT neurotransmission in the brain can be augmented by compounds that function as selective serotonin reuptake inhibitors (SSRIs). These drugs, fluoxetine (Prozac), citalopram (Celexa) and paroxetine (Paxil) and related compounds, increase 5-HT availability in the synaptic cleft by inhibiting SERT.
The intestinal mucosa has an enormous capacity to remove 5-HT from the interstitial space because all of the epithelial cells that line the luminal surface of the gut express SERT.5, 9 Therefore, once 5-HT is released from EC cells and acts on local receptors, it is transported into epithelial cells by SERT. Postprandially, 5-HT also enters the bloodstream where it is rapidly taken up by platelets, which also express SERT.10 In fact, 5-HT found in platelets arises primarily from 5-HT released from EC cells.11, 12
Evidence for altered serotonin signalling in IBS
It is becoming increasingly clear that changes in 5-HT signalling occur in IBS. Due to the importance of 5-HT as an intercellular signalling molecule in intrinsic and extrinsic gut reflexes, various elements of 5-HT signalling have been evaluated to determine whether it is altered in IBS.5, 13–17 These include EC cell numbers, TpH message levels, 5-HT content, 5-HT release, SERT immunoreactivity, SERT message levels and platelet-free serum 5-HT levels. The findings of these studies are summarized in Tables 1–3.
A common feature of the studies conducted to date is that they have all reported changes in one or more aspects of 5-HT signalling in IBS. However, the results of these investigations are not entirely in harmony. For example, various combinations of changes in EC cell populations and 5-HT content have been reported in the different forms of IBS. One of these studies reported that no changes in 5-HT release were detected in IBS-D or IBS-C under basal or stimulated conditions.5 If this finding reflects the physiological nature of 5-HT release in these individuals, it would indicate that the same amount of 5-HT is being released regardless of possible changes in EC cell numbers or 5-HT content, and therefore changes in 5-HT signalling upstream of release may be irrelevant. When interpreting data related to elements of 5-HT signalling downstream of 5-HT release, both consistencies and contradictions exist. In the case of IBS-D, Coates et al.5 reported a decrease in SERT expression, indicating that the capacity of the epithelial cells to sequester locally released 5-HT would be compromised. Consistent with this, Dunlop et al.15 reported an increase in serum 5-HT levels, which was also reported in IBS-D. On the other hand, a decrease in SERT expression was also reported in IBS-C whereas a decrease in serum 5-HT was observed in this form of IBS. Further studies will have to be conducted to validate the findings that have been reported to date and to resolve the discrepancies that exist.
Is decreased SERT expression in IBS due to genetic and/or epigenetic factors?
The 5-HT transporter, SERT, is a highly regulated protein. For example, the genetic make-up of an individual appears to influence their level of SERT expression.18 Furthermore, activation of G-protein-coupled receptors can lead to changes in SERT by modulating its transcription by moving the transporter to or away from the plasma membrane, and by altering the phosphorylation state of the protein. It is therefore possible that genetic and/or epigenetic factors could affect SERT expression and function in the intestinal epithelium.
It is possible that individuals with IBS have a genetic predisposition for decreased SERT expression. The human SERT gene is located on human chromosome 17; it spans 37.8 kb and is composed of 14 exons that encode a 630 amino acid protein.19, 20 One possible explanation for the variability in SERT expression amongst individuals is the SERT gene-linked polymorphic region (SERT-LPR).21 The SERT-LPR, comprising long (l) and short (s) variants, is located 1.2 kb upstream of the transcription start site in exon 1, and may influence the level of transcription. The l variant contains 16 copies of a 20–23 base pair sequence whereas the s variant contains 14 copies of this sequence. Lymphoblast transfection studies with expression vectors containing the long or short variants linked to a reporter gene indicate that the long variant directs higher levels of transcription compared with the short variant [see Murphy et al. (18) for review]. Consistent with this, higher SERT mRNA levels and higher rates of 5-HT uptake were measured in lymphoblasts of l/l homozygotes compared with those containing at least one copy of the s allele.22 The l variant was also associated with higher rates of 5-HT uptake, and higher levels of 5-HT binding in platelets.23–25
It is important to note, however, that not all studies have reported a correlation between SERT-LPR alleles and SERT expression. Analysis of SERT mRNA levels in a large number of permanent lymphoblast cell lines failed to detect a correlation with SERT-LPR genotype.26 In addition, no correlation between SERT-LPR alleles and promoter activity was observed with expression vectors assayed in COS-727, 28 or PC12 cells.22, 27 Furthermore, no statistically significant correlations have been detected between SERT-LPR alleles and binding of SERT ligands in the brains of healthy volunteers.29, 30
A number of studies have examined possible relationships between SERT-LPR alleles and psychiatric disorders, anxiety-related personality traits, suicide, alcoholism and response to antidepressants.18 Although many studies have reported correlations between SERT-LPR alleles and a given clinical phenotype, others have not.
The relationship between SERT-LPR and IBS is unclear at this point, as data from the four studies conducted to date are somewhat contradictory. The first such study31 examined possible associations between SERT-LPR and different clinical patterns of IBS, including IBS-C, IBS-D and IBS with alternating symptoms (IBS-A) in a Turkish population. The distributions of SERT polymorphisims were comparable between the healthy subjects and the IBS population as a whole. However, genotype variations were detected amongst the IBS populations, with the s/s genotype being more predominant in individuals with IBS-C and the l/s genotype being more predominant in individuals with IBS-D. The second study32 investigated genotypes related to SERT-LPR and α2 adrenoreceptors in individuals in the upper Midwestern USA with lower GI symptoms (n = 274) that were divided into IBS-D, IBS-C, IBS-A and chronic pain. In this study, no difference was detected between the SERT-LPR genotypes of control individuals vs. the lower GI symptom population as a whole, or controls vs. IBS-D (n = 128), IBS-C (n = 90) or IBS-A (n = 38). Results of this study did identify combinations of polymorphisms that were associated with high symptom scores indicating that the SERT-LPR genotype may predispose a given individual to lower GI symptoms. The authors concluded that future studies of interactions between candidate genes associated with motor, sensory and/or behavioural functions would be of interest. A third study,33 conducted in Korea, did not detect a relationship between SERT-LPR genotype in controls (n = 56) vs. the IBS population as a whole (n = 33) or the IBS-C (n = 12), IBS-D (n = 15) and IBS-A (n = 6) subgroups. They reported that s/s was the predominant genotype in healthy volunteers (n = 56) as well as in individuals with IBS. The fourth study34 involved comparison of SERT-LPR genotype in a large cohort of North American Caucasian females with IBS-D (n = 194) vs. North American Caucasian female healthy volunteers (n = 448). In this investigation, a significant association was observed between IBS-D and the s/s genotype of SERT-LPR.
Correlative support for a genetic predisposition for decreased SERT expression comes from a study that involved an analysis of SERT-binding kinetics in platelets. Similar to serotonergic neurones and intestinal epithelial cells, platelets express SERT on their outer membranes and they act as sponges for 5-HT that enters the bloodstream. In a study by Bellini et al., SERT-binding kinetics of platelets from individuals with IBS-D were compared with those of healthy volunteers.35 Both the maximal binding capacity and dissociation constant were lower in individuals with IBS-D, suggesting a lower density of SERT of platelets in individuals with IBS-D.
The study demonstrating the most dramatic correlation between IBS and SERT-LPR was the study performed by Hicks and colleagues34 that involved the largest sample size and this study involved a relatively homogeneous population (North American Caucasian females). The finding that a link between SERT-LPR and a specific subtype of IBS, but not in IBS as a whole, or subtypes of IBS in heterogeneous populations supports the concept that a universal aetiology for IBS does not exist. It is likely that a variety of conditions, triggers and contributing factors result in the wide array of symptoms that are grouped into the descriptor, IBS. While it is not yet possible to draw a firm conclusion regarding a genetic predisposition for IBS, this is clearly a promising avenue to explore, and any links that are discovered could possibly lead to more effective treatment plans.
In addition to being influenced by means of genetic polymorphisms, SERT expression and function can also be modulated via second messengers that are activated by homologous (5-HT) and heterologous receptors.8, 36 For example, 5-HT, cholera toxin and forskolin increase SERT expression, while cyclic adenosine monophosphate (cAMP) activators increase SERT activity. SSRIs decrease SERT expression, and PKC activation decreases SERT activity. Furthermore, adenosine receptor stimulation activates two different protein kinase G (PKG)-dependent pathways that increase 5-HT uptake by moving SERT to the cytosolic membrane and by phosphorylating SERT via a mitogen activated protein (MAP) kinase-dependent mechanism.
Various elements of 5-HT signalling are altered in IBD5, 37–39 (see Table 4). In addition, 5-HT content, EC cell numbers, release and SERT expression are altered in a number animal models of intestinal inflammation, including trinitro benzenesulphonic acid (TNBS) colitis40 and ileitis41 in guinea-pig, TNBS colitis in mouse,42Trichinella spiralis ileitis in mouse,43Citrobacter rodentium colitis in mouse44 and dextran sodium sulphate colitis in rat45 (see Table 5).
Table 4. Mucosal serotonin signalling in inflammatory bowel disease
The changes that have been detected in the EC cell population and 5-HT levels in postinfectious IBS (PI-IBS) may be brought on by the inflammatory response associated with the infection. The data from experimental animals, which represent genetically homogeneous populations, indicate that the inflammatory process can lead to changes in EC cells as well as SERT expression by enterocytes. This concept is supported by data we presented at the 2005 meeting of the American Motility Society demonstrating that tumour necrosis factor alpha and γ-interferon decrease SERT function in CACO-2 cells, a human colonic epithelial cell line.46 Furthermore, the decrease in SERT immunoreactivity that is detected in T. spiralis-infected mouse intestine persists following recovery from inflammation.43 It is worth noting that EC cells and mucosal 5-HT levels are consistently found to be elevated in animal models of intestinal inflammation and in PI-IBS (Tables 4 and 5), whereas this trend is either not detected or not as consistently observed in IBD, IBS-D or IBS-C (Tables 1–3). As the animal models and PI-IBS involve relatively short time points, it is possible that EC cell hyperplasia occurs and 5-HT levels are elevated during the inflammatory response and then rebound to normal or below normal levels over time.
The cause and effect relationship of altered serotonin signalling in IBS
Serotonin is clearly an important signalling molecule in the activation of motor and secretory reflexes and in the activation of sensory signals from the gut to the CNS. The findings that are described above demonstrate that 5-HT signalling is altered in IBS-D, IBS-C and PI-IBS, but the cause and effect relationship of epigenetic changes in the elements of 5-HT signalling is unclear. In other words, we do not know whether changes in 5-HT signalling contribute to the alterations in GI function and sensation that are the hallmarks of IBS, and/or if elements of 5-HT are altered in response to disrupted function and sensation.
We do not yet know whether 5-HT signalling changes in response to altered gut function, but several lines of evidence support the concept that altered 5-HT signalling can lead to changes in gut function. For example, transgenic mice lacking the gene for SERT typically exhibit symptoms similar to those of IBS-D, but some mice are more similar to IBS-C, as they have decreased colonic motility.47In vitro studies involving evaluation of propulsive motility in the guinea-pig distal colon also demonstrate that changes in 5-HT signalling can affect motility. For example, administration of low concentrations of the SSRI, fluoxetine (Prozac), increases the rate of propulsive motility at low concentrations and slows motility at higher concentrations.40, 48 Furthermore, administration of desensitizing concentrations of 5-HT decreases propulsive motility in vivo. These findings support the concept that peristaltic reflex is complicated, and while it can be augmented by slight stimulation, over-stimulation can lead to decreased efficiency, possibly through receptor desensitization in the case of increased 5-HT availability.
The relationship between serotonin receptor drugs and the pathophysiology of IBS
Receptors that have been identified for 5-HT include the following: 5-HT1A-E, P, 5-HT2A,B,C, 5-HT3, 5-HT4, 5-HT5, 5-HT6 and 5-HT7. Of these, 5-HT1A, 5-HT1P, 5-HT2, 5-HT3, 5-HT4 and 5-HT7 receptors have been identified in the gut, with 5-HT2 receptors being located primarily on muscle and epithelium, and the others being located primarily on nerves.49
Treatment strategies for IBS involving 5-HT-related compounds are currently directed at the predominant symptoms, and 5-HT3 and 5-HT4 receptors have been the targets of choice. The 5-HT3 receptor is a ligand-gated ion channel,50 whereas the 5-HT4 receptor is a seven-transmembrane domain, G-protein-coupled receptor.51 The 5-HT4 receptor is linked to GS, and activation of this receptor leads to an increase in cyclic AMP. While both types of receptors can undergo desensitization, 5-HT3 receptors exhibit rapid desensitization upon sustained exposure to an agonist.52 This property of 5-HT3 receptors bring into question clinical potential of 5-HT3 agonists, particularly those with full agonist properties, and would make 5-HT3 receptors particularly susceptible to increased availability of 5-HT when SERT is decreased.
In IBS-D, the 5-HT3 receptor antagonists such as alosetron are thought to work primarily by inhibiting the activation of 5-HT3 receptors located on the mucosal processes of intrinsic and extrinsic primary afferents. Inhibiting 5-HT3 receptors located on intrinsic sensory neurones can diminish motor and secretory reflex activity, and decreasing the activation of extrinsic sensory neurones, which transmit signals to the CNS, inhibits sensory signals leading to pain and discomfort.53
5-HT4 receptor agonists such as tegaserod are used as a treatment for IBS-C. It is not yet clear whether 5-HT4 receptors are located on the mucosal processes of intrinsic and extrinsic sensory neurones, but if these receptors are present at this location, 5-HT4 agonists would augment the activation of motor and secretory reflex activity. Electrophysiological and anatomical studies have demonstrated that 5-HT4 receptors are located on nerve terminals throughout the ENS that mediate excitatory synaptic transmission.54–56 Activation of these presynaptic 5-HT4 receptors leads to an increase in the amount of transmitters, such as acetylcholine, that are released from these terminals, and therefore enhances reflex responses.
It is not understood whether 5-HT3 antagonists and/or 5-HT4 agonists have an effect on the 5-HT signalling anomalies that have been detected in IBS, or whether their mechanisms of action simply involve increasing or decreasing reflex activity at sites further downstream in the circuitry. It is possible that increased 5-HT availability contributes to the symptoms of IBS-D, and if this is the case, an antagonist working at the site of 5-HT signalling between EC cells and sensory nerves would be operating at the site of disrupted 5-HT signalling. Conversely, if 5-HT4 receptors are also located on mucosal projections of afferent nerves, as some data suggest,57 5-HT4 agonists may promote the activation of enteric reflexes at the site of disrupted 5-HT signalling, and thereby relieve constipation. The presynaptic action of 5-HT4 agonists on nerve terminals in enteric neural circuitry is downstream from the site of disrupted mucosal 5-HT signalling. In an intriguing report by Tonini et al., 5-HT7 receptors were found to mediate intestinal smooth muscle relaxation and accommodation. Endogenous 5-HT was involved in smooth muscle accommodation in the preparatory phase of peristalsis by direct activation of 5-HT7 receptors. The authors suggest that abnormal stimulation of the 5-HT7 receptor may contribute to certain clinical syndromes like IBS and that the 5-HT7 receptor may be a reasonable candidate for therapeutic intervention in some patients.58
The studies summarized here provide compelling evidence for a role for altered mucosal 5-HT signalling in IBS, as well as in IBD and other GI disorders. While the data reported in these studies represent snapshots of 5-HT signalling in what are generally chronic conditions, it is clear that the sequence of events responsible for 5-HT signalling can be remodelled in response to various physiological and pathophysiological conditions. Although the cause and effect relationship of these changes has not been established, it is quite likely that altered 5-HT signalling contributes to abnormal gut function and heightened sensitivity in IBS. Further studies will be required to gain a more complete picture of the changes that are occurring, and which of these changes have pathophysiological consequences.
One part of the picture that is missing is the status of 5-HT receptors in IBS, and also in IBD. Similar to SERT, 5-HT receptor expression is dynamic and could be affected by the amount of 5-HT that is available and by other factors such as inflammatory mediators. For example, adaptive changes in 5-HT3 receptor expression by enteric neurones have been detected in SERT knockout mice.59 Studies of 5-HT receptor expression in the intestinal mucosa have been hampered by the fact that mRNA for these receptors is in the neuronal cell bodies, and is therefore not acquired in a mucosal biopsy. As effective receptor-selective antisera become available, progress may be made in this area. A comprehensive understanding of mucosal 5-HT signalling in IBS and IBD will require a thorough appreciation of the status of receptors that mediate the actions of 5-HT.
It is highly unlikely that one given defect or alteration is responsible for the various changes in gut function and sensitivity that are encountered in disorders of GI function. In other words, a universal aetiology for IBS is not likely to exist. For example, in addition to the alterations in 5-HT signalling that are summarized here, it is clear that a co-morbidity exists between psychiatric disorders and IBS, and that many cases of IBS-D involve a previous infectious inflammatory event (postinfectious IBS). Furthermore, mounting evidence suggests a role of the stress hormone, corticotropin-releasing factor, in IBS.60 Therefore, the variability reported in many IBS studies may be due to the heterogeneity of the patient population. In line with this, it is unlikely that a given therapy will be highly effective in all individuals with IBS, or even a subtype of IBS. Further elucidation, at the molecular level, of the various changes that contribute to the symptoms of the various forms of IBS will hopefully enhance our ability to treat these individuals effectively. In order to make progress in this regard, it will be crucial for investigators to take care in defining the phenotypes of the individuals that are included in a given study, including their predominant GI symptoms, their psychiatric and past medical histories. Furthermore, DNA samples should be acquired for genotyping of these individuals as we identify candidate genes. This type of approach will improve our understanding of the role of 5-HT and other contributing factors in IBS. While the Rome criteria are adequate for making the diagnosis of IBS, assays for detecting molecular abnormalities that are found to contribute to the disorder may assist in identifying those individuals who are most likely to respond to a given treatment plan. This is a significant goal that would ease the frustration and burden of health care providers and patients alike.
This work was supported by NIH grants DK62267, P20 COBRE grant RR16435 and NS050919, and a research grant from Novartis Pharmaceuticals.