Irritable bowel syndrome (IBS) is a common functional gastrointestinal disorder, characterized by recurrent abdominal pain or discomfort in combination with disturbed bowel habits in the absence of identifiable organic cause. Many medications are used for the treatment of patients with IBS, although their actual benefit is often a matter of debate. In particular, only a few are specifically labeled for IBS. In fact, notwithstanding great progress in our understanding of the pathophysiology of IBS thanks to a large number of preclinical and clinical studies of new drugs, the specific armamentarium of medications available is scant. The aim of this review is to outline the current challenges in drug development for IBS, taking advantage of what we have learned through the Rome process (from Rome I in the 1980s to Rome III published in 2006; Drossman, 2006).

The key questions that will be addressed are: (a) do we still believe in the “magic bullet,” i.e., a very selective drug displaying a single receptor mechanism capable of controlling IBS symptoms? (b) IBS is a “functional disorder” where complex neuroimmune and brain-gut interactions occur and minimal inflammation is often documented: do we need to target gut motility, visceral sensitivity, or minimal inflammation? (c) are there validated biomarkers (accepted by regulatory agencies) for studies of sensation and motility with experimental medications in humans? (d) do animal models have predictive and translational value? (e) in the era of personalized medicine, does pharmacogenomics applied to these medications already play a role? Finally, this review will briefly outline medications currently used or in development for IBS. It is anticipated that a more focused interaction between basic science investigators, pharmacologists, and clinicians will lead to better treatment of IBS.

The key pharmacodynamics, pharmacokinetic, and safety features for drugs to be used in the treatment of IBS are outlined in Table 1.

A selective drug is defined as a compound interacting only with one receptor subtype and leaving other receptors unaffected at concentrations achieved at therapeutic doses. The literature on the treatment of IBS has often resorted to the concept of the “magic bullet,” i.e., a very selective drug displaying a single receptor mechanism capable of controlling IBS symptoms (Camilleri et al., 2006a). This was often considered the key to efficacy avoiding side-effects. This approach is no longer ideal because of several important pitfalls.

First, drug selectivity is always a relative concept, which ignores the basic fact that most molecules, even at therapeutic doses, may have several, sometimes disparate biological effects (i.e., hit a large number of targets in the pharmacological space; Garcia-Serna et al., 2010; Kawasaki and Freire, 2011). These effects may depend on the fact that a single receptor/effector pathway plays a role in different systems, so that even selective compounds have off-target effects (Table 1). In addition, there are many instances when the compound is endowed with additional pharmacological properties that hit the so-called antitargets (i.e., unwanted targets), responsible for side-effects, which are clarified only after the compound has undergone clinical trials. The classical example is provided by the cardiac side-effects due to hERG K⁺ channel blockade by the early 5-HT₄ receptor agonists (Tonini et al., 1999).

The second issue is that the multifactorial pathophysiology of IBS (with multiple brain-gut and neuroimmune interactions) makes it unrealistic to expect that drugs acting on a single receptor may achieve substantial therapeutic gain over placebo in an area where the placebo response rate is substantial (approaching 40% across all randomized controlled trials; Ford and Moayyedi, 2010). As in other fields (Morphy et al., 2004), evidence suggests that a balanced modulation of multiple targets can provide a superior therapeutic effect and side effect profile compared to the action of a selective ligand. Designed multiple ligands that hit a large variety of targets have been produced through rational approaches in which structural features from selective ligands are combined (Morphy et al., 2004). A key challenge in the design of multiple ligands is attaining a balanced activity at each target of interest with a suitable pharmacokinetic profile.

The third issue is that mechanisms underlying symptoms in IBS may differ among patients, hence the need to consider using multiple therapies. With selective drugs, primary clinical endpoints were achieved in less than 70% of patients with the approved agents such as tegaserod or alosetron (Camilleri et al., 2000; Muller-Lissner et al., 2001; Cremonini et al., 2003). On the other hand, it seems reasonable to propose treatment with combination therapy, which is the rule when treating medical conditions such as hypertension or asthma, when monotherapy is no longer adequate. Because of the redundancy of mechanisms controlling neurosensory, neuromuscular, and neuroimmune functions in the gut, it is conceivable that effective treatment of functional gut disorders may require combination therapy.

One example is provided by tachykinin receptor antagonists, which have so far given disappointing results because of inherent differences among animal models and humans: it has been suggested that the analgesic efficacy of multi- or pan-tachykinin receptor antagonists is superior to that of mono-receptor antagonists (Holzer, 2004a).

When drugs address a specific target (e.g., a symptom such as visceral hypersensitivity or motility), heterogeneity in the pathophysiology impacts negatively on the therapeutic gain, if patients are not carefully selected in a clinical trial. Indeed, some of the disappointing results of the past can be ascribed to the lack of understanding of pathophysiology: the same symptom (e.g., diarrhea) does not necessarily depend on the same pathways in all patients.

Thus, new drugs should target a pathophysiological mechanism (provided that it is known!), rather than a specific receptor; on the other hand, recruiting carefully selected patient subgroups may significantly reduce the generalizability of the results of the trial.

Pharmacokinetics may help to achieve gut selectivity and reduce side-effects. This approach is particularly relevant when there are potential actions outside the gut, as it is indeed the case with peripherally restricted opioid receptor antagonists (such as methylnaltrexone and alvimopan), which do not cross the blood brain barrier and, in addition, have very low oral bioavailability (De Ponti, 2002; Holzer, 2004b). An example of minimally absorbed compound is also the guanylate cyclase-C agonist linaclotide (Wensel and Luthin, 2011; Busby et al., 2013), which is now FDA- and European Medicines Agency (EMA)-approved for IBS with constipation.

Another important pharmacokinetic property is the lack of interactions with food or other drugs. Significant interactions with CYP2D6 and CYP3A4 should be predicted in early drug discovery because of their involvement in drug metabolism with important pharmacogenetic aspects.

Finally, as regards safety aspects, apart from the standard safety evaluations, two issues deserve special attention following the experience with cisapride (torsades de pointes associated with QT prolongation; De Ponti et al., 2001) and alosetron (ischemic colitis; Moynihan, 2002). It is clear that even very rare events may negatively impact the risk/benefit balance of drugs that are used to provide symptom improvement of non-serious (though troublesome) diseases such as IBS (De Ponti et al., 2002; Tack et al., 2012). It is remarkable that in IBS with diarrhea, Shah et al. (2012) found that one adverse event resulting in study discontinuation occurred for every 2.3 and 2.6 patients who benefited, respectively, from tricyclic antidepressants and alosetron, i.e., the number needed to harm was approximately 3. This is quite low, considering the numbers needed to treat reported in the literature for drugs in IBS (Brandt et al., 2009; Camilleri and Mayer, 2009; Menees et al., 2012). Shah et al. (2012) conclude that, rather than simply focusing on the number needed to treat, clinicians should be aware of harm when using pharmacotherapy for IBS.

The classical concept of IBS as a functional disorder derives from the fact that no organic cause can be identified and the diagnosis of IBS is one of exclusion after other disorders have been ruled out. In addition, the precise mechanisms underlying symptom generation are unknown.

However, research of the past 20 years has provided significant advances in the understanding of the pathophysiology of IBS, with an emerging consensus that the various clinical manifestations (including non-gastrointestinal comorbid symptoms) of chronic abdominal pain can best be viewed as a dysregulation in the complex interplay between events occurring in the gut lumen (including microbiota), the mucosa, the enteric nervous system (ENS), and the central nervous system (Mayer and Tillisch, 2011; Matricon et al., 2012). This dysregulaton leads to alterations in sensation, motility, brain-gut interactions, and neuroimmune interactions. Considerable evidence documents that sensitizing proinflammatory mediators, mast cells and their products, tryptases, are increased in tissues of patients with colorectal hypersensitivity (Cenac et al., 2007; Balestra et al., 2012; Buhner et al., 2012).

It has been shown that colonic mast cell infiltration and mediator release in proximity to mucosal innervation may contribute to abdominal pain perception in IBS patients (Barbara et al., 2004). Indeed, mucosal mast cell mediators from IBS patients excite rat nociceptive visceral sensory nerves (Barbara et al., 2007). In a recent study (Balestra et al., 2012), mucosal biopsies were obtained from the descending colon of patients with IBS and controls. Mucosal mast cells were identified immunohistochemically. The impact of spontaneously released mucosal mediators on guinea pig electrically stimulated longitudinal muscle myenteric plexus (LMMP) preparations was assessed in vitro by means of selective receptor antagonists and inhibitors. Patients with IBS showed an increased mast cell count compared with controls. Application of mucosal mediators of IBS to LMMPs potentiated cholinergic twitch contractions, an effect directly correlated with mast cell counts and mediated by activation of prostanoid receptors, TRPV1, and P2X receptors. These results support the role of mucosal inflammatory mediators and mast cell activation in altered motor function of IBS.

It is also intriguing that, in patients with IBS, 5-HT spontaneous release was significantly increased irrespective of bowel habit and correlated with mast cell counts and the severity of abdominal pain. This suggests that increased 5-HT release contributes to development of abdominal pain in IBS, probably through mucosal immune activation (Cremon et al., 2011).

Several studies have reported the onset of IBS-like symptoms following established bacterial or viral infections of the GI tract (Barbara et al., 2009). This so-called “postinfectious” IBS occurs in approximately 10% of patients undergoing a documented infectious gastroenteritis, and risk factors to develop symptom persistence are longer duration of the gastroenteritis, female sex, psychosocial stressors at the time of the infection, and psychological factors such as anxiety or depression. Although a causal relationship between abdominal pain and acute or chronic infections cannot be established most of the times, it is tempting to speculate that host-microbial interactions in vulnerable individuals during the early phase of the disorder may lead to permanently altered immune response, which then continues to play a role when symptoms persist in the absence of the infectious organism.

The participation of the gut microenvironment in the pathophysiology of IBS is suggested by studies indicating an interplay between luminal factors including the microbiome, the epithelial barrier, and the mucosal immune system (Stanghellini et al., 2010; Camilleri et al., 2012). In an animal model (McVey Neufeld et al., 2013), microbiota were shown to be necessary for normal excitability of gut sensory neurons and this provides a potential mechanism for the transfer of information between the microbiota and the nervous system.

In postinfectious IBS and in IBS with diarrhea, decreased expression and structural rearrangement of tight junction proteins in the small bowel and colon may lead to increased intestinal permeability. These abnormalities might contribute to the outflow of antigens through the epithelium, causing overstimulation of the mucosal immune system. Accordingly, subgroups of patients with IBS show higher numbers and activation of mucosal immunocytes, especially mast cells. Immune factors, released by these cells, including proteases, histamine, and prostanoids, might also participate in maintaining the permeability dysfunction and contribute to the activation of abnormal neural responses, which, in turn, are involved in abdominal pain perception and changes in bowel habits.

All these mechanisms represent new therapeutic targets in IBS. Here, it is important to remember that probiotics are also currently viewed as an attractive therapeutic option in IBS because of their recognized safety and of their documented biological effects on the host. Preclinical studies have shown that some probiotic strains exhibit potentially useful properties including anti-inflammatory effects, improvement of mucosal barrier homeostasis, beneficial effects on intestinal microbiota, and a reduction of visceral hypersensitivity. However, it remains to be determined to what extent a beneficial effect on these parameter translates to a significant effect on clinical outcomes: although the effect of probiotics on IBS is positive in some randomized, controlled studies, the gain over placebo is small and identification of a tailored probiotic approach for subgroups of patients represents a future challenge.

The complex neuroimmune and brain-gut interactions sometimes associated with minimal mucosal inflammation (Ford and Talley, 2011) and neuroplastic changes in the ENS (Giaroni et al., 1999) pose several questions as regards potential targets for pharmacological intervention: should the therapeutic focus be primarily gut motility, visceral sensitivity, or minimal inflammation? Assessment of these parameters in humans can be undertaken by using a variety of invasive and non-invasive techniques, some well established and others requiring further validation. By using these techniques, alterations in both sensory and motor function have been reported in IBS and our understanding of sensorimotor dysfunction has indeed increased. Thus, inflammatory, immunologic, and other processes, as well as psychosocial factors such as stress, can alter the normal patterns of sensitivity and motility through alterations in local reflex activity or via altered neural processing along the brain-gut axis. A firm relationship between sensorimotor dysfunction and the production of symptoms, however, has been difficult to show. Thus, the clinical relevance of the former requires further research.

In this context, it is important to remember that in 2003 the EMA adopted a document produced by the Efficacy Working Party on the “Points to consider on the evaluation of medicinal products for the treatment of the irritable bowel syndrome” (EMA, 2003). Although the document now needs to be updated (it still refers to Rome II criteria), a key statement is that “The patient’s global assessment of symptoms and abdominal discomfort/pain should be used as the two primary endpoints. Statistically significant changes must be found in both parameters.” Thus, clinical efficacy must rely on clinical endpoints in the patient’s perspective, for instance through the global assessment of multiple symptoms. Mechanistic (pathophysiological) studies provide a rationale for drug development, but do not generally predict symptomatic success and do not necessarily identify the most appropriate dose for clinical trials. An important goal is to develop non-invasive tests that identify important pathophysiological mechanisms and assess symptom pattern in short-term (4 weeks) therapeutic trials that pave the way to longer trials. Notably, the EMA document carefully considers the duration of efficacy trials, stating that they must be long enough to determine whether the response is sustained and to determine the effects of treatment withdrawal. A duration of 6 months of active treatment is considered necessary considering the cyclic and non-life threatening nature of the disease.

For inclusion and exclusion criteria for IBS, the current EMA “Points to consider” document refers to Rome II criteria (current Rome III criteria differ from Rome II and, notably, the Rome IV process is expected to start in 2013). A revised EMA guideline is awaited soon and should come into force by the end of 2013. A key issue that needs to be addressed, apart from the update to Rome III, is the discrepancy with the FDA guidance issued in 2012 (FDA, 2012). Indeed, the EMA document recommends the two co-primary endpoints indicated above, whereas the FDA guidance recommends a primary endpoint that measures the effect of treatment on two major IBS signs and symptoms: abdominal pain and abnormal defecation (stool frequency or stool consistency, depending on subtype of IBS). The FDA guidance also acknowledges that patient-reported outcome (PRO) measures of the signs and symptoms of the condition are the only currently available measures that can adequately define a treatment effect in a clinical trial. In addition, because of the limitations of using a single-item patient-reported rating of overall change as a primary endpoint, the FDA document recommends the development of a multi-item PRO instrument. The PRO measure(s) should capture all the clinically important signs and symptoms of the target population. The ongoing regulatory discussion will certainly help all those involved in clinical trials to plan future research.

Linaclotide may serve as an example of the current regulatory situation, with differences in endpoints recommended by the FDA and the EMA: the efficacy and safety of this agent in patients with IBS-C was evaluated in two randomized, placebo-controlled Phase 3 trials. These trials were designed according to both FDA and EMA guidelines and findings based on FDA-recommended endpoints were reported in two recent studies (Chey et al., 2012; Rao et al., 2012), whereas the findings of a planned, separate analysis of both trials based on the distinct efficacy parameters prespecified for EMA submission were published separately (Quigley et al., 2013).

In closing this section, it should be remembered that, over the past 25 years, the Rome process has insisted on clinical features to diagnose IBS on the assumption that grouping of patients with similar features facilitates identification of patients most likely to respond to a given pharmacological agent, but this is not necessarily so because the same symptom (pain) may have several underlying pathophysiological mechanisms. This explains the criticism raised by some investigators against the Rome criteria (Dang et al., 2012). On the other hand, it must be acknowledged that there are several reasons to establish an accurate diagnosis of IBS: to relieve patient uncertainty and initiate the most appropriate treatment, avoiding the burden of unnecessary medications or diagnostic procedures and surgeries (Mearin and Lacy, 2012). In other words, the Rome criteria try to transform the diagnosis of IBS from one of exclusion into a positive diagnosis based on history, physical examination, use of precise diagnostic criteria in the absence of specific alarm features.

Biomarkers are objectively measurable indicators of normal or pathological processes or pharmacological responses to a therapeutic intervention (Anonymous, 2001). In order to improve development and usage of biomarkers, a score system of different types of biomarkers has been proposed, depending on their impact on drug development (Wehling, 2009).

Positive modifications of biomarkers which imply improvement of the disease can be taken as endpoints during drug development. To provide a more complete picture, a therapeutic target can coincide with the biomarker (e.g., TNFα) or, as a component of the disease mechanism, modulate it (e.g., NF-κB).

Unfortunately, as stated in the 2012 FDA document, no validated and accepted biomarker exists in IBS. In addition, the limited repertoire of clinical manifestations of sensorimotor disorders of the gut such as IBS can actually derive from multiple mechanisms, leading in turn to similar symptoms. In many clinical programs of new drugs for IBS, the emphasis was primarily on symptom assessment of broad groups of patients identified by the Rome criteria. As already discussed above, this approach was not ideal and it not surprising that drugs of potential value have been abandoned.

Certain biomarkers can, in a limited fashion, be used to predict the success of a drug in IBS or to understand its mode of action. These studies may be incorporated in the recommended steps for drug development, but should be viewed only as preliminary/complementary steps of the development program, which must comply with the regulatory guidance quoted in the previous section.

Currently established tests that can be used as potential biomarkers for clinically relevant endpoints in IBS include the following.

Translational medicine defines conditions and prerequisites to transfer more reliable results obtained from preclinical biomedical research to clinical applications, thus improving patient care through timely and efficient promotion of clinical innovation.

A disease model should mimic the clinical disease condition as much as possible and allow the investigation of unclear pathophysiological mechanisms toward the development of new potential therapeutic options. In spite of undeniable differences among species, animal models still represent the major source of information about biological system and provide an invaluable means to study complex physiological and biochemical interactions.

The predictivity of the disease model itself will substantially depend on the efforts toward optimization and scientific validation of the model, both in terms of resemblance and transferability to humans. Pharmacological targets and biomarkers must be reproduced and be susceptible to pharmacological modulation in the animal model, to ensure that the model is predictive of a therapeutic effect in humans. The real translational value of a biomarker is represented by its capacity of being thoroughly informative after its validation in clinical studies which provide the bedside to bench approach.

An animal model used as a screening tool in drug development should ideally reproduce different biomarkers providing multiple endpoints to fully assess the spectrum of activity of a test compound. Accurate selection of endpoints plays a pivotal role to assure the translational value of disease models and should take into account the sensitivity and specificity of each endpoint toward prediction of activity.

Although development of new drugs for treatment of IBS can be facilitated by preclinical animal models (Bulmer and Grundy, 2011; Holschneider et al., 2011), it must be acknowledged that a single sufficiently predictive model of efficacy still does not exist and investigators have the option to use multiple models to assess different aspects of the pathophysiology of IBS. This section is only intended as a brief outline of the most commonly used animal models of visceral pain and disturbed gastrointestinal motility, which are reviewed elsewhere (Camilleri et al., 2006a).

Pharmacogenomics refers to the variability of the expression of individual genes relevant to disease susceptibility as well as drug response at cellular, tissue, individual, or population level. Pharmacogenetics refers more specifically to the study of individual variations in DNA sequence related to drug response. The growing interest in this field is due to the fact that risk/benefit evaluations of drugs are not fully appreciated if one does not fully consider individual variations that may significantly affect pharmacokinetics and pharmacodynamics.

A detailed discussion of all the classes of drugs proposed as potential therapeutic agents in IBS is beyond the scope of this review. The main pharmacological approaches to IBS are summarized in Table 2, which does not include agents traditionally used in IBS, such as laxatives and antidiarrheal agents, respectively for constipation and diarrhea, and probiotics. The gut microbiome as a therapeutic target is covered elsewhere (Floch et al., 2011; Simren et al., 2013). The reader is also referred to a recent detailed review of current and potential pharmacological approaches in IBS (Camilleri, 2012). Herbal preparations used in IBS are covered by another review (Rahimi and Abdollahi, 2012). Brain-gut interactions and possible sites/mechanisms of pharmacological intervention along the brain-gut axis are discussed in a recent review in this Journal (Fichna and Storr, 2012).

Several meta-analyzes of pharmacological treatments for IBS have been published in recent years and there is ongoing debate on their interpretation (Lesbros-Pantoflickova et al., 2004; Brandt et al., 2009; Camilleri and Mayer, 2009).

From the above overview, it is clear that research in the treatment of IBS is still very active. Although in the past decade some innovative pharmacological agents have not fulfilled their promise because of unexpected side-effects, it is likely that new pathophysiological concepts as well as the publication of new regulatory documents by the FDA and the EMA will be of great help for drug developers.

The author declares that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.