Management of functional gastrointestinal disorders
B honorary senior lecturer and associate registrar for public mental health, Royal London Hospital, London, UK, Centre for Neuroscience, Surgery and Trauma, London, UK and Royal College of Psychiatrists, London, UK
Functional gastrointestinal (GI) disorders (eg irritable bowel syndrome and functional dyspepsia) are very common conditions which are associated with very poor quality of life and high healthcare utilisation. They are caused by disorders of GI functioning, namely altered gut sensitivity, motility, microbiota, immune functioning and central nervous system processing. They cause chronic symptoms throughout the gut (eg pain, dyspepsia and altered bowel habit), all of which are made worse by maladaptive patient behaviours, stress and psychological comorbidity. Management involves a biopsychosocial approach involving changes in lifestyle and diet, addressing coexisting psychological comorbidity and using medication to treat underlying pathophysiology. Pharmacological treatment with antispasmodics, neuromodulators, motility agents and antidepressants is effective. Psychotherapy in motivated individuals is equally effective. Success of treatment is increased by a good doctor–patient relationship and so this needs to be taken into account during the consultation.
Functional gastrointestinal disorders (FGID) are a group of disorders characterised by chronic gastrointestinal (GI) symptoms (eg abdominal pain, dysphagia, dyspepsia, diarrhoea, constipation and bloating) in the absence of demonstrable pathology on conventional testing. Historically, they were defined as conditions which had no organic basis, but this definition has evolved with increasing understanding of these conditions and we now know that they arise due to alterations in brain–gut communication. The current classification system (ROME IV) divides them into 33 adult disorders and 20 paediatric disorders, the most common subtypes being irritable bowel syndrome (IBS) which causes abdominal discomfort, altered bowel habit and bloating; and functional dyspepsia (FD) which causes epigastric pain or discomfort, often related to eating which can be associated with fullness and satiety. 1
FGID are very common with a worldwide prevalence of 40%, more common in women than men and this decreases with age. 2 They account for 12% of the workload in primary care and 30% of gastroenterology outpatient consultations. 3,4 More than two-thirds of patients with FGID will have seen a doctor in the last 12 months and 40% will use regular medication. 5 FGID pose a huge economic burden and treating them cost the NHS at least £72.3 million in the year 2014/2015, of which, two-thirds was on prescriptions, community care and hospital treatment. 6
The presence of FGID is often associated with chronic pain (eg fibromyalgia) and other functional syndromes (eg chronic fatigue syndrome), and two-thirds will have psychopathology including anxiety and depression. 7 It is therefore not surprising that these patients have very poor quality of life, worse than other chronic medical conditions (eg grade III congestive cardiac failure and rheumatoid arthritis). 8
Due to the very low quality of life (QOL) of patients with FGID and the fact that they incur a lot of healthcare costs, it is important that they are recognised and managed promptly.
It is now clear that there is abnormal physiological functioning in patients with FGID, thought to be due to underlying alterations in GI motility (either too fast or too slow), visceral hypersensitivity, altered microbiota, increased intestinal permeability, low grade immune infiltration and altered central nervous system processing of sensory input. 9 However, symptoms and healthcare seeking arise for complex reasons involving an interplay between early life events and coping styles, learned behaviour, alterations in GI physiology, and associated psychological morbidity as seen in the biopsychosocial model in Fig 1. 9 Management is therefore not simply directed at the abnormal physiology or symptoms but has to address behaviours, cognitions and beliefs.
Biopsychosocial model of functional gastrointestinal disorders. CNS = central nervous system; ENS = enteric nervous system; FGID = functional gastrointestinal disorders.
The optimal approach involves a holistic assessment starting with a detailed history, taking care to exclude the presence of red flags (weight loss, family history of cancer, nocturnal symptoms, anaemia or GI bleeding) and organic differentials. It is important to ask about diet, lifestyle and psychological status, as this will enable you to target these as part of the management plan. Integral to all this is the doctor–patient relationship and the quality of the consultation (Box 1). This involves being empathic, avoiding jargon, being honest and admitting when you do not have the answers, which takes a lot longer than organising yet another futile test. In our opinion, 10 minutes is not sufficient for this kind of consultation, however, spending time addressing all these factors on the first visit and breaking the diagnosis of a functional disorder will save time (and money) on future visits.
Helpful questions during a consultation
Examination should include an assessment for abdominal masses, and quality of pain as well as a rectal examination. The latter is essential to rule out rectal masses and haemorrhoids, and to assess for anal tone and function. The latter can be assessed at baseline and by asking the patient to squeeze as if they are preventing themselves from emptying their bowels. Anal hypotonia is associated with faecal incontinence and hypertonia can be associated with dyssynergic defecation, itself a cause of constipation.
With the current ROME classification, it is possible to make a positive diagnosis of FGID based on the pattern of symptoms, and so exclusion of all organic disease is not necessary. However, serious differentials that seem feasible after taking a good history (Box 2) should be ruled out.
Differentials for irritable bowel syndrome
All patients should get a basic blood and stool panel including:
full blood count to look for anaemia
urea and electrolytes to look for dehydration and evidence of electrolyte derangements with diarrhoea
C-reactive protein or erythrocyte sedimentation rate to look for underlying inflammation, this should be normal in IBS
thyroid function tests
faecal calprotectin, if diarrhoea is present, to rule out inflammatory causes of this
Helicobacter pylori testing (stool antigen test or urea breath test) for patients with dyspeptic symptoms.
Endoscopy: If a patient has typical IBS symptoms with a normal faecal calprotectin and there are no red flags to suggest a colorectal cancer (see earlier) then a lower GI endoscopy is not needed. There is little yield in performing a gastroscopy for H pylori negative dyspepsia in the absence of alarm symptoms (such continuous pain, vomiting, anaemia and weight loss in patients under the age of 60), so this should not routinely be organised. 10
Abdominal ultrasound: Abdominal ultrasound can be useful in IBS to screen for abdominal causes of pain and, in particular, for ovarian cancer which can cause pain, visible abdominal bloating and altered bowel habit. In dyspepsia, it can be useful to look for gallstones if the history is suggestive (ie colicky pain with fatty meals).
SeHCAT scan. If available, SeHCAT scans should be used to assess for bile salt malabsorption which is present in up to a third of patients with IBS-D. 11 Typical symptoms include watery diarrhoea, often yellow in colour, with or without nocturnal symptoms and faecal incontinence.
GI physiology. GI physiology is rarely indicated in IBS. One situation where it can be helpful is in patients who have severe constipation and are not responding to multiple laxatives. Lower GI physiology testing, particularly a colonic transit study and proctography can be useful at differentiating slow transit from a rectal evacuatory problem and can therefore help in fine tuning the management of constipation. In patients with functional dyspepsia, a gastric emptying study can be useful to look for severely delayed gastric emptying if there is persistent vomiting which is impacting on nutritional status, as this can help with decisions regarding feeding. For all physiological tests, it is important to be aware that medications, particularly opiates and anticholinergics, will alter GI motility and transit.
General and initial approach
Once you have diagnosed a FGID, it is important to put a label on it, as patients often complain that they do not have a diagnosis or that ‘nobody knows what is causing their symptoms’, and then to explain in simple language what FGIDs are; information sheets or online resources can be very useful. In order to manage patient expectations, it is useful to reiterate the incurable nature of FGID and to explain that the aim of management is not to remove symptoms completely or return the patient back to ‘normal’, but to give them more control over their symptoms so that the GI symptoms do not dominate their life.
It is helpful when managing these patients to address the bio-psycho-social factors, in reverse:
social/lifestyle factors: diet, exercise, sleep, and ingestion of caffeine, alcohol and other medication
psychological factors: presence of stress, anxiety, low mood and history of eating disorders
biological factors: physiological abnormalities and medication side effects which are contributing to symptoms.
The heterogeneity of FGID makes it difficult to design an algorithm to fit all patients, however, using a biopsychosocial approach and identifying factors which may have triggered symptoms and which ones are maintaining them enables the clinician to focus on modifying these factors as part of a personalised management plan.
Integrated multidisciplinary clinical care (eg gastroenterologist, nurse, dietitian and psychologist) appears to be superior to gastroenterologist-only care in terms of improving symptoms, psychological state, quality of life and cost of treating functional gastrointestinal disorders, so this should be offered where possible. 12
Exercise can improve bowel function, improve transit time (in females) and, therefore, help with constipation. 13 It can also reduce stress, improve mood and lead to better sleep, all of which impact on GI symptoms. In a randomised controlled trial (RCT), increased physical activity was associated with a greater reduction in IBS symptom severity scores. 14 Patients should be advised to do 20–30 minutes of exercise 3–5 times per week, even if this is just walking.
Sleep disturbances are associated with both upper and lower GI symptoms and worse QOL. 15 Sleep disturbances can be initial (anxiety preventing onset of sleep), middle (typically broken sleep) and late (early morning awakening (can be a sign of depression)). Digital technology (eg ‘Fitbits’ and phone apps) can measure sleep patterns and quality, and then advise interventions to improve this (eg Sleepio app). Patients should be advised on sleep hygiene. If a drug is needed, benzodiazepines should be avoided but melatonin (3 mg nocte) can be used. This has been shown to reduce abdominal pain and rectal hypersensitivity but not bloating or overall IBS symptom scores. 16
Apps can be useful for tracking lifestyle factors and monitoring changes to this. There are hundreds of healthcare apps available so it can be difficult to know what to recommend. Apps which are listed on the NHS Apps Library or ORCHA (www.orcha.co.uk) have been reviewed and can be more confidently recommended. We tend to recommend Sleepio for sleep and Headspace for mindfulness. Bowelle helps track symptoms in IBS. Zemedy is an IBS self-management app that uses a cognitive behavioural therapy (CBT) approach. The Monash University Low FODMAP Diet™ App helps users identify foods with high fermentable oligo-, di-, and monosaccharides and polyols (FODMAP) content and is useful for patients who are on the low FODMAP diet (further details follow later).
Stimulants and depressants
Caffeine increases diarrhoea, alcohol worsens reflux symptoms and heavy use is associated with increased risk of functional dyspepsia. 17 Both caffeine and alcohol can lead to disrupted sleep, so these should be minimised and not taken late at night in symptomatic individuals.
Cannabis misuse, with direct causative links to vomiting syndromes, has latterly been replaced by internet-bought cannabidiol and other plant derivatives. Even if the patient experiences none of the commonly associated GI side effects, such as vomiting and diarrhoea, there is an association with anxiety, fatigue, somnolence, risk of dependence and the potential for drug interactions, so this should not be recommended. 18,19
Diet and nutrition
In FGID, symptoms are frequently associated with food intake, and so dietary assessment and optimisation should be part of the initial management strategy. Patients can be asked to keep a food diary to identify foods which trigger symptoms and to identify eating behaviours, however, the possibility of reporting bias should be noted.
IBS is associated with irregular dietary patterns and reduced diet quality and diversity (ie not having the recommended range and quantity of particular food groups), so education about how to eat as well as what to eat is important. 20,21 National Institute for Health and Care Excellence (NICE) has a useful information sheet for an ‘IBS diet’ which focuses on eating small regular meals, avoiding insoluble fibre, fatty foods, gas producing foods and caffeine.
There is evidence for symptom improvement following reduction in lactose and in high-starch and sugary foods and drinks, so these can be reduced if appropriate. 20
The identification that consumption of foods high in FODMAP can exacerbate symptoms of IBS because of fermentation and osmotic effects (bloating/pain and diarrhoea, respectively) has led to the use of a low FODMAP diet as a dietary intervention for IBS, particularly in IBS-D. 22 However, this diet runs the risk of being overly restrictive and the long-term effects on nutrition and the colonic microbiome are unclear, so it is important that this is implemented by a trained dietitian. 23
In two RCTs in Sweden and the USA, there was a similar reduction (40–50%) in IBS symptom scores in both a ‘low FODMAP diet’ and an IBS/NICE diet. 24,25 Thus, in the absence of dietetic support for the former, it is useful for a clinician to provide generic dietary advice in clinic based on NICE guidance.
The evidence for a gluten free diet is less clear cut. While minimising gluten intake is associated with a reduction in abdominal pain, it is unclear whether this is due to the fact that gluten is a high FODMAP food. 26 Patients can be asked to reduce their gluten intake if this is felt to be helpful but there is no clear evidence for restricting this completely.
Soluble fibre – psyllium or isphagula husk (eg fybogel) – is cheap and useful for symptoms of IBS, particularly in IBS-C. This should be started at low dose and increased slowly. There is no evidence for insoluble fibre or for bran, both of which can exacerbate pain and bloating. 27
Patients with FGID have altered microbiota, and this will be increased if patients have been on long-term or recurrent antibiotics. Microbial alteration with probiotics can be trialled and there is growing evidence for this. A meta-analysis demonstrated that probiotics improve global IBS symptom scores as well as individual symptoms of abdominal pain, bloating and flatulence. 28 Probiotics can improve stool consistency and frequency in both IBS-D and IBS-C but this is not as clear-cut as its effect on pain and bloating. 29 Combination probiotics are more likely to be beneficial but there is no consistent data to suggest what that combination should be, nor what the dose should be. 30 Although the long-term effect of probiotics on the microbiota is unknown, they are generally safe and so trialling them for 2–3 months can be considered early on in the treatment of IBS-D and possibly IBS-C.
There is no evidence to recommend faecal microbiota transfer, prebiotics (dietary supplements that result in specific changes in the composition and/or activity of the GI microbiota) or synbiotics (a mixture of probiotics and prebiotics that act synergistically to promote the growth and survival of beneficial organisms) at the current time.
Psychopathology: identification and psychological treatment
Although it may not be constructive to medicalise stress and distress by imposing diagnoses of anxiety and clinical depression, it is important to identify these as they are treatable and improving them will lead to an improvement in GI symptoms and QOL. Use of simple questionnaires (such as the hospital anxiety and depression scale) are a quick and easy way to identify this in an outpatient setting. Treatment is with pharmacological agents or with psychotherapy, both of which are effective treatments for FGID. 31
Several symptoms of anxiety are similar to those in FGID (eg diarrhoea, vomiting, abdominal cramps and nausea) and patients with multiple functional syndromes often have an element of health anxiety. GI specific anxiety can be measured by the visceral sensitivity index (VSI) and this is the best predictor of IBS symptom severity.
It is important not to blame the GI disorder on low mood but explain how mood will contribute to GI symptoms and how it is treatable. Urgently refer patients with current suicidal ideation or plans to mental health professionals. For others, consider the role of psychotherapy versus pharmacotherapy for treating the low mood. As a general rule, unless there is suicidal ideation, we would consider psychotherapy (further details follow later) as a first line before medication.
Eating disorders and disordered eating
Although FGID are not typically associated with eating disorders, the presence of meal related symptoms can lead to disordered eating and, in severe cases, the development of food phobias. This is due to the conditioned pairing of an unpleasant GI symptom (eg abdominal pain) with specific foods which can then lead to avoidant restrictive food intake disorder (ARFID) which is an emerging category of eating disorders. 32
A meta-analysis has confirmed that psychological therapies including CBT, gut-directed hypnotherapy, dynamic psychotherapy, and relaxation and mindfulness therapy are effective treatments for FGID (number needed to treat (NNT) 3–6). 30 Which of those is chosen will be determined by availability of local services. CBT is probably the easiest type of therapy to access (via Improving Access to Psychological Therapies (IAPT)) and most effective with an NNT of 3, meaning at least one in three referrals to CBT will experience significant reduction in symptoms. 31 Success of psychotherapy may be dependent on the expertise of the therapist; in some studies, psychotherapy delivered in specialised centres appears to have a better outcome. 33 However, CBT and mindfulness is effective even when delivered via the internet which will improve access to psychological therapies. 34,35 Psychotherapy can be time consuming, usually delivered over 12–14 weekly sessions. However, the effects appear to be long lasting and relatively free of side effects, so this form of therapy should be considered in willing patients. 36–39
A meta-analysis has confirmed that antidepressants are also effective treatments for IBS, equally effective to psychotherapy, when patients are compliant with them. 31,40 The use of antidepressants for treating symptoms of FGID are discussed later.
Biological management of FGID involves either treating the underlying pathophysiology (ie neuromodulators to treat visceral hypersensitivity) or treating the symptoms (eg antiemetics to treat nausea or laxatives to treat constipation). Various algorithms exist to help guide this for functional dyspepsia and for IBS (Fig 2). 10,41 For the purposes of this review, a symptom-based approach is presented.
Algorithms for pharmacological management of irritable bowel syndrome and functional dyspepsia. a) Irritable bowel syndrome. Dotted red boxes identify medications that are useful for more than one symptom in IBS-D, dotted blue boxes identify medications that are useful for combination symptoms in IBS-C. b) Functional dyspepsia. Dotted blue boxes identify medicines that target for both nutrient tolerance and pain. PPI = proton pump inhibitor; SNRI = serotonin noradrenergic reuptake inhibitor; SSRI = selective serotonin reuptake inhibitor; TCA = tricyclic antidepressant.
Opiates should be avoided as they are associated with dependence, tolerance and addiction, and can led to narcotic bowel syndrome which causes bloating, constipation, nausea and a paradoxical increase in pain with increasing doses of opiates. 42 Instead, first-line management of pain involves antispasmodics for colicky pain. In more resistant cases, often characterised by chronic burning/neuropathic pain, neuromodulators are used as second-line treatment.
First line: antispasmodics
Antispasmodics are useful for colicky abdominal pain in IBS. There is good evidence for hyoscine (10–20 mg three times a day (tds); NNT 3) and dicycloverine (10–20 mg tds; NNT 4), though these can cause anticholinergic side effects of dry eyes and a dry mouth, and can worsen constipation. There is also evidence for peppermint oil (eg colpermin 2 capsules tds; NNT 4). It can cause heartburn, so may be best avoided for patients with coexisting GORD. 30 The evidence for mebeverine (135 mg tds) is not as good, however, it is well tolerated and can be very effective in a group of patients which do not respond to other antispasmodics. 30 Our practice is to trial a second antispasmodic if the first fails to control colicky abdominal pain.
Second line: neuromodulators
Antidepressants. The most commonly used antidepressants are tricyclic antidepressants (TCAs) and selective serotonin reuptake inhibitors (SSRIs). As a rule, antidepressants improve symptoms, particularly abdominal pain, as well as psychological distress in FGID, and the effect is increased in secondary compared with primary care, which probably reflects the greater psychological and pain comorbidity in the former. 31 For every four patients treated with an antidepressant, one will get better.
Most antidepressants will have GI side effects and can cause either constipation or diarrhoea (supplementary material S1), so choosing your antidepressant wisely can also help treat the altered bowel habit in IBS. TCAs (eg amitriptyline) are useful for patients with diarrhoea by slowing GI transit, whereas SSRIs (eg sertraline, citalopram, fluoxetine) are useful in those with constipation by accelerating transit. There are no RCTs of serotonin noradrenergic reuptake inhibitors (eg duloxetine) in FGID, however, open label studies for patients with IBS and comorbid anxiety and depression show that it is well tolerated at a total dose of 60 mg per day and improves IBS symptoms as well as depression and anxiety. 43,44
Treatment with antidepressants is for an average of 18 months and stops once patients are symptom free for at least 6 months. In order to improve compliance, it is important to clarify to the patients that antidepressants are used primarily to target the IBS symptoms, notably pain, rather than the mood. It is important to pre-warn patients about the side effects and, if necessary, start at a very low (subtherapeutic) dose and work up slowly, in order to reduce adverse effects and therefore improve compliance. We would caution against using more than one antidepressant at a time unless there is expertise in doing this.
Gabapentinoids. Pregabalin and gabapentin are commonly used in chronic pain conditions and have a role in treating visceral hypersensitivity in FGID. Pregabalin is associated with an improvement in IBS symptoms (bloating, diarrhoea and abdominal pain) compared with placebo, so is a good choice for IBS-D. 45 This should be started at a low dose of 50 mg twice a day (bd) and increased gradually in line with symptomatic response, to a maximum of 300 mg bd, although it is not typically necessary to increase beyond 225 mg bd (3 × 75 mg tablets bd). 45 It causes weight gain, so may not be the best choice in obese individuals. Clinicians should be aware that it is addictive and is now considered a drug of abuse, so should be chosen with caution in certain patients. Gabapentin is an alternative but tends to have a worse side effect profile. Although there are no RCTs evaluating the efficacy of gabapentin on IBS symptoms nor the optimal dose for symptom improvement, one study demonstrated that, for patients using 300 mg gabapentin daily, there was a reduction in rectal sensitivity to distension and an increase in the thresholds for abdominal pain, bloating and discomfort. 46
First line: loperamide
Loperamide (Imodium) reduces stool frequency and improves stool consistency in IBS-D, however, it is not effective at reducing abdominal pain or bloating, so it is poorly tolerated in IBS patients. 30 Loperamide–simethicone chewable product (Imodium plus) is much better tolerated and results in quicker relief from diarrhoea and greater relief from abdominal discomfort compared with loperamide alone or a placebo and, therefore, this should be preferentially recommended to patients who have IBS-D. 47 Patients can be asked to take two tablets to start with and then one tablet after each unformed stool until the diarrhoea stops (up to a maximum of eight tablets per day).
Second line: ondansetron
Ondansetron, a 5HT3 antagonist, improves stool consistency, frequency, urgency and bloating but not abdominal pain compared with placebo or to mebeverine for patients with IBS-D, so it is especially useful for patients who are more troubled with the altered bowel habit than the pain. 48 As it is an antiemetic, it can also improve nausea for patients who have overlapping dyspepsia. Alosetron was the precursor to ondansetron for IBS-D, however, it was associated with ischaemic colitis and therefore withdrawn. In a meta-analysis of patients with IBS-D and IBS-M, patients on 5HT3 antagonists did better than those on eluxadoline and rifaximin (see later), so it is worth considering the use of ondansetron early for patients with IBS-D. 49
Third line: rifaximin and eluxadoline
Rifaximin is a non-absorbable antibiotic used in the treatment of GI disorders. A meta-analysis demonstrated that rifaximin is superior to placebo for reducing symptoms of diarrhoea and bloating in non-constipated patients with IBS, however, it had no effect on pain. 30,50 Recent unpublished evidence suggests that with 550 mg bd for 2 weeks, this effect may be sustained for at least 12 weeks after taking the antibiotic. However, rifaximin remains unlicensed for use in IBS in the UK and the long-term effects on the microbiota are unknown, so it is difficult to make evidence-based recommendations about this antibiotic at the current stage.
Eluxadoline is an opioid receptor antagonist which reduces stool frequency and improves global IBS symptoms, however, because of the association with pancreatitis, it is avoided for patients with a risk factor for acute pancreatitis (eg a prior history of pancreatitis, previous cholecystectomy, gallstones or alcohol). 30 The RELIEF RCT demonstrated that, in non-responders to Imodium who have an intact gallbladder, eluxadoline can be helpful in reducing stool frequency and improving pain over a 12-week period compared with placebo. A typical dose is 100 mg bd but this can be reduced to 75 mg bd in the case of side effects (eg nausea, abdominal pain, constipation and vomiting). 51
First line: osmotic laxatives
Polyethylene glycol (PEG)-based laxatives (such as Movicol and Laxido) as well as lactulose help to draw water into the bowel to soften the stool. Although they are associated with an increase in frequency of bowel movements, they do not alleviate pain in IBS. Lactulose can make bloating worse. In practice, it is useful to use osmotic laxatives to improve stool frequency for patients with constipation/IBS-C but this would need to be coupled with other agents to help with other symptoms such as pain.
Second line: prucalopride
Prucalopride is a highly selective 5HT4 agonist which acts as a prokinetic in the gut. An RCT demonstrated its efficacy in all patients with chronic constipation and in women with chronic constipation in whom laxatives have failed to provide adequate relief. 52 According to NICE, it is licensed for women in whom treatment with two laxatives have failed. A typical dose is 2 mg daily. A higher dose of 4 mg daily will result in improvement in straining than in stool frequency but will also be associated with greater side effects (such as headache, nausea and diarrhoea). 52,53 Patients who do not respond in the first 4 weeks are unlikely to do so with more treatment, so this can then be stopped. Prucalopride also acts as a gastric prokinetic, so it can be useful for patients with dyspepsia/gastroparesis type symptoms and would be a good choice for patients who have IBS-C and functional dyspepsia overlap.
Third line: secretagogues
If the above measures fail, then the third-line option involves the use of secretagogues (linaclotide and lubiprostone) which improve bowel frequency and overall IBS symptoms. Some patients experience unwelcome diarrhoea with this so it may be worth reducing the dose in that case.
Linaclotide results in increased chloride and bicarbonate secretion into the gut lumen which leads to increased fluid secretion and intestinal transit. It improves bowel frequency and reduces bloating compared with placebo; this effect is similar even if the dose (290 μg) is reduced to 72 μg, which reduces diarrhoea (a common cause for discontinuation) and therefore may improve compliance. 54,55
Lubiprostone at a dose of 8 μg bd improves abdominal pain, bloating and stool frequency in IBS-C more than placebo but can be associated with nausea. 56
This is the second most common FGID and can be divided into epigastric pain syndrome (EPS), characterised by epigastric pain and burning unrelated to meals, and postprandial distress syndrome (PDS) which causes early satiety, postprandial fullness, nausea and epigastric bloating. A proportion of patients with FD will also have mild to moderate delays in their gastric emptying (Fig 2b).
First line: proton pump inhibitors and H pylori eradication therapy
Patients who are proton pump inhibitor (PPI) responsive should be continued on the lowest dose needed to manage symptoms, and if this is not effective, it should be stopped. 10
Second line treatment: H2 blockers and prokinetics
H2 blockers. In a meta-analysis of treatments for functional dyspepsia, H2 antagonists (such as ranitidine) are comparable to, if not more effective than, PPIs. This is not surprising in view of the increasingly important role played by mast cells and histamine in the development of functional dyspepsia. 10 Translating this to clinical practice, it is worth trialling H2 blockers for patients with FD even when PPIs have failed, although, in the current climate, the supply issues with these medications may make this difficult in practice.
Prokinetics. Prokinetics can be used in PDS, particularly if there is delayed gastric emptying. They improve symptoms but not quality of life. 57 Many prokinetics represented in RCTs are not available in the UK (eg itopride, acotiamide, cisapride and mosapride), however, these were no more effective than domperidone, which is available in the UK. Domperidone has been associated with increased QT intervals which has limited its use, however, for patients with a normal QT interval, domperidone can be safely trialled, although it is important to recheck electrocardiography once the patient is established on domperidone. There are no trials assessing other prokinetics (such as metoclopramide and erythromycin) in FD, however, theoretically these can be used in the short term.
Third line: neuromodulators
The evidence for antidepressants in functional dyspepsia is less clear but there seems to be a role for low dose TCAs such as amitriptyline (10–30 mg nocte) or imipramine (25 mg daily for 2 weeks then 50 mg daily) for patients with epigastric pain (ie epigastric pain syndrome). 58,59 Patients need to be warned about anticholinergic side effects, which can reduce compliance. Mirtazapine is increasingly gaining popularity, especially for patients with postprandial symptoms (ie discomfort, fullness and nausea post-meals). It improves symptoms and nutrient tolerance even in the absence of coexisting depression and anxiety and causes weight gain, so would be ideal for patients with symptoms of postprandial fullness who are underweight. 60 It also improves sleep and mood which will have a beneficial effect on global symptoms. Patients should start at 15 mg nocte and increase monthly to a maximum of 45 mg. A meta-analysis has demonstrated that there is no role for SSRIs (eg sertraline, fluoxetine or citalopram) in FD. 58 Buspirone is an anxiolytic which can improve gastric accommodation and can be useful for patients with postprandial fullness and early satiety but is associated with poorly tolerated side effects of dizziness and somnolence so not the first choice of neuromodulator. 10
FGID are common but complex disorders which are associated with a lot of morbidity and associated psychopathology. As the aetiology of these disorders is still incompletely understood there remains no cure for them. Treatment involves a good therapeutic relationship and a holistic approach to treat the patient rather than disease, using a biopsychosocial model.
Additional supplementary material may be found in the online version of this article at www.rcpjournals.org/clinmedicine:
S1 – Pharmacological agents that can be used for functional gastrointestinal disorders.
The gastrointestinal tract – a central organ of cannabinoid signaling in health and disease
In ancient medicine, extracts of the marijuana plant Cannabis sativa were used against diseases of the gastrointestinal (GI) tract. Today, our knowledge of the ingredients of the Cannabis plant has remarkably advanced enabling us to use a variety of herbal and synthetic cannabinoid compounds to study the endocannabinoid system (ECS), a physiologic entity that controls tissue homeostasis with the help of endogenously produced cannabinoids and their receptors. After many anecdotal reports suggested beneficial effects of Cannabis in GI disorders, it was not surprising to discover that the GI tract accommodates and expresses all the components of the ECS. Cannabinoid receptors and their endogenous ligands, the endocannabinoids, participate in the regulation of GI motility, secretion, and the maintenance of the epithelial barrier integrity. In addition, other receptors, such as the transient receptor potential cation channel subfamily V member 1 (TRPV1), the peroxisome proliferator-activated receptor alpha (PPARα) and the G-protein coupled receptor 55 (GPR55), are important participants in the actions of cannabinoids in the gut and critically determine the course of bowel inflammation and colon cancer. The following review summarizes important and recent findings on the role of cannabinoid receptors and their ligands in the GI tract with emphasis on GI disorders, such as irritable bowel syndrome, inflammatory bowel disease and colon cancer.
The endocannabinoid system in the GI tract
Cannabis has a long history as a traditional therapeutic agent for the treatment of abdominal pain and gut dysfunction. This beneficial effect is based on the fact that the gastrointestinal (GI) tract is endowed with cannabinoid (CB) receptors and their endogenous ligands. Together they make up the endocannabinoid system (ECS), a physiologic entity that controls homeostasis in the gut. There is also a wide range of cannabinoid compounds of exogenous origin. Next to herbal cannabinoids, such as Δ 9 -tetrahydrocannabinol (Δ 9 -THC), cannabidiol, tetrahydrocannabivarin, cannabichromene, cannabigerol and others, there is a large array of synthetic cannabinoids. In general, cannabinoid compounds can be divided into five distinct classes, i.e. classical cannabinoids (e.g., Δ 9 -THC); non-classical cannabinoids (e.g., CP-55,940); indoles (e.g., WIN55,212), eicosanoids, and antagonist/inverse agonists (e.g., rimonabant) (1). For a detailed description of the ECS in the gut, the reader is referred to more comprehensive reviews (2,3).
In short, the ECS consists of the CB receptors 1 and 2 (CB1, CB2), their endogenous ligands (“endocannabinoids”) as well as their degrading and synthesizing enzymes. CB1 receptors can be found throughout the GI tract. There, they are predominantly located in the enteric nervous system (ENS) (4) and the epithelial lining (5). Additionally, CB1 is found in extrinsic fibers of the ENS, plasma cells, and in smooth muscle cells of blood vessels within the colonic wall (6,7). Within the ENS, the CB1 receptor is expressed prejunctionally in cholinergic, but not nitrergic neurons, explaining why CB1 activation can depress excitatory transmitter release (8). CB2 receptors are mainly present in immunocytes, myenteric plexus neurons, and in epithelial cells during ulcerative colitis (7,9). In addition to CB receptors, the orphan G-protein coupled receptor 55 (GPR55) and the transient receptor potential cation channel subfamily V member 1 (TRPV1) are endocannabinoid-responsive receptors and may be responsible for non-CB1/CB2 receptor effects of cannabinoids in the GI tract and are therefore regarded as part of an expanded ECS (10,11). PPAR receptors, in particular PPARα and PPARγ, are also responsive to herbal, synthetic and endogenous cannabinoids and may mediate many of the analgesic and anti-inflammatory effects observed in cannabinoid treatment [rev. in (12)]. The abovementioned receptors are present in the GI tract, e. g. on nerve terminals of extrinsic primary afferents (TRPV1) (2), and the ENS and enterocytes (PPARα, GPR55) (2,13).
Endocannabinoids are short-lived bioactive lipids and produced “on demand”. Arachidonoyl ethanolamine (anandamide) and 2-arachidonoylglycerol (2-AG) are among the best characterized endocannabinoids and are synthesized by N-acyl phosphatidylethanolamine phospholipase D (NAPE-PLD) and diacylglycerol lipases (DAGL), respectively. They are degraded by specific enzymes: anandamide primarily by fatty acid amide hydrolase (FAAH) and 2-AG by monoglyceride lipase (MGL; or monoacylglycerol lipase, MAGL) (rev. in (3)). In the GI tract, FAAH and MGL were shown to be expressed in epithelial cells, the ENS, and in immune cells during ulcerative colitis (6,7,14). Endocannabinoids may be also degraded by cyclooxygenase-2 (COX-2) and lipoxygenase to give rise to prostaglandin ethanolamides, glyceryl prostaglandins, hydroxyeicosatetraenoic acid and hydroperoxyeicosatetraenoic acid derivatives (15,16). In contrast to the degrading enzymes, the synthesizing enzyme of anandamide, NAPE-PLD, and of 2-AG, DAGL α and β, have been observed in epithelial, myenteric plexus and lamina propria cells, and also in the smooth muscle layer (7).
Acylethanolamides other than anandamide, like palmitoylethanolamide (PEA) and oleoylethanolamide (OEA), can be classified as endocannabinoid-like compounds. They do not directly activate CB receptors but they can activate GPR55 (predominantly PEA) and GPR119 (only OEA) and are able to influence the signaling of anandamide via an entourage effect (17). PEA and OEA also activate PPARα and are present in high levels within the gut. Both of them are degraded by FAAH, however, PEA is preferentially degraded by another amidase, N-acylethanolamine-hydrolyzing acid amidase (NAAA), which is strongly expressed in immune cells and active particularly in the intestine, suggesting a potentially pathophysiological role in the GI tract (rev. in (17)). In summary, the GI tract is able to locally produce its own endocannabinoid ligands according to its physiological needs and may rapidly react to disturbances in the gut to maintain homeostasis.
Cannabinoids in GI motility and secretion
Cannabinoids affect gut motility mainly by activating CB1 receptors present on enteric neurons (18). Activation of CB1 receptors results in the inhibition of acetylcholine release which consequently causes a decrease of intestinal smooth muscle contractility and peristalsis (19). Early studies demonstrated that the plant-derived CB receptor agonist Δ 9 -THC, the main component of Cannabis, decreases intestinal transit and inhibits electrically evoked contractions in guinea pig explants (20,21). Synthetic CB receptor agonists likewise reduce gastric emptying, upper GI transit, and colonic propulsion (reviewed in (2)). In contrast, rimonabant (SR141716), an inverse agonist of CB1 receptors, increased electrically-evoked contractions and peristalsis in isolated intestinal segments (22,23), as well as intestinal motility in vivo (24). Although CB2 receptors are expressed in the ENS, they are suggested to play a minor role in the regulation of gut motility under basal conditions but might become important under pathophysiological settings (9). Indeed, JWH-133, a CB2 receptor agonist, but not arachidonyl-2′-chloroethylamide (ACEA), a CB1 receptor agonist, attenuated gut transit dose-dependently in the inflamed gut of rats, an effect that was prevented by a CB2 receptor antagonist (25). There is also increasing evidence that GPR55 is involved in the regulation of gut motility since its agonist O-1602 was able to slow down whole gut transit in mice (13). Both PEA and OEA inhibit intestinal transit in mice but the mode of action is unclear because neither CB receptors nor PPARα seem to be involved in that process (26,27); however, in a mouse model of postinflammatory IBS (mustard oil-induced), inhibition of transit by PEA could be blocked with a CB1 receptor antagonist, but was not significantly modified with a PPARα antagonist (28).
Acute inhibition of endocannabinoid-synthesizing or – degrading enzymes also modulates intestinal motility. Thus, inhibition of DAGLα was able to normalize gut motility in a mouse model of genetically-induced constipation (29). Pharmacological inhibition of FAAH or MGL led to a decrease in gut motility through mechanisms that involved a rise in anandamide or 2-AG levels, respectively, and the activation of CB1 receptors (14,30,31). Interestingly, FAAH-deficient mice did not show alterations in basal gut motility; however, pharmacological inhibition or genetic deletion of FAAH normalized endotoxin-induced hypermotility (31). Taschler et al. demonstrated that MGL-deficient mice did not show alterations in basal gut motility but that they were insensitive to CB receptor agonist treatment due to desensitization of intestinal CB1 receptors (30).
It has to be noted that also the gut brain-axis may account for the regulation of gut motility by cannabinoids. For instance, intracerebroventricular injection of the CB receptor agonist WIN55,212-2 attenuated whole gut transit in mice (32). Additional evidence that gut motility might be regulated by central CB receptors was provided by Vianna et al. who showed that deletion of CB1 receptors specifically in the vagal nerves of mice caused an increase in GI motility (33). Similar to rodents, CB1 receptors are functionally present in the human small and large intestines (34–36). Thus, WIN55,212-2 and ACEA inhibited electrically-evoked contractions in a healthy human colon and this effect was completely blocked by rimonabant (37). Also 2-AG and anandamide were shown to inhibit acetylcholine-induced contractions in explants of human colonic longitudinal and circular muscle, however, this effect could not be blocked with CB1 or CB2 antagonists (38). The authors suggested a non-cannabinoid or alternative cannabinoid pathway mediating this effect (38). It is possible that the non-CB effects by anandamide may have been brought about by GPR55 which causes relaxation in the murine colon (13).
There is evidence that cannabinoids play an important role in the regulation of gastric and intestinal secretion in rodents and humans. Studies revealed that cannabinoids reduce the production of gastric acid secretion by activating CB1 receptors (19). In mice, intestinal hypersecretion induced by cholera toxin was reduced by CB1 receptor activation (39). In another study, pharmacological inhibition or genetic deletion of FAAH provided beneficial effects against diclofenac-induced gastric irritation (40). In contrast, enhanced secretion was observed in humans treated with the CB1 antagonist rimonabant (41). In summary, a large body of evidence demonstrates that (endo-) cannabinoids affect physiologic functions of the gut, a property that could be therapeutically exploited. Activation of CB1 receptors by increased levels of endocannabinoids and, as a consequence, a slowed gut motility might have beneficial effects for patients with symptoms of hypermotility. On the other hand, inhibition of endocannabinoid synthesis or blockade of CB1 receptors might enhance gut motility in GI disorders associated with constipation. If central side effects of cannabinoids could be overcome, modulation of cannabinoid levels would certainly represent a valuable pharmacological approach for the treatment of GI disorders. Another possibility could be the use of non-psychotropic cannabinoids like cannabidiol, which has been described as a ligand of many receptors including GPR55, TRPV2, PPARγ and 5-HT1A but not of CB receptors (but might modulate their actions) (42). Cannabidiol has shown relaxant effects on croton oil- and sepsis-induced hypermotility in mice (43,44).
Cannabinoids in emesis and nausea
The dorsal vagal complex (DVC) in the brainstem is the site responsible for the vomiting reflex while the neural circuitry responsible for nausea is less known. CB receptors and particularly FAAH and MGL are present in the DVC and area postrema suggesting an important role of endocannabinoids in the control of emesis (45–47). Cannabis has been traditionally used as an antiemetic agent, and exogenous cannabinoids are presently prescribed for people with chemotherapy-induced nausea and vomiting (48). However, due to central side effects, cannabinoids are not used as first line drugs.
The endocannabinoids anandamide and 2-AG have been shown to reduce emesis in experimental models (46). Drugs that can raise endocannabinoid levels without causing the typical cannabinoid agonist-induced central side effects are therefore potential options to treat emesis. The FAAH inhibitor URB597 reduced LiCl-induced emesis via CB1 and CB2 receptors (46). Reduction of emesis by the MGL inhibitor JZL184 was shown to be sensitive to CB1 antagonism (49). Also cannabidiol showed anti-emetic and anti-nausea effects in animal models, the effects were brought about by indirect agonism of 5-HT1A somatodendritic autoreceptors in the dorsal raphe nucleus (50).
The role of endocannabinoids has been investigated more recently in detail in the conditioned gaping model in rats and results indicate that 2-AG and the visceral insular cortex (VIC) could play an important role in nausea (51). Exogenous 2-AG, but not exogenous anandamide, applied by bilateral intra VIC infusion, dose-dependently suppressed conditioned gaping (51). The effect could not be blocked with the CB1 antagonist AM251, but instead with the COX inhibitor indomethacin (51). Interestingly, bilateral VIC infusion with the MGL inhibitor MJN110 also suppressed conditioned gaping but here, the effect could be blocked with AM251 (52).
Endocannabinoids have been clearly established as important messengers in the neuronal network that controls vomiting and nausea. Interference with endocannabinoid degradation may represent a valuable therapeutic approach not only against emesis but also against anticipatory nausea in chemotherapy patients.
Cannabinoids and functional bowel disorders
Irritable bowel syndrome (IBS) and functional dyspepsia are the most frequent functional bowel disorders encountered globally. The previous view that functional GI disorders lack histopathological and biochemical alterations has been challenged by studies demonstrating low grade inflammation, increased presence of immune and mast cell, changes in the epithelial barrier, and bacterial overgrowth in IBS patients. These alterations together with a derangement of the gut-brain axis may be involved in the development of visceral hyperalgesia and motility disturbances. The predominant presence of CB1 receptors along the gut-brain axis may allow cannabinoids to positively influence derangements along this axis (3,53). The role of the ECS in IBS has been already described in a previous review by Storr&Sharkey (53). Here, more recent results will be summarized and discussed.
IBS: visceral hypersensitivity and the ECS
Symptoms of IBS, such as abdominal pain, discomfort, and altered bowel habits, have been previously linked with visceral hypersensitivity and aberrant 5-hydroxytryptamine (5-HT) signaling (53). Feng et al. explored the link between 5-HT and the ECS and observed increased levels of 5-HT, but a decrease in anandamide, in the duodenal mucosa of patients with postinfectious IBS (PI-IBS) (54). Using a rat model, they showed that acute luminal administration of 5-HT into the duodenum induced anandamide release via vagal 5-HT3 receptors, whereas chronic 5-HT treatment decreased anandamide levels via 5-HT3, indicating that 5-HT may be involved in the regulation of intestinal anandamide content. In addition, luminally-applied CB1 receptor agonists attenuated 5-HT-induced hyperalgesia (54). In IBS-D (diarrhea-predominant) patients, no changes in anandamide levels but a decrease in PEA was observed in comparison to healthy subjects. The decrease was associated with abdominal pain (55). The IBS-D patients also had an increase in 2-AG while IBS-C patients had higher levels of OEA (55). It is interesting that levels of PEA were also found decreased in a mouse model of inflammation-induced hypermotility (croton oil-induced) (56). The decrease was reduced by a non-psychoactive Cannabis extract, cannabichromene, in a CB receptor-independent manner (56). In contrast, in a mouse model of postinflammatory IBS (mustard oil-induced), PEA slowed gut transit, an effect that was dependent on CB1 receptors (28). By use of a trinitrobenzenesulfonic acid (TNBS)-induced model of visceral hypersensitivity, Iwata et al. showed that a CB2 receptor agonist was effective in improving pain thresholds in a dose-dependent manner without signs of central CB1 receptor activation (40,57). Considering these data it is possible that low levels of endocannabinoids in IBS patients may contribute to hyperalgesia and abdominal pain and cause perturbations in the bowel motility which could be improved by endo- or exocannabinoids via CB- and possibly non-CB receptor pathways. This leads to the idea that FAAH inhibitors could be valuable therapeutics against PI-IBS and possibly other forms of IBS. In accordance with this concept, several studies reported that pharmacological inhibition of FAAH and also MGL significantly reduced visceral nociception in rodent models of colorectal distension and acetic acid-induced abdominal stretching (40,58,59). In this context it is worth to mention the role of mast cells in IBS. Activated mast cells have been shown to correlate with abdominal pain in IBS (60). Since mast cells express CB receptors and are also targets of PEA (61), which is thought to modulate mast cells activation, endocannabinoids may regulate activity of mast cells and hence interfere with IBS symptoms like abdominal pain; however, this remains to be shown.
IBS: stress, pain and the ECS
Chronic stress can induce visceral hyperalgesia via the hypothalamic–pituitary–adrenal axis and probably adds to the pain that IBS patients perceive. Recent work in rat models has shown that chronic stress causes reciprocal changes in 2-AG and COX-2/FAAH levels in L6–S2, but not L4–L5 dorsal root ganglia (DRGs) (62). Moreover, CB1 receptors were downregulated while TRPV1 receptors were upregulated in L6–S2 but not in L4–L5 DRGs, indicating region-specific changes in primary sensory fibers innervating the distal colon (62). A report suggests that epigenetic regulation in the DRG neurons could be responsible for these changes: while chronic stress was associated with methylation in the promoter region of the Cnr1 gene (encodes the CB1 receptor), histone acetylation at the Trpv1 promoter and expression of the TRPV1 receptor were increased (63). These findings point out that reciprocal changes in the endovanilloid and endocannabinoid system occur in visceral sensory fibers and that these changes could contribute to hyperalgesia and abdominal pain.
Stress and visceral pain may be also regulated by the ECS within the CNS. It is known that chronic stress reduces levels of anandamide (but increases 2-AG) in the brain and downregulates CB1 receptors, and that these changes may contribute to the stress response (64). In line with this, both the FAAH inhibitor PF 3845 and the dual FAAH/MAGL inhibitor JZL 195 were effective in inflammatory and mechanically evoked visceral pain models suggesting that an increase in endocannabinoid levels alleviates visceral pain (59). A more thorough description of this topic is given in (65).
IBS: genetic variations and the ECS
Genetic variations of ECS components (CB receptors, synthesizing/degrading enzymes) may be associated with the pathogenesis of functional bowel disorders. Polymorphism in the FAAH gene (C385A) leads to a mutant FAAH enzyme and reduces breakdown of anandamide (66). A study in patients with constipation predominant (C-) IBS, D- and M- (mixed) IBS, with chronic abdominal pain and functional dyspepsia, showed a clear association of the non-wild type FAAH genotype with functional bowel disease phenotypes and with accelerated colonic transit in IBS-D patients (67). However, no statistically significant association between the FAAH genotype and sensation measurements was observed (67). A polymorphism in the CNR1 gene, rs806378, was found to be significantly associated with IBS symptom phenotype, colonic transit in IBS-D, and sensation rating of gas, but not with pain (68). In line with a possible role of CNR1 variants in the development of IBS symptoms, allele frequencies of AAT triplet repeats in CNR1 were observed to be associated with IBS in a study of a Korean population (69). Similar results, namely the detection of eight CNR1 alleles with AAT triplet repeats, were reported in a Chinese IBS cohort, whereas no association could be detected between C385A FAAH polymorphism and IBS pathogenesis (70). Interestingly, FAAH activity was recently determined in whole colon samples from patients who underwent colectomy for slow transit constipation (71).The results revealed a strong decrease in activity in these patients as compared to individuals free of transit dysfunction (71). The FAAH enzyme, therefore, seems to be a key molecule for the regulation of endocannabinoid levels and colon motility, but not for GI pain sensation.
Effect of CB receptor agonists in IBS patients
From animal studies it was rightfully concluded that cannabinoid agonists could improve visceral pain thresholds in humans. In a previous study performed in healthy volunteers to investigate the effect of dronabinol (Δ 9 -THC) on colonic motility and sensation, 7.5 mg dronabinol induced relaxation of colon motility and tone postprandially (72). The effect of dronabinol on visceral perception to rectal distension was then tested in IBS patients (positively diagnosed by Rome II criteria) and healthy subjects in a small trial, but no differences in sensory thresholds and discomfort were observed between the cohorts (73). A different study revealed inhibitory effects of dronabinol on fasting colonic motility and an increase in colonic compliance, particularly in patients with diarrhea predominant forms of IBS, but failed to demonstrate effects on sensation and tone (74). The report also suggested that FAAH and CNR1 variants could have had an impact on the effects of dronabinol (74). In a subsequent trial performed in IBS-D patients, no significant effect of dronabinol on colonic transit was observed; however, in a subset of patients with the CNR1 polymorphism rs806378, dronabinol moderately delayed colonic motility (75).
Thus, it seems that CB receptor activation in IBS has potential therapeutic value, but probably only in IBS-D patients with genetic variations of ECS components.
There is good indication that the ECS may be involved in functional dyspepsia. Tack et al. have previously shown that early satiety and symptoms of functional dyspepsia are caused by a disturbed gastric accommodation (76). In addition, hypersensitivity to gastric balloon distension was observed to be present in a subset of patients with functional dyspepsia (77). A cross-over, randomized, controlled clinical trial in healthy individuals now demonstrated that CB1 receptor antagonist rimonabant was able to inhibit meal-induced gastric accommodation, but did not affect fasting gastric compliance or sensitivity to gastric balloon distension, indicating that gastric accommodation is controlled by endocannabinoids (78). However, it was not clear from the study whether the ECS controls accommodation via centrally-mediated pathways or via the ENS. A new study has recently addressed the question as to whether CB1 receptors in the brain are involved in functional dyspepsia and could demonstrate that increased availability to a CB1 receptor radioligand was predominantly found in brain regions involved in the regulation of visceral pain and satiety (79).These findings would argue for a role of central CB receptors in the regulation of gastric accommodation in humans. It is, therefore, possible that both, central and peripheral CB receptors are involved in the development of functional GI disorders, and that pharmacological manipulation of exclusively peripheral CB receptors may not provide full benefit for patients with these disorders.
Microbiota and the ECS
A change in the microbiotic population of the gut may alter the permeability and lead to metabolic endotoxemia and hence to metabolic disorders associated with obesity. Endocannabinoids are involved in the regulation of energy metabolism and food intake and communicate in this respect with the microorganisms of the gut (80). The epithelial lining expresses CB receptors and they are most likely involved in these mechanisms. 2-AG and PEA cause an increase in epithelial barrier function (“gate keeper”) while anandamide is thought to be a “gate opener” (81). Thus, the intestinal ECS may have an important role in the control of microbial products entering the bloodstream and in the development of metabolic diseases. A detailed review on this topic is given in (81).
Dysbiosis (alteration in the composition of gut microbiota) has been also suggested as one of the potential causes of IBS, especially in the case of PI-IBS (82). It is known that antibiotic therapy provides certain benefits for IBS patients (83), however, it is not quite clear how eradication of bacteria could contribute to symptom relief. In this context it is interesting that Lactobacillus acidophilus NCFM could induce CB2 receptor expression in the rodent gut mucosa (84). When applying NCFM in a model of chronic colonic hypersensitivity, it caused analgesia which was abrogated by i.p. blockade with AM630, suggesting that CB2 receptors may provide a link between gut microbiota and visceral hypersensitivity (84). However, in a human trial, CB2 receptors were not found to be upregulated in colonic mucosal biopsies from persons that were given Lactobacillus acidophilus NCFM over a period of 21 days (85). On the other hand, treatment of mice with antibiotics reduced pain-related responses to i.p. application of acetic acid or intracolonic capsaicin (86). The effect was accompanied by a small rise in CB2 receptor transcripts in colon tissue, as well as a decrease of CB1 and mu-opioid receptors. Additionally, total luminal bacterial counts correlated with CB receptor expression (86) suggesting a possible interaction between microbial products and CB receptors.
Cannabinoids and intestinal inflammation
Chronic inflammatory conditions of the GI tract are known as inflammatory bowel disease (IBD) and occur in two major forms, ulcerative colitis (UC) and Crohn’s disease (CD). IBD is thought to originate from a complex interaction of the gut microbiota (or their products) with the epithelial barrier, based on the genetic background and the immune system of the host (87). To investigate the role of cannabinoids in IBD, mostly animal models that rely on chemically-induced mucosal inflammation have been used.
The endocannabinoid system as a therapeutic target in IBD
Evidence gathered from several studies in rodents points to a therapeutic relevance of the ECS in IBD. As reviewed by Izzo & Sharkey (2) and Alhouayek & Muccioli (88), endocannabinoid signaling is largely enhanced in the inflamed intestine. Expression of CB1 (89) and CB2 receptors (90), and of anandamide (91) were increased, whereas FAAH levels were reduced in the initial phase of colitis (92). Pharmacological strategies to enhance endocannabinoid levels, either by inhibition of endocannabinoid degradation (92–94) or of the transport across the plasma membrane (91,92) ameliorated inflammation. In particular, inhibition of FAAH by PF-3845 (94) and FAAH/COX blockade by ARN2508 (95) dramatically reduced damage in experimental colitis models. In the latter study, raised levels of anandamide, PEA and OEA were measured that most likely contributed to the beneficial effect (95). A recent work by Alhouayek et al showed that inhibition of NAAA, which preferentially degrades PEA, caused significant improvement of experimental colitis suggesting that PEA is an important acylethanolamide in the regulation of intestinal inflammation (96). In accordance, oral administration of PEA (which is interestingly sold as an over-the-counter drug and advertised to mitigate symptoms of GI disorders) exerted anti-inflammatory effects in the gut (97). Experiments on cultured human colonic biopsies derived from UC patients showed that PEA caused a decrease in expression and release of inflammatory mediators which was dependent on PPARα (98).
Activation of the CB1 (89) or CB2 receptor (90,99) with specific agonists also protected from colitis. Accordingly, genetic ablation or pharmacological antagonism of CB1 (89,100) or CB2 receptors (90,100) left mice more susceptible to intestinal inflammation. Moreover, treatment with Δ 9 -THC was reported to reduce colitis in rats (101). The limitations of using Cannabis for treatment of gut inflammation, however, are the psychoactive effects that arise from activation of CB1 receptors in the brain. Investigation of pharmacologically active cannabinoids with low or no affinity for CB1 receptors and of atypical cannabinoids would be therefore of high interest. Indeed, it has been shown that cannabidiol and cannabigerol, two non-psychotropic ingredients of Cannabis, have proven beneficial in various models of intestinal inflammation (101–105). Also, the atypical cannabinoid O-1602 was reported to reduce disease severity in a CB1-/ CB2 receptor-independent way by inhibiting neutrophil recruitment (106). Recently, GPR55, which is part of the “expanded” ECS, has been investigated in experimental colitis. A pro-inflammatory role of GPR55 could be established because genetic deletion of GPR55 and treatment with the GPR55 antagonists CID16020046 or ML-191 alleviated intestinal inflammation (97,106,107). In this context, cannabidiol, which is known to act as a GPR55 antagonist (108), showed inhibition of GI inflammation in an LPS-induced model by targeting enteric reactive gliosis (103). Interestingly, only parts of the beneficial effects of cannabidiol in this model were mediated by PPARγ (103) raising the possibility that GPR55 could have been involved in this effect. Cannabidiol may also exert a protective effect on the intestinal barrier. In a Caco-2 cell monolayer stimulated by EDTA, cannabidiol concentration-dependently caused rapid recovery of the barrier and this effect was inhibited by a CB1 antagonist (109). Since cannabidiol has no affinity to CB1 receptors, the authors argued that cannabidiol could have antagonized CB1-mediated increases in permeability mediated by locally produced endocannabinoids (109). Activation of CB2 receptors also attenuated cytokine-evoked mucosal damage in human colonic explants (110).
Cannabis for the treatment of IBD?
Questionnaires among IBD patients revealed that Cannabis is commonly used as a self-medication to relieve IBD-related symptoms like abdominal pain, diarrhea, and loss of appetite (111,112). A retrospective study reported significant improvements in 21 out of 30 CD patients after Cannabis use (113). In a small prospective placebo-controlled study of CD patients, a beneficial clinical response was achieved in 10 out of 11 subjects in the treatment group (114). A more recent questionnaire confirmed that the use of Cannabis subjectively improved pain and other symptoms in IBD patients, but also pointed out that Cannabis use for more than six months was a strong predictor in CD patients for requiring surgery (115).
Despite these interesting findings, the exact mechanisms how the ECS operates in IBD have not yet been unraveled but evidence gathered so far points to an overall protective role ( Fig. 1 ). The up-regulation of ECS components possibly constitutes an attempt to restore homeostatic balance (3). Cannabinoids have been shown to influence the recruitment of immune cells to the site of intestinal inflammation (93,106,107) and to reduce the release of pro-inflammatory cytokines, i.e. TNF-α, IFN-γ, IL-1β and IL-6 (93,102,103,105). Activation of the CB1 receptor might also lead to enhanced wound closure during colitis (5). Of particular interest are recent findings that gut microorganisms may influence the expression of intestinal ECS components (81). 2-AG and PEA were mostly associated with beneficial effects on the gut-barrier function (81). The crosstalk between gut microbiota and the ECS is therefore worthy to be further examined in future studies.
Expression of receptors and synthesizing/degrading enzymes of the endocannabinoid system (ECS) in the normal and acutely inflamed human gastrointestinal (GI) tract. Data were taken from Wright et al. (5) and Marquéz et al. (7). CB1, CB2, cannabinoid receptors 1 and 2; FAAH, fatty acid amide hydrolase; MGL, monoacylglycerol lipase; NAPE-PLD, N-acyl phosphatidylethanolamine phospholipase D; DAGL, diacylglycerol lipase.
Collectively, cannabinoids show great potential in the treatment of IBD and further research is warranted to gain a better insight into the mechanistic actions of (endo-) cannabinoids.
Cannabinoids and colon cancer
Differential expression of components of the ECS in colorectal cancer (CRC) was first reported by Ligresti et al. (116). In this study, anandamide and 2-AG contents were found to be higher (3-fold and 2-fold, respectively) in CRC lesions as compared to normal colonic mucosa and, interestingly, their levels were higher in adenomatous polyps than in carcinomas (116). Increased endocannabinoid synthesis in CRC was also reported in a more recent study (117). Here, anandamide, as well as its synthesizing enzyme NAPE-PLD, were up-regulated approximately 2-fold in cancer tissues. Intriguingly, mRNA expression and activity levels of FAAH were also increased. Most likely, as a consequence of increased FAAH activity, elevated levels of arachidonic acid, the main product of anandamide and 2-AG degradation, were also detected (117). In another study, the main degrading enzyme of 2-AG, MGL, was also found increased in CRC specimens (118).
Examination of CB1 expression revealed a down-regulation of mRNA levels in 18 out of 19 colon cancer samples as compared to adjacent non-neoplastic colon mucosa (119). The reason for this silencing was found to be DNA hypermethylation at CpG islands around the transcription start site of CNR1. In parallel to the epigenetic regulation, also protein levels of CB1 receptors were reduced in colon cancer specimens as shown by Western blotting (119). These findings were corroborated by Cianchi et al. who reported CB1 receptor expression to be higher in normal colonic epithelium than in colonic tumor tissue (120). However, a comprehensive study describing the correlation between CB1 receptor immunoreactivity and patient outcome conducted in 534 Korean patients found no differences in overall survival between patients with carcinomas of either high or low CB1 receptor immunoreactivity (121). Distant metastasis was found to be lower in patients with high CB1 receptor expression, but there were no differences in lymph node metastasis, tumor invasion, or tumor size. Surprisingly, in stage IV patients, high CB1 immunoreactivity even correlated with a poorer survival rate (121). Similar observations were made in a cohort of 487 Swedish patients (122). There, high CB1 expression was reported to correlate with poorer disease-specific survival in stage II microsatellite stable CRC patients (122). Reduced overall survival has also been reported for patients who were either homo- or heterozygous for the 1359 G/A single nucleotide exchange in the CNR1 gene although it is not yet known how this polymorphism affects cannabinoid signaling (123). CB2 receptor mRNA expression was found in 28.6% of CRC samples and significantly correlated with lymph node involvement (124), however, no consistent data on protein expression were available. So far, the human studies indicate increased endocannabinoid activity in colon cancer while the role of CB receptors remains less clear.
Cannabinoids reduce carcinogenesis in animal models of colon cancer
In mice, colon cancer can be induced either chemically or, for instance, by germline mutation of the adenomatous polyposis coli (Apc) gene. Apc Min/+ mice spontaneously develop multiple polyps in the intestine. Additional knock out of Cnr1 or inhibition of the CB1 receptor with AM251 in these mice caused a strong increase in intestinal polyp burden, whereas activation of CB1 receptors with methanandamide significantly reduced the number of polyps (119). Genetic deletion of Cnr2 (the gene encoding CB2 receptor), had no effect on polyp growth in this model (119). Chemically, colon cancer develops after multiple intraperitoneal injections of the carcinogen azoxymethane (AOM). In this model, anandamide and 2-AG were found increased in the colon of AOM-treated mice (125). In addition, inhibition of FAAH with N-arachidonoyl-serotonin (AA-5-HT) reduced the development of precancerous lesions, and furthermore, the non-selective, synthetic CB1/CB2 receptor agonist, HU210, was able to mimic this effect (125).
Antitumorigenic effects in the AOM model were also observed with non-psychotropic cannabinoids. For instance, cannabidiol was shown to reduce the formation of aberrant crypt foci (ACF), polyps, and tumors in the colon and the AOM-induced up-regulation of p-Akt (126). It also counteracted caspase-3 inactivation. In colorectal carcinoma cell lines, it protected DNA from oxidative damage and it reduced cell proliferation in a CB1-, TRPV1- and PPARγ-antagonists sensitive manner (126). A “cannabidiol botanical drug substance” (a Cannabis sativa extract with high content of cannabidiol) had similar effects in the same model, reducing ACF, polyp and tumor formation via CB1 and CB2 receptor activation (127), whereas treatment with cannabigerol reduced the number of ACFs only (128). In yet another murine model, in which colitis-associated colon cancer was induced through the application of AOM and dextran sulfate sodium (DSS), the atypical cannabinoid O-1602 showed antitumorigenic properties (129). The drug reduced the number and area of tumors by 30% and 50%, respectively. In addition, activation of the oncogenic transcription factor STAT3 was decreased while pro-apoptotic factors p53 and Bax were increased in O-1602 treated mice (129). Perhaps surprisingly, one study showed that antagonism of CB1 receptors with rimonabant reduced the formation of ACFs with 4 or more crypts in mice with AOM-induced colon cancer (130).
Potential applications of cannabinoids and related substances have also been studied in xenograft models. The semi-synthetic cannabinoid quinone HU-331 (131) and the hexahydrocannabinol analogue LYR-8 (132) reduced tumor growth of xenografts derived from HT-29 cells. Likewise, the CB2 receptor agonist CB13 inhibited the growth of DLD-1 derived tumors (120). A “cannabidiol botanical drug substance” (127) and cannabigerol (128) decelerated or even halted the growth of HCT116 xenografts, respectively.
Anticarcinogenic mechanisms of cannabinoids: reduction of cancer cell proliferation and inhibition of angiogenesis and metastasis
Cannabinoids have been shown to exert anti-proliferative effects on colon cancer cells through apoptosis via activation of CB1/CB2 receptors, or through receptor-independent mechanisms (rev. in (133)). The molecular mechanisms underlying the induction of apoptosis upon CB1/CB2 receptor activation have been discussed in detail by Velasco et al. (134). Briefly, de novo synthesis of the pro-apoptotic sphingolipid ceramide (120), down-regulation of the protein survivin (inhibitor of apoptosis) (119), inhibition of PI3K/Akt signaling (135,136), and induction of endoplasmic reticulum stress that leads to autophagy-mediated cell death (136), have all been reported. Notably, cannabinoids with low or no affinity for CB receptors (like cannabidiol and O-1602) are also known to exert anti-proliferative effects, although the underlying mechanisms have not yet been fully clarified (126,127,129). A cannabinoid-like compound LYR-8, for instance, was demonstrated to decrease angiogenesis in a xenograft model using chick chorioallantoic membranes (132). Concomitantly, the expression of factors that modulate the tumor microenvironment, like vascular endothelial growth factor, COX-2, and hypoxia-inducible factor 1α was reduced in this model (132). Inhibition of MGL, either pharmacologically or through silencing with siRNA, attenuated the invasion of colon cancer cells (118), suggesting a role of endocannabinoid degrading enzymes in CRC progression. Importantly, adhesion and migration of highly metastatic colon cancer cells was shown to be diminished after treatment with cannabidiol or a GPR55 inhibitor (108).
In conclusion, data obtained so far point to a deregulation of the ECS in colon cancer that could be interpreted as an attempt to restore the original healthy state. Despite controversial data on the role of the ECS in human colon cancer, promising preclinical data on the reduction in tumor growth by typical and atypical cannabinoid compounds warrant further exploration on the cause of ECS deregulation in colon carcinogenesis. It should be of prime interest to investigate known and hitherto unknown components of the ECS to better understand the complexity of CB receptor signaling by endocannabinoids and the regulation of their synthesizing and degrading enzymes.
Cannabinoids have a long history of being used to treat diseases or to alleviate symptoms. In modern medicine, this is not fully translated, and cannabinoids or cannabinoid-derived drugs are rarely used mainly due to the lack of clinical trials supporting such use. Over the last decades, cannabinoid research was driven by basic scientists who characterized pharmacological actions of cannabinoids, who discovered the ECS with all its constituents, and who taught us how activation or blockade at different sites may be helpful for the treatment of GI diseases. The GI tract is one of the regions where cannabinoid signaling is involved in many physiological and pathophysiological regulatory mechanisms, this is now clearly understood. The last decade has added more translational studies, and we have learned where cannabinoids are involved in pathophysiological states and human disease and where and how cannabinoids alter physiological or pathophysiological conditions. Through a recent meta-analysis we are also better informed on side effects associated with cannabinoid treatment. The analysis revealed that there was an increased risk of short-term adverse events with cannabinoids, mostly dizziness, dry mouth, nausea, fatigue, somnolence, euphoria, drowsiness, but also cardiac (1.42; 0.58-3.48; odds ratio; 95% CI) and hepatobiliary (3.07; 0.12-76.29; odds ratio; 95% CI) disorders were among them (137). Nevertheless, the opportunities are multifold with targeting the numerous involved receptors with agonists and antagonists, and with targeting synthesizing and degrading mechanisms. To harvest the potential therapeutic effects is now challenging, but based on the broad cannabinoid platform built by basic researchers, clinical trials are urgently wanted. From a scientist’s perspective and all the caveats in mind, it seems to be a matter of time when cannabinoid compounds will be used in the treatment of GI disease