Genetics of irritable bowel syndrome

Irritable bowel syndrome (IBS) is the most commonly diagnosed functional gastrointestinal disorder (FGID) with a worldwide prevalence of 10–20 %, predominantly among women [1]. Its clinical appearance varies but is usually characterized by recurrent abdominal pain or discomfort accompanied by changes in bowel habits and diarrhea (IBS-D), constipation (IBS-C), or both (mixed, IBS-M). In the absence of reliable biomarkers or specific laboratory tests, consensus criteria towards a positive diagnosis have been developed: the symptom-based Rome III criteria [2]. Due to a complex and not fully elucidated pathophysiology, there is no cure, and available treatment options can only be directed to amelioration of symptoms in a trial-and-error approach based on patients’ individual symptomatology. Comorbidity with other FGIDs as well as psychological conditions such as depression and anxiety is observed, as well as certain non-gastrointestinal non-psychological comorbidities including fibromyalgia, chronic fatigue syndrome, and chronic pelvic pain [3]. IBS negatively affects quality of life, which reflects in increased work and school absenteeism, decreased work productivity, and higher utilization of the health care system, with considerable repercussions on health and socioeconomic systems [3, 4].

A heritable component of IBS has been demonstrated in twin and family studies, although heritability estimates have varied between 0 and 57 % [10]. Recently, however, a Swedish nationwide survey including more than 50,000 cases showed increased IBS risk among first-, second- and third-degree relatives, clearly indicating that a genetic component exists [11]. Evidence is now accumulating that genetic risk in IBS spans from complex polygenic conditions with combinations of common variants, to cases with rare single gene abnormalities [12]. For the majority of IBS patients, the genetic background will be constituted by a large set of common genetic variants, each contributing a small risk effect. At the same time, there may be subsets of patients where highly penetrant genetic variability in individual genes accounts for most of the phenotype. A good example of this phenomenon recently came from a study at the Mayo Clinic in collaboration with our group [13]: sequencing of the SCN5A gene in 584 IBS patients and 1380 asymptomatic controls identified functionally deleterious mutations in 2.2 % of IBS cases but none among controls. The SCN5A gene encodes the Na_V1.5 ion channel responsible for the pacemaker function of the heart but is also present on interstitial cells of cajal, the ‘pacemaker’ cells of the gut. Most of the mutations identified were loss-of-function and carriers most often constipation-predominant. Of note, a severe IBS-C case with a highly penetrating SCN5A loss-of-function mutation could be successfully treated with mexiletine, a drug known to restore Na_V1.5 channel function. In addition, also common SCN5A single nucleotide polymorphisms (SNPs) were found to affect IBS risk in our pilot genome-wide association study (GWAS) of ~5000 subjects (described further down), as well as in four independent case-control cohorts from Sweden, Italy, Greece, and the USA. These findings hence strengthen the hypothesis that both rare mutations and common variants may be implicated in IBS. However, overall, gene-hunting efforts in IBS have so far been scarce [12], and large-scale efforts remain to be carried out. Recognizing that IBS spans from complex polygenic conditions to rare single-gene forms, we need to adopt different strategies to identify these genetic factors.

An alternative approach to the discovery of IBS relevant genes is to study the endophenotypes, otherwise called intermediate phenotypes of disease. These are usually quantitative traits related to IBS, such as colonic transit time and visceral sensitivity ratings. To focus on biological observations instead of clinical entities for establishing, diagnosis was proposed already 50 years ago, and this alternative strategy has been used successfully for genetic studies in the psychiatric field where illnesses often are, as IBS, complex and heterogeneous [18]. The concept of endophenotypes is to use disease-associated phenotypes in an attempt to reduce the complexity and to increase chances of finding genes important to biological processes underlying disease etiology. For the case of IBS, suitable traits to use for this approach include bowel movement frequency, Bristol stool form scale (a subjective measure of stool consistency, related to colonic transit time), pain and sensation ratings in response to visceral stimuli, and others. Previous work from our group has shown associations between polymorphisms in the neuropeptide S receptor gene (NPSR1) and colonic transit time, as well as gastrointestinal (GI) sensory ratings, such as gas, pain, and urgency [19]. The NPSR1 protein is a receptor for neuropeptide S, a neuropeptide involved in anxiety, response to stress and fear, inflammation, and nociception. Neuropeptides act in the brain-gut axis and have been implicated in IBS before [2, 6]. We also showed that the neuropeptide S (NPS)-NPSR1 system can induce the expression of other neuropeptides in vitro [19, 20]. Hence, this gene was a plausible candidate to investigate also for its potential genetic associations with abdominal pain. In the Swedish BAMSE birth cohort, we performed a candidate-gene study in relation to recurrent abdominal pain (RAP) [21], which occurs frequently among children and is one of the cardinal symptoms of FGIDs. The mechanisms of visceral pain and RAP are not fully understood, a heritable component has been demonstrated, and a few candidate genes proposed. We observed association with RAP at 7/24 tested SNPs, with the strongest signal in correspondence of a putative regulatory region upstream NPSR1 where they may exert their genetic effects through the modulation of gene expression.

Overall, IBS poses some major challenges, and it appears that in order to overcome these and be able to identify true unequivocal risk genes and variants, we need to implement larger scale analyses. Genetic-risk genes and variants have been successfully identified and replicated in a plethora of complex diseases using the powerful and hypothesis-free approach of GWAS studies and their meta-analyses (see the NHGRI-EBI GWAS Catalog; www.ebi.ac.uk/gwas) [22]. For instance in the GI field, a recent study reported the identification of 38 additional susceptibility loci for inflammatory bowel disease (IBD), bringing the total tally of confirmed IBD risk loci to 200 [23]. On the contrary, IBS genetics is very much lagging behind, and no similar GWAS effort had been attempted before our Screening-Across Life Span Twin (SALT) study (see below).

GWA studies are powerful approaches to discovering genetic risk loci in complex diseases [24], though this has never even been attempted in IBS. In order to reach the statistical power necessary to detect meaningful association, very large sample size is required. Unfortunately, this is currently unfeasible in the IBS community, as the number and size of well-characterized cohorts around the world is still surprisingly slim. In fact, only a portion of individuals suffering from IBS according to the Rome III criteria seek medical attention for their symptoms, with numbers reported ranging from 10 to 70 % [1]. In order to reach more individuals, we propose to shift approach and instead make use of general population samples for discovery purposes in IBS genetics. Large biobanks and general population cohorts offer a great opportunity to perform genetic epidemiological association studies with considerably larger sample sizes. When available, Rome questionnaire data may be used to identify IBS cases and asymptomatic controls in these large study populations. In addition, genotype data may also be linked to phenotype in these studies using the International Classification of Diseases (ICD codes) and electronic medical records (EMR). By shifting to a general population approach for the discovery phase, we gain considerable sample size and improve in the definition of controls, as they are co-sampled with the cases and classified with the very same investigative tools (for example questionnaires). However, once IBS-risk genes and variants have been discovered, validation of results should be performed by replication and targeted analyses in case-control cohorts from gastro clinics with well-characterized IBS patients and healthy controls. Furthermore, investigating rare variants/mutations in IBS, next-generation sequencing in selected cases and controls will also be necessary.

Our group recently conducted the very first pilot GWAS of IBS [25]. In this study, Rome II criteria from questionnaires were used to identify 534 IBS cases and 4932 asymptomatic controls from the SALT study of the Swedish Twin Registry. A GWAS was performed and top regions replicated in six independent case-control studies with patients and controls recruited from European and US clinics. Confirmatory, one of the loci (harboring the KDELR2 gene) was successfully replicated in all these samples (total sample size of 8977) with consistent effect of association. Even though sample size was still too small to reach genome-wide significance, this study corroborated our hypothesis that general population-based cohorts with associated genetic and epidemiological/healthcare data provide excellent opportunities to study the genetic architecture of IBS and related GI symptoms.

Overall, the ultimate value of genetic research in IBS is to identify key physiological mechanisms, which will help us understand its pathophysiology. These findings will help improve IBS diagnosis and classification and therefore impact therapeutic strategies and personalized treatments.