Familial aggregation in inflammatory bowel disease: Is it genes or environment?

Abstract

Inflammatory bowel disease (IBD) develops in genetically susceptible individuals due to the influence of environmental factors, leading to an abnormal recognition of microbiota antigens by the innate immune system which triggers an exaggerated immune response and subsequent bowel tissue damage. IBD has been more frequently found in families, an observation that could be due to either genetic, environmental or both types of factors present in these families. In addition to expanding our knowledge on IBD pathogenesis, defining the specific contribution to familial IBD of each one of these factors might have also clinical usefulness. We review the available evidence on familial IBD pathogenesis.

Key Words: Inflammatory bowel disease, Familial aggregation, Familial clustering, Environmental factors, Genetics, Genome wide association studies

INTRODUCTION

Our present understanding of inflammatory bowel disease (IBD) pathogenesis considers ulcerative colitis (UC) and Crohn’s disease (CD) as complex conditions that develop in genetically susceptible individuals due to the influence of various environmental factors[1-6]. An abnormal recognition of certain antigens of the bowel microbiota by elements of the innate immunity is thought to play a key role, leading to an exaggerated immune response, release of pro-inflammatory molecules and, ultimately, bowel tissue damage.

In past decades a greater incidence of IBD cases among UC and CD relatives, referred to as family aggregation or familial IBD, has been clearly demonstrated[7-11]. The reason for such an increased risk is not straightforward, but genes, the environment or a combination of both could, in theory, account for family aggregation, considering their contribution to the development of IBD. A better understanding of the factors leading to familial IBD might result in clinical applications.

FAMILLIAL IBD

Familial aggregation among IBD patients or “familial IBD” is defined by the occurrence of a trait in more family members than expected by chance. Multiple population studies have demonstrated that relatives of an IBD patient have a much higher risk of developing the same condition, compared to the general population[7-11]. The magnitude of familial aggregation depends on several factors, including: (1) type of IBD; (2) population studied; and (3) family relationship.

In respect to type of IBD, familial aggregation has been more frequently reported in CD than UC[7,9]. In first degree relatives, the age-adjusted relative risk of developing the same type of IBD ranges from 2-8 for UC and from 5-10 in the case of CD[7-9]. As elegantly demonstrated by Yang et al[9], affected relatives can develop both forms of IBD, although the greatest risk is associated with the appearance of the same disease type occurring in the index case. Additionally, CD patients tend to have a much higher frequency of relatives affected with UC when compared to UC patients having relatives affected with CD[7].

In respect to different populations, it has been shown that Jewish families present more than twice the number of multiple affected families when compared to the non-Jewish population[9]. However, we cannot rule out the possibility that differences amongst other geographical populations are due to their study design.

Finally, frequency of familial IBD also varies according to the degree and type of kinship. The prevalence of IBD in second-degree relatives appears to be lower than in first-degree relatives, especially in those with discordant disease[7]. In addition, although only a few studies have estimated the age-adjusted risk of IBD in relatives, it seems that the risk of IBD in offspring is higher than in parents and similar, or even slightly higher, than in siblings[8,9,11]. A potential source of bias could result from underreported cases in second-degree relatives and older generations, which might have influenced these differences.

In addition to an increased risk of developing the disease, first-degree relatives of IBD patients also have an increased likelihood of sharing the same phenotype[8,12]. This seems to be partly true in CD owing to the striking clinical concordance in families, in respect to disease location and behavior. On the contrary, literature is scarce and presents mixed results for CD severity and complications[8,9,13-16]. In UC families, the phenotype concordance data is less consistent, but a high concordance rate related to colonic extent and extra-intestinal manifestations has been reported[12,14].

The possibility that IBD develops at an earlier age in offspring than in their parents, a phenomenon known as anticipation, has been controversial[17-21]. Although different studies have reported such differences in age of onset, it seems that multiple biases could account for these findings. Whether familial IBD is a different clinical entity was the subject of debate. However, in the largest population-based study including 654 sporadic and familial IBD patients, a positive familial IBD history did not significantly influence clinical course or risk of developing IBD-related complications[22].

ROLE OF GENES IN FAMILIAL IBD

Evidences supporting the role of genetic factors in IBD pathogenesis

The fact that IBD is a genetically mediated disease was initially derived from the physician’s perception of a higher prevalence of UC and CD cases among the relatives of IBD patients. This hypothesis was initially supported by case report studies showing clustering in IBD families and was subsequently confirmed by several population-based studies[3-7]. In one of these studies, Yang et al[9] described a risk of developing UC and CD among first degree relatives of IBD patients of 1.6% and 5.2%, respectively. The risk of developing IBD also varies according to the ethnic origin of individuals, a fact that is likely to be linked to their genetic background. In that regard, the prevalence of IBD among the Jewish population is 2 to 4 times higher than in any other ethnic group, being greater in Ashkenazi than in Sephardic or Oriental Jewish, with no influence from their geographical location[9,11].

Another source of evidence underlying the key role of genetic factors in IBD steams from twin studies. In these studies the greatest IBD concordance rate was found in monozygotic twins, ranging from 20% to 50% in CD and from 14% to 19% in UC twins, whereas in dizygotic twins concordance rates dropped to 0%-7% in both CD and UC twins[23-26]. The degree of monozygotic-dizygotic twin concordance found in CD point towards a genetically mediated condition with a non-Mendelian inheritance pattern. Of note, the concordance rates observed in CD are greater than these found in type 1 diabetes, asthma or schizophrenia, all of them diseases with a well-established genetic background.

Genes involved in CD pathogenesis

Several genome linkage studies identified a number of CD susceptibility regions in chromosomes 1, 3, 4, 5, 6, 7, 10, 12, 14, 16, 19 and X[27-34]. After subsequent confirmation, regions on chromosomes 16, 12, 6, 14, 5, 19 were named IBD1 to IBD7, respectively[35]. Of these 7 original loci, only IBD1 (chromosome 16q12) was replicated in all studies, whereas another three loci, IBD2 (chromosome 12), IBD3 (chromosome 6) and IBD4 (chromosome 14) were replicated in some of the studies[36]. Later on, several CD susceptibility genes, such as NOD2, NOD1, toll-like receptors (TRLs) genes, and novel organic cation transporter (OCTN) genes were identified using either a candidate gene approach or positional cloning techniques[37].

The first and most relevant CD susceptibility gene described to date is NOD2. The carriage of one or more of the three main NOD2 variants (Arg702Trp, Gly908Arg and Leu1007incC) is found in 25%-45% of CD Caucasian patients and in only 15%-20% of healthy subjects[38]. A clear gene-dose effect has been described for NOD2 in CD patients. While the risk of developing CD is increased by 2-3-fold in subjects carrying one NOD2 variant allele, it reaches a 20-40-fold increase in subjects with two or more NOD2 variant alleles. In addition to increasing CD susceptibility, the NOD2 gene variants can also influence CD behaviour, phenotypes and need of surgery[39]. Similarly, the carriage of NOD2 variants has also been linked to a slight increase in familial CD risk[40]. In spite of being the most powerful CD susceptibility gene found to date, it must be underlined that the NOD2 gene only accounts for a small proportion of the genetic inheritance of CD in Caucasians. Moreover, NOD2 gene variants are infrequent in some geographic areas, such as Scotland, Ireland and Scandinavia, or even completely absent in subjects with an Asian and African-American genetic background[37]. In these cases other genes must account for the genetic predisposition to develop CD. This is in keeping with the highly polygenic nature of this disease.

In the last 3 years, the field of IBD genetics has experienced a dramatic transformation. Completion of the human genome project and development of tools capable of simultaneously studying a great number of genes has resulted in a much higher number of genes influencing CD and UC susceptibility than expected. Several genome-wide association studies (GWAS) have been undertaken in CD patients and healthy controls[41-45] and a meta-analysis has been recently published[46]. In this meta-analysis more than 30 independent loci are found to be convincingly associated with CD, providing an extraordinary insight into CD pathogenesis. Interestingly, most CD susceptibility genes are involved in either recognition of bowel microbiota antigens by the innate immunity, the IL-17/IL-23 pathway or autophagy, suggesting that these molecular mechanisms play a key role in CD pathogenesis.

Several studies have evaluated the frequency of CD-related mutations in affected families[47-49]. Jess et al[47] studied NOD2 mutation frequency in a population of Danish twins with IBD. In this study, a high prevalence of NOD2 mutation was observed in both CD twins and their healthy siblings. A Swedish study on monozygotic twin pairs reported a NOD2 frequency in both concordant CD siblings of only 22%, although the prevalence of NOD2 was indeed higher in concordant than in discordant twin pairs[48]. Joossens et al[49] investigated the prevalence of genetic markers (NOD2, NOD1, TLR4, CARD8) in multiplex and single-case families, healthy relatives and controls. The authors found a significant correlation between the number of genetic mutations per family and an increasing number of first-degree relatives with CD. However, these results could not discriminate between single-case and multiplex families.

Genes involved in UC pathogenesis

Two studies aimed at describing the influence in UC of the well-established CD susceptibility genes reported very interesting findings[50,51]. It became clear there is a genetic overlap between the two forms of IBD, with some genes involved in the development of both CD and UC (3p21.31, NKX2-3, CCNY). On the contrary, other genes have been only associated with UC, but not CD (ECM1 HERC2, STAT3 and PTPN2). In that regard, a very recently published UC GWAS meta-analysis has demonstrated that roughly half of the known CD susceptibility gene loci are shared by UC[52]. In addition, more than 20 exclusive UC loci have been recognized to date including, among others, IL10, ARPC2 and ECM1[52-55]. To summarize our present understanding, we believe that some genetic factors influence the global predisposition to develop IBD, whereas other genes are related to the risk of developing either UC or CD, specifically. Such genetic overlap between UC and CD probably contributes to the existence of a 5% of non-classifiable or indeterminate colitis among the IBD population. Similarly, it also contributes to the fact that first degree relatives of CD or UC patients not only have an increased risk of developing the same type of IBD, but also the other form of IBD, although with a lower frequency in the latter case.

EVIDENCE SUPPORTING AN ENVIRONMENTAL AETIOLOGY IN FAMILIAL IBD

Evidences supporting the role of environmental factors in IBD pathogenesis

The remarkable increase in the IBD incidence in the last few decades cannot be explained by changes in the genetic background of a certain population. Instead, it clearly points towards the existence of potent environmental factors playing a key role in IBD pathogenesis. Several studies performed in Europe that have evaluated IBD prevalence in immigrant populations from low IBD risk areas found that immigrants present a similar or higher risk of developing IBD when compared to the indigenous population[56-58]. These results suggest that differences in prevalence are probably associated with lifestyle and environmental factors, and not with a specific genetic background. It is remarkable though that only very few studies have addressed environmental etiologic factors in respect to familial IBD.

Eating habits, pets and previous infections

One of the largest controlled studies addressing the impact of environmental factors on familial IBD was conducted in Belgium by Van Kruiningen et al[59] who investigated 21 families with 3 or more first-degree relatives affected with CD. Subjects were interviewed using an extensive questionnaire on potential environmental factors. In this study affected families and controls presented some remarkable differences in eating habits, domestic factors and medical history. IBD patients ate fewer oats, rye and bran, consumed more unpasteurized cheese and drank more well water, compared to controls. Additionally, an increased frequency of smoking habit, appendectomy and fecal-oral transmitted infections was found in a subject who later developed CD. In respect to domestic habits, affected families also presented a lower daily contact with pets during childhood. Taken together, these results point towards a role of certain gastro-intestinal infections as triggering events contributing to the development of IBD, whereas the contact with pets during childhood seem to have a protective role favoring the immune system modulation.

Another familial aggregation study by the same Belgian group reported on a large Moroccan family with multiple CD cases[60]. Potential environmental, genetic and serological markers were studied in all family members. No differences in CD susceptibility genes or serological antibodies described in Caucasian populations were found between CD patients and healthy subjects. The study of environmental factors revealed the consumption of a large amount of unpasteurized milk in all family members, which was an environmental factor previously associated with occurrence of familial CD[59].

Another questionnaire-based study on environmental factors in a large monozygotic and dizygotic twin population included more than 300 twin pairs who were discordant for IBD diagnosis[61]. Twins with UC and CD reported recurrent gastrointestinal infections more frequently than their healthy siblings. These findings are in keeping with an increased frequency of fecal-oral transmitted infections reported in multiplex CD families[59] and suggest that past gastrointestinal infections might influence the risk of IBD.

Appendectomy

Appendectomy is associated with a lower risk of developing UC, although the exact mechanisms of this protective role are still not elucidated[62]. While controversial, it seems that the effect of appendectomy requires a certain degree of inflammation (appendicitis or lymphadenitis) and also applies to subjects undergoing appendectomy before the age of 20 years[63]. However, in a recently published study evaluating the usefulness of appendectomy as a therapeutic strategy for distal UC, 40% of patients experienced a complete symptoms resolution after elective appendectomy[64]. Conversely to UC, appendectomy seems to be associated with an increased risk of developing CD, although the studies addressing this issue yield conflicting results[65-70]. These discrepancies might be due to the inclusion of appendectomies performed at CD diagnosis or to methodological differences. More recently, data from large Swedish and Danish cohorts and a meta-analysis have demonstrated that the risk of developing CD is markedly increased only during a short period following appendectomy, disappearing after 5 years[71,72]. This behavior suggests that the association of appendectomy with CD might be a diagnostic bias, instead of a true risk factor.

Tobacco

Smoking habit, particularly cigarette smoking, is the most indisputable example of the influence of the environment on IBD[73,74]. Smoking has striking opposite effects on CD and UC, supporting the notion that distinct mechanisms underlie the pathogenesis of each form of IBD[74]. Subjects who have never smoked and former smokers are at a higher risk of developing UC, whereas present smokers have an increased risk of CD. In addition to the impact on disease susceptibility, smoking habit also modifies the clinical course of disease, increasing the risk of experiencing a relapse and the need for surgery[75-78]. Moreover, it has been demonstrated that tobacco discontinuation improves CD course[79]. Tuvlin et al[80] conducted a survey on tobacco use in UC and CD patients in familial IBD. In this study, smokers had double the risk of CD and former smokers had double the risk of UC, in younger age groups. A Danish case-report study on 2 monozygotic female twins with ileo-colonic CD and their non-affected brother and parents showed that though the healthy father. brother and twins all presented a NOD2 variant related to CD, only the affected twins were smokers, had undergone appendectomy and were on oral contraceptive use[81]. To further evaluate the influence of smoking habit in familial IBD, Bridger et al[82] analyzed 658 IBD patients, including 339 affected sibling pairs of whom 89 were discordant for smoking when diagnosed. Siblings who were discordant for smoking and IBD type almost always show CD in the smoker and UC in the non-smoker patient. The authors also suggest that the protective effect of tobacco on UC is due to a shift towards development of CD, in subjects prone to undergo bowel inflammation, rather than true protection of from the development of UC.

Oral contraceptive and non-steroidal anti-inflammatory drugs

Several studies have addressed the potential contribution of contraceptive pills to the development of IBD. It has been demonstrated that the risk of IBD in women taking oral contraceptives is greater that in controls, although there is no direct evidence for a causal relationship[83-85]. Data from two meta-analysis suggest a modest association between the use of oral contraceptives and development of IBD, with a pooled relative risk adjusted for smoking habit of 1.46 for CD and 1.26 for UC[86,87]. The most recent meta-analysis also suggests that the risk disappears once the medication is discontinued[87]. Other frequently used types of drug that have also been associated with IBD are non-steroidal anti-inflammatory drugs (NSAIDs)[88-95]. Due to their inhibitory action on protective prostaglandins, these drugs could enhance intestinal permeability facilitating disease activity. A growing body of evidence suggests a true association between NSAIDs and IBD activity, although the existence of multiple confounding factors makes it difficult to establish a formal relationship[96]. These confounding factors include selection of inadequate control groups, publication bias and intestinal tissue damage due not only to IBD activity but also to NSAIDs. Unfortunately, there are no studies evaluating the role of oral contraceptives and NSAIDs in familial IBD.

IBD as an infectious disease

Van Kruiningen et al[97] analyzed the pedigrees and time course of IBD development in a group of patients with familial IBD. They found that first-borns and subsequently born siblings were more frequently affected[87,97]. In addition, they described a statistically significant CD clustering that would indicate an infectious etiology supported by which family members were affected and time to develop symptomatic disease. Another study by Van Kruiningen et al[98] assessed 2 French families with multiple CD cases, in an attempt to identify the suspected infectious cause. However, Campylobacter, Yersinia, mycobacteria, mycoplasma, torovirus, coronavirus, Brucella, Influenza and animal enteropathogenic infections were all ruled out and no pathogen could be identified. Even though an abnormal recognition of antigens of the intestinal microbiota by innate immunity is thought to play a key role in IBD pathogenesis, there are no studies addressing specifically the interaction between bowel microbiota and familial IBD.

Intestinal permeability

An abnormal gut barrier function, with an increased intestinal permeability, could contribute to CD pathogenesis[99]. It that regard, it has been demonstrated that small intestinal permeability is increased not only in patients with CD but also in their healthy relatives[100,101]. Peeters et al[100] reported that 25% of healthy first-degree relatives of CD patients had an increased small intestinal permeability. The mechanisms responsible for these disturbances in bowel permeability are not fully elucidated and its pathogenesis is a much debated topic. In their study in familial and sporadic CD, Peeters et al[100] reported no specific genetic pattern accounting for the abnormal permeability found in patients and relatives. In addition, there were no significant differences between families with multiple members affected and families with only one individual affected. Interestingly, almost half of the spouses presented increased intestinal permeability, which clearly suggests that this abnormality is due to environmental and not to genetic mechanisms. In keeping with this finding, Fries et al[102] studied the prevalence of intestinal permeability and some CD genetic markers (including NOD2 main variants) in 23 families of CD patients. Authors found no association between increased intestinal permeability and genetic markers in their population. In contrast, other studies have found an association between an abnormal bowel permeability and the presence of NOD2 variations[101,103].

CONCLUSION

IBD is a complex polygenic disorder modulated by a series of environmental factors, some of which are likely yet to be determined. In spite of the significant progress done in the study of both the genetic and environmental factors associated, the exact contribution of each one of the many factors involved is still largely unknown. From a practical point of view, at present no specific recommendations to IBD families, in respect to genetic or environmental counselling, are available. Large, prospective, population-based studies following IBD patients and their healthy relatives will be necessary to untie the knots in this tangled web of genetic and environmental factors.