Article Text

Download PDFPDF

Genetics of inflammatory bowel disease: a puzzle with contradictions?
  1. Ist Department of Medicine
  2. Christian-Albrechts-University
  3. Schittenhelmstrasse 12
  4. 24105 Kiel, Germany
  5. s.schreiber{at}

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

See article on page 787

Genetic factors have a well established role in the aetiology of inflammatory bowel disease (IBD). The mode of inheritance suggests a polygenic disease with penetrance of the genetic factors being strongly influenced by the lifestyle of an industrialised society. The development of linkage analysis has allowed the generation of important molecular clues to the location of putative disease genes.1

The inheritance of genomic DNA can be traced by identification of highly polymorphic regions (“microsatellites”). Ideally, polymorphic regions should be used which are present in a high number of variations (“alleles”) in the population (that is, are highly informative).

If a phenotype (for example, disease) is inherited, the use of linkage analysis assumes that it will be possible to trace the piece of DNA which contains the putative disease gene. In affected relative pairs of individuals (sharing the same phenotype), certain areas of the DNA which contain disease genes should therefore be coinherited more frequently in comparison with random distribution. In a systematic analysis (“genome scan”), a densely spaced set of microsatellites is used throughout the entire genome to identify the origin of DNA and follow them through the inheritance process. Statistical measures can be used to qualify the degree of over proportional sharing (of polymorphic alleles). This “LOD” score (which roughly corresponds to a logarithmised p value) is usually displayed as a function of the genomic map (for example, fig 1 in Lesage and colleagues2). Linkage analysis typically leads to the description of a genomic area (often as large as one third of a chromosome), which most likely contains a disease gene.

Figure 1

Variations in linkage results in repeated studies are dependent on the genetic impact of a locus and sample size. A population was simulated with a six locus disease, with three contributing loci of high frequency but low impact (scenario F) and three contributing loci with high impact but low frequency (scenario R). LOD score as a function of the number of analysed sib pairs is shown, as calculated by the SIPPAIR program in 60 replicate studies, which were conducted in sib pairs randomly drawn from the population. The analysis of locus 1 under both disease model scenarios (F=frequent, locus specific heritability h2=0.37; R=rare, h2=0.25) is shown. Boxes denote median (interquartile range), and whiskers indicate minimum and maximum values (reprinted from Hampe and colleagues21).

It is obviously easier to detect disease genes in a monogenic disorder in which inheritance follows Mendelian rules. In many cases analysis of a few large families has led to the localisation of disease genes. In polygenic disorders, multiple disease genes are present and therefore linkage analysis has to be conducted using a population based approach. The method has to account for the fact that multiple disease genes may be present in varying combinations.

Siblings should share 50% of their DNA with each of their parents and also with each another. These DNA pieces should be randomly distributed, and sharing (in a sufficiently large sample) should not exceed 50% for any of the regions traced with microsatellites. Using siblings who share a disease will allow assessment of an over proportional amount of sharing which then can be used to describe a region in which disease genes are suspected.

In IBD a series of linkage studies have been conducted. The study by Hugot and colleagues3 described a region of chromosome 16 (“IBD1”) which has been confirmed in a large number of subsequent studies.4-12 Satsangi and colleagues13 found a region on chromosome 12 (“IBD2”) which was also confirmed in other studies.8 ,9 ,14 ,15 We have described IBD3 on chromosome 69 ,16 which was confirmed by Rioux and colleagues,17 and Brant and coworkers have suggested IBD4 (personal communication). In these studies linkage was considered as verified based on the criteria of Kruglyak and Lander18 if the LOD score exceeded 3.6 in a single study and exceeded 2.0 in subsequent populations.

Using stochastic instruments for a population based analysis will always carry a significant amount of statistical uncertainty. Little is known of the theoretical specificity and power of linkage analysis. Therefore, we constructed a computer model of a six locus disease (three strong and three weak disease genes) and simulated transmission through random mating.19-21 Using a virtual population, repeat linkage studies can be simulated and different strategies for finding disease genes can be explored. From these studies it has become apparent that results from repeated linkage studies in the same population will result in LOD scores varying in a certain range.21 The power of linkage to detect disease genes in fairly high and increases with the number of sib pairs used (fig1).21 However, using linkage for fine mapping of disease genes requires a high number of sib pairs, beyond technical and epidemiological feasibility.

In the study presented in this issue ofGut, Lesage and coworkes have conducted a linkage experiment limited to chromosome 12 (see page787).2 They used families from all over Europe. In their study they could not replicate linkage on chromosome 12 which had been described previously by Satsangi and colleagues.13 Taking into account the results from mathematical modelling described above, the differences are not as surprising as they first appear: variance is expected in repetitive linkage studies which is greater in studies with small family numbers and in disease genes with a low impact on the susceptibility to disease development, respectively. Lesage and coworkers used 95 families, which is in the lower range of power in comparison with larger studies.2 All other studies which found linkage to chromosome 12 reported LOD scores much smaller than those described by Satsangi et al in the initial analysis.13 As concluded by Lesage and colleagues2 further studies are needed to determine if chromosome 12 contains a major disease gene or if the putative gene in this region has only a small influence on the risk of developing IBD.

A third possibility is that we are looking at a population specific difference. Although it appears unlikely that the population in the UK is substantially different from that in France or Belgium, only a repeat study in the UK could clarify this issue. Within our own linkage study9 much of the power for the weak linkage seen on chromosome 12 resulted from the sib pairs recruited from the UK which would support this theory (unpublished). A consortial analysis, which drew about 50 sib pairs from each sample collected world wide,22 confirmed linkage to chromosome 12 although at a much lower significance level than that reported initially.

The work of Lesage and colleagues2 further indicates that linkage to chromosome 12 is not as powerful as initially thought. Identification of the first disease genes for IBD is still eagerly awaited. Although most likely in the distant future, only the definition of the causative mutations will allow a precise construction of a disease model and more exact predictions of the number and influence of additional genetic alterations.


Linked Articles

  • Inflammation and inflammatory bowel disease
    S Lesage H Zouali J-F Colombel J Belaiche J-P Cézard C Tysk S Almer M Gassull V Binder M Chamaillard I Le Gall G Thomas J-P Hugot