Contribution of the IL-2 and IL-10 genes to inflammatory bowel disease (IBD) susceptibility

Abstract

Although the importance of genetic susceptibility to IBD has been established by epidemiological studies, the genes involved remain poorly characterized. Important candidate genes include those encoding the immunoregulatory cytokines IL-2 and IL-10. The aim of this study was to assess the contribution of the IL-2 and IL-10 genes to IBD susceptibility. One hundred and ninety-eight pairs of siblings with IBD were genotyped at dinucleotide repeat polymorphisms within the IL-2 and IL-10 genes, and data analysed by the affected sib-pair method of linkage analysis and the transmission disequilibrium test (TDT). A subset of 89 affected sibling pairs was genotyped at markers flanking the IL-2 gene as part of a genome-wide search. The IL-2 polymorphism showed no linkage to IBD overall, but modest evidence for linkage to the ulcerative colitis (UC) data set (P = 0.028). A microsatellite 4 cM distal to the IL-2 gene showed a similar distortion in the ulcerative colitis subgroup (P = 0.006). The TDT showed some distortion of allelic transmission for the IL-2 polymorphism in the UC group, but this did not reach statistical significance (P = 0.09). Results for the IL-10 polymorphism were not significant. Thus the gene encoding IL-2 may contribute to UC susceptibility, but the effect is modest and must await replication in other data sets. The IL-10 gene does not appear to contribute to the risk of developing UC or Crohn's disease.

INTRODUCTION

There is strong evidence from epidemiological and family studies that genetic factors play an important role in determining susceptibility to IBD. Particularly striking are data from the Swedish twin registry showing concordance for Crohn's disease (CD) in 8/18 monozygotic twins but only 1/26 dizygotic twins [1], and the fact that siblings of IBD patients have a considerably higher risk of disease than the general population. λs (risk to siblings/population risk) for CD is 25–35 and for ulcerative colitis (UC) is 10–15 [2,3]. That this increased risk to relatives is maintained over geographical separation, and that non-blood relatives of patients in the same household are at no increased risk, underlines the importance of genetic susceptibility [4].

Inheritance of IBD does not, however, follow a monogenic pattern, at least in the majority of cases [4]: a more appropriate model is one in which several genes lead to increased risk of disease, some contributing to both CD and UC susceptibility and others perhaps determining the disease-specific phenotype.

Genes encoding immunoregulatory molecules clearly constitute important candidate susceptibility loci for IBD, and a number of recent studies have served to highlight the central roles played by IL-2 and IL-10 in orchestrating the immune response.

IL-2, together with interferon-gamma (IFN-γ), is produced by activated Th1 lymphocytes and stimulates the macrophages, natural killer (NK) cells and cytotoxic T cells of the cell-mediated immune response. The presence of large numbers of activated T cells in the involved mucosa of IBD patients suggests that IL-2 is likely to be playing at least some role in stimulating inflammation. Measurement of IL-2 protein and mRNA in vivo, however, has been confounded by methodological problems, and there is great inconsistency of results between studies. Nonetheless, there is an emerging consensus that IL-2 levels are increased in active CD, suggesting that this is a Th1-driven disease, but they are probably normal in UC [5–8].

Further evidence for a role for IL-2 in intestinal inflammation comes from studies with IL-2 knock-out mice. Animals homozygous for the IL-2 gene defect develop normally to 4 weeks of age: about half of the animals then develop a progressive lethal anaemia, whilst the remainder develop chronic diarrhoea and intermittent rectal bleeding [9]. Colonic histology reveals inflammatory changes with some homology to UC: there is mucosal ulceration with goblet cell depletion, crypt abscess formation and crypt distortion, and occasionally epithelial dysplasia. Inflammation is most marked distally and the small bowel is not affected. Interestingly, mice bred in germ-free conditions do not develop colitis.

IL-10 is produced by Th2 lymphocytes and acts to down-regulate macrophages and other antigen-presenting cells (APC). It also inhibits production of cytokines by Th1 lymphocytes in response to these APC, but does not block synthesis of cytokines by Th2 cells. There is a relative paucity of data regarding IL-10 production in IBD, with many of the same methodological difficulties pertaining as for IL-2. The evidence that is available points to increased IL-10 protein and mRNA in UC but not in CD [10].

Targeted disruption of the IL-10 gene has also been achieved in mice, and again results in intestinal inflammation in the presence of an intact gut flora [11]. The animals develop anaemia, weight loss and a chronic enterocolitis—particularly severe in the duodenum, proximal jejunum and proximal colon. This may reflect dis-inhibition of Th1 responses, triggering an immune response to intestinal microflora.

The human genes for both IL-2 and IL-10 have been cloned and sequenced. Fortuitously for the purposes of genetic studies, both contain highly polymorphic dinucleotide repeats or microsatellite markers, although to date there are no reports that the different alleles are associated with differing levels of gene expression. These microsatellite markers were studied in a cohort of families with multiple cases of CD and UC, to assess possible linkage or association of the genes for IL-2 and IL-10 to IBD susceptibility. Data are also presented for microsatellite markers flanking the IL-2 gene on chromosome 4; the IL-10 gene has not been accurately mapped on chromosome 1.

PATIENTS AND METHODS

Ethical approval for this study was given by the Central Oxford Research and Ethics Committee (COREC) in November 1992.

Study patients

One hundred and sixty-five pedigrees composed of 168 nuclear families with two or more affected offspring were recruited with the help of the National Association for Crohn's and Colitis and co-operating physicians. All individuals studied are European Caucasians resident in the UK and less than 5% are Jewish. Both parents were available for genotyping in 80 of the nuclear families, and one parent was available in 48.

The families studied contained a total of 198 pairs of siblings with IBD. In 89 pairs both individuals had CD, in 69 pairs both had UC, and in 40 pairs one had CD and the other had UC (‘mixed’ families). Case notes were reviewed to confirm the diagnosis.

In 147 families two sibs in one generation had IBD; in 15 families three sibs had IBD; and in a further three families there were two affected sib-pairs in successive generations (Table 1).

Table 1.

Description of the family panel used for linkage analysis and association study

CD, Crohn's disease; UC, ulcerative colitis; M, mixed CD and UC; IBD, sum of all families with inflammatory bowel disease.

The subset of this group used for genotyping of markers flanking IL-2 comprised 44 sib-pairs with CD, 30 with UC and 15 mixed pairs: these 89 sib-pairs were genotyped for the markers on chromosome 4 as part of a two-stage genome-wide search for IBD susceptibility loci.

Venous blood was collected into EDTA tubes from all affected individuals, available parents and unaffected siblings. DNA was extracted using a standard phenol/chloroform method, and stored at −70°C prior to genotyping.

Genotyping

PCR oligoprimers flanking a dinucleotide repeat in the 3′ untranslated portion of the IL-2 gene on chromosome 4q26 were identified from the Genome Database (accession no. GDB 344258). Microsatellite markers D4S430 and D4S406 lie 4 cM distal and 6 cM proximal, respectively. For the IL-10 gene on chromosome 1, DNA sequence flanking a dinucleotide repeat within the 5′ promoter region of the gene (accession no. GDB 808945) was used to design a pair of oligoprimers using OLIGO 4.0 software (Fig. 1).

Fig. 1.

Sequences for the two pairs of oligo-primers used to amplify dinucleotide repeats within the IL-2 and IL-10 genes.

Conditions for PCR for each primer pair were optimized over a range of annealing temperatures (53–60°C) and magnesium concentrations (1.0–2.5 mm). Amplification was performed in 96-well plates (Falcon) on MJ Research Machines PTC225. A 4-min denaturation at 95°C was followed by 35 cycles each comprising 45 s at 95°C, 45 s at the optimized annealing temperature and 10 s at 72°C. The 15 μl polymerase chain reaction (PCR) mixture contained 5 μl DNA (10 ng/μl), 0.08 μl Taq polymerase (5 U/μl), 0.6 μl of each primer (66 ng/μl), 1.5 μl 10× buffer, 0.375 μl dNTPs (final concentration 100 μm) and magnesium chloride in distilled water. Optimal conditions for the IL-2 primers were an annealing temperature of 55°C and MgCl₂ concentration of 2 mm, and for IL-10 55°C at MgCl₂ concentration of 1.5 mm. The amplified DNA product was electrophoresed on 6% acrylamide gels, and genotyped using the Applied Biosystems GENESCAN/GENOTYPER method (Fig. 2).

Fig. 2.

An ‘electropherogram’ is shown for the IL-2 microsatellite marker. The four traces correspond to four members of a single family: the top two are the affected offspring, the third is the mother and the fourth is the father. In this case the offspring each receive a different allele from their mother (sizes 117 and 145), but do share the allele from their father (size 135).

Statistical analysis

The genotyping data were analysed for linkage using the SIBPAIR program of Terwilleger & Lathrop [12], and for family-based association using the transmission disequilibrium test (TDT) [13]. Both programs use χ²-based statistics, and for all markers studied the probability of linkage was calculated for IBD overall, and for CD and UC separately.

SIBPAIR detects distortion of allele sharing by affected sibling pairs from that expected under random segregation, i.e. distortion of 0, 1 and 2 allele sharing from the expected ratios of 25%:50%:25%. It uses a maximum likelihood method of affected sibling pair linkage analysis, and its output is equivalent to an identity by descent statistic [12]. Where one or both parental genotypes are missing, the likelihood is given by a sum of terms corresponding to each of the possible parental genotype combinations. These are weighted by the genotype frequencies, calculated under the assumption of Hardy–Weinberg equilibrium, and incorporate information from all available offspring.

The TDT uses a likelihood ratio test for linkage disequilibrium, comparing the frequency with which each allele at the given marker locus is transmitted from heterozygous parents to affected offspring, with the frequency with which it is not transmitted.

RESULTS

Results of the linkage analysis for the IL-2 and IL-10 microsatellite markers are presented in Table 2. Both markers showed high heterozygosity: values are derived from allele frequencies in the populations studied.

Table 2.

Results of linkage analysis, given as P values and lod scores, for the microsatellite markers within the IL-2 and IL-10 genes: derived allele sharing ratios are also given

CD, Crohn's disease; UC, ulcerative colitis; IBD, inflammatory bowel disease (CD + UC + mixed).

There was no evidence for linkage of IL-2 to IBD overall (P = 0.073). When the data are analysed for CD and UC separately, however, some modest evidence for linkage of IL-2 to the UC group is seen (P = 0.028). Supportive evidence comes from the adjacent marker D4S430, which was studied in 89 sib-pairs with IBD as part of a genome-wide search: like the IL-2 marker it does not appear linked to IBD overall (P = 0.41) or CD (P = 0.5), but does show modest evidence of linkage to UC (P = 0.006). Formal tests of heterogeneity for D4S430 suggest that this locus may indeed be UC-specific (P = 0.01)—something which the results for the IL-2 locus itself just fail to achieve (P = 0.09). D4S406 showed no evidence of linkage either to the whole IBD data set, or to CD or UC data sets individually.

Results of the TDT analysis for the IL-2 microsatellite show no evidence of allelic association to IBD overall (P = 0.44; allele transmission shown in Table 3a). In the UC subgroup the ratio of transmission:non-transmission was 7:1 for allele 2 (size 116–118 base pairs with the given oligoprimer set). However, this result was not significant (P = 0.13).

Table 3.

Transmission/non-transmission of alleles of the IL-2 (a) and IL-10 (b) microsatellites from heterozygous parents to individuals with IBD

Results are shown for alleles appearing at least five times for the whole IBD data set. For the ulcerative colitis subgroup some distortion was seen for allele 2 of the IL-2 microsatellite (seven transmissions: one non-transmission; data not shown in table), but this did not achieve statistical significance (P = 0.13).

With regard to the data for IL-10, there was no evidence for linkage at the intragenic marker studied either for IBD overall or for CD or UC individually (respective P values 0.43, 0.45 and 0.5; Table 2), and results from the TDT were likewise negative (P = 0.11; Table 3b).

DISCUSSION

Genes encoding immunoregulatory molecules clearly constitute important candidate susceptibility loci for IBD. The HLA class II region has been implicated particularly in UC [14,15], but results have been less consistent for CD, pointing to the heterogeneity which is likely to exist at the genetic level between these two diseases. Cytokine genes also represent important candidates given their role in immunoregulation, but to date few studies have been published.

Considerable evidence points to an important role for the cytokine IL-2, particularly in Th1 lymphocyte responses, and for IL-10 predominantly in Th2 responses. Work in both human IBD and animal models of intestinal inflammation suggests that both may play an important role in disease pathogenesis. The question is whether any abnormalities of IL-2 or IL-10 expression in IBD are of primary pathogenic significance—determined by abnormalities within the IL-2 or IL-10 genes themselves—or whether they represent ‘appropriate’ sequelae of other pathogenic events.

Do the data presented in the current study help to differentiate between a primary intragenic abnormality and an ‘appropriate’ secondary effect? Clearly the interpretation of the data depends on how much weight one attaches to the modestly positive linkage results for IL-2 and its nearby microsatellite marker D4S430 in the UC group, the suggestion of genetic heterogeneity at D4S430, and the tentative suggestion of allelic association for the IL-2 marker.

Using the stringent criteria for statistical significance applicable to genome searching, the linkage results are essentially negative, but such criteria are probably too conservative in the context of a candidate gene study. Nonetheless, some consideration must be given to the fact that, world-wide, multiple candidates are being analysed for linkage to and association with IBD susceptibility. Given these constraints, a P value of 0.028 in the UC group is at best modest evidence for linkage, albeit supported by the results for D4S430.

The distortion observed in the transmission disequilibrium test for allele 2 in the UC group (7:1 case:control) appears impressive, but fails to achieve statistical significance, partly due to the small numbers involved. The combination of this weakly positive evidence for linkage and hint of allelic association may, however, be pointing to a true effect at the IL-2 locus in those patients with UC.

If one supposes that the distortion does indeed represent a true genetic effect, how can this be reconciled with what is already known of IL-2 in IBD? Given a model in which UC and CD share some susceptibility loci but differ at others—the latter perhaps conferring the disease-specific phenotype—and given that IL-2 levels appear increased in CD but not UC, it could be speculated that a defect in the IL-2 gene prevents up-regulation of its expression and leads to the UC phenotype in individuals already predisposed to IBD due to other genetic and environmental factors. Individuals susceptible to IBD but with an intact IL-2 gene might be more likely to develop a CD phenotype, due to an intact Th1 regulatory pathway. Such a hypothesis is consistent with the development of a UC-like illness in mice rendered IL-2-deficient, and also recognizes the genetic heterogeneity which is likely to exist within IBD.

The data for the IL-10 polymorphism appear more convincingly negative, both for the linkage and TDT results. Given that any distortion occurring in linkage analysis tends to be seen over a wide chromosomal segment, this suggests that the IL-10 gene is unlikely to play a primary role in the pathogenesis of either CD or UC. It should be emphasized that this does not rule out a secondary role for the IL-10 protein in IBD pathogenesis, or an IL-10 receptor defect, or aberrant expression of the gene due to extrinsic regulatory elements. Any such aberrant expression is, however, unlikely to be due to abnormality within the IL-10 gene itself.

In conclusion, the data presented are consistent with an effect of the IL-2 gene in UC susceptibility—an hypothesis which must await testing in further data sets—but suggest that the IL-10 gene is unlikely to play a role in IBD pathogenesis.