Functional Correlates of the Interleukin-1 Receptor Antagonist Gene Polymorphism in the Colonic Mucosa in Ulcerative Colitis

Home , IL1A

Genes and Immunity (2004) 5, 8–15 & 2004 Nature Publishing Group All rights reserved 1466-4879/04 $25.00 www.nature.com/gene Functional correlates of the interleukin-1 receptor antagonist gene polymorphism in the colonic mucosa in ulcerative colitis

MJ Carter1, S Jones1, NJ Camp4, A Cox2, J Mee2, B Warren3, GW Duff2, AJ Lobo1 and FS di Giovine2 1The Gastroenterology and Liver Unit, University of Sheffield, Royal Hallamshire Hospital, Sheffield, UK; 2Division Of Genomic Medicine, University of Sheffield, Royal Hallamshire Hospital, Sheffield, UK; 3Histopathology Department, John Radcliffe Hospital, Oxford, UK; 4Genetic Epidemiology, University of Utah Medical School, UT, USA

Association studies have identified the interleukin-1 receptor antagonist gene allele 2(IL-1RNn2) as a marker of susceptibility in ulcerative colitis (UC). This study investigated the significance of the IL-1RN genotype with respect to protein and mRNA expression in the colonic mucosa. Homogenates of rectal biopsies from 99 UC and 54 controls were assayed for cytokines IL- 1ra, IL-1a and IL-1b using ELISA. IL1RN, IL1A and IL1B genotypes were determined using restriction-enzyme analysis. The ability of the two IL1RN alleles to generate steady-state mRNA accumulation was assessed in the colonic mucosa of seven heterozygous patients. Stepwise linear regression demonstrated that IL-1RN genotype (P¼0.001), diagnosis (Po0.0001) and treatment (Po0.03) were independent factors associated with the IL-1ra protein level whilst IL1RN genotype (P¼0.005) and macroscopic inflammatory grade (Po0.0001) were associated with the IL-1ra/ total IL-1 ratio. The IL1RNn2 correlated with reduced IL-1ra and IL-1ra/IL-1 ratio with a gene dosage effect. In heterozygous UC patients the ratio of allele 1 mRNA / allele 2 steady state mRNA was always greater than 1 (range: 1.2–3.1) (P¼0.018). The IL-1RNn2 is associated with reduced levels of IL-1ra protein and IL-1RN mRNA in the colonic mucosa, providing a biologically plausible explanation for the observed association of the allele with the disease. Genes and Immunity (2004) 5, 8–15. doi:10.1038/sj.gene.6364032

Keywords: interleukin-1 receptor antagonist; cytokines; inflammatory bowel disease; ulcerative colitis; genetics

Introduction trol association studies have demonstrated significant associations with particular alleles in a number of The causes of the chronic idiopathic inflammatory bowel candidate genes although lack of reproducibility may diseases (IBD), ulcerative colitis (UC) and Crohn’s further suggest genetic heterogeneity. disease (CD), remain unknown. Evidence suggests Cytokines have been implicated in the pathogenesis of however that disease susceptibility and severity are chronic inflammatory diseases and have a regulatory determined at least in part, by genetic factors.1 Recent and effector role in the mucosal immune and inflamma- epidemiological data, including family studies,2 twin tory response in IBD.11 Interleukin-1 (IL-1) alpha and studies3 and investigations in different ethnic popula- beta are major proinflammatory cytokines involved early tions,4 suggest that UC and CD represent related, in the inflammatory cascade. The interleukin-1 receptor complex genetic diseases. Associations with genetic antagonist (IL-1ra) is the inhibitor of these IL-1 agonists syndromes and autoimmune disorders, which have a and acts by competitively binding to IL-1 receptors known genetic component, further support this concept.5 without eliciting signal transduction.12 All three proteins Genome-wide linkage scans have identified a number of are coded by genes in the interleukin-1 gene cluster on susceptibility loci suggesting that several genes are the long arm of chromosome 2.13 associated with IBD.6–9 Some of these loci appear to be Enhanced production of IL-1 has been shown in the common to both CD and UC while others are unique to gut tissue of animal models of intestinal inflamma- either disease. Such genetic complexity is postulated to tion14,15 and in both the colonic mucosa16 and by cell arise from epistasis (gene–gene interactions) and from preparations isolated from the intestine of patients with gene–environment interactions.9,10 Individual case–con- IBD.17 Tissue levels correlate with disease activity but the IL-1ra/IL-1 ratio shows the closest correlation with inflammation.18 Rabbit immune complex colitis is atte- Correspondence: Dr MJ Carter, Flat 9, 156 Haverstock Hill, Belsize Park, nuated by the administration of IL-1ra19 and reduction of London NW3 2AT, UK. E-mail: Martyn. [email protected] IL-1ra levels by either gene knockout in mice20 or by This study was financially supported by a grant from The National antibody neutralisation in rabbits21 increases suscept- Association of Colitis and Crohn’s Disease and The Special Trustees ibility to the induction of colitis. An imbalance in the of the Central Sheffield Hospitals. Dr Carter is a Digestive Disorders Foundation Research Training Fellow. IL-1ra/IL-1 ratio may therefore contribute to the chronic Received 07 March 2003; revised 24 September 2003; accepted 26 inflammatory response in ulcerative colitis. The mole- September 2003 cular mechanisms underlying this imbalance in IBD are IL-1 receptor antagonist in ulcerative colitis MJ Carter et al 9 unknown. However, interindividual differences in cyto- flamed controls and UC patients, studied. In this small kine production may be under genetic control.22 cohort of patients, no statistical differences in allele The IL-1 gene cluster contains several polymorphic carriage rates or allelic frequencies were demonstrated genetic markers. The IL-1ra gene (IL-1RN) contains a although a trend was evident for IL-1RN, with carriage variable number of tandem repeats (VNTR) polymorph- of the allele 2 increased in patients with UC compared to ism in intron 2,23 which is in complete linkage disequili- controls was observed (53 vs 39%) as expected. These brium with a single-nucleotide polymorphism (SNP) in patients were genotyped as part of a large study exon 2 ( þ 2018).24,25 SNPs have also been identified in assessing the role of the IL-1RN in genetic susceptibility both the IL-1 beta (IL-1B) and the IL-1a genes (IL-1A).26 to UC.25 The variable degree of linkage disequilibrium that exists across the gene cluster between these markers has Cytokine concentrations in colitis vs normals recently been studied in detail.27 Table 2 shows the tissue cytokine levels (normalised for Polymorphisms within the IL-1 gene cluster have been DNA content) of biopsies from UC and control patients. associated with both susceptibility to and severity of a The tissue levels of IL-1a and IL-1b were low in the number of chronic inflammatory and autoimmune control biopsy specimens. Total IL-1b tissue levels were conditions.26 Allele 2 of the IL-1RN VNTR polymorph- approximately 5–10 fold higher than the IL-1a levels. IL- ism (IL-1RNn 2) was first reported to be associated with 1ra tissue levels were approximately 1000 fold higher UC in 1994.28 Although not all subsequent studies in than total IL-1 levels in biopsies from controls. The IL-1a, Caucasian Northern European patients have confirmed IL-1b and IL-1ra tissue levels were all increased in rectal this association29–34 their lack of statistical significance biopsy samples from patients with UC compared to could reflect inadequate power. This was supported by controls. The IL-1ra/IL-1 ratio in the mucosa was the significant association found in a further well- decreased in UC patients compared to the controls. powered study and an accompanying meta-analysis of this and the seven previous studies in Caucasian North 25 European patients. Two subsequent studies have been Table 2 Cytokine protein steady-state levels and IL-1ra/total IL-1 performed in Northern European populations. One of ratio in colonic mucosa in controls and UC patients (rg cytokine/ these studies confirmed the association35 while the other mg DNA)(mean7s.e.) did not.36 A further study did confirm the genetic association in Jewish and Hispanic populations.37 The UC patients Controls association appears to be greatest in patients with extensive disease and in those requiring surgery for IL-1a 144.8754.9 9.371.7 UC.32,38 Allelic associations of SNPs within the IL-1A IL-1b 1954.37653.3 48.177.5 7 7 and IL-1B with IBD have not been demonstrated, IL-1ra 158250.9 15396.1 43291.5 7337.8 IL-1ra/total IL-1 479.97103.4 1062.97196.7 although this may represent the limited power of these studies.28,32,34,39 However, a recent study did suggest that the IL-1B þ 3954 polymorphism might have a role in Colonic mucosal levels of IL-1 alpha, IL-1 beta, IL-1ra and IL-1ra/ 40 total IL-1 ratio of biopsies from UC and noninflammatory control determining disease behavior. 7 To further investigate the previously described genetic patients. Data are shown as mean ( SEM). Cytokine concentrations association with UC, the functional effect of the IL-1RNn2 are normalised for DNA content of the homogenate (to normalise for cell number). Cytokine concentrations and DNA content of and other polymorphic variants within the IL-1 gene tissue homogenates were measured using ELISA and a novel cluster have been examined in terms of steady-state protein fluorescent nucleic acid stain, Picogreens respectively, as described and mRNA expression in vivo in the colonic mucosa. in Materials and methods. Using stepwise linear logistic regression as described in Statistical methods, diagnosis of UC was independently significantly associated with IL-1a (P¼0.008), IL-1b Results (Po0.0001) and IL-1ra (Po0.0001) colonic protein levels. The IL- 1ra/total IL-1ra level was not significantly reduced in the colonic Patient characteristics/IL-1 gene cluster SNP genotypes mucosa in patients with UC compared to noninflammatory controls Table 1 shows the IL-1RN ( þ 2018), IL-1B (À511 and (P40.05) after degree of inflammation and IL-1RN genotype were þ 3954) and IL-1A ( þ 4845) genotypes for the nonin- accounted for by the regression analysis.

Table 1 Genotypes of UC patients and controls

Controls Ulcerative colitis

1,1 (%) 1,2(%) 2,2(%) 1,1(%) 1,2(%) 2,2(%)

IL-1RN 33 (61.1) 17 (31.5) 4 (7.4) 47 (47.5) 47 (47.5) 5 (5.0) IL-1b -511 22 (40.7) 26 (48.1) 6 (11.1) 40 (40.4) 48 (48.5) 11 (11.1) IL-1b +3954 33 (61.1) 19 (35.2) 2 (3.7) 66 (66.7) 27 (27.3) 6 (6.1) IL-1a +4845 32 (59.3) 18 (33.3) 4 (7.4) 55 (56.6) 37 (37.4) 7 (7.1)

IL-1RN, IL-1B-511, IL-1B+3954 and IL-1A+4845 genotypes for UC and noninflammatory control patients. Genotyping was performed using PCR and restriction enzyme digestion and gel electrophoresis as described in Materials and methods. Allelic carriage rates were compared between cases and controls using w2 statistics. No statistical differences in allele carriage rates were demonstrated although a trend was evident for IL-1RN, with carriage of the allele 2 increased in patients with UC compared to controls was observed (53 vs 39%) as expected.

Genes and Immunity IL-1 receptor antagonist in ulcerative colitis MJ Carter et al 10

Figure 1 Colonic mucosal levels of IL-1ra according to IL-1RN 7 Figure 2 Colonic mucosal IL-1ra/total IL-1 ratio according to IL- genotype. Data are shown as mean ( s.e.m.). Cytokine concentra- 7 tions are normalised for DNA content of the homogenate (to 1RN genotype. Data are shown as mean ( s.e.m.). Cytokine normalise for cell number). Cytokine concentrations and DNA concentrations are normalised for DNA content of the homogenate content of tissue homogenates were measured using ELISA and a (to normalise for cell number). Cytokine concentrations and DNA s content of tissue homogenates were measured using ELISA and a novel fluorescent nucleic acid stain, Picogreen respectively, as s described in Materials and methods. IL-1RN genotypes were novel fluorescent nucleic acid stain, Picogreen respectively, as determined using PCR and restriction enzyme analysis as described described in Materials and methods. IL-1RN genotypes were in Materials and methods. Using stepwise linear logistic regression determined using PCR and restriction enzyme analysis as described as described in Statistical methods the IL-1 RN genotype was in Materials and methods. Using stepwise linear logistic regression independently significantly associated with colonic IL-1ra protein as described in Statistical methods the IL-1RN genotype was level (P¼0.001). independently significantly associated with the colonic IL-1ra/total IL-1 ratio (P¼0.005).

Factors determining colonic cytokine concentrations model had an adjusted R2 of 0.498 (Po0.0001). Variables Linear regression analyses were performed in a stepwise for IL-1A, IL-1B or IL-1RN genotypes were not signifi- fashion. Three variables were significant for inclusion in cant and were not included in the model. the linear regression model for colonic IL-1ra level. In Three variables were significant for inclusion in the order of inclusion into the model, these were diagnosis linear regression model for colonic IL-1b protein level. In (Po0.0001), IL-1RN genotype (P¼0.001) and treatment order of inclusion into the model, these were diagnosis (P¼0.03). The proportion of the total variance accounted (Po0.0001), histological inflammatory grade (P¼0.007) for by these three variables in the final model was 0.388 and macroscopic inflammatory grade (P¼0.02). The final (adjusted R2 0.388; Po 0.0001). Figure 1 shows the tissue model had an adjusted R2 0.565 (Po0.0001). Variables for IL-1ra levels in (a) UC patients and (b) control patients IL-1A, IL-1B or IL-1RN genotypes were not significant according to the IL-1RN genotype. Carriage of IL-1RNn2 and were not included in the model. was associated with reduced protein level with a gene dosage effect. Heterozygous patients had intermediate Ratio of allele 1/allele 2 mRNA in heterozygous patient protein levels while homozygous patients had the lowest colonic mucosa levels. Figure 3 shows a composite phosphor-imaging gel of the Two variables were significant for inclusion in the allele-specific transcript analysis for the seven UC linear regression model for colonic IL-1ra/total IL-1 patients who were heterozygous for the þ 2018 IL-1RN ratio. In order of inclusion into the model, these were polymorphism. The ratio of allele 1 mRNA/allele 2 macroscopic inflammatory activity (Po0.0001) and IL- mRNA was significantly greater than 1 in all cases 1RN genotype (P¼0.005). The final model had an (P¼0.018), ranging from 1.2 to 3.1. Allele 2 was 2 adjusted R of 0.253 (Po0.0001). Figure 2 shows the consistently associated with lower accumulation of tissue IL-1ra/total IL-1 ratio in (a) UC patients and IL-1ra mRNA than allele 1. (b) control patients according to the IL-1RN genotype. Carriage of IL-1RNn 2 was associated with a reduced IL-1ra/total IL-1 ratio with a gene–dosage relationship. Discussion Three variables were significant for inclusion in the linear regression model for colonic IL-1a protein level. In We have studied the functional correlates of polymorph- order of inclusion into the model, these were histological isms within the IL-1 gene cluster in the colonic mucosa in inflammatory grade (P¼0.002), diagnosis (P¼0.008) and vivo, to investigate whether the observed genetic associa- macroscopic inflammatory grade (P¼0.023). The final tion between the rarer allele of the IL-1RN and UC is

Genes and Immunity IL-1 receptor antagonist in ulcerative colitis MJ Carter et al 11

Figure 3 Samples from seven UC patients, heterozygotes for IL-1RN ( þ 2018) were analysed by allele-specific transcript analysis (see Materials and methods). In lane 1 of each gel is the phosphor-imaging acquisition of the undigested RT-PCR product. In lane 2, the sample has been digested with AluI, hence the upper band (a) represents cDNA from mRNA transcribed from allele 2, and the lower band by allele 1. In lane 3, the sample is digested with MspI, the upper band (a) represents cDNA from mRNA transcribed from allele 1, and the lower band (b) being related to allele 2. Following quantitation of each band by phosphor-imaging, the allele 1/allele 2-derived cDNA ratio by both digestions is shown (c). The average ratio for each patient is also shown (d). reflected in functional differences of the genes in the consideration which may provide an explanation for patient population. The results presented here show that these contrasting results. Although our patients with UC the rarer allele of the IL-1RN is associated with reduced had a lower IL-1ra/IL-1 ratio than noninflammatory accumulation of IL-1ra protein within the colonic patients this was not demonstrated to be independent of mucosa, and is also associated with a significant the degree of inflammation and IL-1RN genotype. The reduction in the biologically important IL-1ra/total IL-1 IL-1ra/total IL-1 ratio was however demonstrated to be ratio. Using a allele-specific transcript analysis metho- very highly associated with the degree of inflammation dology in heterozygous individuals with UC, we have as reported previously. further demonstrated that the rarer allele is associated Previous data regarding the functional effects of the with less steady-state mRNA within the colonic mucosa IL-1RN polymorphism on IL-1ra and IL-1 production are compared to the more common allele. This functional conflicting. The rarer allele 2 of the IL-1RN has been association for the disease-related allele provides a associated with both increased and decreased IL-1ra biologically plausible explanation for the association protein production from peripheral blood mononuclear between the allele and UC as the reduced IL-1ra and cells (PBMC) in vitro.43,37 The investigators however concomitant IL-1ra/IL-1 ratio may prevent adequate used different stimuli to elicit IL-1ra protein production control of mucosal inflammation. from the PBMC which makes direct comparison The results presented also demonstrate that the well- difficult. The effect of IL-1RNn2 on serum levels of characterised polymorphisms within the IL-1A and IL-1B IL-1ra has also been assessed in healthy controls and do not appear to be associated with a functional effect in was shown to correlate with increased plasma levels of terms of their respective cytokine expression in the the protein with some coordinate regulation supplied colonic mucosa. This may provide an explanation for the by the IL-1beta gene.44 A further small study in diabetic lack of association between these polymorphisms and patients demonstrated similar findings but was unable ulcerative colitis in previous case–control association to confirm the findings in healthy controls.45 However, studies.28,32,34,39 serum IL-1ra is an acute-phase protein46 and regulation The results presented here also provide further data of serum levels may be under different control assessing factors that are associated with cytokine levels pathways compared to local tissue levels such as the in the colonic mucosa. Patients with ulcerative colitis had colonic mucosa. higher colonic levels of IL-1ra, IL-1a and IL-1b compared A single study has investigated protein expression in to healthy controls and disease status was shown to be the colonic mucosa and suggested that total IL-1ra an independent predictor of these protein levels, which protein was reduced in individuals who carried at least concurs with previous data.16,18,41,42 Inflammation is one copy of the IL-1RNn2.47 Results from this previous associated with marked rises in colonic levels of study represented a combination of patients with UC, proinflammatory cytokines IL-1a and IL-1b which are CD, inflammatory controls and noninflamed controls. normally expressed at low levels. IL-1ra is a constitu- Our results support and extend this later study’s tively expressed cytokine and although levels rise with findings by demonstrating a significant functional effect inflammation, we have not demonstrated this to be of the IL-1RNn2 in the colonic mucosa. We have also independent of the diagnosis of ulcerative colitis and IL- demonstrated that the allele 2 is also associated with a 1RA genotype.18,41,42 We did not demonstrate the reduction in the biologically important ratio of IL-1ra/ previous observation that IBD was an independent total IL-1. In addition, we have shown that the decreased predictor of the IL-1ra/total IL-1 ratio.18,41,42 However, protein expression is associated with a decreased these previous studies did not take genotypic effects in to steady-state mRNA level.

Genes and Immunity IL-1 receptor antagonist in ulcerative colitis MJ Carter et al 12 There are conflicting data regarding the functional were collected prospectively at the time of sampling. In effect of the IL-1 beta polymorphisms on IL-1 beta all, 17 patients were taking corticosteroids, 59 were production.48,49 Indeed, Santtilla’s48 study suggest that taking 5-aminosalicylic acid drugs, 14 were taking the IL-1RNn2 or an unknown allele strongly associated azathioprine and 28 were on no medication at the time with it has a more major role in vitro IL-1 beta of sampling colonic tissue. Median age at diagnosis was production. There are no data on the quantitative 31 years (range 8–75) and median age at biopsy was 41 functional correlates of the IL-1 alpha polymorphism years (range 17–82). 63 patients had nonextensive although it results in protein dimorphism (serine for disease, 52 were female and 14 had a family history of alanine at 114). In this study, in vivo in the colonic mucosa IBD (first- or second-degree relative). we have been unable to demonstrate any functional Healthy controls who were undergoing colonoscopy correlates of these polymorphisms. for investigation of noninflammatory conditions (n¼54) The VNTR polymorphism is in a noncoding region of were also studied. The absence of colonic inflammation the gene in intron 2 and the exon 2 polymorphism is a in these patients was confirmed by histological examina- conservative base-pair change which does not alter the tion of a separate rectal biopsy taken at the same time as protein amino-acid sequence. It is difficult to envisage those taken for assay of cytokine concentrations. Of these how either polymorphism has a direct functional effect in patients 32 were female and 22 were male. Median age at terms of IL-1ra production and indeed unpublished gene biopsy was 53 years (range 23–85). Indication for reporter studies showed that the VNTR itself had no colonoscopy was; 14 polyp surveillance, 10 per rectum effect on gene transcription in vivo (JK Tarlow, personal bleeding, eight altered bowel habit, seven polyp excision, communication). It is therefore likely that both poly- seven iron deficiency anaemia, four abdominal pain, two morphisms are in linkage disequilibrium with a func- colonic carcinoma follow-up, one constipation and one tional polymorphism, as yet undefined. This functional angiodysplasia. genetic change is perhaps most likely to be either in the gene promoter region or in the 30 untranslated region, Genotyping for IL-1 gene cluster polymorphisms affecting gene transcription or post-transcriptional Blood was taken into ethylenediamine-tetraacetic acid events, respectively. Whether this reduced mRNA level (EDTA) tubes (2 Â 10 ml) and stored frozen until DNA is due to decreased transcription or decreased mRNA was extracted by standard methods.26 DNA was stored at stability is under further investigation. Positive results in 41C. Genotyping was performed using PCR and restric- either sense will focus study on the putative promoter tion enzyme digestion and gel electrophoresis. All regions or 30UTR of the IL-1RN where functional variants patients were genotyped for the C/T IL-1RN SNP at in linkage with the IL-1RN allele may be sought. position þ 2018 as previously described,26 a marker The results presented therefore demonstrate that the which is in 100% linkage disequilibrium with the IL-1RN IL-1RNn 2 correlates with reduced levels of IL-1ra VNTR polymorphism.25 All patients were also geno- protein and mRNA in the colonic mucosa providing a typed for the two C/T SNPs in the IL-1 beta gene at biologically plausible explanation for the observed positions À511 and þ 3954, and for the C/T IL-1 alpha association of the allele with ulcerative colitis. Future SNP at position þ 4845, as previously described.26 work will involve the characterisation of DNA structures Negative controls were included with each reaction. in the IL-1 cluster that are responsible for this functional association, and the use of this information to develop Biopsy specimens new therapies including pharmacogenetic approaches. Endoscopic biopsy specimens were obtained during colonoscopy. All specimens were taken from the rectum at approximately 10 cm from the anal margin to ensure Materials and methods uniformity. Three biopsies were taken during endoscopic examination for cytokine assay and a further sample for Patients and controls histopathology. The degree of inflammation at the biopsy This study was approved by the South Sheffield Local site was graded macroscopically by a gastroenterologist Research Ethics Committee in 1994 and was performed according to the Baron grade51 and microscopically by a in accordance with the Declaration of Helsinki. All single gastroenterological histopathologist.52 Biopsies for patients participated in the study after giving fully cytokine analysis were immediately frozen in liquid informed written consent. nitrogen before being transferred to a freezer for storage A total of 99 patients with UC were studied. Disease at À701C prior to homogenization. Biopsies for histolo- diagnosis had been confirmed by colonoscopy with gical examination were formalin fixed and stained by colonic series of biopsies in all patients within 4 years standard methods with haematoxylin and eosin. of entry into the study. Patients were included only if histological assessment was suggestive, highly sugges- Assay of cytokine concentrations tive or diagnostic of UC as outlined in the British Society Biopsy specimens for assay of cytokine concentrations of Gastroenterology guidelines.50 Individuals with histo- and DNA content were homogenised on ice in 750 mlof logical appearances of indeterminate chronic inflamma- PBS pH 7.2 in the presence of proteinase inhibitor tory bowel disease or Crohn’s disease were excluded. (Cocktail Protease Inhibitor, Roche, UK) using a pestle The histological limit of inflammation was used to define and mortar mechanical homogeniser on ice. Insoluble disease extent which was classified as either extensive or material was centrifuged at 10 000 g (Beckman J6, USA) at nonextensive depending on whether the inflammation 41C for 15 min. To avoid repeated freeze thawing, tissue extended beyond the splenic flexure. Clinical details homogenate was divided into separate aliquots for assay including sex, age at diagnosis, age at biopsy, extent of of each cytokine and for DNA content, and were stored colitis, current medication, family history of UC and IBD at À701C prior to use.

Genes and Immunity IL-1 receptor antagonist in ulcerative colitis MJ Carter et al 13 Aliquots of the tissue supernatant were used for the PCR products of 154 bp were divided and allele- determination of IL-1ra, IL-1a and IL-1b by ELISA (R&D specific restriction digestion performed using restriction Systems, Minneapolis, MN, USA). Where protein con- enzymes AluI and MspI. Digested products were centrations exceeded the limit of detection of the assays, electrophoresed on 9% polyacrylamide gels which were concentrations were assayed in appropriate dilutions of then dried and analysed on a Biorad G-250 molecular the tissue homogenate. Total IL-1 agonist concentration imager. MspI digestion of heterozygous cDNA gave was made by addition of IL-1a and IL-1b concentrations. fragments of 154 bp (labelled) corresponding to allele 1 In order to avoid bias for cellularity due to confounding (uncleaved) and 125 bp (labelled) and 29 bp (unlabelled) elements in the extracellular matrix and the presence of from allele 2. AluI digestion of heterozygous cDNA gave oedema, when there are variable degrees of inflamma- fragments of 154 bp (labelled) corresponding to allele 2 tion, we normalised cytokine concentrations to total (uncleaved) and 126 bp (labelled) and 28 bp (unlabelled) DNA content, a method previously used in chronic from allele 1. inflammatory arthritis53 and IBD.54 Radioactive RT-PCR products were quantified by measuring the density of phosphor-imager units in each Assay of DNA content of colonic biopsy specimens band, and after subtraction of background, these results Aliquots of the soluble supernatant were used for the were used to calculate the relative amounts of mRNA determination of DNA content. DNA content was derived from each allele. Such analysis accounted for assayed using a novel fluorescent nucleic acid stain, confounding factors such as interindividual variability Picogreens(PicoGreent, Molecular Probes, Leiden) that and loading and/or pipetting errors. Moreover, by using selectively binds double-stranded DNA and provides restriction enzymes specific for each allele, the ratio was accurate estimates in complex solutions.55 DNA content measured in two opposite ways simultaneously in the of tissue homogenates was assayed using a LS50B same patient, providing an internal control. Experiments scanning fluorimeter (Perkin-Elmer, Warrington, UK). were repeated three times in all seven individuals from Fluorescein excitation and emission wavelengths of the initial RNA preparation. samples were compared to DNA standard curves and values obtained by interpolation. All tissue homogenates were assayed in triplicate. Statistical methods Allelic carriage rates were compared between cases and Allele-specific transcript analysis (ASTA) controls using w2 statistics. To analyse which factors were The IL-1RN þ 2018 polymorphism was used as an associated with the individual cytokine levels and the IL- exonic marker to recognise mRNA copies that had been 1ra/total IL-1 ratio, a multiple stepwise linear regression transcribed from either allele of the IL-1RN gene, using a was performed. Natural logarithms of cytokine previously described method of allele-specific transcript concentrations and the IL-1ra/Total IL-1 ratio were analysis developed in our laboratory.24 This method has entered as the dependent variable. Factors included in recently been revisited and validated.56 all the analyses were diagnosis (UC or control), histolo- Rectal mucosal biopsies were obtained from seven gical inflammatory grade (0–9), macroscopic inflamma- patients with UC who were heterozygous for the SNP tory grade (0–3), treatment with anti-inflammatory (C/T at þ 2018) in exon 2 of the IL-1RN gene. Biopsies medication (yes or no), sex (male or female) and were taken as described above but were stored in liquid genotypes (1,1, 1,2 and 2,2) of the four single-nucleotide nitrogen until homogenisation. Total cellular RNA was polymorphisms. For the allele-specific transcript analy- extracted after tissue homogenisation on dry ice using a sis, the ratio of allele 1 mRNA to allele 2 mRNA was modification of the acid guanidinium–phenol–chloro- compared using the Wilcoxon signed-rank test for paired form method (RNAzol; Biogenesis).57 data. Approximately 1 mg RNA was reverse-transcribed with random hexamers and AMV reverse transcriptase (Pro- mega) in a final reaction volume of 10 mlat421C for 1 h. This cDNA was then subjected to PCR with a forward References primer 50-TTC.TAT.CTG.AGG.AAC.AAC.CAA.CTA. GTA.GC-30 adjoining the polymorphic site and reverse 1 Satsangi J, Jewell DP. The genetics of inflammatory bowel primer 50-CAC.CAG.ACT.TGA.CAC.AGG.ACA.GGC. disease. Gut 1997; 40: 572–574. ACA.TC-30. The forward primer contained a single-base 2 Orholm M, Munkholm P, Langholz E, Nieleson OH, Sorensen TIA, Binder V. Familial occurrence of inflammatory bowel mis-match two bases from the 3’ end that creates a disease. N Engl J Med 1991; 324: 84–88. restriction site for AluI in allele 1 of the exon 2 3 Tysk C, Lindberg E, Jarnerot G, Floderus-Myrhed B. Ulcera- polymorphism (AG3CTGG). The disease-associated al- tive colitis and Crohn’s disease in an unselected population of lele 2 itself completes a restriction site for MspI. After 28 monozygotic and dizygotic twins. A study of heritability and cycles of PCR excess 32P-end-labelled reverse primer was the influence of smoking. Gut 1988; 29: 990–996. added and allowed to anneal. Data from optimisation 4 Yang H, McElree C, Roth M-P, Shanahan F, Targan SR, Rotter experiments suggested us to use 28 cycles of PCR to JI. Familial empiric risks for inflammatory bowel disease: ensure suboptimal cycling so that products had not differences between Jews and non-Jews. Gut 1993; 34: 517–524. reached plateau phase (data not shown). The radioactive 5 Hayward PAR, Satsangi J, Jewell DP. Inflammatory bowel disease and the X chromosome. QJMed1996; 89: 713–718. reverse primer (added at the start of the last PCR cycle) 1 6 Hugot J-P, Laurent-Puig P, Gower-Rousseau C et al. Mapping was then extended for 5 min at 72 . Thus, the only of a susceptibility locus for Crohn’s disease on chromosome radioactive products were from the final extension and 16. Nature 1996; 379: 821–823. were not likely to be heteroduplexed and subsequently 7 Satsangi J, Parkes M, Louis E et al. Two-stage genome-wide resistant to restriction digestion. search in inflammatory bowel disease provides evidence for

Genes and Immunity IL-1 receptor antagonist in ulcerative colitis MJ Carter et al 14 susceptibility loci on chromosomes 3, 7 and 12. Nat Genet 1996; (ed). Cytokine Molecular Biology. Oxford University Press: 14: 199–202. Oxford, 2000, pp 21–46. 8 Hampe J, Schreiber S, Shaw SH et al. A genome-wide analysis 27 Cox A, Camp NJ, Nicklin MJH, di Giovine FS, Duff GW. An provides evidence for novel linkages in inflammatory bowel analysis of linkage disequilibrium in the IL-1 gene cluster disease in a large European cohort. Am J Hum Genet 1999; 64: using a novel grouping method for multiallelic markers. Am J 808–816. Hum Genet 1998; 62: 1180–1188. 9 Cho JH, Nicolae DL, Gold LH et al. Identification of novel 28 Mansfield JC, Holden H, Tarlow JK et al. Novel genetic susceptibility loci for inflammatory bowel disease on chromo- association between ulcerative colitis and the anti-inflamma- somes 1p, 3q and 4q: evidence for epistasis between 1p and tory cytokine interleukin-1 receptor antagonist. Gastroenterol- IBD1. Proc Natl Acad Sci 1998; 95: 7502–7507. ogy 1994; 106: 637–642. 10 Cho JH, Brant SR. Genetics and genetic markers in inflam- 29 Louis E, Satsangi J, Roussomoustakaki M et al. Cytokine gene matory bowel disease. Curr Opin Gastroenterol 1998; 14: polymorphisms in inflammatory bowel disease. Gut 1996; 39: 283–288. 705–710. 11 Sartor RB. Cytokines in intestinal inflammation: pathophysio- 30 Bioque G, Bouma G, Crusius JBA et al. Evidence for genetic logical and clinical considerations. Gastroenterology 1994; 106: heterogeneity in IBD: the interleukin-1 receptor antagonist in 533–539. the predisposition to suffer from ulcerative colitis. Eur J 12 Arend WP. Interleukin-1 receptor antagonist. Adv Immunol Gastroenterol Hepatol 1996; 8: 105–110. 1993; 54: 167–227. 31 Hacker UT, Gomolka M, Keller E et al. Lack of association 13 Nicklin M, Weith A, Duff GW. A physical map of the region between an interleukin-1 receptor antagonist gene poly- encompassing the human interleukin-1 alpha, interleukin-1 morphism and ulcerative colitis. Gut 1997; 40: 623–627. beta and interleukin-1 receptor antagonist genes. Genomics 32 Heresbach D, Alizadeh M, Dabadie A et al. Significance of 1994; 19: 382–384. interleukin-1 beta and interleukin-1 receptor antagonist 14 Cominelli F, Nast CC, Clark BD et al. Interleukin-1 (IL-1) gene genetic polymorphism in inflammatory bowel disease. Am J expression, synthesis and effect of specific IL-1 receptor Gastroenterol 1997; 92: 1164–1169. blockade in rabbit immune complex colitis. J Clin Invest 33 Andus T, Caesar I, Vogi D, Scholmerich J, Gross V. Association 1990; 86: 972–980. of HLA-DR15, P-ANCA and IL-1 receptor antagonist allele 2 15 McCall RD, Haskill S, Zimmerman EM, Lund PK, Thompson with ulcerative colitis. Gastroenterology 1995; 108: A770. RC, Sartor RB. Tissue interleukin-1 and interleukin-1 receptor 34 Stokkers PCF, van Aken BE, Basoski N, Reitsma PH, Tytgat antagonist expression in enterocolitis in resistant and suscep- GNJ, van Deventer SJH. Five genetic markers in the tible rats. Gastroenterology 1994; 106: 960–972. interleukin 1 family in relation to inflammatory bowel disease. 16 Ligumsky M, Simon PL, Karmeli F, Rachmilewitz D. Role of Gut 1998; 43: 33–39. interleukin-1 in inflammatory bowel disease—enhanced 35 Vijgen L, Van Gysel M, Rector A et al. Interleukin-1 receptor production during active disease. Gut 1990; 31: 686–689. antagonist VNTR-polymorphism in inflammatory bowel 17 Mahida YR, Wu K, Jewell DP. Enhanced production of disease. Genes Immun 2002; 3: 400–406. interleukin-1 beta by mononuclear cells isolated from the 36 Craggs A, West S, Curtis A et al. Absence of a genetic association mucosa with active ulcerative colitis or Crohn’s Disease. Gut between IL-1RN and IL-1B polymorphisms in ulcerative colitis 1989; 30: 835–838. and Crohn’s disease in multiple populations from northeast 18 Casini-Raggi V, Kam L, Chong YJT, Fiocchi C, Pizarro TT, England. Scand J Gastroenterol 2001; 36: 1173–1178. Cominelli F. Mucosal imbalance of IL-1 and IL-1 receptor 37 Tountas NA, Casini-Raggi V, Yang H et al. Functional antagonist in inflammatory bowel disease: a novel mechanism and ethnic association of allele 2 of the IL-1 receptor of chronic intestinal inflammation. J Immunol 1995; 154: antagonist gene in ulcerative colitis. Gastroenterology 1999; 2434–2440. 117: 806–813. 19 Cominelli F, Nast CC, Duchini A, Lee M. Recombinant 38 Roussomoustakaki M, Satsangi J, Welsh K et al. Genetic interleukin-1 receptor antagonist blocks the proinflammatory markers may predict disease behaviour in patients with activity of endogenous interleukin-1 in rabbit immune colitis. ulcerative colitis. Gastroenterology 1997; 112: 1845–1853. Gastroenterology 1992; 103: 65–71. 39 Bouma G, Crusius JBA, Garcia-Gonzalez MA et al. Genetic 20 Melani L, Hirsch E, Guanzon M, Pizarro TT, Hirsch D, markers in clinically well-defined patients with ulcerative Cominelli F. Deletion of the IL-1 receptor antagonist gene colitis (UC). Clin Exp Immunol 1999; 115: 294–300. increases the susceptibility to experimental colitis in mice 40 Nemetz A, Nosti-Escanilla MP, Molnar T et al. IL1B gene (abstract). Gastroenterology 1997; 112: A1040. polymorphisms influence the course and severity of inflam- 21 Ferretti M, Casini-Raggi V, Pizarro TT, Eisenberg SP, Nast CC, matory bowel disease. Immunogenetics 1999; 49: 527–531. Cominelli F. Neutralization of endogenous IL-1 receptor 41 Nishiyama T, Mitsuyama K, Toyonaga A, Sasaki E, Tanikawa antagonist exacerbates and prolongs inflammation in rabbit K. Colonic mucosal interleukin-1 receptor antagonist in immune colitis. J Clin Invest 1994; 94: 449–453. inflammatory bowel disease. Digestion 1994; 55: 368–373. 22 Westendorp RGJ, Langermans JAM, Huizinga TWJ et al. 42 Dionne S, D’Agata ID, Hiscott J, Vanounou T, Seidman EG. Genetic influence on cytokine production and fatal meningo- Colonic explant production of IL-1 and its receptor antagonist coccal disease. Lancet 1997; 349: 170–173. is imbalanced in inflammatory bowel disease (IBD). Clin Exp 23 Tarlow JK, Blakemore AIF, Lennard A et al. Polymorphism in Immunol 1998; 112: 435–442. human IL-1 receptor antagonist gene intron 2 is caused by 43 Danis VA, Millington M, Hyland VJ, Grennan D. Cytokine variable numbers of an 86-bp tandem repeat. Hum Genet 1993; production by normal human monocytes: inter-subject varia- 91: 403–404. tion and relationship to an UIL-1 receptor antagonist (IL-1Ra) 24 Clay FE, Tarlow JK, Cork MJ, Cox A, Nicklin MJH, Duff GW. gene polymorphism. Clin Exp Immunol 1995; 99: 303–310. Novel interleukin-1 receptor antagonist exon polymorphisms 44 Hurme M, Santtilla S. IL-1 receptor antagonist (IL-1Ra) plasma and their use in allele-specific mRNA assessment. Hum Genet levels are co-ordinately regulated by both IL-1Ra and IL-1b 1996; 97: 723–726. genes. Eur J Immunol 1998; 28: 2598–2602. 25 Carter MJ, di Giovine FS, Jones S et al. Association of 45 Mandrup-Poulsen T, Pociot F, Molvig J et al. Monokine the interleukin-1 receptor antagonist gene with ulcerative antagonism is reduced in patients with IDDM. Diabetes 1994; colitis in Northern European Caucasians. Gut 2001; 48: 43: 1242–1247. 461–467. 46 Cominelli F, Pizarro TT. Interleukin-1 receptor anatagonist: a 26 di Giovine FS, Camp NJ, Cox A et al. Detection and population novel acute phase protein with anti-inflammatory properties. analysis of IL-1 and TNF gene polymorphisms. In: Balkwill F J Clin Invest 1997; 99: 2813.

Genes and Immunity IL-1 receptor antagonist in ulcerative colitis MJ Carter et al 15 47 Andus T, Daig R, Vogl D et al. Imbalance of the interleukin-1 52 Ruddell WSJ, Dickinson RJ, Dixon MF, Axon ATR. Treatment system in colonic mucosa-association with intestinal inflam- of distal ulcerative colitis (proctosigmoiditis) in relapse: mation and interleukin-1 receptor antagonist genotype 2. comparison of hydrocortisone enemas and rectal hydrocorti- Gut 1997; 41: 651–657. sone foam. Gut 1980; 21: 885–889. 48 Santtilla S, Savinainen K, Hurme M. Presence of the IL-1RA 53 Tagariello G, di Giovine FS. Interleukin-1 in haemophilic allele 2 (IL1RNn2) is associated with enhanced IL-1b produc- arthritis. Thromb Haemost 1996; 75: 979–980. tion in vitro. Scand J Immunol 1998; 47: 195–198. 54 Silva MA, Menezes J, Wizman S, Gendron R, Oligny L, 49 Pociot F, Molvig J, Wogensen L, Worsae H, Nerup J. A TaqI Seidman EG. Cytokine tissue levels as markers of disease polymorphism in the human interleukin-1 beta (IL-1 beta) activity in pediatric Crohn’s disease. Pediatr Res 2003; 54: gene correlates with IL-1 beta secretion in vitro. Eur J Clin 1–6. Invest 1992; 22: 396–402. 55 Ahn SJ, Costa J, Emanuel JR. PicoGreen quantitation of DNA: 50 Jenkins D, Balsitis M, Gallivan S et al. Guidelines for the initial effective evaluation of samples pre- or post-PCR. Nucleic Acids biopsy diagnosis of suspected chronic idiopathic inflamma- Res 1996; 24: 2623–2625. tory bowel disease. The British Society of Gstroenterology 56 Uejima H, Lee PW, Cui H, Feinberg AP. Hot-stop PCR: a Initiative. J Clin Pathol 1997; 50: 93–105. simple and general assay for linear quantitation of allele 51 Baron JH, Connell AM, Lennard-Jones JE. Variation between ratios. Nat Genet 2000; 25: 375–376. observers in describing mucosal appearances in proctocolitis. 57 Chomezynsky P, Sacchi N. Acid guanidinium–phenol– Br Med J 1964; 1: 89–92. chloroform method. Ann Biochem 1987; 162: 156–159.

Genes and Immunity