Genes and Immunity (2000) 1, 185–190 2000 Macmillan Publishers Ltd All rights reserved 1466-4879/00 $15.00 www.nature.com/gene Cytokine (TNF␣,LT␣ and IL-10) polymorphisms in inflammatory bowel diseases and normal controls: differential effects on production and allele frequencies K Koss1,2,3, J Satsangi1, GC Fanning3, KI Welsh3 and DP Jewell1 1Gastroenterology Unit, Radcliffe Infirmary, Oxford, UK; 2Gastroenterology Department, Medical University of Gdansk, Poland; 3Transplant Immunology, Nuffield Department of Surgery, Churchill Hospital, Oxford, UK The influence of biallelic polymorphisms in the tumour necrosis factor-alpha (TNF␣), lymphotoxin-alpha (LT␣) and interleukin-10 (IL-10) genes on stimulated TNF␣ and IL-10 production was studied in ulcerative colitis (UC) patients, Crohn’s disease (CD) patients and in healthy controls. A polymerase chain reaction sequence-specific primer (PCR-SSP) system was developed to type nine biallelic polymorphisms, three in each of the TNF␣,LT␣ and IL-10 genes. Production of the TNF␣ and IL-10 was measured by ELISA in lipopolysaccharide (LPS) stimulated whole blood. Four haplotypes of the TNF␣ gene, three haplotypes of LT␣ and three haplotypes of IL-10 were identified. No significant differences in haplotype frequencies were found between patients and controls overall. On subgroup analysis however, haplotype TNF-2 was more frequent in women with extensive colitis compared to distal colitis (31% vs 12%; P = 0.028). This difference was even greater for the combined TNF-2-LT␣-2 haplotype (56% vs 21%; P = 0.0007). The TNF-2 and LT␣-2 haplotypes were associated with higher TNF␣ production in CD patients, and the TNF-4 haplotype was associated with lower TNF␣ production in UC patients. The A allele in the IL-10 promoter region at position −1082 was associated with decreased IL- 10 production in CD patients and controls (P = 0.005, P = 0.015 respectively). These data provide evidence that the effect of TNF␣,LT␣ and IL-10 gene polymorphisms on cytokine production differ in CD, UC patients and controls. Genes and Immunity (2000) 1, 185–190. Keywords: TNF␣;LT␣; IL-10; Crohn’s disease; ulcerative colitis; polymerase chain reaction Introduction important producers of TNF and IL-10 but not LT␣, are known to be in a state of activation. Indeed the pro- ␣ Tumour necrosis factor-alpha (TNF ) is a potent pro- duction of TNF␣ has led to clinical strategies for its block- inflammatory and immunoregulatory cytokine that plays ade by anti-TNF␣ antibodies in Crohn’s disease (CD).8 a key role in the initiation, regulation and perpetuation IL-10 also has proven relevance to IBD since knockout of of the inflammatory response, whereas interleukin-10 (IL- the IL-10 gene induces chronic enterocolitis in mice.9 The 10) inhibits the production of the pro-inflammatory cyto- concentration of IL-10 has been reported to be increased ␣ ␥ kines (IL-1, TNF , IFN- and IL-6) through its action on in patients with both active ulcerative colitis (UC) and 1–3 Th1 cells and macrophages. The close interrelation- active CD when compared to healthy controls.10 This ␣ ships between TNF and IL-10 have been confirmed in study tests the hypothesis that the cytokine profile may 4,5 several studies. Another possible interacting cytokine is be imbalanced in IBD due to both monocyte activation ␣ lymphotoxin-alpha (LT ), which shares the same recep- and genetic polymorphism. tors with TNF␣ and also maps in the HLA class III region on the short arm of chromosome 6. Stable TNF␣ pro- duction, varying less than 15%, has been described, when Results single individuals are sampled over time.6 This makes comparisons between individuals more meaningful. TNF␣ and LT␣ polymorphisms Studies on monozygotic twins and their first-degree rela- Four haplotypes of the three polymorphisms in the TNF␣ tives, using ex vivo endotoxin stimulated whole blood gene and three haplotypes of the three polymorphisms in ␣ samples, provide evidence that 60% of TNF and 75% the LT␣ gene were identified in the examined population of IL-10 variation in production capacity appears to be (Table 1). There were no significant differences in fre- 7 genetically determined. quencies of the biallelic polymorphisms or whole haplo- In inflammatory bowel disease (IBD), monocytes, types between controls, CD and UC patients (Table 2). When subgroups of patients were analysed, haplotype TNF-2 was more frequent in women with extensive com- Correspondence: Dr Konrad Koss, House 29F, New Staff Residence, City pared to distal colitis (31% vs 12%, P = 0.028). When ana- General Hospital, Stoke on Trent ST4 6QG, UK. ␣ E-mail: konradkossȰyahoo.com lysed for the combined haplotype TNF-2-LT -2, the Received 21 June 1999; revised 7 September 1999; accepted 24 Sep- association became stronger (21% vs 56%, P = 0.0007). tember 1999 There was strong linkage disequilibrium between TNF-2 Genetics of TNF, LT and IL-10 in IBD K Koss et al 186 Table 1 Defined haplotypes of TNF␣,LT␣ and IL-10 gene polymorphisms observed in examined population of 435 subjects. Positions of biallelic polymorphisms (+488, −238, −308, +720, +365, +249, −1082, −819, −592) are calculated from the mRNA start codons TNF␣ +488 −238 −308 LT␣ +720 +365 +249 IL-10 −1082 −819 −592 (4538) (AA26) (AspHI) (NcoI) TNF-1 G G G LT␣−1 C C A IL10-GCC G C C TNF-2 G G A LT␣−2 A G G IL10-ACC A C C TNF-3 A G G LT␣-3 C G A IL10-ATA A T A TNF-4 G A G Table 2 TNF␣,LT␣ and IL-10 haplotypes frequencies in Crohn’s effect of steroids was observed in UC patients but the disease (CD), ulcerative colitis (UC) patients and controls. Haplo- difference was not significant, perhaps due to the fact types frequencies were calculated in the group of 236 healthy con- that all patients were receiving 5-ASA. trols, 111 UC, and 91 CD patients. No significant difference in geno- type or haplotype frequencies were observed in those groups examined Effect of CD, UC and smoking on TNF and IL-10 production ␣ Haplotypes Controls UC patients CD patients The LPS-stimulated TNF production was higher in con- (n = 236) (n = 111) (n = 91) trols than in the group of 47 CD or 49 UC patients with- out steroid treatment (Controls: 8958 ± 1978 pg/mL vs TNF-1 0.67 0.72 0.75 CD: 4791 ± 1819 pg/mL, P = 0.0045 or vs UC: 6028 ± 2393 TNF-2 0.2 0.17 0.15 pg/mL, P = 0.015). Also LPS-stimulated IL-10 production TNF-3 0.08 0.06 0.04 was higher in controls than in UC (P = 0.012) or CD TNF-4 0.06 0.05 0.06 patients (Controls: 235 ± 108 pg/mL vs UC: 164 ± 69 LT␣-1 0.38 0.37 0.36 = ± = LT␣-2 0.36 0.41 0.36 pg/mL; P 0.012 or vs CD: 123 43 pg/mL, P 0.023). LT␣-3 0.26 0.23 0.27 Smokers with CD showed higher stimulated IL-10 pro- TNF-2/LT␣-2 0.18 0.16 0.14 tein production in whole blood than non-smokers (208 ± IL10 GCC 0.52 0.55 0.58 76 pg/mL vs 105 ± 34 pg/mL, P = 0.011). This was not IL10 ACC 0.23 0.21 0.20 apparent in healthy controls (281 ± 111 pg/mL vs 220 ± IL10 ATA 0.25 0.24 0.22 78 pg/mL) and UC patients (176 ± 76 pg/mL vs 158 ± 68 pg/mL). There was a tendency towards lower LPS-stimulated TNF␣ production in UC patients with extensive colitis haplotype and LT␣-2 haplotype, which both together compared with distal colitis (4024 ± 1760 vs 6584 ± 1665 form the most frequent haplotype in healthy controls and pg/mL), and a tendency towards lower TNF␣ production IBD patients. in CD patients with colonic involvement compared with ileitis (3784 ± 1671 vs 5340 ± 1891 pg/mL). However, IL-10 polymorphisms these differences did not reach statistical significance. The three biallelic polymorphisms of the IL-10 gene occurred in only three haplotypes (Table 1). No signifi- TNF␣ and LT␣ polymorphisms and TNF␣ production cant differences in allele or haplotypes frequencies were Since steroids have been shown to suppress TNF␣ and observed between the patients and controls (Table 2). IL-10 release from circulating mononuclear cells, func- tional analysis of TNF␣,LT␣ and IL-10 polymorphisms Effect of medication on TNF␣ and IL-10 production were performed in patients who had not received steroid TNF␣ and IL-10 production was measured in lipopoly- treatment for at least a month prior to assay. saccharide (LPS)-stimulated blood taken from a ran- In CD patients, TNF-2 or LT␣-2 haplotypes were asso- domly selected group of 52 controls and 136 IBD patients. ciated with higher TNF␣ production (TNF-2: 6383 ± 828 Large intra-individual differences in LPS-stimulated vs 3890 ± 1544 pg/mL, P = 0.003; LT␣-2: 6039 ± 1322 vs whole blood production of cytokines were observed. No 3487 ± 2142 pg/mL, P = 0.002). However, in controls the effect on the number of white blood cells (WBC) in the LT␣-2 haplotype was associated with lower TNF␣ pro- blood samples and stimulated TNF␣ or IL-10 production duction (8458 ± 1810 vs 10731 ± 2044 pg/mL, P = 0.012). was observed. In UC patients a lower production of TNF␣ was seen in In IBD patients, steroid treatment significantly association with the TNF-4 haplotype (Table 3). decreased LPS-stimulated TNF␣ secretion (1418 ± 1799 vs 5612 ± 1972 pg/mL, P = 0.013). No significant effect on IL-10 polymorphism and IL-10 production LPS-stimulated TNF␣ secretion was observed in patients Due to the stimulating effect of 5-ASA, SLSZ and the sup- treated with 5-aminosalicylic acid (5-ASA) or sulphasala- pressing effect of prednisolone on IL-10 production, data zine.