Is Interleukin-6 Important in Inflammatory Bowel Disease?

Home , Interleukin 6

K Koss1,2,3, J Satsangi1, KI Welsh2 and DP Jewell1 1Gastroenterology Unit, Radcliffe Inﬁrmary, Oxford, UK; 2Transplant Immunology, Nufﬁeld Department of Surgery, Churchill Hospital, Oxford, UK; 3Gastroenterology Department, Medical University of Gdansk, Poland

The IL-6 gene maps to an area of chromosome 7 known to be significant for susceptibility to inflammatory bowel disease. The functional effects of interleukin-6 (IL-6) polymorphisms in the 4th intron and in the 3Ј flanking region of IL-6 gene were studied in 192 inflammatory bowel disease patients and healthy subjects. A polymerase chain reaction with sequence-specific primers (PCR-SSP) was used to determine a G to A polymorphism (* at position 4470 in intron 4 of IL-6 gene). Four alleles in the 3Ј flanking region were studied using a variable number of tandem repeats PCR (VNTR- PCR) amplification. Production of IL-6 was measured in lipopolysaccharide (LPS) stimulated whole blood samples by an enzyme-linked immunosorbent assay (ELISA). A modest increase in the frequency of the IL-6*G allele was noted in Crohn’s disease (CD) patients (50%1) and ulcerative colitis (UC) patients (46.1%) as compared to controls (39.8%1, P1 = 0.025). We were unable to find any significant functional effect of the IL-6 polymorphisms tested on IL-6 protein production. We postulate that the IL-6 polymorphisms investigated here may be in linkage disequilibrium with a susceptibility gene and that they may be utilised as genetic markers. Genes and Immunity (2000) 1, 207–212.

Keywords: genotyping; interleukin-6; Crohn’s disease; ulcerative colitis; polymerase chain reaction

Introduction specific primers (PCR-SSP) system to detect a new biallelic G to A polymorphism in the 4th intron of the IL-6 Results of a genome-wide search in inflammatory bowel gene, which is relevant to the reported MspI restriction disease (IBD) provided evidence that there may be sus- enzyme polymorphism.9 Following evidence from gen- 1 ceptibility genes on chromosomes 3, 7 and 12. One of etic and functional studies on the relevance of IL-6 to IBD the candidate genes lying close to the linked region on we have studied both the distribution of IL-6 polymor- 2 chromosome 7 is the gene encoding interleukin 6, which phisms and their functional effects in IBD patients and is a multifunctional cytokine synthesised by both lymph- healthy controls in an attempt to relate production of IL- oid and non-lymphoid cells. In IBD, the activated mono- 6 to individual polymorphisms as well as to ascertain cytes, macrophages and epithelial cells all produced whether any of these polymorphisms are associated 3 interleukin-6 (IL-6), which has many pertinent biological with IBD. activities including B cell differentiation, stimulation of immunoglobulin secretion, induction of T cell prolifer- ation and acute phase protein production by Results hepatocytes.4 IL-6 is thus a strong candidate for an IBD susceptibility gene. IL-6 allele frequencies The human IL-6 gene is approximately 5 kb in length Genotype and allele frequencies of the IL-6 4470*G/A and consists of five exons and four introns.5 Allelic vari- biallelic polymorphism in the 4th intron and the VNTR ation has been described in the 3Ј flanking region of the polymorphism in 3Ј flanking region of the IL-6 gene were gene due to the insertion of a variable number of AT tan- studied in 88 Crohn’s disease (CD), 104 ulcerative colitis dem repeats (VNTR). This system has been previously (UC) patients and 225 healthy controls. Allele IL-6*G was studied, both by restriction fragment length polymor- found to be more frequent in CD patients (50%) than in phism (RFLP)6 and by amplification of the IL-6 VNTRs healthy controls (39.8%; P = 0.025). No other significant and fractionation on agarose gels.7,8 Here we have difference in genotypes or allele frequencies for examined developed a polymerase chain reaction with sequence- polymorphisms were observed between those groups examined. Although not statistically significant allele IL- 6*G was also more frequent in UC patients when com- Correspondence: Dr Konrad Koss, House 29F, New Staff Residence, City pared to controls (46.1% vs 39.8%) (Table 1). General Hospital, Stoke on Trent, ST4 6QG, UK. Ȱ E-mail: konradkoss yahoo.com IL-6 protein production in subgroups of healthy Konrad Koss was funded by the David and Elizabeth Cross Chari- controls and patients table Trust, Oxford University Medical School and The Royal Society. Since steroids have been shown to suppress the IL-6 10,11 Received 18 June 1999; revised 7 September 1999; accepted 23 Sep- release from circulating mononuclear cells, therefore, tember 1999 IL-6 protein production was only measured in IBD Is IL-6 important in inflammatory bowel disease? K Koss et al 208 Table 1 Genotype and allele frequencies (AF) of biallelic G/A polymorphism in the 4th intron of IL-6 and IL-6 VNTR polymorphism in 88 Crohn’s disease (CD) patients, 104 ulcerative colitis (UC) patients and 225 healthy controls. Allele IL-6*G was more frequent in CD patients than in healthy controls (P = 0.025). No other significant differences in genotypes or allele frequencies were observed between those groups examined

Controls (n = 225) CD patients (n = 88) UC patients (n = 104)

Genotypes (%) AF Genotypes (%) AF Genotypes (%) AF

IL-6*GG 39 (17.3) 0.398 21 (23.9) 0.5 23 (0.221) 0.461 IL-6*GA 101 (44.9) 46 (52.3) 50 (0.481) IL-6*AA 85 (37.8) 0.602 21 (23.9) 0.5 31 (0.298) 0.538 IL-6-VNTR*11 0.0 0.015 0.0 0.0 0.0 0.025 IL-6-VNTR*31 2 (1.0) 0.0 1 (0.01) IL-6-VNTR*33 16 (7.8) 0.287 6 (7.6) 0.282 7 (0.07) 0.248 IL-6-VNTR*41 4 (2) 0.0 4 (0.04) IL-6-VNTR*43 83 (40.7) 32 (41) 35 (0.35) IL-6-VNTR*44 99 (48.5) 0.698 40 (51.3) 0.718 54 (0.53) 0.728

patients that had not received steroid treatment for at subgroups of patients defined by disease activity, need least a month prior to assay. Plasma IL-6 level was meas- for surgery or presence of extraintestinal manifestations. ured in duplicates in 44 healthy controls, 32 CD and 35 UC patients. Liposaccharide (LPS)-stimulated IL-6 pro- IL-6 polymorphisms and IL-6 protein production tein production in whole blood samples was higher in No significant differences in stimulated IL-6 protein pro- healthy controls (41.9 ± 9.7 ng/mL) than in UC patients duction (whether related to the total number of mono- (33.2 ± 11.3 ng/mL; P = 0.01) and CD patients (30.3 ± 11 cytes or not) were observed between individuals with dif- ng/mL; P = 0.02). These proportions were preserved fering biallelic IL-6*G/A or VNTR IL-6 polymorphisms when stimulated IL-6 plasma concentrations were cor- (Table 2). rected for the total number of monocytes. Relationship between IL-6 protein production and Effect of gender: Stimulated blood samples from healthy tumour necrosis factor-alpha (TNF␣) or IL-10 protein males produced more soluble IL-6 protein than from production healthy females (44.3 ± 9.4 ng/mL vs 37.6 ± 8 ng/mL; P A correlation was observed in healthy individuals = 0.019). The same tendency was observed in CD males between stimulated IL-6 and TNF␣ production (r = 0.441, compared to females (35.3 ± 12.4 ng/mL vs 27.3 ± 9.2 P = 0.0021) and also between IL-6 and IL-10 production ng/mL; P = 0.044). However, when the stimulated IL-6 (r = 0.385, P = 0.0082). We have not observed any corre- plasma level was corrected for the total number of mono- lation between IL-6 and IL-10 or TNF␣ production in UC cytes, no significant differences were observed between patients. In CD patients the correlation between stimu- males and females. lated IL-6 and IL-10 was preserved (r = 0.465, P = 0.0146).

Effect of smoking: No differences in stimulated IL-6 pro- Discussion duction were observed between smokers and non- smokers in the group of IBD patients or healthy controls. The present study provides evidence that the IL-6 gene When, however, the stimulated IL-6 plasma level was is involved in determining susceptibility and disease recalculated for the total number of monocytes, smoking was associated with higher IL-6 production in CD patients (P = 0.045). Table 2 Functional effect of biallelic G/A polymorphism in the 4th intron of IL-6 and IL-6 VNTR polymorphism on IL-6 protein production in whole blood samples stimulated with 10 ␮g/mL LPS for Effect of disease activity, localisation, surgery, extraintestinal 24 h. Ulcerative colitis (UC) and Crohn’s disease (CD) patients were manifestation: In Crohn’s disease affecting the colon, not undergoing any steroid treatment for at least a month prior stimulated IL-6 production was higher than when the to assay disease was limited only to the small bowel (37.4 ± 6.6 ng/mL vs 22.3 ± 11 ng/mL; P = 0.049). The higher IL-6 IL-6 secretion (pg/ml) production in these CD patients was also accompanied by higher numbers of monocytes in blood samples (0.59 Healthy controls CD patients UC patients = = = ± 0.14 × 109/L vs 0.49 ± 0.14 × 109/L). When stimulated (n 44) (n 32) (n 35) IL-6 plasma levels were corrected for the total number of IL-6*GG 37.3 ± 5.2 (4) 30.2 ± 8.9 (9) 35.0 ± 11.8 (10) monocytes, there was no association seen with location ± ± ± of disease. In UC patients with proctitis compared to IL-6*GA 43.2 7.1 (17) 30.1 12.7 (16) 35.0 11.9 (12) IL-6*AA 40.9 ± 11.5 (23) 30.9 ± 11.0 (7) 30.3 ± 11.1 (13) extensive disease the number of monocytes in blood IL-6VNTR*33 38.9 ± 5.8 (4) 47.4 (1) 27.1 ± 16.2 (3) 9 samples was lower (0.44 ± 0.12 × 10 /L vs 0.52 ± 0.13 × IL-6VNTR*41 40.9 ± 6.2 (2) — 41.4 (1) 109/L). However, no differences were observed in IL-6 IL-6VNTR*43 42.9 ± 10.1 (21) 30.5 ± 11.8 (13) 31.8 ± 11.5 (14) production according to UC location. No signiﬁcant dif- IL-6VNTR*44 38.4 ± 9.3 (16) 29.5 ± 10.5 (17) 35.1 ± 11.1 (17) ferences in stimulated IL-6 production were observed in

Genes and Immunity Is IL-6 important in inflammatory bowel disease? K Koss et al 209 phenotype in Crohn’s disease patients. A modest increase production was observed. These findings emphasise the in the frequency of the less common IL-6*G allele was complicated balance between cytokines both in controls noted in Crohn’s disease when compared to ulcerative and disease. colitis patients and healthy controls. However, using In conclusion, these preliminary data suggest that there LPS-stimulated whole blood samples from healthy con- may be an association between IBD, especially CD, and trols and IBD patients we were unable to find any effect an IL-6 polymorphism. This awaits confirmation in a of biallelic G to A polymorphism in the 4th intron or larger and independent data set. VNTR polymorphism in the 3Ј flanking region of the IL- 6 gene on IL-6 protein production. We postulate that the Materials and methods IL-6 polymorphisms investigated here may be in linkage disequilibrium with a susceptibility gene and that they Subjects may be utilised as genetic markers in CD. A recent study A group of 104 UC and 88 CD patients with no family has shown that there is a genetically determined differ- history of IBD was randomly selected from Caucasian ence in the degree of the IL-6 response to stressful stimuli adult patients attending the Gastroenterology Clinic at between individuals and that the G/C polymorphism in the John Radcliffe Hospital in Oxford. Clinical data was the 5Ј flanking region may determine production of obtained by questionnaire, personal interviews and plasma IL-6.12 review of case records. The diagnosis of Crohn’s disease Several groups have reported that circulating IL-6 con- or ulcerative colitis was made on the basis of clinical, centrations are higher in Crohn’s disease compared with radiological and histopathological data, according to ulcerative colitis patients or healthy controls.13–16 Other standard criteria. UC disease activity was assessed using studies have shown that IL-6 production by isolated the Sutherland score,30 whereas Crohn’s disease activity monocytes in vitro, both spontaneously and after LPS index (CDAI)31 helped define CD patients. The extent of stimulation, is similar in IBD patients and healthy con- disease was defined using the most resent investigations trols.10,17 When these results were re-analysed accounting (barium enema or colonoscopy) and was classified in UC for the number of monocytes per 10 ml, the stimulated patients as ‘extensive’ when it was proximal to the IL-6 production was increased in CD compared to UC splenic flexure or as ‘distal’ when it was distal to the and healthy controls.17 We were unable to confirm any splenic flexure. Three patients with ‘indeterminate’ colitis of these observations since our results showed mar- were excluded from this study. Demographic data and ginally lower IL-6 production in CD patients. The more clinical characteristics of patients with inflammatory likely explanation is a methodological one. Our use of a bowel disease are summarised in Table 3. Less than 5% whole blood system to study cytokines produced by a of these patients were of Jewish ancestry. variety of cells is possibly the closest to the in vivo situ- ation and the results reflect the biological complexity of Table 3 Clinical characteristics of patients with inflammatory cytokine and cell network interactions. Such network bowel disease: ulcerative colitis (UC) and Crohn’s disease (CD) control may not be exerted in the isolated cell populations studies by others. A further potentially con- CD patients UC patients founding factor is the use of a single stimulus, namely LPS. Although this is a highly relevant stimulus for both Number of patients female 49 54 UC and CD, it is likely that other stimuli, exerting either male 39 50 an up-regulatory or down-regulatory effect on IL-6 production, exist in vivo. Age, years (range) 41.5 (17–72) 51.5 (19–75) We have shown that stimulated IL-6 production in whole blood samples from healthy individuals and CD Localisation ileo-colonic n = 28 distal n = 75 patients is higher in males than in females. This obser- = = vation is in agreement with previous reports showing colonic n 27 extensive n 29 ileal n = 33 that IL-6 secretion by bone marrow cells is suppressed by ovarian hormones (oestrogen).18,19 However, when Activity active n = 24 active n = 28 results were re-expressed in relation to the total number = = of monocytes, the differences between males and females inactive n 64 inactive n 76 were not significant in our study. Medication It is well known that the production of IL-6 is regulated prednisolone 23 15 by a variety of molecules such as IL-1, TNF␣, IL-2, INF␤ azathioprine 8 4 20–25 MTX 5 0 and PDGF, which induce its production. In contrast, 5-ASA 23 36 IL-4, IL-10 and IL-13 inhibit IL-6 production in mono- sulphasalazine 8 39 cytes.26,27 On the other hand IL-6 inhibits IL-1 and TNF␣ and, like IL-1 and TNF␣, it is able to stimulate IL-10 pro- Extraintestinal manifestations 28,29 present 31 18 duction by macrophages in vitro. absent 60 93 We were able to support some of these observations in a whole blood assay. We did find correlations between Smoking IL-6 and IL-10 production and between IL-6 and TNF␣ smokers 34 18 non-smokers 57 93 production following LPS stimulation in healthy individuals. However in UC such correlation was absent sug- n = Number of patients; 5-ASA, 5-aminosalicylic acid; MTX, metho- gesting some specific interference with the cytokine inter- trexate. The following extraintestinal manifestations were observed: actions in the disease. In CD the TNF␣ IL-6 correlation monoarticular arthritis, ankylosing spondylitis, sacroilitis, uveitis, was again absent but correlation between IL-6 and IL-10 erythema nodosum.

Genes and Immunity Is IL-6 important in inflammatory bowel disease? K Koss et al 210 A group of 225 unrelated UK controls: 56 healthy indi- Table 4 Primer sequences used to detect biallelic polymorphism in 4th intron of IL-6 gene by PCR-SSP and VNTR polymorphism in viduals (26 females; 30 males) belonging to the Oxford- Ј shire population and 169 cadaver donors (69 females; 100 3 -flanking region of IL-6 gene. Biallelic polymorphism in 4th intron (position 4470) of IL-6 gene was typed by sequence specific anti- males) from the Oxford Transplant Centre, Churchill sense primers P1 or P2 and sense primer P3. Primers P4 and P5 Hospital served as controls. No two members from one were used to amplify 3Ј-flanking region of IL-6 gene family were included. On the afternoon of blood sam- pling, all healthy individuals had normal oral tempera- Primer Locus in Sequence Working ture and had not undergone intensive physical exercise IL-6 gene conc. in the previous 24 h. P1 4th intron 5Ј-TGTTGAAAGACCACTGATCT 1.5 ␮m Whole blood stimulation and DNA extraction P2 4th intron 5Ј-TGTTGAAAGACCACTGATCC 2.5 ␮m Ј ␮ Four and a half millilitres of peripheral venous blood P3 4th intron 5 -AGGAACAAGCCAGAGCTGTG 1.4 m were collected into siliconised ‘Vacutainer’ tubes contain- P4 3Ј flanking 5Ј-GCAACTTTGAGTGTGTCACG 1.5 ␮m ing 0.105 m sodium citrate. Whole blood was incubated region ° ␮ P5 3Ј flanking 5Ј-GACGTGATGGATGCAACAC 1.7 ␮m at 37 Cin5%CO2 for 24 h with 10 g/ml of LPS (E. coli serotype 055:B5). All samples were gently inverted dur- region ing incubation. Plasma was separated from whole blood cultures following stimulation for 24 h and stored at ␮ −70°C until assayed by ELISA. DNA from the remaining volume of 13 l in 96 well plates under conditions ident- 32 white blood cells was extracted using a conventional pro- ical to those described for HLA Phototyping. Primer ␮ ␮ teinase K (Boehringer Mannheim, Mannheim, Germany) mixes (5 l) overlaid with 10 l of mineral oil were pre- − ° digestion of isolated cells, followed by phenol extraction pared in advance and stored at 20 C until use. The final and ethanol precipitation. concentration of reaction components were as follows: ␮ 200 m of each dNTP, primer mixes, 2 mm MgCl2,67mm Identification of biallelic polymorphism in the 4th Tris Base pH 8.8, 16.6 mm ammonium sulphate, 0.01% ␮ intron (v/v) Tween 20, between 0.01–0.1 g DNA and 0.2 units Using the MspI restriction enzyme, which recognises the Taq polymerase (Advanced Biotechnology, London, UK). sequence 5Ј C/CGG 3Ј, a two-allele polymorphism with Each plate was centrifuged briefly before being placed fragments of either 6.3 kb or 2.6 and 3.9 kb in length has into a PCR machine. PCR amplifications were performed been identified.9 On the basis of a genomic map of the on a MJ Research 96V machines under the following cyc- ° ° IL-6 gene encompassing nucleotides 1 to 5961 (Genbank ling conditions: 1 min at 96 C; 5 cycles of 25 s at 96 C, ° ° ° Nucleotide Query accession Y00081),5 we have identified 45 s at 70 C, 45 s at 72 C; 21 cycles of 25 s at 96 C, 50 s ° ° ° ° five potential MspI cutting sites at positions: 1178/9, at 65 C, 45 s at 72 C; 4 cycles of 25 s at 96 C, 60 s at 55 C, ° 1178/9, 1387/8, 1618/9 and 1893/9 respectively. We have 120 s at 72 C. ␮ assumed that the next polymorphic MspI recognition site The PCR-SSP products (13 l) with loading dye were must be 2.6 or 3.9 kb downstream from the 1893/9 MspI loaded onto a 1% agarose gel containing ethidium bro- × cutting site and that, because the MspI polymorphism is mide in 0.5 TBE and electrophoresed at 4.5 V/cm for not detected by other enzymes, it is likely to be due to a 30 min. site-specific base pair substitution. Using the published The PCR-VNTR products were loaded into a 2% aga- nucleotide sequence of the IL-6 gene, and assuming that rose gel and electrophoresed at 100 V/cm for 3–4 h. Gels it contains one biallelic mutation of a MspI recognition were photographed (b/w) under UV light (320 nm) site, we have searched for possible sequence candidates (Figures 1 and 2). Gels with PCR-SSP products were in the regions 2.6 ± 0.1 kb and 3.9 ± 0.1 kb, and found three potential mutation sites CCAG (4450–3), CCAG (4468–71), CTGG (4516–20) in the first region and one CCAG (5795–9) in the second region. All four candidates were tested by PCR-SSP (as described below) using 10 different DNA samples. One (position 4470) was polymorphic and consisted of the expected A/G substitution.

Typing biallelic and VNTR IL-6 polymorphisms Determination of the 4470 A to G polymorphism was performed by two PCR-SSP reactions. In the first reaction, designed to detect the A allele at position 4470, we used a primer mix of antisense primer P1 (specific to A) and sense primer P3 together with control primers in the working concentrations as shown in Table 4. In the second reaction, designed to detect the G allele at the same position, primer P1 was replaced by primer P2. The control primers amplify a 256 bp amplicon and were derived from exon 15 of the adenomatous polyposis gene.32 The amplification of the 3Ј flanking region containing the Figure 1 The PCR products of variable number of AT tandem IL-6 VNTR polymorphism was performed using a primer repeat (VNTR) in the IL-6 3Ј flanking region. Band of 610 bp rep- 8 mix P4 and P5 (Table 4) as described by Titenko et al. resents IL-6 VNTR*4, 640 bp band represents IL-6 VNTR*3 and PCR-SSP and PCR-VNTR were carried out in a final 760 bp band represents IL-6 VNTR*1.

Genes and Immunity Is IL-6 important in inflammatory bowel disease? K Koss et al 211 6 Bowcock AM, Kidd JR, Lathrop GM et al. The human “interferon-␤2/hepatocyte stimulating factor / interleukine-6” gene: DNA polymorphism studies and localisation to chromosome 7p21. Genomics 1988; 3: 8–16. 7 Bowcock AM, Ray A, Erlich H, Sehgal PB. Rapid detection and sequencing of alleles in the 3Ј flanking region of interleukin-6 gene. Nucleic Acids Res 1989; 17: 6855–6864. 8 Titenko NV, Kurth JH, Bowcock AM, Cavalli-Sforza LL. Poly- merase chain reaction polymorphisms in HLA-DQ alpha and IL-6 from Mongoloid and Caucasoid populations. Biomedical Science 1991; 2: 175–179. 9 Fugger L, Morling N, Bendtzen K et al. MspI polymorphism in the human interleukin 6 (IL-6) gene. Nucleic Acids Res 1989; 17: 4419. 10 Andus T, Gross V, Caesar I. Activation of monocytes during inflammatory bowel disease. Pathobiology 1991; 59: 166–170. 11 Bruin KF, Hommes DW, Jansen J et al. Modulation of cytokine release from human monocytes by drugs used in the therapy of inflammatory bowel diseases. Eur J Gastroenterol Hepatol 1995; 7: 791–795. 12 Fishman D, Faulds G, Jeffery R et al. The effect of novel polymorphisms in the interleukine-6 (IL-6) gene on IL-6 transcrip- Figure 2 The PCR-SSP products of biallelic G to A polymorphism tion and plasma IL-6 levels, and an association with systemic- in the 4th intron of the IL-6 gene, which is relevant to the reported onset juvenile chronic arthritis. J Clin Invest 1998; 102: 1369–1376. MspI restriction enzyme polymorphism. 13 Mahida YR, Kurlac L, Gallagher A et al. High circulating concentrations of interleukin-6 in active Crohn’s disease but not ulcerative colitis. GUT 1991; 32: 1531–1534. scored for the presence or absence of an allele-specific 14 Duclos B, Reimund JM, Lehr L et al. Interleukin 1 (IL1), 6 (IL-6) band (921 bp) when the PCR control amplicon (256 bp) and tumor necrosis factor (TNF) serum levels in Crohn’s disease was present. (CD). Gastroenterology 1991; 100: A576. 15 Gross V, Andus T, Caesar I et al. Evidence for continuous stimu- Measurement of IL-6, TNF␣ and IL-10 secretion lation of interleukin-6 production in Crohn’s disease. Gastro- Plasma soluble IL-6, IL-10 and TNF␣ were measured by enterology 1992; 102: 514–519. R&D Systems ELISA kits Quantokine IL-6 (range of 16 Lobo AJ, Jones SC, Evans SW et al. Plasma interleukin-6 in detection 3.12–300 pg/mL), Quantokine IL-10 (range of inflammatory bowel disease. GUT 1990; 31: A1194. detection 7.8–500 pg/mL) and Quantokine TNF␣ (range 17 Mazlam MZ, Hodgson HJF. Interrelations between interleukin- ␤ of detection 15.6–1000 pg/mL) as described by the manu- 6, interleukin-1 , plasma C-reactive protein values, and in vitro facturer. Plasma samples were diluted 1:200 for IL-6, 1:40 C-reactive protein generation in patients with inflammatory for TNF␣ and 1:20 for IL-10 with Quantokine Calibrator bowel disease. GUT 1994; 35: 77–83. 18 Passeri G, Girasole G, Jilka RL, Manolagas SC. Increased Dilutents. All samples were analysed in duplicate. interleukin-6 production by murine bone marrow and bone cells after estrogen withdrawal. Endocrinology 1993; 133: 822–828. Statistical analysis 19 Bismar H, Diel I, Ziegler R, Pfeilschifter J. Increased cytokine Results of IL-6 production are expressed as median with secretion by human bone marrow cells after menopause or dis- semi-interquartile range. Differences in cytokine continuation of estrogen replacement. J Clin Endocrinol Metab production between groups were examined using the 1995; 80: 3351–3355. Kruskal–Wallis test. Fisher’s exact test was performed for 20 Content J, De Wit L, Poupart P et al. Induction of a 26 kD protein the comparison of genotypes or allele frequencies. Spear- mRNA in human cells treated with an interleukin-1 related, leu- man’s test was used to calculate correlation coefficients kocyte-derived factor. Eur J Biochem 1985; 152: 253–257. between IL-6 and IL-10 or TNF␣ production. A P value 21 May LT, Helfgott DC, Sehgal PB. Anti-beta-interferon antibodies less than 0.05 was considered statistically significant. inhibit the increased expression of HLA-B7 mRNA in tumor necrosis factor-treated human fibroblasts: structural studies of the beta 2 interferon involved. Proc Natl Acad Sci USA 1986; 83: References 8957–8961. 22 Wong GHW, Goeddel DV. Tumor necrosis factor alpha and beta 1 Satsangi J, Parkes M, Louis E et al. Two stage genome-wide inhibit virus replication and synergize with interferons. Nature search in inflammatory bowel disease provides evidence for sus- 323 ceptibility loci on chromosome 3, 7, 12. Nat Genet 1996; 14: 1986; : 819–822. 199–202. 23 Zilberstein A, Ruggieri R, Korn JH, Revel M. Structure and 2 Ferguson-Smith AC, Chen Y-F, Newman MS et al. Regional expression of cDNA and genes for human interferon beta 2 a localisation of the interferon ␤2/B- cell stimulatory factor 2/ distinct species indictable by growth-stimulatory cytokines. hepatocyte stimulating factor gene to human chromosome 7p15- EMBO J 1986; 5: 2529–2537. p21. Genomics 1988; 2: 203–208. 24 Kohase M, May LT, Tamm I et al. A cytokine network in human 3 Kusugami K, Fukatsu A, Tanimoto M et al. Elevation of interleu- diploid fibroblasts: interactions of beta-interferons, tumor kin-6 in inflammatory bowel disease is macrophage- and epi- necrosis factor, platelet-derived growth factor, and interleukine- thelial cell- dependent. Dig Dis Sci 1995; 40: 949–959. 1. Mol Cell Biol 1987; 7: 273–280. 4 Hirano T. IL-6 and its relation to inflammation and disease. Clin 25 Chomarat P, Rissoan MC, Pin JJ et al. Contribution of IL-1, CD14, Immunol Immunophathol 1992; 62: 60–65. and CD13 in the increased IL-6 production induced by in vitro 5 Yasukawa K, Hirano T, Watanabe Y et al. Structure and monocyte-synoviocyte interactions. J Immunol 1995; 155: 3645– expression of human B cell stimulatory factor-2 (BSF-2/IL-6) 3652. gene. EMBO J 1987; 6: 2939–2945. 26 Gibbons R, Martinez O, Matli M et al. Recombinant IL-4 inhibits

Genes and Immunity Is IL-6 important in inflammatory bowel disease? K Koss et al 212 IL-6 synthesis by adherent peripheral blood cells in vitro. Lym- 30 Sutherland LR, Martin F, Greer S et al. 5-Aminosalicylic acid phokine Res 1990: 9: 283–293. enema in the treatment of distal ulcerative colitis, proctosigmo- 27 te-Velde AA, Huijbens RJ, Heije K et al. Interleukine-4 (IL-4) iditis, and proctitis. Gastroenterology 1987; 92: 1894–1898. inhibits secretion of IL-1 beta, tumor necrosis factor alpha, and 31 Best WR, Becktel JM, Singelton JW et al. Development of Crohn’s IL-6 by human monocytes. Blood 1990; 76: 1392–1397. disease activity index: national co-operative Crohn’s disease 28 Minty A, Chalon P, Derocq JM et al. Interleukine 13 is a new study. Gastroenterology 1976: 70: 439–444. human lymphokine regulating inflammatory and immune 32 Bunce M, O’Neill CM, Barnardo MCNM et al. Phototyping: responses. Nature 1993; 362: 248–250. Comprehensive DNA typing for HLA-A, B, C, DRB1, DRB3, 29 Takeshita S, Gage JR, Kishimoto T et al. Differential regulation DRB4, DRB5 and DQB1 by PCR with 144 primer mixes utilizing of IL-6 gene transcription and expression by IL-4 and Il-10 in sequence-specific primers (PCR-SSP). Tissue Antigens 1995; 46: human monocytic cell lines. J Immunol 1996; 156: 2591–2598. 355–367.

Genes and Immunity