sign in sign up
<<
Home , Immunohistochemistry

Gut 1993; 34:1705-1709 1705

Location of tumour factor a by immunohistochemistry in chronic inflammatory Gut: first published as 10.1136/gut.34.12.1705 on 1 December 1993. Downloaded from bowel

S H Murch, C P Braegger, J A Walker-Smith, TT MacDonald

Abstract clinical benefit, and it is therefore important that This study determined the location and the extent of local production is assessed more density of cells immunoreactive for tumour directly. In this study we have used immuno- necrosis factor a (TNF a) in intestinal speci- histochemistry to determine the distribution and mens from 24 patients with chronic inflamma- density ofTNF a containing cells, in the mucosa tory bowel disease (15 with Crohn's disease, and submucosa of patients with chronic inflam- nine with ulcerative colitis) and 11 controls. matory bowel disease. There was significantly increased density of TNF a immunoreactive cells in the lamina propria of both ulcerative colitis and Crohn's Methods disease specimens, although the distribution ofthese cells differed in the two conditions. In PATIENTS ulcerative colitis most of the TNF a immuno- Fifteen surgically resected specimens from reactivity was seen in the subepithelial macro- patients with Crohn's disease were studied. Four phages, with comparatively less in the deep specimens were from the jejunum and 11 from lamina propria, while in Crohn's disease terminal ileum or ascending colon. Twelve of immunoreactive cells were distributed evenly these patients were receiving corticosteroid throughout the lamina propria. Increased sub- treatment and three elemental diet at the time of mucosal immunoreactivity was found only in surgery. Nine further specimens, including two Crohn's disease, in which TNF a positive colonoscopic biopsy specimens, were obtained tended to cluster around from patients with ulcerative colitis, all ofwhom arterioles and venules, often infiltrating and were receiving steroid treatment. The diagnosis disrupting vascular endothelium. It is sug- of Crohn's disease or ulcerative colitis had pre- gested that this degree of TNF a production viously been made on the basis of repeated http://gut.bmj.com/ probably contributes significantly to the patho- colonoscopy, histological tests, and contrast genesis of both Crohn's disease and ulcerative radiography, and all specimens showed histo- colitis, by impairing the integrity of epithelial logical evidence of . The Table and endothelial membranes, increasing inflam- gives the clinical details of these patients and matory cell recruitment, and by prothrombotic treatments received. Six of 11 control specimens effects on the vascular endothelium. were resected from adults (aged 18-64), whose (Gut 1993; 34: 1705-1709) on September 24, 2021 by guest. Protected copyright. Clinical details ofpatients studied indices Treatment Activated macrophages are prominent in the Activity inflammatory infiltrate in both ulcerative colitis ESR Steroid (mm in CRP (mgl and Crohn's disease.'-' There is known to be Age Sex first hour) (mgldl) kg) ASA Location local production of the derived Crohn's disease cytokines interleukin-1 (IL-1) and tumour 1 11 F 35 45 E - Colon necrosis factor a (TNF a),"'0 in inflammatory 2 11 M 12 30 0-2 + Colon 3 13 F 18 26 0-6 + Colon bowel disease, and these may be detected at 4 11 M 70 115 E - Ileum raised concentrations in the serum in active 5 14 F 45 35 1-6 + Colon 6 14 M 26 34 0 4 + Colon disease." Serum concentrations of interleukin-6 7 14 M 23 32 0-6 - Jejunum are also raised in active Crohn's disease but 8 16 F 20 6 E - Jejunum 9 14 M 20 5 0-8 - Jejunum surprisingly not in ulcerative colitis.'2 The 10 17 M 35 40 E - Jejunum importance of macrophage products as agents of 11 9 F 42 28 0 3 - Colon 12 16 M 21 36 10 + Ileum intestinal inflammation has been shown in 13 13 F 22 14 0 75 + Colon Department of Paediatric animal studies by the induction of intestinal 14 14 M 24 80 0 7 + Ileum necrosis TNF a and factor'3 15 13 M 35 42 10 + Ileum Gastroenterology, by platelet activating Ulcerative colitis St Bartholomew's and the protective effect of IL-I blockade in 1 2 F 18 14 1-2 + Colon Hospital, London immune colitis. 14 In studies of 2 13 M 35 10 0-2 - Colon S H Murch complex previous 3 11 F 45 40 0 3 - Colon C P Braegger the possible role of TNF a in inflammatory 4 4 M 30 23 10 + Colon J A Walker-Smith bowel disease we have shown that much higher 5 11 M 56 56 0-6 + Colon T T MacDonald 6 11 F 25 5 0-2 + Rectum concentrations are found in the stools during 7 13 M 49 23 1-5 + Colon Correspondence to: to 5-10 TNF than in serum, 8 10 F 10 5 0 4 + Colon Dr S Murch, Department of relapse'5 (up [ig a/g) 9 9 F 40 18 0 5 + Colon Paediatric Gastroenterology, thus implying high concentrations in inflamed St Bartholomew's Hospital, mucosa. If such high concentration mucosal E= Elemental diet; steroid-prednisolone dose (mg/kg/day); London EClA 7BE. ASA= treatment with aminosalicylic acid (sulphasalazine or Accepted for publication production does occur, it is probable that specific mesalazine); ESR=etythrocyte sedimentation rate; 22 April 1993 treatment to reduce TNF a production will be of CRP=C reactive . 1706 Murch, Braegger, Walker-Smith, MacDonald

indications for surgery included polyposis, heterogeneous mucosal , with pre- chronic constipation, and (two speci- dominant subepithelial immunoreactivity, we

mens were ofterminal ileum, the rest colonic; all subdivided the lamina propria arbitrarily into Gut: first published as 10.1136/gut.34.12.1705 on 1 December 1993. Downloaded from were histologically normal). The other five were subepithelial (within one high power field of the histologically normal colonoscopic biopsy speci- epithelium) and deep (within one high power mens taken from children investigated for field of the muscularis mucosa). We excluded abdominal pain. All biopsy specimens included graticule points overlying the crypts from the muscularis mucosae; however only one con- initial count and subtracted the density of endo- tained sufficient submucosa for analysis of sub- genously positive cells found in the matched mucosal staining. Specimens were frozen in control slides (median 29 cells/mm2 (range 4-78) liquid nitrogen and stored at -70°C until analy- in disease specimens; median 3-5 cells/mm2 sis, when sections cut at 7 [tm were applied to (range 0-18) in controls). Reproducibility of poly-l-lysine coated microscope slides and fixed counts on blinded testing was within 10%. in acetone before .

STATISTICAL ANALYSIS USED As the resultant percentages within groups were CB0006, an IgGl x purified mouse monoclonal not always normally distributed, group medians to human TNF a (a kind gift from were calculated with 95% confidence intervals Celltech Ltd, UK) was used throughout (derived by the Wilcoxon method) and compari- (dilution 1/100). This antibody, which neutral- son was made between groups using the two ises TNF a cytotoxicity on L929 cells and blocks tailed Mann-Whitney U test. TNF a induced MHC class I expression has been given therapeutically to humans suffering from septic shock,'6 and is used to TNF a by Results linked immunosorbent assay (ELISA).'7 It displays no cross reactivity with TNF P, LAMINA PROPRIA interleukin-1 or interleukin-6 in ELISA or cold TNF a+ cells were present in the lamina propria competition assay.'4 Monoclonal antibodies to of six of the 11 normal controls, but these were HLA-DR (1/100, Dako Ltd, Slough), CD25 few in number and largely confined to the (1/100 Dako), and CD3 (UCHT1, an undiluted subepithelial zone (median density 5 5 cells/ supernatant) were used to stain serial sections. mm2, 95% confidence intervals 1-6 to 20 9). Control slides were stained with either TRIS Subepithelial TNF a+ cell density was signific- buffered saline or an isotype matched mono- antly increased compared with controls in both clonal antibody (anti-, 1/100, Dako). ulcerative colitis (145-0 cells/mm2, 110-3 to Figure 1: Density oftumour After incubation with primary antibody for 184-5, p<0-0001) and Crohn's disease (148-0 http://gut.bmj.com/ necrosisfactor a (TNF a) one hour at room temperature, staining was cells/mm2, 1 18-9 to 162, p<0 0001) (see Figs 1A, immunoreactive cells, as developed using the anti- 2). These large cells had the phenotypic appear- determined by point counting with a Lennox graticule, in alkaline phosphatase (APAAP) technique with ance of macrophages, often showing extending tissuefrom normal controls fast red substrate, as previously described.'8 processes (Fig 2), and similar cells on serial andpatients with Crohn's sections stained HLA-DR+, CD25+, CD3-. disease or ulcerative colitis (UC) in (A) subepithelial In addition the showed lamina propria, (B) deep QUANTITATION weak cross reactivity with epithelium. on September 24, 2021 by guest. Protected copyright. lamina propria, and (C) The density of TNF a+ cells within lamina The pattern of staining differed in the deep submucosa. Thefilled propria and submucosa was determined with squares were biopsy a lamina propria, where the density of TNF a+ specimens and the open point counting technique, using a Lennox cells was significantly greater in Crohn's disease squares surgically resected. graticule over at least five randomly chosen high (118 2 cells/mm2, 85 I to 159) than in ulcerative Among the Crohn's disease power fields within each region and counting as specimens, the open circles colitis (28-8 cells/mm2, 10-9 to 64-3, p<0 0005) representjejunal tissue, the positive graticule points overlying positively or controls (1-9 cells/mm2, 0 to 3-8, p<0 0001 - diamonds terminal ileum, stained cells. 19 As many specimens showed Fig IB). TNF a+ cells were still however higher and the squares ascending colon.

300 300 r- 250FA FB 0 C 0 0 0 250 - 250 - 200 - 0 N 200 H E 0 oD 200 F- oRorsi 0 0 150 - 00 0 oo 0 0 C.) 150 F- op 150 F U 0 U 0 0 0 +H 100 _ 00 1- 0 100 F- 0* 0 100 H 0 z 0 00 0 oo 50 _ 00 0 0 0 50 I- 0 0 50 0 8 0 0so 0' 0L Controls Crohn's disease UC Controls Crohn's disease UC Controls Crohn's disease UC (n = 12) (n = 13) (n = 9) (n = 9) (n = 14) (n = 9) (n = 7) (n = 14) (n = 7) Location oftumour necrosisfactor a by immunohistochemistry in chronic inflammatory boweldisease 1707 Gut: first published as 10.1136/gut.34.12.1705 on 1 December 1993. Downloaded from

t I Figure 2: Tumour necrosisfactor (TNF) a immunoreactive 4 cells in chronic inflammatory bowel disease. Top left - lamina http://gut.bmj.com/ propria in Crohn's disease, showing a moderately dense infiltrate ofred stained TNF a+ cells (magnification x200); middle left - subepithelial TNF a+ macrophages, characteristic large cells with processes, in ulcerative colitis (magnification x400). The basal epithelial staining is a result ofweak cross reactivity (also seen in the crypt in top leftpanel; top right- a dense infiltrate ofTNF a+ cells with macrophage morphology in Crohn's disease submucosa (magnification x 100); middle right - a higher power view ofsubmucosal on September 24, 2021 by guest. Protected copyright. TNF a+ macrophages in Crohn's disease (magnification x400); bottom left - vasculopathy in Crohn's disease submucosa. The small arteriole in the centre is surrounded by TNF a+ cells (magnification x400).

in ulcerative colitis than controls (p<0005). The was found in Crohn's disease but not in ulcera- morphology of TNF a+ cells was more variable tive colitis. TNF a+ cell density was signific- within the deeper lamina propria, particularly in antly increased in Crohn's tissue (median density Crohn's tissue (Fig 2). In addition to those with 137-0 cells/mm2, 95% confidence intervals 946 macrophage morphology, a population of to 190-2) compared with both controls (2-30 smaller and more densely stained cells, could be cells/mm2, 0 to 15-5, p<0 0002) and ulcerative identified. These cells were probably lympho- colitis specimens (7 0 cells/mm2, 2 1 to 14-7, cytes, as similar cells in serial sections were p<0 0002) (Figs IC, 2). All of the Crohn's found to be CD3+, CD25+. In addition smaller disease specimens showed evidence of trans- numbers of TNF a+ cells were identified on mural TNF a immunoreactivity, with increased morphological grounds as polymorphonuclear numbers ofTNF a+ macrophages also found in leucocytes. TNF a+ cells were also seen in muscularis externa and serosa. Crohn's granulomas and in the germinal centres There was evidence in all Crohn's specimens of lymphoid follicles: the second were large of accumulation of TNF a+ cells, pre- CD3- cells and were presumably tingible body dominantly CD3- cells with macrophage macrophages. No evidence of TNF a immuno- morphology, around submucosal and serosal reactivity was found elsewhere within the blood vessels (Fig 2). This perivascular TNF a+ follicles. accumulation was frequently associated with disruption of vascular smooth muscle and endo- thelium, and thrombosed or recanalised arteries SUBMUCOSA were seen in eight of 15 specimens. Such peri- Increased submucosal TNF a immunoreactivity vascular concentration of TNF a+ cells was not 1708 Murch, Braegger, Walker-Smith, MacDonald

seen in ulcerative colitis or in Crohn's lamina Although TNF a infusion causes general activa- propria. tion of procoagulant pathways,42 its action seems

preferentially targeted towards dividing endo- Gut: first published as 10.1136/gut.34.12.1705 on 1 December 1993. Downloaded from thelial cells - that is, at sites of injury, where Discussion much lower concentrations will induce this We have found TNF a immunoreactive cells at spectrum of prothrombotic change.43 There is high density within the lamina propria in both now substantial evidence that vascular disrup- Crohn's disease and ulcerative colitis and also tion and focal thrombosis, related to macrophage within Crohn's submucosa. The extent of this infiltration, occurs in Crohn's disease.' The TNF a immunoreactivity confirms that the very production of both the procoagulant TNF a and high concentrations we have detected in stools the potent vasoconstrictor endothelin-145 by during relapse of inflammatory bowel disease perivascular macrophages will contribute power- give a more accurate estimate of local inflamma- fully towards thrombosis in this situation. tion than does the modest increase in serum In view of the high density of TNF a produc- TNF a. Although the difficulty ofobtaining true ing cells in active inflammatory bowel disease, serial sections in frozen tissue prevents formal and the well characterised effects of this cyto- confirmation of the lineage of these TNF a+ kine, it is probable that TNF a makes an cells, we have found similar patterns ofimmuno- independent contribution to inflammatory tissue reactivity in preliminary work with a second damage. If this can be confirmed in animal monoclonal antibody 52 B 83 (also from models ofintestinal inflammation, clinical use of Celltech), and have confirmed, with a reverse anti-TNF a monoclonal antibodies or thera- haemolytic plaque assay, that most mucosal peutic inhibition ofits production by agents such TNF a secreting cells are macrophages (E J as transforming growth factor I,46 interleukin- Breese, T T MacDonald, submitted data). We 1047 or thalidomide48-50 may provide potent and suggest that, whatever the lineage of the TNF a selective alternatives to current immunosuppres- producing cells, the effects of such high local sive treatment. concentrations are likely to be adverse. TNF a produced by subepithelial cells prob- Simon Murch is supported by Action Research, Christian Braegger by the Swiss National Fund, and Thomas MacDonald by ably impairs epithelial integrity,20 particularly in the Wellcome Trust and the Crohn's in Childhood Research the presence of interferon y.2122 Interferon y may Association. We would like to thank Dr Sue Stephens for her advice in the preparation of this manuscript and Mr William also synergise with TNF a in several proinflam- Shand for his assistance in the collection ofthe surgical matory ways, increasing expression of adhesion specimens. molecules23 and MHC class I and II,2324 and 1 Selby WS, Poulter LW, Hobbs S, Jewell DP, Janossy G. upregulating macrophage free radical produc- Heterogeneity of HLA-DR-positive histiocytes in human intestinal lamina propria: a combined histochemical and tion.25 immunohistochemical analysis. J Clin Pathol 1983; 36: 379-

There is clearly the potential within the 84. http://gut.bmj.com/ 2 Allison MC, Cornwall S, Poulter LW, Dhillon AP, Pounder intestine for direct activation of subepithelial RE. Macrophage heterogeneity in normal colonic mucosa macrophages by lipopolysaccharides from the and in inflammatory bowel disease. Gut 1988; 29: 1531-8. 3 Mahida YR, Patel S, Wu K, Jewell DP. Interleukin 2 receptor luminal flora, particularly after any breach of expression by macrophages in chronic inflammatory bowel epithelial integrity. In Crohn's disease epithelial disease. ClGn Exp Immunol 1988; 74: 382-6. 4 Mahida YR, Wu KC, Jewell DP. Respiratory burst activity of permeability is increased,26 and it is noteworthy intestinal macrophages in normal and inflammatory bowel that faecal diversion may be protective.27 The disease. Gut 1989; 30: 1362-70. 5 Choy MY, Walker-Smith JA, Williams CB, MacDonald TT. subepithelial predominance of TNF a immuno- Differential expression of CD25 (interleukin-2 receptor) on on September 24, 2021 by guest. Protected copyright. reactivity in ulcerative colitis is additionally lamina propria T cells and macrophages in the intestinal lesions in Crohn's disease and ulcerative colitis. Gut 1990; concordant with findings ofepithelial deposition 31: 1365-70. of immunoglobulin and complement,6 now 6 Halstensen TS, Mollnes TE, Garred P, Fausa 0, Brandtzaeg P. Epithelial deposition of immunoglobulin GI and known to be related to expression of the 40 kD activated complement (C3b and terminal complement putative autoantigen,28 as well as the sub- complex) in ulcerative colitis. Gastroenterology 1990; 98: 1264-71. epithelial localisation of procollagen RNA pro- 7 Malizia G, Calabrese A, Cottone M, Raimondo M, duction.29 Trejdosiewicz LK, Smart CJ, et al. Expression of leukocyte adhesion molecules by mucosal mononuclear phagocytes in The perivascular production of TNF a in inflammatory bowel disease. Gastroenterology 1991; 100: Crohn's disease, also concordant with known 150-9. 8 Pullman WE, Elsbury S, Kobayashi M, Hapel AJ, Doe WF. patterns of complement deposition,30 is likely Enhanced mucosal cytokine production in inflammatory to be particularly damaging. It binds to high bowel disease. Gastroenterology 1992; 102: 529-37. 9 MacDonald TT, Hutchings P, Choy M-Y, Murch S, Cooke A. affinity endothelial cell surface receptors,3' and Tumour necrosis factor-alpha and interferon-gamma pro- acts to increase inflammatory cell recruitment by duction measured at the single cell level in normal and inflamed human intestine. Clin Exp Immunol 1990; 81: upregulating expression of both adhesion 301-5. molecules32 and the chemotactic cytokine 10 Cappello M, Keshav S, Prince C, Jewell DP, Gordon S. Detection ofmRNAa for macrophage products in inflamma- interleukin-8.3 The vascular effects of TNF a tory bowel disease by in situ hybridisation. Gut 1992; 33: include increase in permeability2034 and powerful 1214-9. 11 Murch SH, Lamkin VA, Savage MO, Walker-Smith JA, promotion of thrombosis35 by reduction of anti- MacDonald TT. Serum concentrations of tumour necrosis coagulant and activation of procoagulant mecha- factor a in childhood chronic inflammatory bowel disease. Gutl1991; 32: 913-7. nisms. It reduces the production of thrombo- 12 Gross V, Andus T, Caesar I, Roth M, Schmolmerich J. modulin,36 a natural anti-coagulant, which Evidence for continuous stimulation of interleukin-6 pro- inactivates factor Va3537 and increases synthesis duction in Crohn's disease. Gastroenterology 1992; 102: of plasminogen activator inhibitor-1.38 TNF a 13 Sun X-M, Hsueh W. Bowel necrosis mediated by tumour necrosis factor is mediated by platelet-activating factor. additionally induces expression of the power- J Clin Invest 1988;- 81: 1328-31. fully procoagulant tissue factor,39 released only at 14 Cominelli F, Nast CC, Duchini A, Lee M. Recombinant interleukin-l receptor antagonist blocks the proinflamma- sites of endothelial damage4" where it activates tory activity of endogenous interleukin-l1 in rabbit immune the extrinsic limb of the coagulation cascade.4' colitis. Gastroenterology 1992; 103: 65-71. Location oftumournecrosisfactor a by immunohistochemistry in chronic inflammatory boweldisease 1709

15 Braegger CP, Nicholls S, Murch SH, Stephens S, MacDonald expression ofa neutrophil chemotactic factor by TNFa, LPS TT. Stool TNF as a marker of intestinal inflammation. and IL-1fi. Science 1989; 243: 1467. Lancet 1992; 339: 89-91. 34 Brett J, Gerlach H, Nawroth P, Steinberg S, Godman G, Stern

16 Exley AR, Cohen J, Buurman W, Owen R, Hanson G, Lumley D. Tumor necrosis factor/cachectin increases permeability Gut: first published as 10.1136/gut.34.12.1705 on 1 December 1993. Downloaded from J, et al. Monoclonal antibody to TNF in severe septic shock. of endothelial cell monolayers by a mechanism involving Lancet 1990; 335: 1275-6. regulatory G .JExp Med 1989; 169: 1977-91. 17 Engelberts I, Moller A, Schoen GJM, van der Linden CJ, 35 Clauss M, Ryan J, Stern D. Modulation of endothelial cell Buurman WA. Evaluation of measurement of human TNF hemostatic properties by TNF: insights into the role of in plasma by ELISA. Lymnphokine Cytokine Res 1991; 10: 69- endothelium in the host response to inflammatory stimuli. 76. In: Beutler B, ed. Tumor necrosis factors: the molecules and 18 Isaacson PG, Wright DH. Imunocytochemistry of lympho- their emerging role in . New York: Raven Press, 1992: recticular tumours. . In: Polak JM, 49-63. van Noorden S, eds. Practical applications in and 36 Moore K, Esmon C, Esmon N. Tumor necrosis factor leads to biology. Bristol: John Wright, 1983: 249. the internalization and degradation of thrombomodulin 19 Lennox B. Observations on the accuracy of point counting from the surface of bovine aortic endothelial cells in culture. including the description of a new graticule. J Clin Pathol Blood 1989; 73: 159-65. 1975; 28: 99-103. 37 Nawroth P, Stern D. Modulation of endothelial cell haemo- 20 Mullin JM, Snock KV. Effect of tumor necrosis factor on static properties by tumor necrosis factor. J Exp Med 1986; epithelial tight junctions and transepithelial permeability. 163: 740-5. CancerRes 1990; 50: 2172-6. 38 SchleefR, Bevilacqua M, Sawdey M, Gimbrone M, Loskutoff 21 Deem RL, Shanahan F, Targan SR. Triggered human D. Cytokine activation of vascular endothelium. Effects on mucosal T cells release tumour necrosis factor-alpha and tissue-type plasminogen activator and type 1 plasminogen interferon-gamma which kill human colonic epithelial cells. activator inhibitor.J Biol Chem 1988; 263: 5797-803. Clin Exp Immunol 1991; 83: 79-84. 39 Bevilacqua M, Pober J, Majeau G, Fiers W, Cotran R, 22 Schiller J, Bittner G, Storer B, Willson JKV. Synergistic Gimbrone M. Recombinant TNF induces procoagulant antitumor effects of tumor necrosis factor and y-interferon activity in endothelium. Proc Natl Acad Sci USA 1986; 83: on human colon carcinoma cell lines. Res 1987; 47: 4533-7. 2809-13. 40 Ryan J, Brett J, Kisiel W, Stern D. Tumor necrosis factor- 23 Doukas J, Pober JS. IFN-gamma enhances endothelial induced endothelial cell tissue factor is principally associated activation induced by tumor necrosis factor but not IL-1. with the subendothelial matrix. Blood 1990; 82 (suppl 76): J Immunol 1990; 145: 1727-33. 361A. 24 Kvale D, Brandtzaeg P, Lovhaug D. Up-regulation of the 41 Bach R, Nemerson Y, Konigsberg W. Purification and expression of secretory component and HLA molecules in a characterisation of bovine tissue factor. J Biol Chem 1981; human colonic cell line by tumour necrosis factor-a and 156: 8324-31. gamma interferon. Scandj Immunol 1988; 28: 351-7. 42 van der Poll T, Buller HR, ten Cate H, Wortel CH, Bauer KA, 25 Ding A, Nathan CF, Stuehr DJ. Release of reactive nitrogen van Deventer SJH, et al. Activation of coagulation after intermediates and reactive oxygen intermediates from necrosis factor to normal mouse peritoneal macrophages: comparison of activating administration of tumor subjects. cytokines and evidence for independent production. NEngljMed 1990; 322: 1622-7. i Immunol 1988; 141: 2407-12. 43 Gerlach H, Liebermann H, Brett J, Bach R, Godman G, Stern 26 Hollander D. Crohn's disease - a permeability disorder of the D. Growing/motile endothelium shows enhanced respon- tight junction? Gut 1989; 29: 1621-4. siveness to tumor necrosis factor/cachectin.J Exp Med 1989; 27 Rutgeerts P, Goboes K, Peeters M, Hiele M, Penninckx F, 170: 913-31. Aerts R, et al. Effect offaecal stream diversion on recurrence 44 Wakefield AJ, Dhillon AP, Rowles PM, Sawyerr AM, Pittilo of Crohn's disease in the neoterminal ileum. Lancet 1991; RM, Lewis AAM, et al. of Crohn's disease: 338: 771-4. multifocal gastrointestinal infarction. Lancet 1989; 334: 28 Halstensen TS, Das KM, Brandtzaeg P. Epithelial deposits of 1057-62. immunoglobulin Gl and activated complement colocalise 45 Murch SH, Braegger CP, Sessa WS, MacDonald TT. with Mr 40 kD putative autoantigen in ulcerative colitis. Increased endothelin immunoreactivity in Crohn's disease Gastroenterology 1993; 34: 650-7. and ulcerative colitis. Lancet 1992; 339: 381-5. 29 Matthes H, Herbst H, Schuppan D, Stallmach A, Milani S, 46 Ding A, Nathan CF, Graycar J, Derynck R, Stuehr DJ, Stein H, et al. Cellular localization of procollagen gene Srimal S. Macrophage deactivating factor and transforming transcripts in inflammatory bowel disease. Gastroenterology growth factors-,6l, -132, and -f13 inhibit induction of macro- 1992; 102:431-42. phage nitrogen oxide synthesis by IFN-y. J7 Immunol 1990; http://gut.bmj.com/ 30 Halstensen TS, Mollnes TE, Brandtzaeg P. Persistent comple- 145: 940-4. ment activation in submucosal blood vessels of active 47 Bogdan C, Vodovotz Y, Nathan C. Macrophage deactivation inflammatory bowel disease: immunohistochemical byinterleukin 10. J ExpMed 1991; 174: 1549-55. evidence. Gastroenterology 1989; 97: 10-9. 48 Sampaio EP, Sarno EN, Galilly R, Cohn ZA, Kaplan G. 31 Nawroth P, Bank I, Handely D, Cassimeris J, Chess L, Stern Thalidomide selectively inhibits tumor necrosis factor a D. Tumor necrosis factor/cachectin interacts with endo- production by stimulated human monocytes. J Exp Med thelial cell receptors to induce release ofinterleukin 1 .J Exp 1991; 173:699-703. Med 1986; 163: 1363-75. 49 Vogelsgang GB, Farmer ER, Hess AD, Altamonte V, 32 Gamble JR, Smith WB, Vadas MA. TNF modulation of Beschorner WE, Jabs DA, et al. Thalidomide for the

endothelial and neutrophil adhesion. In: Beutler B, ed. treatment of chronic graft-versus-host disease. NEnglJ Med on September 24, 2021 by guest. Protected copyright. Tumor necrosis factors: the molecules and their emnerging role in 1992; 326: 1055-8. medicine. New York: Raven Press, 1992: 65-86. 50 Larsson H. Treatment of severe colitis in Behcet's syndrome 33 Strieter RM, Kunkel SL, Showefl HJ, et al. Endothelial with Thalidomide (CG-217). J IntMed 1990; 228: 405-7.