JOURNAL OF SURGICAL RESEARCH 64, 26– 31 (1996) ARTICLE NO. 0301 P-Selectin and TNF Inhibition Reduce Venous Thrombosis Inflammation 1 THOMAS W. WAKEFIELD, M.D., ROBERT M. STRIETER, M.D., LAMIERE J. DOWNING, M.D., AMY M. KADELL, B.S., CAROL A. WILKE, B.A., MARIE D. BURDICK, B.S., SHIRLEY K. WROBLESKI, B.S., M. LAURIE PHILLIPS,PH.D., JAMES C. PAULSON,PH.D., DONALD C. ANDERSON, M.D., AND LAZAR J. GREENFIELD, M.D. Department of Surgery and Department of Medicine, University of Michigan Medical Center, Ann Arbor, Michigan 48109; Cytel Corporation, San Diego, California 92121; and Pharmacia & Upjohn, Kalamazoo, Michigan 48176 Submitted for publication February 27, 1996 leads to vein wall and valvular damage and the even- Venous thrombosis induces a detrimental inflam- tual development of the syndrome of chronic venous matory response in the vein wall. The cytokine tumor insufficiency. Cytokines [tumor necrosis factor (TNF)] necrosis factor-a (TNF) and the adhesion molecules, and adhesion molecules (selectins) have been found to selectins, have been found to be important in mediat- be important in mediating stimulation of inflammatory ing inflammatory cell stimulation and leukocyte– en- cells as well as leukocyte – endothelial cell adhesion in dothelial cell adhesion, respectively. This study as- many models of inflammation [2]. We hypothesize that sesses the role of TNF and P-selectin in the inflamma- TNF and P-selectin are synergistic in inducing in- tory events associated with venous thrombosis. Rats flammation during venous thrombosis and that inhibi- were passively immunized with neutralizing anti-TNF tion of these molecules will decrease the inflammatory serum alone, anti-TNF plus anti-P-selectin antibody, response and preserve vein wall architecture. We dem- anti-P-selectin antibody alone, control serum, or con- onstrate that combined inhibition of TNF and P-selec- trol anti-P-selectin antibody. Antibodies or control tin provides the lowest neutrophil and total inflamma- sera were given prior to occlusion and at Days 2 and tory cell influx into the vein wall in response to venous 4 postocclusion. Rats were sacrificed at Days 1– 6 and thrombosis. Day 13 after occlusion for inferior vena caval (IVC) wall histopathology and TNF analysis. Differences in METHODS the extent of inflammatory cell infiltrate into the vein wall were found on Days 2, 6, and 13. TNF levels were Induction of inferior vena caval (IVC) thrombosis and tissue analy- elevated in the vein wall of the three groups not given sis. Rats were anesthetized with isoflurane and IVC thrombosis anti-TNF antibody. The levels of TNF at Day 6 posi- was induced by IVC occlusion by ligature or clip just below the level tively correlated with both total inflammatory cell (r of the renal veins with ligation of all the side branches draining into Å 0.53, P õ 0.05) and neutrophil presence (r Å 0.72, the IVC [1], as a modification of previous models of IVC thrombosis [3– 5]. Ligation was performed using nonreactive monofilament pro- P õ 0.01). The lowest IVC wall neutrophil and total lene suture material. At Days 1, 2, 3, 4, 5, 6, and 13 after thrombus inflammatory cell count at Days 2 and 6 and the lowest induction, rats were sacrificed (n Å 4 per time point) and the IVC neutrophil count at Day 13 were found in the anti-TNF was grossly inspected for thrombosis. The IVC was further processed plus anti-P-selectin antibody group. Monocyte influx in the following manner: (a) one-half was fixed in formalin for micro- was also inhibited at Day 13 in this group. These re- scopic examination; (b) one-half was removed, the clot separated from sults suggest a role for combined neutralization of TNF the wall, and then the wall immediately placed into liquid nitrogen and P-selectin in the attenuation of inflammation in- for TNF measurement by ELISA. Animals in the Day 13 group un- derwent reversal of IVC occlusion with removal of the microvascular duced by venous thrombosis.  1996 Academic Press, Inc. clip (that was placed at the time of occlusion) on Day 4 after thrombus induction to allow for spontaneous recanalization. In all other groups, occlusion was by permanent ligature. INTRODUCTION Twenty-eight rats were passively immunized with neutralizing anti-TNF antibody serum alone (1 ml/dose), 28 with a combination of anti-TNF serum (1 ml/dose) plus antibody to P-selectin (PB1.3, 2 Venous thrombosis is associated with a significant mg/kg), 28 with antibody to P-selectin alone (PB1.3, 2 mg/kg), 28 inflammatory response in the vein wall and at the with control serum (1 ml/dose), and 28 with a control anti-P-selectin thrombus/vein wall interface. Inflammation involves antibody (P7, 2 mg/kg). All antibodies were given before thrombus an infiltrate of neutrophils followed by monocytes/mac- induction and then at Days 2 and 4 after induction. We chose not to give antibody after Day 4 to be certain that the animals would not rophages [1]. This inflammatory response potentially develop serum sickness. The period chosen for administration should span the time of the initiation and amplification of the inflammatory 1 This work was supported by NIH HL53355 (T.W.W.), NIH response to venous thrombosis. All anti-rat antibodies were endo- toxin free [6]. PB1.3 (CY-1747) is a monoclonal murine IgG1 antibody HL50057 (R.M.S.), and The Research Advisory Committee, Depart- ment of Surgery, University of Michigan. raised against human P-selectin with a half-life of 3 to 7 days that 26 0022-4804/96 $18.00 Copyright  1996 by Academic Press, Inc. All rights of reproduction in any form reserved. AID JSR 4864 / 6n12$$$141 07-15-96 11:46:50 srga AP: Surg Res