Tumor necrosis factor production during human renal allograft rejection is associated with depression of plasma protein C and free protein S levels and decreased intragraft thrombomodulin expression

Abstract

Fibrin deposition is a common accompaniment of renal allograft rejection, indicating disruption of the normal physiologic balance between procoagulant and anticoagulant pathways. In vitro, tumor necrosis factor (TNF) induces endothelial expression of the procoagulant, tissue factor, and downregulation of thrombomodulin, a key component of the thrombomodulin/protein C (PC)/protein S (PS) pathway, which normally maintains an anticoagulant state by inactivating thrombin, preventing further thrombin formation by degrading factors Va and VIIIa, and decreasing plasminogen activator inhibitor activity. Raised levels of TNF were recently demonstrated within the blood of patients during episodes of renal allograft injection, and may be an early and discriminatory marker of rejection. This led us to investigate prospectively whether monitoring of serum TNF levels was of value clinically, and was associated with effects on circulating PC and PS levels, or alterations in intragraft thrombomodulin expression. Plasma samples (n = 454) were collected three times/week from all patients (n = 25) undergoing renal transplantation during a 9-month consecutive period, and assayed by ELISA and functional assays for TNF, PC, and free PS (FPS). Portions of renal biopsies, taken to evaluate episodes of acute deterioration of renal function, were evaluated by immunoperoxidase labeling for the presence and distribution of TNF, thrombomodulin, PC, PS, thrombin, fibrin, and factors V and VIII. Comparison of 78 plasma samples collected during 26 episodes of biopsy-proven acute cellular rejection with samples collected during periods of stable renal function (n = 349) showed that TNF levels rose significantly (390 ± 242 pg/ml, p < 0.01) above background levels 3 days before rising serum creatinine concentrations, and peaked (2,426 ± 978 pg/ml) on the day of clinical rejection. PC-antigen (Ag) concentrations also decreased 3 days before rejection (68 ± 13%, p < 0.05), and were maximally depressed (49% ± 16%, p < 0.001) on the day of rejection. FPS levels were normal until the day before rejection (63% ± 8%, p < 0.01) and, like PC, were maximally depressed (43 ± 10%) at rejection. Plasma TNF levels were significantly and inversely correlated with PC-Ag (p < 0.001) and FPS (p < 0.005) levels during rejection, regardless of whether such rejection episodes were steroid responsive or required OKT3 monoclonal antibody therapy. TNF, PC, and FPS levels were normal during episodes of cyclosporine toxicity and viral infection. Immunoperoxidase studies showed that rejection was associated with a intrarenal production of TNF, decreased microvascular labeling for thrombomodulin, and widespread endothelial and interstitial deposition of PC, PS, thrombin, fibrin, and factors V and VIII. These findings suggest that TNF is an important mediator of renal allograft rejection, causing depression of the intragraft thrombomodulin/PC/PS pathway and thereby contributing to intragraft fibrin deposition. Moreover, these results indicate that monitoring of circulating concentrations of TNF, PC, and FPS has relevance to the early detection and differential diagnosis of human acute renal allograft rejection.