Effect of Tumor Necrosis Factor on the Human Fibrinolytic System By Paula Silverman, George H. Goldsmith, Jr, Thomas R. Spitzer, Esther H. Rehmus, and Nathan A. Berger We report the effects on the fibrinolytic system of intravenous (IV) recombinant tumor necrosis factor-a (rTNF-a) infusions in patients with advanced cancers. During a phase I clinical trial of rTNF-a, the plasma fibrinolytic system was closely monitored, measuring tissue-type plasminogen activator (tPA) antigen, plasminogen activator (PA) inhibitor activity, and plasma fibrinolytic activity. Thirteen patients with refractory malignancies received 40, 80, or 160 /g/m 2 rTNF-a as 2-hour IV infusions. After a 1-week rest, the same dose was repeated daily for 5 days every 3 weeks for a maximum of four courses. The serum rTNF-a levels peaked at the completion of the IV infusion and rapidly declined thereafter, becoming unmeasurable within 1 hour in all patients. rTNF-a infusion markedly alters the plasma fibrinolytic system. During the 2-

T

UMOR

NECROSIS

factor/cachectin

(TNF-a), a protein produced primarily by

hour infusion, significant increases in the tPA antigen and plasma fibrinolytic activity were seen. After the infusion, PA inhibitor activity increased, neutralizing the plasma fibrinolytic activity. The increase in PA inhibitor activity was maximal 6 hours after the onset of the rTNF-a infusion. Fibrinolytic properties returned to pretreatment values within 24 hours. Daily rTNF-a infusions caused changes in plasma tPA antigen and PA inhibitor similar to those of single infusions. We conclude from these observations that the administration of rTNF-a in vivo to cancer patients causes profound alterations of endothelial cell-derived components of the fibrinolytic system. J Clin Oncol 8:468-475. © 1990 by American Society of ClinicalOncology.

inhibitor (PA inhibitor- 1 ).'-9 Lesser contributors of PA inhibitor-1

to plasma may be platelet 0

macrophages, has been implicated as a mediator in the host response to sepsis and neoplasia.1, 2 In vitro, TNF-a is cytotoxic for tumor cells, 3 -5 and

alpha-granule stores or hepatocytes.'

in vivo causes hemorrhagic necrosis of trans-

with its inhibitor)"11'12 and plasma PA activity

planted tumors in mice. 6 The availability of TNF-a prepared by recombinant DNA technology has facilitated investigation of its biologic properties as well as its therapeutic potential. The in vitro effects of recombinant TNF-a

(rTNF-a) are the subject of intense evaluation. Human clinical trials with rTNF-a are ongoing. Human vascular endothelial cells play a criti-

cal role in the plasma fibrinolytic system through the modulated synthesis and secretion of tissuetype plasminogen activator (tPA) and its fastacting inhibitor, type 1 plasminogen activator

From the Ireland Cancer Center, University Hospitals of Cleveland; and The Department of Medicine, Case Western Reserve University, Cleveland, OH. Submitted April 21, 1989; accepted September 21, 1989. Supported in part by a grantfrom Knoll Pharmaceuticals, Whippany, NJ, and the Judith Graham Pool Postgraduate Research Fellowship. Address reprint requests to PaulaSilverman, MD, Department of Medicine, University Hospitals of Cleveland, 2074 Abington Rd, Cleveland,OH 44106. © 1990 by American Society of Clinical Oncology. 0732-183X/90/0803-0004$3.00/0

468

Specific

assays for both plasma tPA antigen (reflecting the sum of free tPA and inactive tPA in complex (reflecting uncomplexed tPA) have recently been described.13,14 Precise regulation of these fibrin-

olytic proteins is required for normal hemostasis as alterations in plasma tPA or PA inhibitor-1 -have been shown in patients with thrombotic conditions or disseminated intravascular coagu5 lation. -17 Recent studies have indicated that TNF-a is a potent mediator of endothelial functions relevant to thrombosis

and hemostasis.

For example,

rTNF-a induces the expression of tissue factor procoagulant activity on the surface of cultured endothelial cells.' 8 "19In vitro studies have shown that rTNF-a modulates the endothelial fibrinolytic system. Decreases in tPA antigen and increases in PA inhibitor-1 were demonstrated

by Schleef et al after human umbilical vein endothelial cells in culture were incubated 6 hours with 20 to 2,500 U/mL rTNF-a. 20 In

similar experiments, van Hinsbergh et a12 1 also reported increased PA inhibitor activity production in human umbilical artery and vein endothelial cells and in human foreskin microvascular

Journalof Clinical Oncology,Vol 8, No 3 (March), 1990: pp 468-475

Downloaded from ascopubs.org by DALHOUSIE University DAL 11762 on April 20, 2019 from 129.173.072.087 Copyright © 2019 American Society of Clinical Oncology. All rights reserved.

EFFECT OF TNF ON THE FIBRINOLYTIC SYSTEM endothelial cells exposed to rTNF-a. However, when microvascular endothelial cells were exposed to high concentrations (500 to 2,500 U/ mL) of rTNF-a, a consistent increase in tPA antigen was found. Recently, rTNF-a has been shown to cause the induction of PA inhibitor genes and the suppression of the tPA gene in a fibrosarcoma cell line (HT1080 cells). 22 After injection of high doses of rTNF-a into rats, blood fibrinolytic activity decreases 23 and plasma PA inhibitor increases.21 While conducting a phase I clinical trial of rTNF-a in patients with refractory malignancies, we examined our patients for changes in the fibrinolytic system. We have found that infusions of rTNF-a cause a marked alteration in plasma fibrinolytic properties in patients receiving this drug. MATERIALS AND METHODS Preparationof rTNF-oa rTNF-a is produced by recombinant DNA technology and supplied by Knoll Pharmaceuticals, Whippany, NJ. rTNF-a is a nonglycosylated protein with a molecular weight of approximately 17,500 daltons and greater than 95% pure as determined by sodium dodecyl sulfate polyacrylamide gel electrophoresis. The specific activity is 6-10 x 106 U/mg protein in a dactinomycin-free biologic tumor cell (L-929) cytotoxicity assay. 24 The rTNF-a was diluted in 150 mL of normal saline with 0.5% albumin added to prevent adherence of protein to delivery apparatus. Infusions were delivered through a central or peripheral IV catheter.

Clinical Trial Design Adult patients with advanced malignancies refractory to treatment were eligible. Thirteen patients were treated with intravenous rTNF-a on a phase I dose-escalation trial. Cohorts received 40, 80, or 160 Atg/m 2 rTNF-a IV over 2 hours in the single-dose phase of the trial. After a 1-week rest, a 2-hour infusion of rTNF-a at the same dose was administered daily for 5 consecutive days (multiple dose phase). These treatment courses were repeated every 3 weeks for a maximum of four courses. Patients with hemorrhagic or coagulation disorders, and patients taking anticoagulant, thrombolytic, or nonsteroidal antiinflammatory medications were excluded from the trial. The research protocol was approved by the Institutional Review Board of University Hospitals of Cleveland and all patients provided informed consent.

Sample Collection In the single-dose phase, venous blood was collected in plastic syringes and immediately transferred to plastic tubes containing 'so volume 0.5 mol/L Na citrate (pH 5.0) that were kept on ice. Plasma was prepared by immediate centrifugation of samples at 2,800 g at 40C for 15 minutes2 5 and

469 aliquots either assayed immediately or stored at -70 0C. Samples were collected immediately before the rTNF-a infusion, midinfusion, at the completion of the infusion, and 1, 4, and 24 hours after the rTNF-a infusion. Serum samples for pharmacokinetic data were obtained during the singledose phase before infusion, at 1 and 2 hours during the infusion, and at 5, 10, 20, 30, and 45 minutes, and 1, 2, 4, 8, and 24 hours following the end of the rTNF-a infusion. During the multiple dose phase, daily blood samples were collected as stated above, before and 30 minutes after the completion of the rTNF-a infusion. Euglobulin fractions of plasma were prepared by precipitation in 0.01 mol/L Na acetate, pH 4.8 at 40C.26 Euglobulin precipitations were performed immediately after separation of fresh plasma from blood samples. The platelet count, prothrombin time, partial thromboplastin time, and fibrin(ogen)-related antigens were determined in all patients before and 24 hours after each rTNF-a infusion in both phases of the trial and more frequently in selected patients.

FibrinolyticAssays Plasma tPA antigen was measured using Imubind-5 (American Diagnostica, Inc, Greenwich, CT), a commercially available enzyme-linked immunosorbent assay (ELISA) that measures both tPA and tPA-PA inhibitor complexes""•.. 27 with a normal range of tPA antigen in plasma of 9-12 ng/mL. The assay coefficient of variation for all standard curves and plasma samples was 5 10%. Euglobulin samples were thawed, appropriately diluted, and assayed for tPA antigen in an identical fashion. PA inhibitor activity was quantitatively determined by the extinction of tPA activity added to plasma." The tPA activity standard used was two-chain melanoma activator (745 IU/ pg, American Diagnostica Inc). Briefly, thawed plasma samples were prepared for assay by the addition of an equivalent volume of tPA activity standard (100 IU/mL) and incubated for 10 minutes at room temperature. The reaction was stopped and plasmin inhibitors eliminated by acidification with sodium acetate buffer (lmol/L, pH 3.9) and freezing at -20 0 C. Samples were assayed for residual tPA activity in a chromogenic assay in microtiter plates. This assay measures the plasmin-dependent amidolysis of the synthetic substrate H-D-Valyl-L-leucyl-L-Lysine-p-nitroanilide dihydrocholoride (S-2251, Kabi Diagnostics; Helena Laboratories, Beaumont, TX). Plasminogen was prepared from human plasma by affinity chromotography on lysineSepharose28 and Sephadex-G150 gels, treated with 10 mmol/ L diisopropylfluorophosphate (DFP) to remove trace plasmin activity, and diluted to 1.0 mg/mL. Fibrin monomer (2.8 mg/mL) was prepared by DFP treatment of human fibrinogen (Imco, Stockholm, Sweden) followed by treatment with Reptilase-R (Bothrops Atrox, Abbott Laboratories, North Chicago, IL), and stored in 3.5 M urea. The PA activity assay buffer was Tris, 0.05 mol/L, pH 8.3, containing 0.01% Tween 80. The assay reaction mixture consisting of 12.5 gL of fibrin monomer, 10 uL of plasminogen, 30 gL of S-2251 (3 mmol/L), and 60 fL Tris assay buffer was placed in each well of a microtiter plate, 100 pL sample added, and the plate incubated at 370 C for 90 minutes. The reaction was terminated with the addition of 50 ML 20% acetic acid to each well and the absorbance at 405 nm was measured.

Downloaded from ascopubs.org by DALHOUSIE University DAL 11762 on April 20, 2019 from 129.173.072.087 Copyright © 2019 American Society of Clinical Oncology. All rights reserved.

470

SILVERMAN ET AL

PA inhibitor activity standards consisted of duplicate tubes of acidified normal plasma to which were added either tPA activity standard (100 IU/mL) or buffer and then frozen. The standard to which tPA was added represented 0 inhibitory IU/mL while acidified normal plasma without tPA represented 100 inhibitory IU/mL. One inhibitory international unit was then defined as the inhibition of one international unit of tPA under the conditions used. By proper mixing of the standards, final standards in the range of 0 to 80 inhibitory IU/mL were obtained. Sample PA inhibitor activity was then determined from a curve generated from the PA inhibitor standards. The normal range of PA inhibitor activity in plasma as measured by this assay in our laboratory was 7 to 20 inhibitory IU/mL. The assay coefficient of variation was

Effect of tumor necrosis factor on the human fibrinolytic system.

We report the effects on the fibrinolytic system of intravenous (IV) recombinant tumor necrosis factor-alpha (rTNF-alpha) infusions in patients with a...
577KB Sizes 0 Downloads 0 Views