The Specific Activity of Inhibitor-1 in Disseminated Intravascular With Acute Promyelocytic Leukemia

By Yoichi Sakata, Toshinobu Murakami, Atsushi Noro, Kazuya Mori, and Michio Matsuda

In disseminated intravascular coagulation (DIC) with acute t-PA to the sample was measured by a sandwich solid-phase promyelocytic leukemia (APL) in the absence of severe enzyme-linked immunosorbent assay using anti-PAL1 and infection, marked was noted in comparison with anti-t-PA monoclonal . The assay to measure total normal levels of antithrombin 111, which is a major inhibitor of PAI-1 antigen used three monoclonal anti-PAL1 antibodies the coagulation system. Increased plasminogen activator and had similar sensitivities to free active, latent, vitronectin- inhibitor-1 (PAI-1)antigen levels in plasma from patients with bound and t-PA-bound PAI-1. The specific activity of PAL1 septicemia decreased the ratio of the plasma clot lysis rate decreased in patients with DIC (43.7% f 30.6%) and in DIC induced by an anti4- inhibitor monoclonal anti- cases with APL (10.3% f 6.0%) in comparison to patients body to the tissue-type plasminogen activator (t-PA) concen- with septicemia (83.7% f 20.2%) or normal controls tration. This decrease was not as prominent in plasma from (85.8% f 27.3%). In DIC associated with APL, degraded patients with DIC, especially those with APL. To explore the forms of PAL1 were detected in plasma by immunoblotting. character of PAL1 in these plasmas, we measured the spe- These results suggest that a decrease in the specific activity cific activity of PAL1 by determining the ratio of active PAL1 of PAL1 and an increase in secondary fibrinolysis result in a Downloaded from http://ashpublications.org/blood/article-pdf/77/9/1949/605395/1949.pdf by guest on 28 September 2021 antigen to t-PA-unbound PAI-1 antigen. To calculate the hyperfibrinolytic state in DIC patients with APL. amount of active PAL1 antigen, the amount of t-PA/PAI-1 6 1991by The American Society of Hematology. complex before and after the addition of a fixed amount of

EMORRHAGE is the second most common fatal PAI-1 is difficult to determine, especially in patient plasma. H complication in patients with acute promyelocytic Therefore, we constructed an enzyme-linked immunosor- leukemia (APL). The bleeding diathesis is due mainly to bent assay (EL1SA)-based system for measurement of both disseminated intravascular coagulation (DIC) initiated by total PAI-1 antigen and t-PA/F’AI-1 complexes. The assay the release of procoagulant activity from abnormal promy- specifically measures active PAI-1 in crude biologic fluids elocytes, although primary fibrinolysis mediated by elastase- and determines total antigen levels independent of the like may also play a part.’ Although form of PAI-1 present in the samples. Our results suggest generation in DIC with APL was reported, a beneficial that decreased functional activity of PAL1 and increased effect of for treatment still remains contr~versial.~’ plasmin generation by formation contribute to the Many patients with APL have low leukocyte counts because observed hyperfibrinolysis in patients with DIC and APL. of therapy against the abnormal promyelocytes, which results in severe infection in these patients, and virtually MATERIALS AND METHODS any gram-negative organism can produce severe DIC.4,s Reagents. The following materials were purchased: Freunds’ Therefore, we selected DIC patients with APL who did not adjuvant, polyethylene glycol 1540, dimethyl sulfoxide, 2,6,10,14- have severe infection and examined the parameters relating tetramethylpentadecane (Pristane), and Tween 20 (Wako Chemi- cal, Osaka, Japan); penicillin and streptomycin (GIBCO, Grand coagulation and fibrinolysis. Because the fibrin(ogen) deg- Island, NY); calf serum, Dulbecco’s modified Eagle’s medium radation products (FDP) in the plasmas were remarkably (DMEM), and 96-well polyvinyl chloride microtiter plates (Titer- high relative to the decrease in c6 or antithrombin teck) (Flow Laboratories, North Ryde, Australia); fetal calf serum I11 (as described later), we analyzed this abnormal hyperfi- (FCS; Filtron, Altona, Australia); tissue culture flasks (Corning brinolysis. Glass Works, Corning, NY); gentamycin, bovine serum Initiation of fibrinolysis is mediated by plasminogen (BSA, essentially fatty acid-free, -free), theophylline, and activators and regulated by plasminogen activator inhibi- prostaglandin E, (Sigma, St Louis, MO); solid-phase lactoperoxi- tor-1 (PAI-l).’ PAI-1 inhibits both tissue-type plasminogen dase-glucose oxidase (Enzymobeads) and horseradish peroxidase- activator (t-PA) and by forming a 1:l stoichiomet- conjugated goat antimouse Ig (BioRad, Richmond, CA); CNBr- ric complex that resists dissociation by sodium dodecyl sulfate (SDS). PAI-1 is a relatively unstable molecule and rapidly decays into a latent form that can be converted into From the Division of and Hemostasis, Institute of an active molecule after treatment with denaturants8 In Hematology, Jichi Medical School, Tochigi-ken,Japan. Submitted April 27,1990; accepted December 20,1990. contrast, PAI-1 in plasma seems to exist primarily in a Supported in part by a Research Grant for Cardiovascular functionally active form, and thus a good correlation has (62-C) from the Ministry of Health and Welfare of Japan, and been found between PAI-1 activity and antigen in pla~ma.~ Grant-in-Aid for Scienti$c Research on Priority Areas No. 63616512 In addition, plasma vitronectin (VN) appears to bind to from the Ministry of Education, Science and Culture of Japan. PAI-1 in plasma, resulting in an approximately twofold Address reprint requests to Yoichi Sakata, MD, Division of Throm- enhancement of stability.” bosis and Hemostasis, Institute of Hematology, Jichi Medical School, Although several assays have been developed to measure Minamikawachi-machi Tochigi-ken,329-04 Japan. total PAI-1 antigen”.” or PA1 activity12-1sin biologic fluids, The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked the current assay systems do not detect the various forms of “advertisement” in accordance with 18 U.S.C. section 1734 solely to PAI-1 with similar efficiency. Furthermore, the functional indicate this fact. assays do not distinguish between the different PAIS, so 6 1991 by The American Society of Hematology. that the specific activity (activity/antigen) of PA-unbound 0006-4971I91 /77O9-0OO7$3.OOlO

Blood, Vol77, No 9 (May 1). 1991: pp 1949-1957 1949 1950 SAKATA ET AL activated Sepharose 4B, protein A-Sepharose CL-4B, and JTAI-2, and JTAI-3) against PAI-1 for the same epitope on PAL1 Sepharose PDlO (Pharmacia-LKB Biotechnology, Uppsala, Swe- was tested as previously described.m Peroxidase-coupled MoAbs den); collagen (Niko Bioscience, Tokyo, Japan); and 2,2‘ asino-di were prepared by coupling the antibodies to horseradish peroxi- [3-ethyl-benthiazoline sufonate] (ABTS) (Kirekegaard & Perry dase using m-periodate by the method of Nakane and Kawaoi?6 Laboratories, Gaithersburg, MD). The mouse myeloma cell line Clinical groups. DIC patients (n = 58) in this study were P3-X63-Arg8-U1 (P3U1) was provided by Dr T. Watanabe (Ky- separated into two groups: groups with (n = 15) or without APL ushu University, Fukuoka, Japan). Aprotinin and single-chain (n = 43). The presence of DIC was diagnosed for each patient by urokinase-type plasminogen activator (scu-PA) were gifts from marked reductions in counts, , antithrombin 111 Mochida Pharmaceutical (Tokyo, Japan). (except in APL without ), plasminogen, and a2-plasmin t-PA was purified from medium conditioned by human mela- inhibitor levels and elevated levels of fibrin degradation products. noma cells with aprotinin as described previously16 and had a In this study, DIC patients with APL and without sepsis (n = 13) specific activity of 550,000 U/mg protein according to the t-PA were carefully chosen and separated from the two cases with severe International Standard (83/517). It migrated as a single band in sepsis (see Fig 5, lanes g and h). The DIC patient group without SDS-polyacrylamidegel electrophoresis (SDS-PAGE) under reduc- APL included 23 patients with leukemia, 12 patients with lung ing and nonreducing conditions. The t-PA International Standard cancer, 6 patients with gastric cancer, and 2 patients with biliary was supplied by the National Institute for Biological Standards and tract infection. Thirty-two patients had either high- or low-grade Downloaded from http://ashpublications.org/blood/article-pdf/77/9/1949/605395/1949.pdf by guest on 28 September 2021 Controls (London, England). Glu-plasminogen was purified from fever. A sepsis group without DIC included five patients with in the presence of 100 inactivating malignant tumor, two with respiratory tract infection, and five with units (KIU)/mL aprotinin as described.I6 Fibrinogen was prepared peritonitis. according to the method of Blomback and B10mback.l~Vitronectin Plasma. Blood was collected from the antecubital veins of was initially prepared from fresh frozen plasma according to the healthy volunteers and patients into 0.1 vol 3.8% sodium citrate method of Dahlback and Podack.’* Immunopurified VN was also with or without theophylline (final 0.001 mol/L) and prostaglandin prepared from fresh frozen plasma by ammonium sulfate precipita- E, (final 0.5 mg/L) with a 30-second handcuff at about 9 AM. The tion (50% saturation) and affinity chromatography on a column collected blood was immediately cooled with ice water and centri- containing anti-VN polyclonal immobilized on Sepharose. fuged at 2,000g for 20 minutes at 4°C to prepare platelet-poor Human PAI-1 was purified from the conditioned media of cultured plasma (PPP). All plasma was stored at -80°C until use. Gel- human umbilical vein endothelial cells basically as described by filtered and platelet-rich plasma were prepared as de- Hekman and Lo~kutoff.’~Except for immunopurified VN, the ~cribed.’~PAI-l-depleted plasma was prepared from pooled healthy purified appeared to be greater than 98% homogeneous volunteers’ plasma by passage through rabbit anti-PAI-1 IgG- on SDS-PAGE. Immunopurified VN had a significant amount of immobilized Sepharose (5 mg IgG/mL Sepharose), which had been material that smeared from the top of the gel to the 65,000-dalton equilibrated with a mixture of 0.9 vol0.15 mol/L NaCl and 0.1 vol polypeptide on SDS-PAGE. Protein concentrations were deter- 3.8% sodium citrate. Single time point plasma samples were taken mined from their respective extinction coefficients at 280 nm.*” from patients with DIC. Protein concentrations of t-PA and PAI-1 were determined by Plasma clot lysis induced by MoAb against &‘-plasmin inhibitor amino acid analysis by the phenylthiocarbamyl method as de- (&’-PI). Plasma was supplemented with 2 pmol/L Glu-plasmino- scribed” using the PICO.TAG system (Waters Associates, Milford, gen and 1.2 pmol/L of antilu2-PI MoAb JTP-1, which induces MA) after hydrolysis for 24 hours in HCI vapor at 110°C. The spontaneous plasma clot lysis.= a-Thrombin (final 0.1 U/mL) and quantities of t-PA and PAL1 were calculated based on their amino CaCI, (final 0.018 mol/L) were added, and the solution was allowed acid composition as deduced from their cDNA sequence^^^^^^ and to clot for 15 minutes at 37°C in a glass tube. The clots were wound assumed molecular weights of 70,000 and 50,000, respectively. onto a bamboo stick and squeezed to remove as much fluid as Monospecific antisera against t-PA, scu-PA, PAI-1, and VN possible. Clots were subsequently resuspended in the same sera. were prepared in rabbits as described previously.I6 The t-PA and Fibrinolysis was monitored by measuring the amount fibrin PAI-1 complex was purified as follows: Confluent cultures of of degradation products in the supernatant after 3 hours of incuba- human endothelial cells (1.5 X lO’/cm’) were washed three times with serum-free medium 199 and then incubated in serum-free tion at 37°C using counter-immunoelectrophoresis as described?’ medium containing 1 p,g/mL t-PA for 6 hours. This conditioned The amount of FDP before incubation was subtracted from this medium was applied to an affinity column containing anti-t-PA value. antibody immobilized on Sepharose. Bound antigen was eluted Solid-phase ELISA for total PAI-I. Polyvinyl chloride microtiter with 3 mol/L KSCN and was further purified using anti-PAL1 plates were coated with a mixture of two MoAbs against PAI-1 (10 antibody-immobilized Sepharose as described.I6 The quantity of &nL JTAI-1 and JTAI-2) for 16 hours at 4°C as described.” The PAI-1 present as a complex was calculated from the amount of plates were emptied and the wells were treated for 2 hours at 25°C t-PA present, assuming a 1:l stoichiometric complex. The amount with 0.05 mol/L Tris-HC1, 0.1 mol/L NaCI, pH 7.4 (TBS buffer) of t-PA was measured by ELISA using anti-t-PA monoclonal containing 1% BSA. After washing two times with TBS containing antibodies (MoAbs) JTA-1 and JTA-2 as described.I6 In this 0.05% Tween 20 (TBS-Tween) and 0.1% BSA, samples were ELISA, the amount of t-PA complexed with PA-1 can be measured diluted in the same buffer and aliquots were added to the coated by using the same standard curve constructed for free t-PA as wells. When samples were plasma, the samples were diluted with described.I6 The level of t-PA was also confirmed by the commer- the same buffer four times and the standard curves were con- cial ELISA kit Imulyse t-PAZ4(Biopool AB, Umea, Sweden). structed with various concentrations of purified PAI-1 diluted with MoAbs against PAJ-1, using either t-PAiPAI-1 complex or latent a final 25% PAI-1 immunodepleted plasma. Variability due to PAI-1 as an antigen, were raised by standard methods using dilution of plasma proteins was adjusted by using PAI-1 immunode- BALB/c mice and the mouse myeloma cell line P3U1 as de- pleted plasma. The plates were then incubated for 10 hours at 4°C. scribed.” The dissociation constants (kd) of ’zI-labeled MoAbs for After the plates were washed with cold TBS-Tween, peroxidase- the latent PAI-1 and t-PAPAI-1 complex were assessed by a anti-PAI MoAb JTAI-3 was added to each well and incubated for solid-phase assay as described,” according to the Frankel and another 10 hours at 4°C. After washing twice with cold TBS- Gerhard method.” The competition between MoAbs (JTAI-1, Tween, hydrolysis of the ABTS substrate by adsorbed JTAI-3 was SPECIFIC ACTIVITY OF PAL1 IN DIC WITH APL 1951 monitored at 25°C on an ELISA analyzer ETY-96 (Toyo Sokki, MA). The labeled MoAbs JTAI-1, -2, -3, -4, and -5 had specific Tokyo, Japan) for 30 minutes. activities of 6.8 x lo5, 7.2 x l@,7.7 x lo5, 8.3 x 1@,6.3 x l@,and Solid-phase ELZSA for t-PAIPAZ-1 complex. The ELISA for 8.8 x l@cpdkg, respectively. The concentration of a2-PI/plasmin t-PAPAI-1 complex in samples was performed basically as de- complex in plasma was measured using the ELISA kit TD-80C scribed above using JTAI-1 as the coating antibody and peroxidase- (Teijin Co, Tokyo, Japan)?3 anti-t-PA MoAb JTA-1. To assess the level of active PAI-1 antigen in a sample, single-chain t-PA (final 75 ng/mL) and aprotinin (final 500 U/mL) were added to the sample and the mixture was RESULTS incubated for 30 minutes at 37°C. After incubation, samples were Coagulation and fibrinolysis in DIC patients with APL. diluted four times with TBS-Tween containing 0.1% albumin, and DIC associates frequently with APL even in the absence of differences due to interference of plasma proteins were adjusted severe infection and results in a high mortality from with PAI-l-depleted plasma added just before ELISA. The quan- hemorrhage, especially intracranial bleeding.’ Although tity of active PAI-1 was calculated from the amount of t-PA/PAI-1 complex formed after correction for the amount of complex in the thrombin generation was reported in DIC with APL,34 original sample and assuming a 1:l stoichiometry. neither activity and antigen levels of protein C6 nor an-

SDS-PAGE. SDS-PAGE in slab gels was performed using tithrombin I11 levels decreased in our DIC patients with Downloaded from http://ashpublications.org/blood/article-pdf/77/9/1949/605395/1949.pdf by guest on 28 September 2021 resolving gels of 10% acrylamide and stacking gels of 4% according APL in the absence of sepsis. Because sepsis is one of the to Laemmli.MSDS-PAGE in tube gels was performed using gels of major triggers of DIC, we carefully selected DIC patients 7% acrylamide according to Weber and Osborn.” Molecular for those with APL and without sepsis. weight standards were obtained from Bio-Rad. To observe the As shown in Table 1, the a2-PI antigen and fibrinogen profile of PAI-1 from plasma, plasma was immunoadsorbed with levels in cases of DIC with APL were significantly lower and anti-PAL1 polyclonal antibody immobilized on Sepharose. The the level of FDP was significantly higher than observed in column was eluted with sample buffer for SDS-PAGE containing the group of DIC patients with other . These results 2% SDS, and samples were fractionated by SDS-PAGE. The gel was processed for reverse fibrinautography or Western blotting. suggested that a marked increase in secondary fibrinolysis Western blotting of the samples was performed after electrical occurs in DIC patients with APL. However, in these transfer to nitrocellulose, using anti-PAI-1 IgG, as previously patients the consumption of antithrombin I11 and protein described.w C, and the increase in PAs and plasmin-&-PI complex Miscellaneous. Human umbilical vein endothelial cells were levels were not as great as those found in the control DIC isolated from human umbilical cord veins by the method of Jaffe et group. ai3*and were cultured as described.16 Plasma concentrations of Plasma clot lysis induced by an MoAb against &?-PI. In scu-PA were measured basically as described previously by ELISA DIC, “secondary fibrinolysis” that occurs on the fibrin clots using anti-urokinase MoAb JS-1 (final 10 &mL) as a coating plays a role not only in the removal of formed clots but also antibody, peroxidase MoAb JS-2 as a second antibody, and purified in the prevention of fibrin deposition. To test the capacity scu-PA as the standard urokinase-type plasminogen activator (u-PA). This ELISA measures the amount of single-chain u-PA, for initiation of fibrinolysis in these patients, we observed two-chain u-PA that was activated by plasmin, and u-PA com- secondary fibrinolysis by adding a small amount of throm- plexed with PAL1 by using the same standard curve for free bin and 1.2 Fmol/L MoAb against a2-PI (MoAb JTPI-1) as scu-PA (data not shown). Purified PAI-1 was reactivated with described in Materials and Methods. This MoAb induces either 8 moVL urea or 4 mol/L guanidine hydrochloride at pH 5.5 complete spontaneous clot lysis when incubated at 37°C for for 30 minutes at 37”C, and dialyzed for 12 hours at 4°C against 0.05 10 hours after clotting, as described?’ Because the concen- mom sodium acetate pH 5.5 containing 0.1 moVL NaCl and 0.05% tration of plasminogen in plasma affected this spontaneous Tween 20 and for 2 hours at 4°C against TBS-Tween pH 7.4. PA1 clot lysis and this effect leveled off in the presence of more activity in a sample was assessed by the method as describedl3.l6 than 2 p,mol/L plasminogen (data not shown), we added 2 (conventional functional assay of PAI) using single-chain t-PA, FmoVL Glu-plasminogen to each plasma sample before plasminogen (1 p,mol/L), CNBr-digested fibrinogen (1 p,mol/L), and S-2251 (Kabi Diagnostica, Stockholm, Sweden) (0.5 mmol/L). clotting. As shown in Fig 1, the ratios of the amount of fibrin One unit of PAI is defined as the amount of PA1 that neutralizes 2 degradation products produced after 3 hours of+incubation ng of single-chain t-PA in 10 minutes at room temperature. MoAbs at 37°C relative to the antigen level of t-PA were signifi- against PAL1 were radioiodinated using lactoperoxidase-glucose cantly different in patients with different disease states. The oxidase and Na lzsI (20 Ci/mg) (New England Nuclear, Boston, ratios in patients with septicemia were remarkably low and

Table 1. Various Parameters Relating to Coagulation and Fibrinolysis in DIC Patients

DIC (APL-) (N = 43) DIC (APL) (N = 13) Sepsis (N = 12) Normal (N = 40)

AT 111 (%) 68.8 -c 26.4 109.2 12.3 98.6 e 10.3 100 t 10 Fibrinogen (mg/100 mL) 277.8 f 158.4 80 f 40.7 414.1 f 152.0 250 -c 50 FDP (pg/mL) 78.2 f 68.1 217.1 f 205.0 < 10 < 10 a2-PI (pg/mL) 76.3 e 21.2 42.9 e 11.2 95.5 f 12 100 t 10 PIC (pg/mL) 5.2 ? 6.3 2.2 t 3.1 <0.5 < 0.5 U-PA (ng/mL) 2.9 f 1.6 3.3 f 2.1 2.6e 1.7 2.2 f 1.8 t-PA (ng/mL) 78.2 ? 68.1 11.0 2 6.4 19.3 * 15.0 8.92 2 7.89 Plasma was analyzed as described in Materials and Methods. Values are mean f SD. Abbreviations: APL-, without APL; ATIII, antithrombin 111; FDP; fibrinogen and/or fibrin degradation products; PIC, plasma a2-PI complex; U-PA, urokinase-type plasminogen activator. 1952 SAKATA ET AL

800 t-PA/PAI-1 complex with a similar efficiency, standard curves were constructed using PAI-l-depleted pooled nor- mal plasma supplemented with either purified PAI-1 or t-PA/PAI-1 complex. In any combination of two MoAbs, the reactivities toward free PAI-1 and the t-PA/PAI-1 complex were at least one and a half times different (data not shown). Therefore, we used a mixture of JTAI-1 and 600 JTAI-2 as the coating antibodies and peroxidase-labeled JTAI-3 to develop an assay for total PAI-1 antigen. Using c-- this system, free PAI-1 and t-PA/PAI-1 complex were E \ recognized equivalently when diluted in normal PAI-1- 0 depleted plasma (Fig 2A). However, perhaps due to the Y - double antibody-coating, this assay was not so sensitive (about 0.6 ng/mL). Because plasma had to be diluted at -5 400 Downloaded from http://ashpublications.org/blood/article-pdf/77/9/1949/605395/1949.pdf by guest on 28 September 2021 -> least 32 times to minimize the interference of plasma 0 proteins and the interdilution coefficient of variation of .-S four serial dilutions (4, 8, 16, and 32 times) of 20 different L P samples was 58%, the samples were diluted with PAI-1- L depleted plasma. Under these modified conditions, the interdilution coefficient of variation of 20 different samples 200 was 5.1%. When adding purified latent PAI-1 or t-PA/ PAI-1 complex to PAI-1-depleted plasma of final concen- trations of 10,50, or 100 ng/mL, recoveries from this assay were 110% 5 5%, 95% k 4%, and 101% k 6% for latent PAI-1 and 95% ? 7%, 102% ? 7%, and 90% ? 4% for t-PA/PAI-1 complex, respectively (mean ? SD, n = 3). In- ter-assay and intra-assay coefficients of variation were 7.9% 0 1'1'1'1' and 5.2% (n = lo), respectively. 25 50 75 100 1 5 To develop an ELISA for the t-PA/PAI-1 complex, JTAI-1 was used as the coating antibody and peroxidase- t-PA (ng/ml) labeled anti-t-PA MoAb (JTA-1) was used to detect t-PA antigen as it recognizes both free t-PA and the t-PA/PAI-1 Fig 1. Fibrinolytic potential in DIC patient plasmas. The fibrinolytic complex with a similar efficiency.16 Figure 2B illustrates a potential of the plasmas estimated by JTPI-1 induced fibrinolysis dose-response curve of purified t-PA/PAI-1 complex di- ([FDP] after 3 hours' incubation - [FDP] before incubation) was com- luted to varying concentrations in normal PAI-1-depleted pared with the concentration of t-PA in plasma as described in Materials and Methods. Symbols represent the following: (A),normal plasma. In this sandwich ELISA, addition of a greater than control; (0).DIC associated with APL in the absence of sepsis; (01, 50-fold molIL excess of free latent PAI-1 to the samples just DIC without APL; (A),sepsis without DIC. Data represent the mean of before assay did not interfere with the measurement of triplicate assays. t-PA/FAI-1 complex in the samples (data not shown), and the assay sensitivity was 0.2 ng/mL. The samples were those in DIC patients without APL were relatively low in diluted with PAI-1-depleted plasma to measure both t-PA/ comparison with the level of t-PA. Interestingly, the ratios PAI-1 complex and total PAI-1 antigen. When adding in DIC plasma with APL were not lower than those in purified t-PA/PAI-1 complex to final concentrations of 10, normal plasma. Because this assay was intended to observe 50, or 100 ng/mL, recoveries were 105% ? 6%, 101% k 5%, the regulation of the fibrinolysis in the initiation step by and 98% ? 6%, respectively (mean ? SD, n = 3). Inter- neutralizing antiplasmin activity, we examined the antigen assay and intra-assay coefficients of variation were 6.8% and activity level of PAI-1. and 4.9%, respectively (n = 10). Development of ELISA for total PAI-1 and t-PAIPAI-1 Development of modified ELISA to measure active PAI-1 complex. We selected three MoAbs against PAI-1 (JTAI-1, antigen. t-PA forms a 1:l stoichiometric complex with -2, and -3) whose kds were similar for free PAI-1 or the active PAI-1 but does not complex with latent PAI-1," and t-PA/PAI-1 complex: JTAI-1, kd = 9.7 x mol/L for the amount of '251-labeledt-PA/PAI-1 complex, generated latent PAI-1 and 4.8 X mol/L for complex; JTAI-2, by the addition of 1251-labeledt-PA to the sample, correlates kd = 7.8 x molIL and 11.2 x mol/L; and JTAI-3, well with the conventional functional assay of PAI.16 There- kd = 4.7 x mol/L and 12.3 x mol/L, respectively. fore, we determined the amount of active PAI-1 antigen by Furthermore, these three MoAbs did not compete with one adding a fixed amount of t-PA to biologic fluids and another for binding to free PAI-1 or t-PNPAI-1 complex in measuring the newly generated t-PA/PAI-1 complex by the a competition assay" (data not shown). ELISA method described above. To determine the opti- To determine whether an ELISA using any pair of the mum t-PA concentration and incubation time, various MoAbs could detect both t-PA-unbound PAI-1 and the concentrations of t-PA and aprotinin were added to sam- SPECIFIC ACTIVITY OF PAL1 IN DIC WITH APL 1953

2, i AI E C 1 0 in e ZI n 10 0 noy = 0.59x+1 .O m- r2 = 0.90 (n=24) I n I u I,I' u' 07 I 1 I I I I I111 JFW- 1 0 40 50 0 10 20 30 40 50 io 20 30 Concentration (ng/ml) PAI-1 in complex (ng/ml) Downloaded from http://ashpublications.org/blood/article-pdf/77/9/1949/605395/1949.pdf by guest on 28 September 2021

y =O. 57x +7.8 r2=0.83 (n=40) E E 3 C 0 ev) 1- D 0

0 2 4 6 81012 0 1 00 200 3 10 Concentration (ng/ml) PAI-1 in Complex (ng/ml)

Fig 2. Dose-responsecurves of total PAL1 and t-PNPAI-1 complex and correlation between PA1 activity and active PAI-1 antigen. (A) Various concentrations of purified latent PAL1 (0)or t-PNPAI-1 (0) complex were added to normal PAl-l-depleted plasma and their differential reactivities were determined by the ELISA for total PAL1 as described in Materials and Methods. The abscissa shows the amount of latent PAI-1 and calculated PAI-1 in the complex assuming a 1:l complex between t-PA and PAI-1. (B) Various concentrations of purified t-PNPAI-1 complex were added to normal PAl-l-depleted plasma and concentrations of complex were determined as described in Materials and Methods. The data represent the mean of three experiments. (C) Urea- or guanidine-activated PAI-1 samples (two each) were diluted with either TBS-Tween containing 0.1 % BSA (0)or PAl-l-depleted plasma (0)as described in Results, and active PAL1 levels were determined by a conventional method (ordinate) and by our modified ELISA (abscissa]. One unit of PA1 is defined as the amount of PA1 that neutralizes 2 ng of single-chain t-PA. (D) Ten sepsis patient plasmas without DIC, 16 DIC patient plasmas, and 14 normal plasmas were collected as described in Materials and Methods, and active PAL1 levels were determined by ELISA as described above. ples that contained activated PAI-1 or to three different 10 nonpregnant patients with sepsis without DIC, 16 DIC plasma samples and the amount of t-PNPAI-1 complex was plasmas, and 14 normals, there was also a good correlation measured after various incubation times. Added t-PA between the two assay methods (Fig 2D). rapidly complexed with PAI-1 in samples and the complex Effect of complexation of PAI-1 to vitronectin and the decay was stable at least 60 minutes at 37°C. A final concentration of active PAI-1 to the latentform on the measurement of PAI-1 of 75 ng/mL single-chain t-PA and 500 U/mL aprotinin was antigen. Because plasma vitronectin binds to active PAI-1 sufficient to minimize proteolytic degradation of PAI-1 and and apparently stabilizes PAI-1 activity twofold to threefold dilution of samples (data not shown). Therefore, 75 ng/mL in blood," it was critical to know to what degree the ELISA t-PA (final concentration) was added to samples and the for total PAI-1 measures vitronectin-complexed PAI-1. mixtures were incubated for 30 minutes at 37°C. Furthermore, because there may be free active and latent To test the correlation between our assay and the PAI-1 in biologic fluids, it was also important to test the conventional method using t-PA, plasminogen, fibrinogen reactivities of the ELISA toward active and latent PAI-1. fragments, and S2251 as de~cribed,'~,'~purified PAI-1 was To these ends, total PAI-1 antigen and the time-dependent activated with urea (n = 2) or guanidine chloride (n = 2) decay of active antigen levels were determined in the for assay. Samples were adjusted to a similar level of activity presence or absence of exogenous vitronectin (Fig 3A). The as determined by the conventional method and diluted decay rate of activated PAI-1 (40% of the antigen was several times with either TBS Tween containing 0.01% activated by measuring activity and antigen using the BSA or PAI-1-depleted plasma. As shown in Fig 2C, there conventional activity method and our total PAI-1 method), was a good correlation between PAI-1 antigen calculated purified from conditioned media of endothelial cells, was from the generated complex and PA1 activity determined compared with that of inherent active PAI-1 present in by the conventional method. Futhermore, in the plasma of conditioned media. The results were essentially the same 1954 SAKATA ET AL

120- using either form of PAI-1, although active PAI-1 in the conditioned media was 8.3% 2 2.1% (mean 2 SD, n = 3) of total PAI-1 antigen. The level of active PAI-1 antigen decreased gradually in the absence of vitronectin (half-life: 1.5 hours), while in the presence of vitronectin the half-life was prolonged (3 hours) at 37°C. By using a sandwich immunoradiometric assay we could detect complexes of PAI-1 and vitronectin in the presence of exogenous vitronec- tin. In contrast, total PAI-1 antigen did not change even if 20 - activated PAI-1 decayed into the latent form, or if activated 7 PAI-1 complexed with vitronectin (Fig 3A). 00Because PAI-1 in plasma or platelet releasates were 0 1 2 3 4 5 supposed to bind to VN, the effect of decay of active PAI-1 Time (h) in these samples on the measurement of total PAI-1 antigen 120 by this assay was determined at 37°C. The specific activity of Downloaded from http://ashpublications.org/blood/article-pdf/77/9/1949/605395/1949.pdf by guest on 28 September 2021 PAI-1 from platelet releasates decayed with a half-life of 3 100 hours, whereas total PAI-1 antigen did not change (Fig 3B). - The releasate from the different methods for aggregating -8 platelets did not show any distinct differences in their rates 80 of loss of PAX-1 activity. At 3TC, PAI-1 activity also r I declined at a similar rate (half-life: 3 hours) for normal 6o plasma, DIC patient plasmas, or for normal citrated whole blood (Fig 3C). 40 Evaluation of plasma PAI-I in patients. Total PAI-1 1 levels in plasma from DIC patients and from patients with 20 3sepsis were higher than those from normal controls (Fig 0 1 2 3 4 5 4A). However, the specific activity of PAI-1 (ratio of active Time (h) PAI-1 to t-PA unbound PAI-1) in plasma from DIC patients, especially those with APL, was greatly reduced in

120, ~ I comparison with that observed from sepsis patients or normal controls (Fig 4B). The level of t-PA-unbound PAI-1 was calculated by subtracting the value of 5/12 X (t-PA/PAI-1 complex) from the total PAI-1 concentration. During DIC, platelets are continuously activated by throm- bin or other agonists in vivo and released PAI-1 from platelets may contribute to the increase in PAI-1 antigen. The specific activity of platelet PAI-1 determined by our ELISA method was about 6%, a result almost identical to that described by Declerck et al (about 5%).'" These results suggest that an increasing level of platelet PAI-1 released into the plasma may be one explanation for the observed 0 10 20 low PAI-1 specific activities in DIC patients. However, Time (hl because the APL patients had low platelet counts and an extremely low PAI-1 specific activity, the profiles of PAI-1 Fig 3. Effect of incubation at 37°C on the levels of total PAL1 and antigen in patient plasmas were examined by Western active PAI-1 antigen in purified PAI-1, platelet releasates, PPP, or blood. (A) Purified guanidine-activated PAL1 (final concentration, blotting (Fig 5). As shown in Fig 5, lane f, PAI-1 appeared active: 40 ng/mL and total: 100 ng/mL) was incubated at 37°C for to be cleaved, presumably by some (s) present in various times alone (0)or mixed with purified VN (0)(final concentra- the patient plasma. Additionally, in DIC associated with tion 25 kg/mL) in TBS-Tween containing 0.1% BSA. At the indicated APL and sepsis (Fig 5, lane h), the high molecular weight times, aliquots were removed and immediately cooled. The levels of of total PAL1 antigen (closed symbols) and active PAI-I antigen (open (MW, - 110,000) form PAI-1, which also reacted with symbols) were measured as described in Materials and Methods. (6) anti-t-PA antibody by Western blotting, and degraded Platelet releasates were obtained from gel-filtrated platelets by forms of PAI-1 were present as well as an apparent overall treatment with thrombin (0.5 U/mL) (0).collagen (40 pg/mL) (0).or increased level of PAI-1 antigen. freezing and thawing (A)and centrifugation at 3,OOag for 10 minutes at 4°C. The supernatants were incubated at 37°C for various times. At the indicated times, aliquots were removed and the levels of total DISCUSSION PAI-1 (closed symbols) and active PAL1 (open symbols) antigen were measured as described in Materials and Methods. (C) Change in the APL has been characterized by proliferation of morpho- levels of active PAL1 antigen (open symbols) and total PAI-1 antigen logically abnormal promyelocytes and a severe bleeding (closed symbols) in normal citrated plasma (0).in DIC patient plasma diathesis. The bleeding diathesis is due mainly to DIC (0).and in citrated whole blood containing antiplatelet agents (A) during incubation at 37°C was monitored by ELISA as described in initiated by release of procoagulant activity from abnormal Materials and Methods. granules in the promyelocytes.' Thrombin generation34and SPECIFIC ACTIVITY OF PAL1 IN DIC WITH APL 1955

200 the beneficial effect of heparin in the management of A patients with APL were reported.' However, if we carefully selected thc DTC group with APL to eliminate patients with sepsis, which often accompanies the reduction of leukocyte count by chemotherapy against abnormal promyelocytes, the level of antithrombin I11 (Table 1) and protein Ch did 1 oc not decrease as much as in DIC with other diseases. In this group, marked hyperfibrinolysis was noted (Table l), and heparin had no beneficial effect on the bleeding diathesis in these patients (data not shown). Although we did not examine the possibility that abnormal promyelocytes pro- duce antithrombin I11 and , these results may imply that this hyperfibrinolysis is not simply explained by an increase in secondary fibrinolysis. To simulate "second- N=40 N=13 N=12 N=30 ary fibrinolysis," thrombin and CaCI, were added to each Downloaded from http://ashpublications.org/blood/article-pdf/77/9/1949/605395/1949.pdf by guest on 28 September 2021 patient plasma and the fibrinolytic potential in plasma was analyzed by simple clot lysis assay using MoAb JTPI-1, which neutalizes 132-pI activity and induces spontaneous 100 85.8k27.3 clot lysis.2n132-pI activity was neutralized in this assay so D that the ratio of fibrinolysis to t-PA (FP ratio) reflects the 80 PA1 activity in plasma. Because PAI-1 is the primary physiologic inhibitor of t-PA in clot lysis, we measured the amount of PAI-1 in patient plasma to explain the different 60 FP ratios in the DIC groups with APL or without APL, in the patient group with septicemia, and in normal controls (Fig 1). Because PAI-1 in biologic fluids exists in a free 40 active, latent, VN-bound and t-PA-bound forms, total PAI-1 antigen levels have been difficult to measure using 20 murine MoAbs, although a number of assay systems have been developed."*'2 For example, the Imulyse PAI-1 anti- gen assay from Biopool detects both active and latent forms 0 of PAI-1, whereas the t-PNPAI-1 complex is recovered N=40 N=13 N=12 N=38 with about 10-fold lower efficiency according to the manu- facturer's specifications. By using two different MoAbs as Fig 4. Total PAL1 antigen and specific activity of PAL1 in DIC patient plasma. The levels of total PAL1 antigen, active PAL1 antigen, the first antibodies in our assay system, we can measure free and t-PA/PAI-1 complex were measured as described in Materials and active, latent, t-PA-bound and VN-bound PAI-1 with Methods. Patient plasmas: DIC without APL (W); DIC with APL similar efficiencies (Figs 2 and 3). without sepsis (0);sepsis without DIC m), and normal volunteer However, because it is almost impossible to purify only plasma a.Values are mean ? SD. Mr X10-3 110

84 Fig 5. PAL1 antigen in DIC F patient plasma. Conditioned me- dia of cultured human endothe- lial cells (a), and plasma for nor- mal (b), sepsis (c), DIC with sepsis (d), and DIC with lung cancer (e), DIC with APL in the absence of sepsis (f), and DIC with APL in the presence of sepsis (g and h). Plasma samples were immunoad- 33 sorbed by anti-PAl-1 Sepharose and the eluate was fractionated and subjected to Western blot- ting using a mixture of three Mo- Abs (JTAI-l, -2, and -3) as de- 24 scribed in Materials and Methods. a bcdefgh 1956 SAKATA ET AL inherently active PAI-1 from the conditioned media of we used the present modified ELISA method to analyze the HUVEC, we cannot exclude the possibility that the rela- total PAI-1 antigen, t-PA/PAI-1 complex, and active PAI-1 tively low specific activity of the PAI-1 in the conditioned antigen in the patient plasmas. media may lead to the invisibility of an existing differential The PAI-1 antigen levels increased in DIC patients and reactivity of this assay between inherent active PAI-1 and in patients with septicemia (Fig 4A), and the FP ratios in the other forms of PAI-1. A number of assay systems have these patients were lower than those in normal controls been developed that estimate PAI activity in biologic (Fig 1). These results may suggest the retardation of clot fluid.”-l5For example, measurement of residual PA activity, lysis in vivo by the elevated PAI-1 in these diseases.39The either directly or by measuring plasminogen activating ratio in DIC patients was higher than that in patients with activity in the presence of fibrin fragments after adding a septicemia. This difference may be explained by the fact fixed amount of PA to the samples, is one of the typical PA1 that the specific activity of PAI-1 in patients with DIC activity assays. However, these assays cannot discriminate decreased significantly in comparison to patients with PAI-1 activity from other PA inhibitors, in particular PAI-2 septicemia or in normal controls (Fig 4B). In DIC patients activity during pregnancy and PA13 in the presence of the PAI-1 antigen levels may increase because of release heparin. We can measure the concentration of active PAI-1 from platelets by thrombin and other agonists, or increased Downloaded from http://ashpublications.org/blood/article-pdf/77/9/1949/605395/1949.pdf by guest on 28 September 2021 antigen specifically by measuring t-PAPAI-1 complex after production by endothelial cells after endotoxin, interleu- adding a fixed amount of t-PA to the samples. A similar kin-1, or ischemic stimulation. In general, the plasma from approach for the measurement of active PAI-1 has also a DIC patient without APL contained a small amount of been described by several gro~ps.3~”’ degraded PAI-1 as observed on Western blots. However, Any measurement of PAI-1 levels in biologic fluids must the majority of the PAI-1 migrated as an intact Mr 50,000 rely on an appropriate and stable standard material for band. Because PAI-1 in the DIC plasma decayed into the accurate determinations. For functional PAI-1 assays, the latent form at a rate similar to PAI-1 in normal plasma (Fig choice of a standard is particularly complicated because it is 3C), it is unlikely that the low specific activity results from a trace plasma protein whose activity decays into latency. In an instability of the active PAI-1. However, it is possible addition, PAI-1 is present primarily in the latent form in that this inactive PAI-1 originated from activated platelets, most conditioned media from cultured cells’ and in plate- assuming increased platelet activation during DIC. Al- lets.” The current assay measures active PAI-1 by virtue of though the level of total PAI-1 antigen in DIC patient its ability to bind a known amount of added t-PA, and the plasma with APL was higher than that in normal controls, standard was purified t-PA/F’AI-1 complex. This design the FP ratio was similar to that in normal control. This may obviated the need to activate latent PAI-1 for a standard, be due to the low specific activity of PAI-1 in DIC patient and the complex is inherently more stable than t-PA- plasma with APL (Fig 4B). unbound active PAI-1 because the complex is acyl-enzyme In DIC patient plasma with APL, marked hyperfibrino- complex. ‘251-labeledt-PNPAI-1 complex dissociated some- lysis in vivo was suggested by the increase of FDP and the what (from about 5% to 35%) under nonreduced condi- decrease of fibrinogen and a2-PI activity, and the level of tions according to the size of the gel and the time of the plasmin/a2-PI complex did not increase (Table 1). Because electrophoresis by Laemmli’s SDS-PAGE method, and this the antigen levels of PAS were not greatly elevated, hyperfi- may be due to nucleophilic attack on the acyl-enzyme brinolysis in DIC with APL may be due to at least two complex by the positively charged Tris ion at high pH as factors: (1) proteolytic cleavage of fibrinogen, &-PI,” and described.38However, less than 5% of the complex dissoci- PAI-1 (Fig 5) by leukocyte elastase or other proteases from ated by the Weber and Osborn SDS-PAGE method or by abnormal promyelocytes; and (2) increased secondary fibrin- gel-filtration chromatography using a TSK-GEL G3000 SW olysis. There are several report^'^,^' of decreased levels of column and a CCP-8010 HPLC system (Toyo Soda, Tokyo, either a2-PI or PAI-1, and the patients had bleeding Japan) after 18 hours at room temperature or 2 months of tendencies. In these cases, the antigen levels of t-PA were storage at -20°C (data not shown). We cannot rule out the normal and treatment of these individuals with epsilon- possibility that a small amount of degraded t-PNPAI-1 aminocaproic acid was effective. This suggestion is further complex may be generated by adding t-PA to the samples. supported by the degraded forms of PAI-1 observed in In fact, the slope of the calibration curve between active plasmas from DIC patients associated with APL and sepsis PAI-1 levels measured with the current assay and those (Fig 5, lanes g and h). In one case (Fig 5, lane h), the with the conventional method in a purified system (Fig 2C: combined effects of endotoxin and APL protease on the slope = 0.65) or in patient plasmas (Fig 2D: slope = 0.57) marked increase of t-PA and PAI-1 and of was smaller than the estimated one (slope = 0.7) (Fig 2C). PAI-1 might result in the appearance of t-PAPAI-1 com- However, the correlation was satisfactory enough, so that plex and the degraded forms of PAI-1.

REFERENCES 1. Groopman J, Ellman L Acute promyelocytic leukemia. Am J effect of heparin in the management of patients with APL. Br J Hematol7:395,1979 Haematol68:283, 1988 2. Goldberg MA, Ginsberg D, Mayer PJ, Stone RM, Maguire 4. Yoshikawa T, Tanaka KR, GuZe LB: Infection and DIC. M, Rosenthal DS, Antin JM: Is heparin administration necessary Medicine 50:237, 1971 during induction chemotherapy for patients with acute promyelo- 5. Preston FE, Malia RG, Sworn MJ, Blackburn EK Intravascu- cytic leukemia? Blood 69:187,1987 lar coagulation and E Coli septicemia. J Clin Pathol26:120,1973 3. Hoyle CF, Swisky DM, Freedman L, Hayhoe FGJ: Beneficial 6. Mimuro J, Sakata Y, Wakabayashi K, Matsuda M: Level of SPECIFIC ACTIVITY OF PAL1 IN DIC WITH APL 1957 protein C determined by combined assays during disseminated ing endothelial-cell type plasminogen activator inhibitor. Proc Natl intravascular coagulation and oral anticoagulation. Blood 69: 1704, Acad Sci USA 83:6776,1986 1987 24. Ranby M, Bergsdorf N, Nillson T, Mellbring G, Winblad B, 7. Sprengers ED, Kluft C: Plasminogen activator inhibitors. Bucht G: Age dependence of tissue plasminogen activator concen- Blood 69:381,1987 trations in plasma, as studied by an improved enzyme-linked 8. Heckman CM, Loskutoff DJ: Endothelial cells produce a immunosorbent assay. Clin Chem 32:160,1986 latent inhibitor of plasminogen activators that can be activated by 25. Frankel ME, Gerhard W The rapid determination of denaturants. J Biol Chem 260:11581,1985 binding constants for antiviral antibodies by a radioimmunoassay. 9. Urden G, Hamsten A, Wiman B: Comparison of plasminogen An analysis of the interaction between hybridoma proteins and activator inhibitor activity and antigen in plasma sample. Clin Chim influenza virus. Mol Immunol16101,1979 Acta 169:189,1986 26. Nakane PK, Kawaoi A: Peroxidase-labeled antibody. A new 10. Declerck PJ, Mol MD, Alessi M, Baudner S, Paques E, method of conjugation. J Histochem Cytochem 22:1084,1974 Preissner KT, Muller-Berghaus G, Collen D: Purification and 27. Sakata Y, Loskutoff DJ, Gladson CL, Hekman CM, Griffin characterization of a plasminogen activator inhibitor 1 binding JH: Mechanism of protein C-dependent clot lysis: Role of plas- protein from human plasma. J Biol Chem 263:15454,1988 minogen activator inhibitor. Blood 68:1218,1986

11. Kruithof EKO, Nicolosa G, Bachman F: Plasminogen activa- 28. Sakata Y, Eguchi Y, Mimuro J, Matsuda M, Sumi Y: Clot Downloaded from http://ashpublications.org/blood/article-pdf/77/9/1949/605395/1949.pdf by guest on 28 September 2021 tor inhibitor-1: Development of a radioimmunoassay and observa- lysis by a monoclonal antibody against a2-plasmin inhibitor. Blood tions on its plasma concentration during venous occlusion and after 74:2692,1989 platelet aggregation. Blood 70:1645,1987 29. Aoki N, Saito H, Kamiya K, Sakata Y, Kobakura M: 12. Declerck PJ, Alessi MC, Verstreken M, Kruithof EKO, Congenital deficiency of a2-plasmin inhibitor associated with Juhan-Vague I, Collen D: Measurement of plasminogen activator severe hemorrhagic tendency. J Clin Invest 63:877,1979 inhibitor 1 in biologic fluids with a murine monoclonal antibody- 30. Laemmli UK Cleavage of structural proteins during the based enzyme-linked immunosorbent assay. Blood 71:220,1988 assembly of the head of bacteriophage T4. Nature 227:680,1970 13. Chmielewska J, Ranby M, Wiman B: Evidence for a rapid 31. Weber K, Osborn M: The reliability of molecular weight inhibitor to tissue plasminogen activator in plasma. Thromb Res determinations by dodecyl sulfate-polyacrylamide gel electrophore- 31:427,1983 sis. J Biol Chem 244:4406, 1969 14. Aznar J, Gilabert J, Estelles A, Espana F: Fibrinolytic 32. Jaffe EA, Hoyer JW, Nachman RL Synthesis of antihemo- activity and protein C in preeclampsia. Thromb Haemost 55:314, philic factor antigen by cultured human endothelial cells. J Clin 1986 Invest 52:2757,1973 15. Stief TW, Lenz P, Becker U, Heimburger N: Determination 33. Mimuro J, Koike Y, Sumi Y, Aoki N: Monoclonal antibodies of plasminogen activator inhibitor capacity of human plasma in to discrete regions in a2-plasmin inhibitor. Blood 69:446, 1987 presence of oxidants: A novel principle. Thromb Res 50:559, 1988 16. Sakata Y, Okada M, Nor0 A, Matsuda M: Interaction of 34. Kenneth A, Rosenberg BD, Rosenberg RD: Thrombin tissue-type plasminogen activator and plasminogen activator inhib- generation in acute promyelocytic leukemia. Blood 64:791,1984 itor 1 on the surface of endothelial cells. J Biol Chem 263:1960, 35. Schleef RR, Sinha M, Loskutoff DJ: Immunoradiometric 1988 assay to measure the binding of a specific inhibitor to tissue type 17. Blomback B, Blomback M: Purification of human and bovine plasminogen inhibitor. J Lab Clin Med 106:408, 1985 fibrinogen. Arkh Kemi 10415,1956 36. Amiral J, Plassart V, Grosley M, Mimilla F, Contont G, 18. Dahlback B, Podack ER: Characterization of human S Guyadert M: Measurement of tPA and tPA-PAI-1 complexes by protein and inhibitor of the membrane attack complex of comple- ELISA, using monoclonal antibodies: Clinical relevance. Thromb ment. Demonstration of a free reactive thio group. Biochemistry Res (suppl7):99,1988 262:9935,1987 37. Dcclerck PJ, Collen D: Measurement of plasminogen activa- 19. Hekman CM, Loskutoff DJ: Bovine plasminogen activator tor inhibitor 1 (PAI-1) in plasma withvarious monoclonal antibody- inhibitor 1: Specificity determinations and comparison of the based enzyme-linked immunosorbent assay. Thromb Res (suppl active, latent and guanidine-activated forms. Biochemistry 27:2911, 10):3, 1990 1988 38. Mauro B, Schleef RR, Espana F, Loskutoff DJ, Griffin JH: 20. Wakabayashi K, Sakata Y, Aoki N: Conformation-specific Interaction of type 1 plasminogen activator inhibitor with the monoclonal antibodies to the -induced structure of protein enzymes of the contact activation system. J Biol Chem 264:11738, C. J Biol Chem 261:11097,1986 1989 21. Bidilingmeyer BA, Cohen SA, Talvin TL: Rapid analysis of 39. Colucci M, Paramo JA, Collen D: Generation in plasma of a amino acids using pre-column derivatization. J Chromatogr 336:93, fast-acting inhibitor of plasminogen activator inhibitor in response 1984 to endotoxin stimulation. J Clin Invest 75:818, 1985 22. Ny T, Elgh F, Lund B: The structure of the human 40. Brower MS, Harpel PC Proteolytic cleavage and inactiva- tissue-type plasminogen activator gene: Correlation of intron and tion of a2-plasmin inhibitor and C1 inactivator by human polymor- exon structures to functional and structural domains. Proc Natl phonuclear leukocyte elastase. J Biol Chem 257:9849, 1982 Acad Sci USA 81:5355,1984 41. Schleef RR, Higgins DL, Pilemer E, Levitt U:Bleeding 23. Ny T, Sadwey M, Lawrence DA, Millan JL, Loskutoff DJ: diathesis due to decreased functional activity of type 1 plasminogen Cloning and sequence of a cDNA coding for the human P-migrat- activator inhibitor. J Clin Invest 83:1747,1989