Tissue Factor Pathway Inhibitor: The Carboxy-Terminus Is Required for Optimal Inhibition of Factor Xa

By Robin Wesselschmidt, Karen Likert, Thomas Girard, Tze-Chein Wun, and George J. Broze Jr

Tissue factor pathway inhibitor (TFPI) is a multivalent Kunitz- type protease inhibitor that binds to and inactivates factor Xa directly, and in a factor Xa-dependent fashion inhibits the factor Vlla/tissue factor catalytic complex. TFPl is a slow, tight-binding, competitive, and reversible inhibitor of factor Xa, in which the formation of an initial encounter complex between TFPl and factor Xa is followed by slow isomerira- tion t o a final, tightened complex. Wild-type recombinant TFPl (rTFPI), expressed in mouse C127 cells, separates into two forms on heparin-agarose chromatography that elute at 0.3 mol/L and 0.6 mol/L NaCI. Western blot analysis shows that both forms contain the N-terminus of full-length TFPI, but only rTFPI(0.6) is recognized by an antibody directed against the C-terminus. rTFPl(0.3) and rTFPI(0.6) inhibit fac- tor Xa with 1:l stoichiometry and inhibit factor Vlla/tissue factor equally in an endpoint-type assay. However, rTFPI(0.6) is a more potent inhibitor than rTFPl(0.3) of coagulation in normal plasma induced by either factor Xa or tissue factor.

LASMA CONTAINS an inhibitor of tissue factor- P induced coagulation, previously referred to as lipopro- tein-associated coagulation inhibitor (LACI) or extrinsic pathway inhibitor (EPI) and most recently termed tissue factor pathway inhibitor (TFPI).’.* TFPI binds to and inhibits factor Xa directly, and inhibits the factor VIIa/ tissue factor complex in a factor Xa-dependent manner.3 TFPI is an approximately 40-Kd glycoprotein that contains a cluster of acidic amino acids at its amino-terminus, followed by three tandem Kunitz-type protease inhibitory domains, and a cluster of basic amino acids near its carboxy-terminu~.~ The first Kunitz domain of TFPI is required for the inhibition of the factor VIIa/tissue factor complex and the second Kunitz domain of TFPI is responsi- ble for the inhibition of factor Xa.’

TFPI is a slow, tight-binding, competitive, and reversible inhibitor of factor Xa.6 Typical of Kunitz-type protease inhibitors, its mechanism of action appears to conform to the kinetic model:

d O W

E + I e E1 e EI*

From the Division of HematologylOncology, The Jewish Hospital at Washington University Medical Center, St Louis, MO; and Monsanto Corporate Research, Chesterfield, MO.

Submitted June 28, 1991; accepted December 9, 1991. Supported inpart by Grants No. HL-34462 and HL-I4147from the

National Institutes of Health, Bethesda, MD, and The Monsanto Chemical Corporation, St Louis, MO.

Presented at the Intemationnl Society on Thrombosis and Haemo- stasis meeting in Amsterdam, The Netherlands, July, 1991.

Address reprint requests to George J. Broze, Jr, MD, Division of HematologyiOncology, The Jewish Hospital of St Louis, 216 S Kingshighway Blvd, St Louis, MO 63110.

The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. section I734 solely to indicate this fact.

0 I992 by The American Society of Hematology. 0006-4971 1921 7908-0032$3.00/0

The initial inhibition of factor Xa (<5 seconds) produced by rTFPI(0.6) is several-fold greater than that produced by rTFPl(0.3). presumably reflecting a lower Ki of the immediate encounter complex between factor Xa and TFPI. The differen- tial effect of these forms of TFPl on tissue factor-induced coagulation in normal plasma appears to be directly related to their ability to inhibit factor Xa. To confirm the role of the C-terminal region of TFPl in optimal factor Xa inhibition, a carboxy-terminal mutant of rTFPI, which is truncated after leucine 252 and thus lacks the basic sequence K T K R K R K K Q R V K (residues 254-265). was expressed in C127 cells. This form of rTFPl elutes from heparin-agarose at 0.28 mol/ L NaCl and inhibits factor Xa at a rate that is slower than rTFPl(O.3). The Ki(final)s for factor Xa inhibition by rTFPI(O.6). rTFPl(O.3). and rTFPI,,,, are 3.1 k 0.6,19.6 z?: 0.8, and 19.6 f 3.0 pmol/L, respectively. 0 1992 by The American Society of Hematology.

where E = enzyme, I = inhibitor, and E1 represents an immediate encounter complex described by a Ki(initial), which slowly isomerizes to a final tightened-complex EI* characterized by a Ki(final).6 The kinetic term “tight- binding” refers to the fact that inhibition occurs at concen- trations of inhibitor approaching that of the enzyme itself and “slow” means that the final degree of inhibition does not occur immediately.

While investigating TFPI’s interaction with heparin, we found that purified recombinant TFPI (rTFPI), produced in mouse C127 cells, separates into two forms on heparin- agarose chromatography and that these forms possess markedly different inhibitory potencies against factor Xa and tissue thromboplastin-induced coagulation of normal plasma. This study examines the possible structural basis for the functional differences between these forms of rTFPI.

MATERIALS AND METHODS

Proteins

rTFPI was immunoaffinity purified from the conditioned media of mouse C127 cells that had been transfected using a bovine papilloma virus containing wild-type TFPI cDNA.’ In brief, serum- free conditioned media was concentrated (YM 30; Amicon, Dan- vers, MA) and fractionated by ammonium sulfate precipitation. Proteins precipitated between 23% and 90% saturation of ammo- nium sulfate were applied to a monoclonal anti-TFPI agarose column and, after washing of the column, the rTFPI was eluted with 1.5 mol/L NaSCN. The pooled rTFPI was dialyzed against 0.15 mol/L NaC1, 0.02 mol1L Na,SO, and frozen at -70°C. Thrombin (2,500 Ulmg) was produced from purified prothrombin? Recombinant factor VI1 and factor VIIa were from Novo-Nordisk (Gentofe, Denmark). Bovine and human factor X were isolated as previously described, and activated using the factor X coagulant protein (XCP) from Russell’s Viper venom, which had been insolubilized on agarose beads? XCP was purified from crude viper venom as previously described.” Crude, ethylenediamine tetraace- tic acid (EDTA)-washed, tissue thromboplastin from human brain was used as the source of tissue factor? The concentration of rTFPI was determined by analysis of amino acid composition; the addi-

2004 Blood, Vol79, No 8 (April 15), 1992: pp 2004-2010

C-TERMINALLY TRUNCATED TFPI 2005

tion of norleucine served as the internal control for recovery. Based on the predicted amino acid sequence of TFPI, rTFPI(0.6) was assumed to contain 34 Asx (asparagine + aspartic acid resi- dues), and the carboxy-terminal truncated forms, rTFPI(0.3) and rTFPI,.,,, were assumed to contain 33 Asx.4 The concentrations of bovine and human factor Xa were determined by active-site titration.’’

Heparin-Agarose Chromatography

A 1 x 3 cm heparin-agarose (Sigma Chemical Co, St Louis, MO) column was equilibrated in 50 mmol/L Tris-HCI, pH 7.5/100 mmol/L NaCl (TS), then 30 pg of rTFPI was applied and the column was developed with a linear NaCl gradient from 0.1 mol/L to 1.0 mol/L, at 15 mL/h over 10-column volumes. The two forms of rLACI, eluting at 0.3 mol/L and 0.6 mol/L NaCl, were pooled separately and concentrated using Centricon microconcentrators (Amicon).

Assays

Factor Xa-Induced Coagulation of Plasma In a fibrometer (BBL, Cockneysville, MD), 50 pL of rabbit brain

cephalin, prepared as described by the manufacturer (Sigma), 50 pL of CaC1, (25 mmol/L), and 50 pL of human factor Xa (1 nmol/L) are incubated at 37°C. After 30 seconds, 100 pL of a 50/50 mixture of TFPI sample and normal human (George King Biomed- ical, Overland Park, KS) is added, and the degree of apparent factor Xa inhibition is determined by comparison of the clotting time to a standard curve of factor Xa. The results are expressed as the mean of duplicate measurements.

Tissue Factor-Induced Coagulation of Plasma In a fibrometer, 50 pL of rabbit brain cephalin, 50 pL of CaCI,

(25 mmol/L), 50 pL human brain thromboplastin (1:50 dilution), and 50 pL of TFPI sample in 50 mmol/L Tris-HCl, pH 7.5/100 mmol/L NaCl/O.l% bovine serum albumin (TBSA) are incubated at 37°C. After 30 seconds, 50 pL of normal human plasma is added, and the degree of apparent tissue factor inhibition is determined by the comparison of the clotting time to a standard curve of thromboplastin. The results are expressed as the mean of duplicate measurements.

Thrombin-Induced Coagulation of Plasma In a fibrometer, 50 pL rabbit brain cephalin, 50 pL CaCI, (25

mmol/L), 50 pL thrombin (2.5 U/mL), and 50 pL of TFPI sample in TBSA are incubated at 37°C. After 30 seconds, 50 pL normal human plasma is added, and the degree of apparent thrombin inhibition is determined by the comparison of the clotting time to a standard curve of thrombin. The results are expressed as the mean of duplicate measurements.

XCP-Induced Coagulation of Plasma In a fibrometer, 50 pL of rabbit brain cephalin, 50 pL of CaCI,

(25 mmol/L), 50 pL XCP (0.63 nmol/L), and 50 pL of TFPI sample in TBSA are incubated at 37°C. After 30 seconds, 50 pL of normal human plasma is added and the degree of inhibition is determined by comparison of the clotting time to a standard curve of XCP. The results are expressed as the mean of duplicate measurements. TFPI did not inhibit XCP activation of factor X in control experiments based on the cleavage of ’H-factor X, in which XCP was substituted for factor VIIa/tissue factor (see below).

Stoichiometry of Factor Xa-TFPI Binding Active-site-titrated bovine factor Xa (1 nmol/L) is incubated

with TFPI (0.1 to 0.6 nmol/L). After 60 minutes, the remaining factor Xa activity is determined by measuring the initial dA, after the addition of methoxy carbonyl-D-cyclohexylglycyl-glycyl-arginine- p-nitroanilide acetate (Spectrozyme Xa; American Diagnostica Inc, Greenwich, CT) to a final concentration of 100 pmol/L.

Factor Xa Inhibition by TFPl In a fibrometer, 50 pL of rabbit brain

cephalin, 50 pL of CaCI, (25 mmol/L), 50 pL of human factor Xa (1 nmol/L), and 50 pL of the TFPI sample in TBSA are incubated at 37°C. After various periods of time (5 seconds to 10 minutes), 50 pL of factor X-deficient human plasma (George King Biomedical) was added, and the degree of apparent factor Xa inhibition is determined by comparison of the clotting time to a standard curve of factor Xa.

In a cuvette, rTFPI (2.5 nmol/L) in TBSA and Spectrozyme Xa (100 Fmol/L) are mixed. The reaction is initiated by the addition of human factor Xa (0.5 nmol/L), and the b5 is continuously plotted. Heparin (1 U/mL; Elkins-Sinn, Inc, Cherry Hill, NJ) was added in certain reactions.

Coagulation assay.

Continuous chromogenic assay.

Factor VIIa 1 Tissue Factor Inhibition by TFPI This assay is a modification of a previously

described procedure and uses the release of ’H-labeled activation peptide of human factor X to measure residual factor VIIa/tissue factor activity after incubation with TFPI? Twenty-five microliters of TFPI sample in TBSA and 50 pL of a mixture containing factor VIIa (8 nmol/L), human factor X (40 nmol/L), CaC1, (8 mmol/L), and tissue thromboplastin (1%) are incubated for 30 minutes at room temperature. Then, 25 pL ’H-factor X (400 nmol/L) is added and exactly 10 minutes later the reaction is stopped by the addition of 200 pL of ice-cold 7.5% trichloroacetic acid (TCA), mixed, and placed on ice. After centrifugation at 14,OOOg for 5 minutes, 200 pL of the supernatant is added to 4 mL of Cytoscint (ICN, Irvine, CA) and counted in a beta-counter. The results are expressed as the mean of duplicate measurements.

To test the effect of the forms of rTFPI on the initial rate of factor VIIa/tissue factor inhibition, 300 pL of TFPI (2 to 20 nmol/L) sample in TBSA, 300 pL factor VI1 or VIIa (20 nmol/L), 300 pL of thromboplastin (2%), and 300 pL ’H-factor X are mixed and incubated at room temperature. At certain times (0.5 to 30 minutes), 100 pL samples of the reaction mixture are added to 200 pL cold 7.5% TCA and the release of the factor X activation peptide determined as detailed above.

Endpoint assay.

Initial rate assay.

Determination of &(final) The Ki(fina1) for each form of TFPI was determined using the

method of Bieth.” Human factor Xa (100 pmol/L) and TFPI (50 to 250 pmol/L) were incubated at room temperature in 950 pL of TBSA. After 60 minutes, 50 pL of the chromogenic substrate, Spectrozyme Xa, was added to a final concentration of 500 pmol/L, and the dA, was followed spectrophotometrically until it was constant. The fractional protease activity remaining (a) was plotted against inhibitor concentration (Io) and the data were fitted by nonlinear regression analysis (Sigma Plot; Jandel Scientific, Corte Madera, CA) to the equation:

a = 1 - E, + I, + Ki,,, - [(E, + I, + Ki,,,)’ - 4EoI,]’”/2E,

This led to the direct estimation of Kiapp. Ki values were then

2006

0.4

0 00

2

0

/

IO 15 20 25 FRACTION NUMBER

Fig 1. Heparin-agarose chromatography of rTFPl purified from mouse C127 calls. The chromatography was performed as described in Materials and Methods; Am and the NaCl gradient are shown.

derived by correction for the concentration of substrate using the following equation:

Ki = Kiapp/(l + S/K,,,)

where S is the substrate concentration and its K,,, for human Xa was experimentally determined to be 130 pmol/L. The results are expressed as the mean f standard deviation of three measure- ments.

Immunoblot analysis Western blotting after sodium dodecyl sulfate (SDS)-polyacryl-

amide gel (15%) electrophoresis was performed as previously described’ using primary antibodies raised in rabbits against whole TFPI or one of the following synthetic peptides: D S E E D E E H T I, TFPI,,, (NTP); and K I A Y E E I F V K N M, TFPIm,,,, (CTP). The peptides were crosslinked to keyhole limpet hemocyanin using N-succinimidyl bromoacetate’* before their use as immunogens, and antibodies were raised in rabbits using the method of Viatukai- tis.” A murine monoclonal anti-TFPI antibody,I4 linked to agarose (2 mg/mL), was used to immunoprecipitate rTFPI from the conditioned media of mouse C127 cells. Twenty microliters of anti-TFPI agarose beads was added to 500 )LL of conditioned media and the mixture was rocked for 2 hours. The beads were then collected by centrifugation, washed with saline, and the bound TFPI eluted by the addition of 60 )LL of SDS-polyacrylamide gel electrophoresis sample buffer and heating to 100°C for 5 minutes. The beads were pelleted by centrifugation and the supernatant sample used for Western blotting.

Construction and Expression of rTFPI,.252 A plasmid, pUCGB9R1, containing the cDNA of TFPI, which

had been modified for insertion into a mammalian expression vector,” was used to construct this mutant. Briefly, the 3’ untrans- lated region and the last 75 bp of the open reading frame of TFPI cDNA was removed by restriction digestion of pUCGB9R1 with Stu I and HindIII. The synthetic double-stranded DNA (pCCTAT- GATGATAGA, GGATACTACTATC’ITCGAp) was ligated into the Sfu IIHindIII-digested plasmid, yielding a TFPI cDNA that encodes a TFPI molecule that is truncated after leucine,,. The sequence of the construct was confirmed using dideoxy-chain termination analysis.’6 The mutant-TFPI cDNA was cloned into the pMON1123 expression vector and cotransfected with pSV2neo

IOC

75

5c

25

WESSELSCHMIDT ET AL

A

LAC1 (nM)

B

I I

2oL Oh5 d.1 0.12 0.5 1.0

LAC1 (nM)

Fig 2. Interactions of the forms of rTFPl with factor Xa and the factor Vlla/tissue factor complex. (A) Stoichiometry of factor Xa:TFPI binding. Bovine factor Xa (1 nmol/L) and rTFPl (0.1 to 0.6 nmol/L) were incubated for 1 hour at room temperature and the remaining factor Xa activity was determined by chromogenic assay. (A) Stock C127 TFPI; (0) rTFPl(O.3); (0) rTFPI(0.6). (6) Inhibition of factor Vlla/tissue factor by rTFPl in an endpoint assay (see Materials and Methods). The Wfactor X activation peptide released (y axis) is directly related t o remaining factor Vlla/tissue factor activity. (0) rTFPl(O.3); (0) rTFPI(0.6).

Table 1. Inhibition of Coagulation in Plasma by rTFPl

rTFPI(O.6) rTFPI(0.3) rTFPI,.,,,

1 nmollL 4 nmol/L 1 nmol/L 4 nmol/L 1 nmol/L 4 nmol/L

Factor Xa 15 46 0 6 0 5 Tissuefactor 20 59 0 6 0 5 XCP 23 61 0 8 0 6

Percent apparent inhibition of factor Xa, tissue factor, and XCP induced coagulation of plasma by the forms of rTFPl (see Materials and Methods). Each rTFPl was tested at concentrations of 1 and 4 nmol/L.

C-TERMINALLY TRUNCATED TFPI 2007

TIME (min)

Fig 3. Inhibition of factor Xa by rTFPl in a coagulation assay. The time course of human factor Xa inhibition by the forms of rTFPl was determined as described in Materials and Methods. (A) Control (no TFPI); (0) rTFPl(O.3) 1 nmol/L; (0) rTFPl(0.3) 4 nmol/L; (*I rTFPI(0.6) 1 nmol/L; (0) rTFPI(0.6) 4 nmol/L.

into C127 mouse mammary tumor cells by CaPO, precipitation." After selection in G418, a stable clone was isolated and expanded. The rWPI,.,, was isolated from conditioned media using immunoaf- finity chromatography.'

RESULTS

Separation of rTFPl Into Two Forms by Heparin-Agarose Chromatography

rTFPI, immunoaffinity purified from the conditioned media of mouse C127 cells,' separates into two protein peaks after heparin-agarose chromatography (Fig 1). These peaks elute at 0.3 mol/L and 0.6 moI/L NaCl and are

0.2

m 0 8 0.1

0

referred to as rTFPI(0.3) and rTFPI(0.6), respectively. Each form of TFPI binds to factor Xa with a 1:l stoichiom- etry (Fig 2A) and they inhibit factor VIIa/tissue factor to the same extent in an endpoint assay (Fig 2B). rTFPI(0.6), however, is a considerably more potent inhibitor than rTFPI(0.3) of coagulation in plasma induced by factor Xa or tissue factor when measured in single-stage coagulation assays (Table 1). Neither form of TFPI inhibited thrombin- induced coagulation of plasma (data not shown).

Closer examination of the rate of inhibition of factor Xa by rTFPI(0.3) and rTFPI(0.6) in coagulation assays shows that the initial degree ( < 5 seconds) of factor Xa inhibition by rTFPI(0.6) is several-fold greater than that of rTFPI(0.3) (Fig 3). Studies using the continuous cleavage of a chromogenic substrate (Spectrazyme Xa) to measure remaining factor Xa activity confirm this phenomenon (Fig 4A), and demonstrate a disparate response to heparin (1 U/mL) for the inhibition of factor Xa by the two forms of rTFPI. While heparin enhances the rate of inhibition of factor Xa by rTFPI(0.3), it appears to modestly slow the inhibition produced by rTFPI(0.6) (Fig 4B). The Ki(fina1)s for rTFPI(0.3) and rTFPI(0.6) against human factor Xa are 19.6 f 0.8 and 3.1 f 0.6 pmol/L, respectively."

The greater inhibition of tissue factor-induced coagula- tion in normal plasma produced by rTFPI(0.6) as compared with rTFPI(0.3) in the single-stage coagulation assay ap- pears to be due to its superior ability to inhibit factor Xa rather than a differential effect on factor VIIa/tissue factor. For the individual forms of TFPI, the relative inhibition of coagulation in plasma is similar whether clotting is induced by factor Xa or tissue factor (Table 1). Further, coagulation induced by the factor X coagulant protein of Russell's Viper venom (XCP) is inhibited by each form of rTFPI in a nearly identical manner as tissue factor-induced coagula- tion (Table 1). However, additional experiments have failed to detect a difference between the rTFPI(0.3) and rTFPI(0.6) in the initial rates of factor VIIa/tissue factor inhibition in the presence of factor X (data not shown).

1

IO 20 0 IO 20 0 10 20 TIME (min)

Fig 4. Inhibition of factor Xa by rTFPl in a continuous chromogenic assay. Reactions were initiated by adding human factor Xa (0.5 nmol/L final Concentration) to rTFPl (2.5 nmol/L final) and Spectrozyme Xa (100 pmol/L final). In each panel (0) denotes cleavage of the substrate in the absence of rTFPI. (A) (0) rTFPl(O.3); (0) rTFPI(0.6). (B) Same as (A), except the reactions were performed in the presence of heparin (1 U/mL). (C) rTFPI,, in the absence (AI and presence (A) of heparin (1 U/mL).

2008 WESSELSCHMIDT ET AL

2 3

97 - 66 -

I

31 - 1 21 - ; 14- 1

Fig 5. SDS-PAGE of rTFP1. Each lane contains 10 pmol of rTFPI. Lane 1, stock C127 rTFPI; lane 2, rTFPl(0.3); lane 3, rTFPI(0.6). The gel was developed by silver staining.

Structural Differences Between rTFPl(O.3) and rTFPI(0.6) and the Expression of a Recombinant Carboxy-Truncated Form of TFPI

Rabbit polyclonal antibodies, directed against the whole rTFPI protein or specific synthetic peptides based on the amino-terminal (NTP) and carboxy-terminal (CTP) amino acid sequence of TFPI, were used to examine the structural differences between rTFPI(0.3) and rTFPI(0.6). SDS- polyacrylamide gel electrophoresis (SDS-PAGE) (Fig 5) and Western blot analyses (Fig 6) shows that TFPI(0.3) migrates slightly faster than TFPI(0.6) and that while both forms of rTFPI are recognized by antibodies against the

A B 1 2 3 1 2 3

amino-terminus of the TFPI molecule, only rTFPI(0.6) is detected by antibodies against the carboy-terminus (Fig 6). Similar reduction in size and loss of carboxy-terminal staining occur when purified rTFPI(0.6) is added to the media of C-127 cells in tissue culture (Fig 7). It has not been determined whether the apparent proteolytic truncation of full-length TFPI that produces rTFPI(0.3) occurs through specific cleavage at a single-site or multiple cleavages at alternative, though closely-spaced sites.

To confirm the role of the carboxy-terminus of TFPI in the optimal inhibition of factor Xa, an rTFPI, truncated following Leu,, (rTFPI,.,,), was expressed in mouse C127 cells. This form of rTFPI lacks the last 24 amino acids of full-length TFPI, which includes the basic sequence K T K R K R K K Q R V K (TFPI residues 254-265). rTFPI,.,, binds factor Xa with 1:l stoichiometry and produces factor VIIa/TF inhibition equivalent to the other forms of rTFPI in the end stage assay (not shown). rTFPI,.,, elutes from heparin-agarose at a slightly lower NaCl concentration (0.28 mol/L) than rTFPI(0.3) (not shown); inhibits coagula- tion in plasma induced by factor Xa, tissue factor, or XCP to an extent that is slightly less than rTFPI(0.3) (Table 1); and inhibits factor Xa at a rate somewhat slower than rTFPI(0.3) (compare Fig 4A with C). Its Ki(fina1) against factor Xa is 19.6 2 3.0 pmol/L and the rate of inhibition of factor Xa by rLACI,.,,, like that produced by rTFPI(0.3). is enhanced in the presence of heparin (1 U/mL) (Fig 4C).

DISCUSSION

These studies show that the carboxy-terminus of TFPI is required for optimal inhibition of factor Xa. The relative affinities of rTFPI,.,,, rTFPI(0.3), and rTFPI(0.6) for heparin-agarose parallel the rates at which they inhibit factor Xa. This suggests that both of these properties are related to the number of basic amino acids retained at the carboy-terminus of the TFPI molecule. It also implies that it is the highly basic sequence (TFPI,,,) within the

C 1 2 3

i 110 - 84 - 47 -

33 -

24 - 16 -

Fig 6. Western blot analysis of rTFPI. In the immunoblots shown In (A), (B), and (C) each lane contains 10 pmol of rTFPI. Lane 1, stock Cl27 rTFPI; lane 2, rTFPl(O.3); lane 3, rTFPl(O.6). Rab- bit polyclonal anti-TFPI antibod- ies were used as follows: (A), anti-whole TFPI; (B), anti-CTP; (C), anti-NTP.

C-TERMINALLY TRUNCATED TFPI 2009

Fig 7. Truncation of rTFPl(O.6) In the media of cultured Cl27 cells. C127 cells were grown to confluence in the wells of a 24-well cluster plate. The media was then replaced with serum-free media or serum-free media containing rTFPI(0.6) (25 pmollwell). A t the indi- cated time, the rTFPI in the conditioned media from a well was immunoprecipitated (see Materials and Methods). Half the sample was used for immunoblotting with rabbit polyclonal anti-whole TFPI (A) and the other half of the sample was used for immunoblotting with rabbit polyclonal anti-CTP (6). Lane 1, media to which no rTFPI(0.6) had been added; lane 2, media with rTFPl(O.6) after 48 hours of culture; lane 3, media with rTFPl(0.6) after 16 hours of culture; and lane 4, media with rTFPl(0.6) incubated in the absence of cells.

carboxy-terminus that may be required for the enhance- ment of the rate of factor Xa inhibition.

The greater factor Xa inhibition produced immediately (<5 seconds) after the addition of rTFPI(0.6), as opposed to rTFPI(0.3), to factor Xa (Fig 3), suggests that the difference in the ability of these forms of rTFPI to inhibit factor Xa-induced coagulation is in large part due to a lower Ki(initia1) for the immediate encounter complex of full-length TFPI with factor Xa. The mechanism for this effect remains to be defined, but it could be related to a direct interaction between factor Xa and the carboxy- terminus of TFPI. Alternatively, it may involve a particular conformation of the TFPI molecule, which requires the presence of the carboxy-terminus.

It is intriguing that the factor Xa inhibition by carboxy- terminal truncated rTFPI, but not full-length rTFPI, is enhanced by heparin. The reason for this disparate effect of heparin is not clear at present and is the subject of further

investigations. The previously reported &(final) for factor Xa of endogenously produced HepG2 TFPI is 30-fold greater than that reported here for mouse C127-derived rTFPI.6 This disparity could be related to intrinsic differ- ences in the posttranslational processing of TFPI by these two types of cells. In view of the results presented here, however, unrecognized proteolytic degradation of the HepG2 TFPI during cell culture and/or during the purifica- tion of HepG2 TFPI is an equally plausible explanation for this difference.

Because factor Xa binding is a prerequisite for the inhibition of factor VIIa/tissue factor by TFPI, we expected the more rapid inhibition of factor Xa produced by rTFPI(0.6) to be associated with a faster rate of inhibition of the factor VIIa/tissue factor complex. However, at the concentrations of rTFPI tested (0.5 to 5.0 nmol/L), no difference between the initial rates of factor VIIa/tissue factor inhibition by rTFPI(0.6) and rTFPI(0.3) or rTFPI,.,, could be detected in a purified system using factor VI1 or factor VIIa, tissue factor, and factor X. Furthermore, all forms of the rTFPI inhibited XCP-initiated coagulation in the same fashion as they inhibited tissue factor-induced coagulation, providing strong evidence that it is an effect on factor Xa rather than on the factor VIIa/tissue factor complex that mediates the apparent inhibition of tissue factor-induced coagulation in the one-stage assay.

Partial carboy-terminal truncation of the rTFPI ex- pressed by transfected C127 cells can be detected in their conditioned media (unpublished observations), and similar apparent cleavage of full-length rTFPI(0.6) occurs when it is added to cultures of nontransfected C127 cells (Fig 7). The enzyme(s) responsible for this apparent proteolytic truncation of rTFPI has not been identified. Further, it is conceivable that at least a portion of the truncated rTFPI found in tissue culture may be produced by posttransla- tional processing within the host C127 cells themselves, perhaps through the action of a dibasic endoprotease." Carboxy-terminal truncation of rTFPI likely explains the recent observation that rTFP1 produced in baby hamster kidney cells is a less potent inhibitor of coagulation in one-stage assays than heparin-releasable TFPI in plasma,'* which has been shown to be full-length.14

If a similar proteolytic process occurs in vivo, it could provide a means of modulating TFPI activity. Furthermore, consideration of the differential inhibitory activities of full-length and truncated TFPI against factor Xa may be important in studies investigating the therapeutic value of TFPI," because one might expect full-length TFPI to be a more potent antithrombotic agent in vivo.

REFERENCES 1. Broze GJ Jr, Warren LA, Girard JJ, Miletich JP: Isolation of

the lipoprotein-associated coagulation inhibitor produced by HepG2 (human hepatoma) cells using bovine factor Xa affinity chromatog- raphy. Thromb Res 48:253,1987

2. Rao LVM, Rapaport SI: Studies of a mechanism inhibiting the initiation of the extrinsic pathway of coagulation. Blood 69645, 1987

3. Broze GJ Jr, Warren LA, Novotny WF, Huguchi DA, Girard JJ. Miletich JP: The lipoprotein-associated coagulation inhibitor

that inhibits the factor VII-tissue factor complex also inhibits factor Xa: Insight into its possible mechanism of action. Blood 71:335,1984

4. Wun TC, Kretzmer KK, Girard TJ, Miletich JP, Broze GJ Jr: Cloning and characterization of a cDNA coding for the lipoprotein- associated coagulation inhibitor shows that it consists of three tandem Kunitz-type inhibitory domains. J Biol Chem 263:6001, 1988

5. Girard TJ, Warren LA, Novotny WF, Likert KM, Brown SG,

2010 WESSELSCHMIDT ET AL

Miletich JP, Broze GJ Jr: Functional significance of the Kunitz- type inhibitor domains of lipoprotein-associated coagulation inhib- itor. Nature 338:518, 1989

6. Broze GJ Jr, Girard TJ, Novotny W F Regulation of coagula- tion by a multivalent Kunitz-type inhibitor. Biochemistry 29:7539, 1990

7. Day KC, Hoffman LC, Palmier MO, Kretzmer KK, Haung MD, Pyla EY, Spokas E, Broze GJ Jr, Warren TG, Wun TC: Recombinant lipoprotein-associated coagulation inhibitor inhibits tissue thromboplastin-induced intravascular coagulation in the rabbit. Blood 76:1538,1990

8. Esmon C Function of factor V in prothrombin activation. Doctoral thesis, Washington University, St Louis, MO, 1974

9. Broze GJ Jr, Majerus PW: Purification and properties of human coagulation factor VII. J Biol Chem 255:1242,1980

10. Chase T Jr, Shaw E: Titration of trypsin, plasmin, and thrombin with p-nitrophenyl p’-guanidinobenzoate HCI. Methods Enzymol 19:20, 1970

11. Bieth JG: In vivo significance of kinetic constants of protein proteinase inhibitors. Biochem Med 32:387, 1984

12. Bernatowicz MS, Matsueda GR: Preparation of peptide- protein immunogens using N-succinimidyl bromoacetate as a heterobifunctional crosslinking reagent. Anal Biochem 155:95, 1986

13. Vaitukaitis J L Production of antisera with small doses of immunogen: Multiple intradermal injections. Methods Enzymol 73:46,1981

14. Novotny WF, Palmier M, Wun TC, Broze GJ Jr, Miletich JP: Purification and properties of heparin-releasable LACI. Blood 78:394, 1991

15. Girard TJ, Warren LA, Novotny WF, Bejcek BE, Miletich JP, Broze GJ Jr: Identification of the 1.4 kb and 4.0 kb messages for the lipoprotein-associated coagulation inhibitor and expression of the encoded protein. Thromb Res 55:37,1989

16. Sanger F, Nicklen S, Coulson AR: DNA sequencing with chain-terminating inhibitors. Proc Natl Acad Sci USA 74:5463, 1977

17. Barr PJ: Mammalian subtilisins: The long-sought dibasic processing endoproteases. Cell 66:1,1991

18. Lindahl AK, Abildgaard U, Larsen ML, Staalesen R, Ham- mer AKG, Sandset PM, Nordfang 0, Beck TC: Extrinsic pathway inhibitor (EPI) released to the blood by heparin is a more powerful coagulation inhibitor than is recombinant EPI. Thromb Res 62:607,1991

19. Haskell BJ, Torr SR, Day KC, Palmier MO, Wun TC, Soble BE, Abendschein DR: Prevention of arterial re-occlusion after thrombolysis with recombinant lipoprotein-associated coagulation inhibitor. Circulation 84:821,1991