Estrogen regulation of a tissue factor-like procoagulant in the immature rat uterus.

0013-7227/90/1261-0176$02.00/0 Endocrinology Copyright © 1990 by The Endocrine Society

Vol. 126, No. 1 Printed in U.S.A.

Estrogen Regulation of a Tissue Factor-Like Procoagulant in the Immature Rat Uterus* ELENA E. JAZINt, HERBERT W. DICKERMAN, AND KATHERINE P. HENRIKSON Wadsworth Center for Laboratories and Research, New York State Health Department and State University of New York Graduate School of Public Health, Albany, New York 12201-0509

are consistent with identification of the procoagulant as tissue factor. mRNA for tissue factor was increased in the uterus 3 h

ABSTRACT. An estrogen-responsive procoagulant activity is present in the plasma membrane fraction of immature rat uterus.

This procoagulant has many of the properties of tissue factor, a widely occurring, integral membrane protein which initiates the intrinsic pathway of coagulation. Procoagulant activity was demonstrated to activate prothrombin in rat uterus, to activate human coagulation factor X, and to cause clot formation by human plasma. Procoagulant activity could be solubilized from the plasma membrane by the detergent octyl glucoside and had an apparent mol wt of 20,000-40,000 by gel filtration. Procoagulant activity was increased 4-fold within 3 h after immature rats were injected with estradiol. The increase was tissue- and hormone specific and was not affected by a warfarininduced vitamin K deficiency. Coagulation factor VII was required for clot formation by the procoagulant. These properties

T

HE IMMATURE rat uterus responds to a single injection of estrogen with immediate increases in wet weight and protein content (3-5 h), nucleic acid and protein synthesis, and the initiation of cell division within 18-24 h. An increase in prothrombin content is one of the early responses (1). It occurs within 3 h of estrogen administration and coincides with the imbibition of water and influx of plasma proteins. We studied an estrogen-responsive uterine proenzyme of a serine proteinase (2, 3) and have proved in the preceding paper that it is prothrombin (1). Increased levels of prothrombin were detected because an endogenous activator also increased in response to estrogen treatment of the animal. The activator is an integral plasma membrane protein (4). After showing that the proenzyme was prothrombin, we decided to examine the proteins of the Received September 5, 1989. Address all correspondence and requests for reprints to: Katherine P. Henrikson, Ph.D., Wadsworth Center for Laboratories and Research, New York State Department of Health, Empire State Plaza, P.O. Box 509, Albany, New York 12201-0509. * This work is taken in part from a thesis presented to the State University of New York Graduate School of Public Health in partial fulfillment of the requirements for the degree of Doctor of Philosophy in Biomedical Sciences. It was supported in part by NSF Grant DCB8710087. t Present address: Fundacion Campomar, Antonio Machado 151, 1405 Buenos Aires, Argentina.

after estrogen stimulation.

In the preceding paper we showed that prothrombin is increased in the immature uterus within 3 h of estrogen stimulation. The presence of increased amounts of a tissue factor-like procoagulant in the same time period suggests a functional relationship between these two proteins and a possible role for both in uterine development. Thrombin is a growth factor in fibroblasts and endothelial cells. We propose that after estrogen stimulation, prothrombin enters the uterus with the influx of plasma proteins and is activated by the procoagulant to thrombin. We suggest that thrombin might act as a paracrine factor early in the estrogen-stimulated development of the uterus. (Endocrinology 126: 176-185, 1990)

coagulation cascade to identify the estrogen-regulated activator. We show here that the activator has many of the properties of tissue factor (TF) (5), the widely occurring integral membrane protein which initiates the intrinsic pathway of coagulation. To describe the role of the uterine membrane component in prothrombin activation we must review the early steps in coagulation (see Fig. 1) (6, 7). TF is normally masked on the cell surface. When it is exposed, TF binds factor Vila and factor X in a ternary complex. Factor Xa is formed in a calcium-dependent reaction, dissociates from the complex, and associates in a second membranebound complex with factor Va, calcium, and prothrombin. In this prothrombinase complex, factor Xa catalyzes the formation of thrombin. TF is unique among coagulation factors because it is not exposed to circulating coagulation factors, and it has no 7-carboxyl glutamate (Gla) residues as do prothrombin and factors VII and X. TF can be described as a membrane bound cofactor for factor VII, since it has no enzymatic activity of its own. It has a wide tissue distribution; brain and placenta have relatively high concentrations of TF (6). Expression of TF is modulated in mouse 3T3 cells by platelet-derived growth factor or by the addition of serum to serumstarved cells where increased TF mRNA was present

176

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 13 August 2016. at 15:59 For personal use only. No other uses without permission. . All rights reserved.

ESTROGEN-RESPONSIVE UTERINE PROCOAGULANT early in the growth cycle of stimulated cells and was superinduced by cycloheximide (8). Estrogens have not been implicated previously as regulators of TF. The presence of TF as a procoagulant can be verified by the use of coagulation factor-deficient plasmas. TF is equally effective in clotting whole plasma or plasma deficient in factors not involved in its pathway (such as factor XII), but it cannot clot plasma deficient in factor VII (6, 7). In addition, the TF-factor Vila complex can activate factor X in a reaction measured independently of coagulation. We examined uterine membrane fractions from control and estrogen-treated rats for their ability to generate thrombin, form clots, and activate factor X to Xa. Our results demonstrate the presence of an estrogen-regulated integral plasma membrane procoagulant with many of the properties of TF.

Materials and Methods N-(p-Tosyl)L-arginine-3[H]methyl ester ([3H]TAME) was purchased from Amersham (Arlington Heights, IL) and purified as described (1). Liquifluor was obtained from New England Nuclear (Boston, MA); steroid hormones were from Steraloids, Inc. (Wilton, NH); glycerol was from J. T. Baker (Phillipsburg, NJ); 2-mercaptoethanol from Eastman Kodak (Rochester, NY); 3-(cyclohexylamino)propane sulfonic acid from Calbiochem (La Jolla, CA); Bradford protein reagent and silver stain kit from Bio-Rad Laboratories (Richmond, CA). Benzoyl-Ile-Glu-Gly-Arg-p-nitroanilide (S-2222) was purchased from Helena Labs (Beaumont, TX). Rabbit brain TF (thromboplastin) was obtained from Sigma Chemical Co. (St. Louis, MO). The full-length cDNA probes for mouse and human TF were generous gifts from Dr. Daniel Nathans of the Johns Hopkins University (Baltimore, MD) and Dr. William Konigsberg of Yale University (New Haven, CT). Lubrol PX was obtained from Pierce (Rockford, IL); Nonidet P-40 from BRL (Gaithersburg, MD); 3-[(3-chlolamidopro-

THROMBIN

t

177

pyl)dimethylammonio]l-propanesulfonate from Calbiochem; octylglucoside (OG) and lauryl sarcosine from Sigma Chemical Co.; and Triton X-100, Triton X-114, and Tween-20 from BioRad. To prevent autooxidation, all detergent solutions contained 1 mol of a free radical scavenger, butylated hydroxytoluene, per 500 mol detergent (9). Phosphatidylethanolamine from E. coli and phosphatidylcholine and phosphatidylserine from bovine brain were obtained from Avanti Polar Lipids (Birmingham, AL). The lipids were dissolved in chloroform. A lipid mixture (phosphatidylcholine-phosphatidylethanolamine-phosphatidylserine, 1:1:0.25) was prepared in polypropylene tubes. The chloroform was evaporated to dryness in a stream of N2, buffer (20 mM Tris, pH 7.7; 50 mM KC1; 0.5% OG; 5% glycerol; and 0.05 mM EGTA) was added so the final concentration of lipid was 10 mg/ml, and the lipids were dispersed by sonication while the samples were kept in ice and under a stream of nitrogen. Human standard, factor XH-deficient, and factor VH-deficient plasmas were purchased from Calbiochem. The coagulation factor-deficient plasmas contained less than 1% of the normal concentration of the depleted factor. Purified human factor X was a gift from Dr. Michael J. Fasco of the Wadsworth Center for Laboratories and Research. All other reagents were purchased from Sigma Chemical Co. or Fisher Scientific (Rochester, NY). Female Wistar rats, 19-21 days old, were supplied by the Wadsworth Center's Griffin Laboratories, maintained on rat chow and water ad libitum, and used within 24 h of delivery. Preparation of cytosols and membrane fractions from uterus and other tissues Nineteen- to 21-day-old rats were injected ip with 4.5 /xg 17/3estradiol (E2) in a 1% ethanol-0.14 M NaCl solution, while control animals were inoculated with vehicle. In some experiments, 30-50 rats/group were injected with 4.5 ng of other steroid hormones. After 3 h (or other specified times) the rats were killed by cervical dislocation, and the uteri were removed. In some experiments other tissue samples were collected. These tissues were prepared in the same way as uterine samples. Cytosol and membrane fractions were prepared as described previously (3). In brief, the postmitochondrial supernatant was loaded onto a discontinuous gradient that consisted of layers of 20% glycerol and 50% glycerol in 0.04 M Tris (pH 7.7), 0.1 M KC1, and 0.1 mM EGTA and centrifuged at 116,000 x g for 1 h. The 20% glycerol layer and the interface between the two glycerol layers were removed and pooled as the crude membrane fraction, which was the source of the membrane procoagulant. The supernatant above the glycerol layers was defined as cytosol and used as the source of the uterine prothrombin as previously described (2, 4). The membrane fraction was concentrated and freed of cytosolic contaminants by passage through a second glycerol discontinuous gradient. After centrifugation, the supernatant was discarded, and the glycerol layers were pooled (0.6 ml) and designated the concentrated membrane fraction. Detergent solubilization

FIG. 1. The extrinsic pathway of coagulation. PT, Prothrombin; Va, VII, and Xa, coagulation factors.

Aliquots of concentrated membranes (1-2 mg protein/ml in 0.04 mM Tris, pH 7.7; 0.1 M KC1; and 35% glycerol) were passed


ESTROGEN-RESPONSIVE UTERINE PROCOAGULANT

178

through a dextran-coated P10 column to remove the glycerol, adjusted to a final concentration of 5% glycerol, and mixed with different concentrations of detergent. The mixture (200 tx\) was incubated overnight at 4 C on a shaker, and then loaded onto a discontinuous gradient (0.15 ml 20% glycerol and 0.3 ml 50% glycerol in 0.8-ml Beckman minitubes). After a 30-min centrifugation at 151,000 X g, the activator activity was assessed in the starting material (total activity), in the supernatant (soluble activity), and in the combined 20% glycerol fraction and the 20-50% glycerol interface (insoluble activity). The soluble fraction was adjusted to 20% glycerol for storage. Results were expressed as a percentage of the initial activity recovered in each fraction.

Endo• 1990 Voll26«Nol

RNA extraction, slot blot preparation, and hybridization RNA extraction was performed by the guanidine thiocyanate method (15), and poly(A) RNA was isolated by oligo(dT) chromatography. RNA was denatured with methyl mercury, and samples were blotted onto nylon membranes (GeneScreen, New England Nuclear-DuPont, Boston, MA). Oven-dried membranes were prehybridized, then hybridized in 50% formamide overnight at 42 C according to previously described procedures (16). The cDNA probes for TF were labeled with [32P]dCTP by nick translation.

Results

Solubilization and characterization of the membrane acUterine procoagulant assay: activation of prothrombin and hy- tivator drolysis of TAME Earlier studies (3) established that the prothrombin The membrane activator was quantitated using a modificaactivator was an integral plasma membrane protein. Our tion of the assay described for the enzyme (2, 4). The procoattempts to solubilize it with detergents were successful agulant was mixed with an excess of partially purified uterine only with OG. Our criterion of solubilization was the prothrombin and incubated with 2 mM CaCl2 and 10 mM Tris, pH 7.7, in a total volume of 100 /*1 at 37 C for 30 min to allow retention of activity in the supernatant after discontinactivation of prothrombin. Enzyme was assayed in triplicate as uous gradient ultracentrifugation of the detergentpreviously described (2). One thousand counts per min of treated membranes (17). Gradients were designed so that radioactivity represented 11 pmol product formed. The specific particles with a sedimentation coefficient smaller than activity of the activator is reported as picomoles of product 30S remained in the supernatant. Figure 2 shows that formed by thrombin per min and per ng procoagulant protein. 50% of the initial activity was recovered in the soluble fraction after incubation of the protein with OG in a Protein assay ratio of 1:22. Attempts to solubilize the activator with Protein concentrations of samples were determined by the other detergents, including Lubrol PX, Triton X-100 and microprocedure of Bradford (10), with BSA as the standard. X-114, 3-[(3-chlolamidopropyl)dimethylammonio]l-proOG did not interfere with the assay (11). panesulfonate, sarkosyl NL-40, and Nonidet P-40, were unsuccessful. Chromogenic assay for human factor X activation The size of the solubilized species was assessed by size Activation of human coagulation factor X by uterine memexclusion fast protein liquid chromatography (18). Actibranes was determined in a two-stage assay using the chromovation of uterine prothrombin was measured in all fracgenic substrate S2222 (12). In the first stage, membrane samples or a TF standard in buffer (0.05 M Tris, pH 8.3, and 0.02 110M CaCl2) were mixed with 5 /x\ purified human factor X (7 mg/ ml) in a final volume of 150 /A. The mixture was incubated at 37 C with gentle agitation. After 10 min (the activation was linear for at least 30 min), 0.1 ml 1 mM S2222 was added, and the sample was incubated 15 min longer at the same temperature (product formation was linear for at least 30 min). The reaction was stopped with 0.1 ml 20% HC1O4, and the product was detected by either the increase in absorbance at 405 nm or a colorimetric reaction for diazotizable amines (13). Coagulation assay Procoagulant activity was quantitated by a one-stage clotting assay. Uterine membrane preparations or rabbit TF standard (0.1 ml) were added to a disposable coagulation cup that contained 0.1 ml 30 mM CaCl2and warmed to 37 C for 3 min. After the addition of normal human plasma (diluted 2-fold), the time required for fibrin clot formation was measured with a fibrometer (Becton Dickinson, Cockeysville, MD). In some experiments normal human plasma was replaced with human plasma deficient in specific coagulation factors (14).

-10

23 DETERGENT / PROTEIN ( WT / WT )

FIG. 2. Solubilization of the uterine activator with OG. Aliquots of concentrated membranes (1-2 mg protein/ml) were incubated with 0.54% OG at 4 C overnight. The samples were loaded on discontinuous glycerol gradients and centrifuged at 151,000 x g for 30 min. Activator activity was assessed with the TAME assay for thrombin generated in the material loaded on the gradients (total activity; O) and in the supernatants obtained after centrifugation (soluble activity; A).


ESTROGEN-RESPONSIVE UTERINE PROCOAGULANT tions after chromatography. The apparent mol wt of the peak of activator was estimated from the relative migration of marker proteins chromatographed under the same conditions. The activator migrated with an apparent mol wt of 20,000-40,000 (Fig. 3). A 3-fold increase in activator activity in animals treated with E2 was observed relative to the control value. Gel filtration resulted in a 17-fold increase, with a 30% recovery relative to the crude membranes. Hormone and tissue specificity of uterine procoagulant If the membrane activator were a procoagulant, its activities in the activation of human factor X and in the activation of rat uterine prothrombin would change in parallel. The relationship between tissue distribution of uterine prothrombin activator and factor X activator was investigated (Fig. 4). Various tissue samples were obtained 3

FIG. 3. Gel filtration of uterine plasma membrane proteins solubilized with OG. Concentrated membrane fractions obtained from E2-treated rats or controls were solubilized overnight at 4 C with 2% OG. Insoluble material was removed by ultracentrifugation. The soluble fraction (200 jil; 1 mg/ml) was loaded onto a Superose column. Fractions from the column were assayed for activator ( ) and protein (A28ol )•

179

h after in vivo E2 administration, and crude membrane fractions were prepared from each tissue. The samples were assayed for both uterine prothrombin activator and factor X activator. The prothrombin activator was increased 5-fold in the uterus. Spleen samples gave erratic results in the prothrombin activator assay because of a high background level of TAME hydrolysis. Factor X activation was increased nearly 2-fold in the uterus after estrogen treatment, while it remained very low in the other tissues. It was surprising to find a higher level of factor X activator in the untreated uterus than in other tissues. To investigate hormone specificity, groups of rats were injected with different steroid hormones. After 3 h crude membrane fractions were prepared and used as sources for both uterine prothrombin activator and factor X activator. In both cases activation was increased 2- to 3fold compared to untreated controls after administration of E2 or estrone, but not after treatment with estriol,

CONTROL

J



180

WET WEIGHT

E n d o • 1990 Vol 126 • No 1

SOLUBLE PROTEIN

MEMBRANE PROTEIN

(yg)

110

44

i—1

100

40

|—|

r—|

600

i—i

90

36

r—|

540 32

80

i—i

480

—|

28

24

FlG. 4. Tissue specificity of prothrombin activator and factor X activator. Groups of 50 rats were injected with E2 or vehicle. Samples from different tissues were extracted 3 h after treatment. Membranes were prepared and assayed for prothrombin activator (activation of prothrombin and hydrolysis of TAME) and for factor X activator (activation of human factor X and hydrolysis of S2222) as described in Materials and Methods.

60

360

20

300 0

1 2

3

5

PROTHROMBIN (pmols/jjg-min)

50 0

1 2

PROTHROMBIN ACTIVATOR (pmols/pg-min) 160

testosterone, dexamethasone, or progesterone (data not shown).

3

5

0

FACTOR X ACTIVATOR (pmols/jjg-min)

1 2

3

5

FACTOR X (pmols/jjg-min)

1

l

JL

JL

JL

1

Time course We reported the time course of the response to E2 of combined uterine prothrombin and its activator in a previous study (2). A new study was performed of the separated activator, prothrombin, and factor X to determine whether the response of proenzymes and activators to E2 stimulation followed the same time course. Uterine wet weight, soluble and membrane-bound protein, activators of factor X and uterine prothrombin, and endogenous levels of factor X and prothrombin measured at intervals after stimulation with E2 are all given in Fig. 5. Our data on uterine wet weight and protein levels confirm earlier data (19). After 3 h uterine wet weight was nearly doubled, and soluble protein was increased 50%. Membrane protein was increased 30% after 5 h. Both prothrombin activity in the cytosol and activator in the membranes reached a peak of activity 3 h after hormone treatment and then began to decline. The amount of factor Xa activity in the cytosol increased at 2 h after injection and reached a peak at 3 h. However, X activator was close to its maximum at 2 h, an hour earlier than the other increases. A pronounced water imbibition and plasma protein infiltration into the uterus occur 3 h after estrogen treatment (19). Increases in wet weight, soluble protein, and prothrombin are consistent with an influx of plasma. The early increase in factor X activator without a similar increase in prothrombin activator would occur if prothrombin activation required a soluble factor [such as

70

420

pi

100 0.36 80

0.28

0.20

60

Ti

0 1 2 3 5

1

700

40 0 1 2 3 5

0 1 2 3 5

0 1 2 3 5

FIG. 5. Time course of E2 response. Groups of 30-34 animals were injected with E2, and uteri were collected at 0, 1, 2, 3, and 5 h. Uterine wet weight (grams), soluble and membrane-bound protein (micrograms), endogenous prothrombin and factor X activity (picomoles of product formed per fig and min), and prothrombin and factor X activators (picomoles of product formed per fig activator and min) were measured at each time. Wet weight and protein levels are given per rat. SE bars are shown for enzyme and activator determinations in triplicate. With the exception of two points, all SEs were 10% or less.

factor VII(a)]1 that did not increase in the uterus until 3 h after stimulation.

Concanaualin-A (Con-A) binding of the activator

Since TF is partially glycosylated, we determined whether the uterine procoagulant had a similar property. 1 By convention, factor Vila is identified as the activated coagulation factor, and factor VII is its inactive precursor. Whether factor VII also has a low level of activity is a matter of controversy (28, 31). The designation VII(a) is intended to reflect the possibility that both factors VII and Vila may be active in the reaction.



Aliquots of the solubilized activator fraction were incubated overnight in a suspension of Con-A-Sepharose (20). After the samples were centrifuged, the supernatants were assayed for activation of human factor X and for activation of uterine prothrombin. In some experiments, the samples were reconstituted with a phospholipid mixture before the assay. The Con-A pellet was eluted with a-methylmannoside (AMM), and the eluate was assayed. The results are given in Table 1. The factor X activator and the prothrombin activator were partially glycosylated, as shown by the ability of 66% and 40% of the activities to bind to Con-A. No activity bound in the presence of 250 mM AMM. About 57% of the bound prothrombin activator and 48% of the factor X activator were eluted by AMM. Reconstitution of the eluted fraction with lipids was required to recover factor X activator, while lipids had no effect on the prothrombin activator. The apparent differences in the lipid requirements of the two activators probably arose because of differences in the assay; in the X activation, a highly purified human factor X preparation which contained little or no lipid was added, while in the prothrombin activation a partly purified uterine prothrombin preparation was added which presumably contained sufficient lipid to accelerate the reaction. Electrophoresis of activated factor X

The products of the activation of factor X by the

181

factor VII(a). An active fragment of 54,000 mol wt (factor Xa) is obtained. In the presence of lipids, factor Xa digests itself to a second active fragment of 50,000 mol wt. Sodium dodecyl sulfate-gel electrophoresis showed that fragments with mol wt of 53,000 and 48,000 were obtained from factor X whether it was activated with venom or with uterine procoagulant (Fig. 6). Differences in the intensities of bands may be due to different amounts of lipids in the venom sample and in uterine membranes, since the extent of proteolysis of the factor X would depend on the amount of lipids in the preparations. Warfarin treatment and membrane procoagulant activity By definition, a procoagulant catalyzes the formation of a fibrin clot. The time required for clot generation from normal plasma (or from fibrinogen) is inversely related to procoagulant content. If the procoagulant is TF, extended clotting times will be observed with plasma deficient in factor VII, the precursor of the proteinase requiring TF as a cofactor. TF will be equally effective in clot formation with whole plasma or with any plasma deficient in another coagulation factor that has a normal level of factor VII. We measured coagulation times in various human plasmas with crude membranes from untreated and E2stimulated rat uteri as the procoagulant. Normal and factor XH-deficient plasmas had equivalent clotting

uterine procoagulant were investigated by gel electropho-

times, while factor VH-deficient plasma had longer times

resis. Human factor X aliquots were incubated with Russell's viper venom or uterine membrane preparations. After 1 or 4 h, samples were boiled and loaded on sodium dodecyl sulfate-polyacrylamide gels. The size of the fragments produced by activation was apparent after electrophoresis and staining with Coomassie blue, shown in Fig. 6. Russell's viper venom activates human factor X (72,000 mol wt) in the same way as the human procoagulant (21), which is a membrane-bound complex of TF-

(data not shown). We reasoned that the rat procoagulant preparations might have some rat factor VII associated with them that allowed clotting to occur and that rat factor VII levels would be decreased by treatment of the rats with warfarin. Coagulation factors VII, X, and prothrombin (but not TF) have Gla residues (7). Gla occurs as the result of a vitamin K-dependent posttranslational protein modification. To determine which components in the pro-

TABLE 1. Both factor X activator and prothrombin activator bind to Con- A-Sepharose

Initial activity Con-A-retained activity Eluted by AMM Bound in the presence of AMM

Factor X activator

Factor X activator + lipids

Prothrombin activator

100 66 4 0

100 72 35 0

100 40 23 0

Concentrated membrane fractions were solubilized with 2% OG. The soluble fraction was incubated with Con-A-Sepharose (in 1% OG, 20 mM Tris-HCl, pH 7.7, and 1 mM MnCl2, MgCl2, and CaCl2). After overnight at 4 C, the samples were centrifuged, and the supernatants were assayed for X activator (activation of human factor X and subsequent hydrolysis of S2222) and for uterine activator (activation of uterine prothrombin and hydrolysis of TAME). In some experiments, the samples were reconstituted with lipids before the assay (phosphatidylcholine-phosphatidylethanolamine-phosphatidylserine, 1:1:0.25); 10 mg/ml in 0.5% OG). The Con-A pellet was eluted with 250 mM AMM for 3.5 h at 4 C and then centrifuged and assayed as described above. Activity is expressed as a percentage of the initial activity. Con-A-retained activity is the difference between initial activity and unbound activity.



182

Endo • 1990 Vol 126 «No 1

FACTOR X

+

4-

+

-

+

4-

—

RVV

— _

Ih —

4h _

4h _

Ih

4h

4h

E2 MEMBRANES

STANDARDS

FIG. 6. Human factor X activation; sodium dodecyl sulfate-polyacrylamide gel electrophoresis of the activation products. Human factor X samples (14 fig) were incubated with 2 ng Russell's viper venom (RVV) or 2 ng membranes obtained from estrogen-treated uteri (E2 membranes) at 37 C for 1 or 4 h. Sample buffer was added at the end of the activation, and samples were boiled, loaded on a 10% polyacrylamide gel, and stained with Coomassie blue after electrophoresis.

thrombin activation scheme contained Gla, we investigated the effect of a warfarin-induced vitamin K deficiency on prothrombin, prothrombin activator, factor X activator, and uterine procoagulant (clot formation) activities. While warfarin decreases the activity of Glacontaining proteins, it has no effect on TF activity. Groups of 50 rats were injected with 1 mg/kg warfarin in PBS, while control groups received PBS only. The animals were maintained with or without 4 mg/liter warfarin in the drinking water. After 2 days groups were injected with E2 or vehicle, uteri were removed 3 h later, and crude membranes and cytosols were prepared. Table 2 shows that the prothrombin activation was unaffected by warfarin. The factor X activator was increased by E2 to similar levels independent of previous TABLE 2. Effect of warfarin on the prothrombin activator and factor X activator

Control-control Control-estrogen Warfarin -control Warfarin-estrogen

Prothrombin activator (pmol/jig-min)

Factor X activator (nmol/^g-min)

65.7 159.6 68.4 174.0

0.8 1.6 0.2 1.1

Groups of 50 rats were injected with 1 mg/kg warfarin in PBS or with vehicle. The warfarin groups had 4 mg/liter warfarin in the drinking water. Two days later groups were injected with 4.5 ng E2 or vehicle, they were killed and dissected 3 h later, and uterine preparations were tested for prothrombin activator, uterine prothrombin, and factor X activator activity.

warfarin treatment. At the same time, the activity of the uterine prothrombin was reduced 5-fold by the warfarin treatment (1). The apparent reduction of factor X activator after warfarin treatment is probably due to a loss of factor VII(a), the Gla-containing enzyme that binds to TF. In this experiment as in others, rat factor VII would be present as an impurity in the partly purified uterine prothrombin preparation used to assay the activator. Procoagulant levels in normal and warfarin-treated animals were also determined by measuring the coagulation times on the same samples (Fig. 7). Coagulation times decrease logarithmically with increases in procoagulant concentration (22). Both control and warfarintreated animals had a 30-40% reduction in coagulation times with normal plasma after E2 treatment, indicating that a procoagulant was increased by E2. The clotting times of factor VH-deficient plasma were increased 50% with warfarin-treated membranes as procoagulant because those membrane preparations had reduced levels of endogenous factor VII. Therefore, the procoagulant increased by E2 stimulation was not factor VII. TF gene expression in the immature rat uterus To determine whether there is TF mRNA in the uterus, total RNA and poly(A) RNA were prepared from uteri before or after an estrogen injection and hybridized to a human or mouse TF cDNA. As a positive control,


ESTROGEN-RESPONSIVE UTERINE PROCOAGULANT 160 140

C D NORMAL PLASMA M VII DEFICIENT PLASMA ESS XII DEFICIENT PLASMA

120 100 80 60 ]

40 20CC

CE,

we

WE'

FlG. 7. Procoagulant activity of uterine membrane fractions from normal and warfarin-treated animals. Animals were treated as described in the text. CC, Controls; CE2) E2 only; WC, warfarin only; WE2, warfarin and E2. Uterine membranes were assayed for their ability to clot normal, factor Xll-deficient, and factor VH-deficient human plasmas.

total RNA from rat brain (a good source of TF) was prepared. In preliminary experiments blots with the human TF cDNA probe, pL5BclO, were positive, indicating the presence of a mRNA related to human TF. The mouse TF clone pGEM 482 was used for further experiments because of the closer relationship of mice and rats. The similarity of mouse and human TF sequences is 67% (23). Figure 8 shows that the mouse probe gave a positive signal in the estrogenized uterus at least as intense as that in brain. A weak signal was seen in the controls. Northern blots with the same RNAs and probe gave three bands, as previously described for human TF (24) (data not shown).

Discussion These investigations of the uterine prothrombin activator addressed two goals: characterizing the activator

and attempting to identify it. The activator was a glycosylated integral membrane component with an apparent mol wt of 20,000-40,000. It was solubilized using the detergent OG. To show that the activator was a procoagulant, we measured procoagulant activity in three different ways: activation of uterine prothrombin, activation of human factor X, and coagulation times with human plasmas. We reasoned that coordinated responses in these assays would be consistent with their origin in a common enzymatic reaction. Both prothrombin activator and factor X activator activities were hormone- and tissue-specific and bound to Con-A-Sepharose. Procoagulant activity, measured by the ability of the uterine membranes to coagulate human plasmas, was increased in the E2-treated membranes compared to that in the controls. In the characterization of the procoagulant, it was important to distinguish between factor VII(a) and TF, the inert cofactor for factor VII(a). Variations in procoagulant activities may reflect different levels of either. Rats made deficient in vitamin K by the administration of warfarin have deficiencies in Gla proteins, including prothrombin and factors VII, IX, and X (25, 26). Warfarin treatment had no effect on the activity of uterine prothrombin activator, a result that was consistent with the activator being TF. Human and bovine TF have been purified (5); human TF was recently cloned and sequenced. The sequence of a cDNA for mouse TF and the deduced amino acid sequence were reported recently (23). We used the mouse cDNA probe to show that TF mRNA was present in the uterus in readily detectable amounts after E2 treatment and in smaller amounts in untreated uteri. This latter observation suggests that low level expression of TF occurs in the rat uterus before estrogenization. TF mRNA has been detected in human adipose, placental, adrenal, small intestine, and kidney tissues by

POLY A RNA FIG. 8. TF mRNA levels in uterus. Total RNA and poly(A) RNA were prepared from the uteri of controls and animals treated with E2 for 2, 3, or 5 h and from rat brain. The poly(A) slots contain 100, 200,400, and 800 ng RNA; the total RNA samples have 200, 400, 800, and 1600 ng RNA/slot. They were hybridized with 5 x 106 cpm pGEM482 DNA overnight in 50% formamide at 42 C, washed at 55 C, dried, and autoradiographed with Kodak X-Omat AR film.

183

TOTAL RNA

CONTROL

CONTROL

ESTROGEN TREATED 3hrs

ESTROGEN TREATED 2hrs ESTROGEN TREATED 5hrs BRAIN


184


Northern blot hybridization (27). However, the level of message was below the limits of detection in other tissues, such as pancreas, liver, and spleen (24). Published immunocytochemical studies of human tissues using monoclonal antibodies also showed differential expression on the surface of the cells; arteries and veins had no staining, and epithelial cells of skin, oropharynx, and bladder stained strongly, as well as renal glomeruli, lung alveolar septa, cerebral cortex, and placental villi (28). Vascular endothelial cells and monocytes synthesize TF within 2-6 h of exposure to agents such as endotoxin, interleukin-1, tumor necrosis factor, phorbol esters, or a-thrombin (29, 30). The presence of TF in the uterus has not been reported previously. In addition, this is the first report of the regulation of TF by estrogen. Our data are consistent with an activator which is TF complexed with factor VII(a). Specifically, we have shown that the procoagulant is an integral plasma membrane protein, it can activate factor X to Xa, and it can clot plasma. It has a requirement for factor VII(a), as shown by the experiment with warfarin. Estrogen might act indirectly to stimulate uterine growth, through the generation of secondary signals that act as mitogens. It has been suggested, for example, that expression of the growth response requires physical interaction of the uterine epithelium with the underlying stroma (32). An example of a secondary signal may be the estrogen response of epidermal growth factor (EGF) and the EGF receptor protein in the uterus, which increase 3-fold after estrogen treatment (33). Immunocytochemical studies revealed the presence of EGF precursor in the membrane of uterine epithelium; estrogentreated mice showed an increase in prepro-EGF mRNA (34). The researchers propose that estrogens might stimulate processing of the membrane-bound EGF precursor to generate EGF, which could interact with its receptors (also increased after the estrogen treatment) and stimulate proliferation. Thus, it is possible that peptide growth factors may cooperate with estrogens to produce cell growth. TF might also be involved in such localized production of the growth factor thrombin. Besides playing a central role in the coagulation cascade, a-thrombin is also a potent mitogen (35) that initiates proliferation of a variety of cells, including fibroblasts (36) and spleen cells (37). It also acts synergistically with other growth factors to stimulate proliferation of mammary tumor cells and endothelial cells (38, 39). Thrombin exposure induces endothelial cells and platelets to express c-sis mRNA coding for platelet-derived growth factor and to release mitogenically active platelet-derived growth factor into the culture medium (40). Thus, thrombin regulates the release of a mitogen that can induce division of surrounding cells.

Endo • 1990 Vol 126* No 1

The increase in TF follows a time course which is consistent with an E2-stimulated increase in transcription of the gene. Olson and Lane (41) have recently described the time course of the posttranslational activation of the EGF, insulin, and acetylcholine receptors. The three receptors have half times of 90-180 min for acquisition of ligand binding capacity and 1.5-3 h for appearance on the cell surface. These times are consistent with our experimental results for the appearance of procoagulant in the uterine plasma membranes. Our evidence is consistent with the proposal that E2 stimulation of uterine growth may be modulated by thrombin. Thrombin generation in the uterus is possible because of the E2-induced increase in a procoagulant that has many properties of tissue factor.

Acknowledgments We thank Dr. Daniel Nathans, Johns Hopkins University, for giving us the mouse TF plasmid pGEM 482 and for permitting us to review a paper before publication. We thank Dr. William Konigsberg, Yale University, for sending us the human tissue factor plasmid pLSBclO, with which some preliminary experiments were performed. We thank Drs. Brian Pentecost and S. A. Kumar for helpful discussions.

References 1. Henrikson KP, Jazin EE, Greenwood IA, Dickerman HW 1990 Prothrombin and thrombin are increased in the estrogen treated immature rat uterus. Endocrinology 126:167 2. Henrikson KP, Dickerman HW 1983 An estrogen-stimulated, calcium-dependent tosylarginine methyl ester (TAME) hydrolase in immature rat uterus. Mol Cell Endocrinol 32:143 3. Henrikson KP, Jazin EE, Dickerman HW 1987 Separation and identification of two components of an estrogen-responsive, calcium-dependent arginine esteropeptidase. J Steroid Biochem 26:189 4. Jazin EE, Dickerman HW, Henrikson KP 1988 Estradiol stimulates a uterine plasma membrane protease activator. Endocrinology 122:500 5. Bach RR 1988 Initiation of coagulation by tissue factor. Crit Rev Biochem 23:339 6. Mann KG, Jenny RJ, Krishnaswamy S 1988 Cofactor proteins in the assembly and expression of blood clotting enzyme complexes. Annu Rev Biochem 57:915 7. Furie B, Furie BC 1988 The molecular basis of blood coagulation. Cell 53:505 8. Lau LF, Nathans D 1987 Expression of a set of growth related immediate early genes in BALB/C 3T3 cells: coordinate regulation with c-fos or c-myc. Proc Natl Acad Sci USA 84:1182 9. Helenius A, McCaslin DR, Fries E, Tanford C 1979 Properties of detergents. Methods Enzymol 56:734 10. Bradford MM 1976 A rapid and sensitive method for quantitation of microgram quantities of protein using the principle of proteindye binding. Anal Biochem 72:248 11. Fanger BO 1987 Adaptation of the Bradford protein assay to membrane-bound proteins by solubilizing in glucopyranoside detergents. Anal Biochem 162:11 12. Husten EJ, Esmon CT, Johnson AE 1987 The active site of blood coagulation factor Xa. J Biol Chem 262:12953 13. Powers CA, Nasjletti A 1983 A kininogenase resembling glandular kallikrein in the rat pituitary pars intermedia. Endocrinology 112:1194 14. Tsao BT, Fair DS, Curtiss LK, Edgington TS 1984 Monocytes can be induced by lipopolysaccharide-triggered T lymphocytes to ex-



15. 16. 17.

18. 19. 20.

21. 22. 23. 24. 25.

26. 27.

press functional factor VH/VIIa protease activity. J Exp Med 159:1042 Chomczynski P, Sacchi N 1987 Single step method of RNA isolation by acid guanidinium thiocyanate-phenol-chloroform extraction. Anal Biochem 162:156 Pentecost BT, Teng CT 1987 Lactotransferrin is the major estrogen inducible protein of mouse uterine secretions. J Biol Chem 262:10134 Limbird LE 1986 The preparation and study of detergent-solubilized receptors. In: Limbird LE (ed) Cell Surface Receptors: A Short Course on Theory and Methods. Martinus Nijhoff, Boston, pl33 Welling GW, Slopsema K, Welling-Wester S 1986 Size-exclusion high-performance liquid chromatography of Sendai virus membrane proteins in different detergents. J Chromatogr 359:307 Katzenellenbogen BS, Gorski S 1975 Estrogen actions on syntheses of macromolecules in target cells. In: Litwack G (ed) Biochemical Actions of Hormones. Academic Press, New York, p 187 Van Den Besselaar AM, Bertina RM 1984 Interaction of thromboplastin apoprotein of different tissues with Concanavalin-A. Evidence for heterogeneous glycosylation of the human apoprotein. Thromb Haemostas 52:192 DiScipio RG, Hermodson MA, Davie EW 1977 Activation of human factor X (Stuart factor) by a protease from Russell's viper venom. Biochemistry 16:5253 Pitlick FA, Nemerson Y 1976 Purification and characterization of tissue factor apoprotein. Methods Enzymol 45:37 Hartzell S, Ryder K, Lanahan A, Lau LF, Nathans D 1989 A growth factor-responsive gene of murine 3T3 cells encodes a protein homologous to human tissue factor. Mol Cell Biol 9:2567 Fisher KL, Gorman CM, Vehar GA, O'Brien DP, Lawn RM 1987 Cloning and expression of human tissue factor cDNA. Thrombosis Res 48:89 Smith GF, Neubauer BL, Sundboom IL, Best KL, Goode RL, Tanzer LR, Merriman RL, Frank ID, Herrman RG 1988 Correlation of the in vivo anticoagulant, antithrombotic, and antimetastatic efficacy of warfarin in the rat. Thrombosis Res 50:163 Suttie IW 1985 Vitamin K-dependent carboxylase. Annu Rev Biochem 54:459 Nemerson Y 1988 Tissue factor and hemostasis. Blood 71:1

185

28. Drake TA, Morrissey JH, Edgington TS 1988 Initial immunocytochemical localization of tissue factor in human tissues using epitope-defined monoclonal antibodies. FASEB J 2:A141O 29. Bevilacqua MP, Pober JS, Majeau GR, Cotran RS, Gimbrone MA 1984 Interleukin-1 induces biosynthesis and cell surface expression of procoagulant activity in human vascular endothelial cells. J Exp Med 160:618 30. Nawroth PP, Stern DM 1986 Modulation of endothelial cell hemostatic properties by tumor necrosis factor. J Exp Med 163:740 31. Rao LVM, Rapaport SI 1988 Activation of factor VII bound to tissue factor. Proc Natl Acad Sci USA 85:6687 32. Inaba T, Wiest WG, Strickler RC, Mori J 1988 Augmentation of the response of mouse uterine epithelial cells to estradiol by uterine stroma. Endocrinology 123:1253 33. Lingham RB, Stancel GM, Loose-Mitchell DS 1988 Estrogen regulation of epidermal growth factor receptor mRNA. Mol Endocrinol 2:230 34. DiAugustine RP, Petrusz P, Bell GI, Brown CF, Korach KS, McLachlan JA, Teng CT 1988 Influence of estrogens on mouse uterine epidermal growth factor precursor protein and mRNA. Endocrinology 122:2335 35. Carney, DH, Cunningham DD 1978 Role of specific cell surface receptors in thrombin-stimulated cell division. Cell 15:1341 36. Perdue JF, Lubenskyi W, Kivity E Sonder SA, Fenton II JW 1981 Protease mitogenic response of chick embryo fibroblasts and receptor binding/processing of human alpha-thrombin. J Biol Chem 256:2767 37. Chen LB, Teng NNH, Buchanan JM 1976 Mitogenicity of thrombin and surface alterations on mouse splenocytes. Exp Cell Res 101:41 38. Medrano EE, Cafferata EG, Larcher F 1987 Role of thrombin in the proliferative response of T-47D mammary tumor cells. Exp Cell Res 172:354 39. Gospodarowicz D, Brown KB, Bridwell CR, Zetter BR1978 Control of proliferation of human vascular endothelial cells to fibroblast growth factor and thrombin. J Cell Biol 77:774 40. Daniel TO, Gibbs, VC, Milfay DF, Garovoy MR, Williams LT 1986 Thrombin stimulates c-sis gene expression in microvascular endothelial cells. J Biol Chem 261:9579 41. Olson TS, Lane MD 1989 A common mechanism for posttranslational activation of plasma membrane receptors. FASEB J 3:1618


Estrogen regulation of creatine kinase-B in the rat uterus.

Estrogen induced expression of the C-jun proto-oncogene in the immature and mature rat uterus.

Prothrombin levels are increased in the estrogen-treated immature rat uterus.

Separation of "estrogen-induced" protein from phosphoprotein phosphatase activity of immature rat uterus.

Selective Estrogen Receptor Modulation by Larrea nitida on MCF-7 Cell Proliferation and Immature Rat Uterus.

Identification of the 'estrogen-induced protein' in uterus and brain of untreated immature rats.

Estrogen receptor-binding affinity of tamoxifen analogs with various side chains and their biologic profile in immature rat uterus.

Expression of tissue factor procoagulant activity: regulation by cytosolic calcium.

Diethylstilbestrol stimulates persistent phosphatidylinositol lipid turnover by an estrogen receptor-mediated mechanism in immature mouse uterus.

Alpha-fetoprotein: secondary estrogen binder in immature rat uterine cytosol.

The effect of 6-nitro-1,3,8-trichlorodibenzofuran as a partial estrogen in the female rat uterus.

Tissue-specific regulation of vasoactive intestinal peptide messenger ribonucleic acid levels by estrogen in the rat.

Oestriol-stimulated synthesis of ribonucleic acid in the uterus of the immature rat.

Glucocorticoid and progestin regulation of eosinophil chemotactic factor and complement C3 in the estrogen-treated rat uterus.

Tissue factor procoagulant expression by rat alveolar epithelial cells.

Effects of in vivo gonadotropin treatment on estrogen levels in the testis of the immature rat.

Contrasting subcellular responses to monohydroxytamoxifen and oestradiol benzoate in the immature rat uterus [proceedings].

Estrogen induction of a small, putative K+ channel mRNA in rat uterus.

Effect of colchicine on estrogen action. I. Inhibition of 17 beta-estradiol-induced water and potassium uptake in the immature rat uterus.

Estrogen induces c-fos expression specifically in the luminal and glandular epithelia of adult rat uterus.

Estrogen receptors in the rat uterus. Retention of hormone-receptor complexes.

Qualification of estrogen binding variables in tissues of the immature rat.

Serum estrogen binding proteins in tissues of the immature rat: quantitation by radioimmunoassay.

Follicle-stimulating hormone stimulates ovarian synthesis of proteoglycans in the estrogen-stimulated hypophysectomized immature female rat.