Mechanism of action of angiostatic steroids: Suppression of plasminogen activator activity via stimulation of plasminogen activator inhibitor synthesis

JOURNAL OF CELLULAR PHYSIOLOGY 155568-578 (1993)

Mechanism of Action of Angiostatic Steroids: Suppression of Plasminogen Activator Activity

Via Stimulation of Plasminogen Activator In hi bitor Synthesis

FRANCINE BLEI, E. LYNETTE WILSON, PAOLO MICNATTI, AND DANIEL B. RIFKIN* Departments of Pediatrics (F.B.) and Cell Biology (E.L.W., P.M., D.B.R.j, New York

University Medical Center, Kaplan Cancer Center and Raymond and Beverly Sdckler laboratory (D.B.R.), New York, New York 10016

Recently, a novel class of angiostatic steroids which block angiogenesis in several systems has been described. Since the elaboration of proteases is believed to be an important component of angiogenesis, we tested whether these steroids blocked the fibrinolytic response of endothelial cells to the angiogenic protein, basic fibroblast growth factor [bFGF]). Cultured bovine aortic endothelial (BAE) cells were incubated with bFCF and/or medroxyprogesterone acetate (MPA), an angiostatic steroid which has been shown to inhibit vascularization, collagenolysis, and tumor growth. When bFGF ( 3 ng/ml) was added to confluent monolayers of BAE cells, plasminogen activator (PA) activity in the medium was increased threefold. In contrast, MPA at 1 OP6 M, lo-' M, 10 M, and 10 -' M decreased PA levels in the medium by 83%, 83%, 75%, and 39%, respectively. The stimulation of PA levels in BAE cells by bFGF ( 3 ng/ml) was abrogated by the presence of loph M MPA. This decrease in PA activity was found to be mediated by a significant increase in plasminogen activator inhibitor type-1 (PAI-1) production. MPA, therefore, negated one of the important enzymatic activities associated with the angiogenic process. In contrast to the decreased levels of secreted PA in cultures exposed simultaneously to MPA and bFGF, cell-associated PA levels remained high, consistent with earlier observations indicating that PAL1 does not inhibit cell-associated PA. Thus, angiostatic steroids may exert their inhibitory effects on angiogenesis by increasing the synthesis of PAI-1. This, in turn, inhibits PA activity and, therefore, plasmin generation, which is essential for the invasive aspect of angiogenesis. 0 1993 Wiley-Liss, Inc

Physiological angiogenesis is characterized by tightly controlled regulation of endothelial cell growth and migration in a limited number of processes, while abnor- mal, unregulated vascular proliferation is a predomi- nant feature of several diseases including proliferative retinopathies, hemangiomas, certain forms of arthritis, and solid tumor growth (Goldie, 1970; Newell, 1986; Folkman, 1987; Mulliken and Young, 1988). New blood vessel formation appears to originate from existing venules via a series of steps, including dissolution of the endothelial cell basement membrane mediated by proteolytic enzymes, endothelial cell migration, and endothelial cell proliferation (Ausprunk and Folkman, 1977; Gross et al., 1982,1983; Folkman, 1984; Folkman and Klagsbrun, 1987). These processes appear to be positively regulated by angiogenic factors such as basic fibroblast growth factor (bFGF) (Moscatelli et al., 1986; Folkman and Klagsbrun, 1987; Moscatelli and Rifkin, 1988).

The ability to inhibit any one or combination of these steps has been of interest in order to provide new thera- peutic modalities for the treatment of diseases in which pathological angiogenesis occurs. A class of angiostatic 0 1993 WILEY-LISS. INC.

steroids has been described (Crum et al., 1985) which, in the presence of heparin, blocked angiogenesis in the chicken embryo chorioallantoic membrane (CAM) assay. The antiangiogenic effect appeared to be indepen- dent of the mineralocorticoid or glucocorticoid effect, as avascular zones were observed in the CAM assay even with compounds that had no mineralocorticoid or glucocorticoid activity. In addition, other steroid preparations in the presence of heparin have been shown to inhibit vascularization in the CAM as well as tumor growth (Gross et al., 1981; Oikawa et al., 1988). In all of these studies, the mode of action of these agents was not identified.

The present study was designed to explore the possible mechanisms of inhibition of angiogenesis by angio-

Received February 4,1992; accepted January 19, 1993. "To whom reprint requestslcorrespondence should be addressed. This manuscript is dedicated to the memory of Elisabeth Suenson, M.D.

569 EFFECT OF ANGIOSTATIC STEROIDS ON FA AND PAI-1

static steroids. Since it has been suggested that proteases secreted by endothelial cells may play an essential role as an early and rate-limiting step of neovascularization (Gross et al., 1982, 1983; Moscatelli and Rifkin, 19881, we analyzed the effect of steroids on the regulation of endothelial cell protease production and activity. We hypothesized that these steroids might act by inhibiting the initial plasminogen activator (PA)-mediated proteolytic events necessary for the es- tablishment of the angiogenic process. To test this hypothesis, we have studied the ability of these steroids to block induction of PA and to increase the production of plasminogen activator inhibitor type-1 (PAI-1) in bovine aortic endothelial (BAE) cells exposed to the angiogenic factor bFGF. Data presented here suggest that the angiostatic effect of these steroids may derive from their ability to induce inhibitors of PA.

MATERIALS AND METHODS Cell culture

BAE cells were cultured using alpha minimal essential medium (a-MEM; Gibco Laboratories, Grand Is- land, New York) containing 1% (vol/vol) penicillin, streptomycin, and 2 mM glutamine, and 5% (volivol) calf serum (Gibco). Primary endothelial cell cultures were initiated from thoracic aortas of freshly slaughtered cows according to the method of Schwartz (1978). Cells used were between passages 12-19. Bovine capillary endothelial (BCE) cells were initiated from the adrenal cortex of freshly slaughtered cattle and culti- vated under previously described conditions (Folkman et al., 1979; Gross et al., 1982). The cells were seeded in 24-well culture dishes (Flow Labs, McLean, VA) or 35 mm culture dishes (Falcon, Lincoln Park, NJ) and grown to confluence.

Cell treatments At confluence, the BAE cells were incubated for

18-24 h in a-MEM containing antibiotics and glutamine in the absence of serum in order to decrease the basal levels of PA and PAI-1. The cells were cultured for an additional 18-24 h in serum-free medium in the presence of bFGF (3 ngiml), transforming growth factor beta (TGF-P, 1 ngiml), or steroids (medroxyprogesterone

Abbreviations

a-MEM BAE BCE bFGF BSA CAM EDTA MPA MW PA PAGE PAI-1 PBS scuPA SDS TCA

TGF-p TIMP u-PA

t-PA

alpha minimal essential medium bovine aortic endothelial bovine capillary endothelial basic fibroblast growth factor bovine serum albumin chicken chorioallantoic membrane ethylenediamine tetraacetic acid medroxyprogesterone acetate molecular weight plasminogen activator polyacrylamide gel electrophoresis plasminogen activator inhibitor type-1 phosphate buffered saline single-chain urokinase-type plasminogen activator sodium dodecyl sulfate trichloroacetic acid tissue-type plasminogen activator transforming growth factor beta tissue inhibitor of metalloproteinase urokinase-type plasminogen activator

acetate [MPA], hydrocortisone, dexamethasone, 17-alpha-hydroxyprogesterone, progesterone, pregnenolone, tetrahydro-S; Sigma, St. Louis, MO) a t various doses. Where indicated, heparin (from porcine intestinal mu- cosa; Sigma) was added to a final concentration of 10 pgiml. Recombinant human bFGF (in 10 mM NaPO,, pH 7, 0.3 M glycerin, 0.92 mgiml) was prepared and kindly donated by Synergen, Inc. (Boulder, CO). Re- combinant TGF-f3 was kindly donated by Genentech, Inc. (South San Francisco, CAI.

Preparation of conditioned medium and cell extracts for P A and PAI-1 activities

Twenty-four hours after the addition of growth factors or steroids, serum-free conditioned medium (1 mll Linbro well) was removed and centrifuged at 1,000 g for 10 min to pellet cellular debris. Tween-80 (0.01% final concentration) was added to the supernatant to prevent adsorption of PA to the tube. The cell monolayers were washed twice with cold phosphate buffered saline (PBS), extracted with 0.5% Triton X-100 in 0.1 M Tris- HCl, pH 8.1, and scraped with a rubber policeman. Pro- tein content of the cell extract was determined using the BCA protein assay reagent (Pierce, Rockford, IL) with bovine serum albumin (BSA) as a standard. All samples were stored at -20°C until used.

Purification of fibrinogen and plasminogen Fibrinogen was purchased from KabiVitrum (Stock-

holm, Sweden). Plasminogen was removed from the fibrinogen by lysine-Sepharose chromatography (Deutsch and Mertz, 1970). Plasminogen was prepared from out- dated frozen human plasma by lysine-Sepharose chromatography according to the method of Deutsch and Mertz (1970).

PA assay The PA in the conditioned media and cell extracts

was determined by the '"I-fibrin plate assay (Gross et al., 1982). Aliquots of conditioned medium and cell extract (approximately 1 pg cellular protein) were mixed with 0.1 M Tris-HCL, pH 8.1, containing BSA (250 pgiml) (as a carrier protein) and human plasminogen (8 kg/ml). This mixture was incubated at 37°C in dishes coated with ""I-fibrin. Aliquots from each well were removed and counted in a Packard gamma scintillation counter. The amount of lZ5I released during the incubation corresponds to soluble fibrin fragments released by plasmin generated by PA present in each well. The BAE cell PA activities were compared to a standard curve of human urokinase-type plasminogen activator (u-PA) (Leo Pharmaceuticals, Denmark) included in each assay. Final results are expressed in Ploug units.

Fibrin zymography for P A activity of BAE cell conditioned medium and cell extracts

Zymography was performed according to the method of Granelli-Piperno and Reich (1978). Briefly, aliquots of conditioned medium and cell extract were separated by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) (3% stacking gel/9% resolving gel) under nonreducing conditions. The SDS gel was washed three times (20 min each) in 2.5% Triton X-100 to remove the SDS, and placed on a fibrin-agar indica-

570 BLEI ET AL

tor gel containing agar, plasminogen, fibrinogen, penicillin/streptomycin, and thrombin in 0.1 M Tris- HC1, pH 8.1 (adaptation of the method of Granelli- Piperno and Reich, 1978). The gels were incubated overnight a t 4°C to enable the bands of PA activity to diffuse into the fibrin-agar gel. Zones of lysis which develop following incubation at 37°C (1-5 h) indicate PA activity. To terminate the reaction, the gel was removed and the fibrin-agar gel was stained in 1% amido black in 70% methanol and 10% acetic acid for 1 h and destained in 70% methanol/lO% acetic acid with several changes until the solution was clear.

Reverse fibrin zymography for PAI-1 activity Aliquots of conditioned medium and cell extracts

were separated by SDS-PAGE (3% stacking ge1/9% resolving gel) under reducing conditions (Erickson et al., 1984). The fibrin-agar indicator gel contained the same components as those described above plus u-PA (25 m u / ml). In this system, PAI-1 activity is evident as bands of opaque material against a clear, completely lysed back- ground. The gels were processed as described for the PA gels.

Immunoprecipitation of PAI-1 BAE cells were prepared as described above. After

incubation for 24 h in serum-free a-MEM, the medium was changed to serum-free, methionine-free RPMI medium (Gibco Labs) (supplemented with glutamine [2 mM1 and methionine [4 mM]). 35S-methionine (NEN DuPont Research Products, Boston, MA) (40 pCi/well), in the presence or absence of steroids and/or growth factors, was added to wells and incubated a t 37°C. Sub- sequent steps followed a modification of the procedures of Moscatelli (1986) and Loskutoff (personal communi- cation). After 13-16 h incubation, conditioned medium (1 ml/well) was collected. The cell monolayer was washed twice with cold PBS and extracted with a deter- gent-containing buffer (0.15 M NaC1, 50 mM Tris-HC1, pH 7.4,1% Triton X-100, 1% deoxycholate, 0.1% SDS, 1 mM ethylenediamine tetraacetic acid [EDTA], 2 mM phenylmethyl-sulfonylfluoride). For measurements of soluble PAI-1, aliquots of the conditioned medium were added to the extraction buffer. Antiserum to PAI-1 was added to each sample, mixed at room temperature, and incubated overnight at 4°C. Polyclonal antibody raised in rabbits to bovine PAI-1 was a generous gift of D. Loskutoff (Scripps Clinic and Research Foundation, La Jolla, CAI. The samples were added to a 1: l suspension of protein A-Sepharose (Sigma) in extraction buffer containing aprotinin 10 U/ml (Sigma) incubated at room temperature for 1 h with continuous mixing. The protein A-Sepharose suspensions were washed twice by centrifugation and the supernatant discarded. The Sepharose beads were layered into new tubes containing extraction buffer in 10% glycerol, and the sample was centrifuged through the glycerol. The supernatant was discarded, and the beads were resuspended in reducing electrophoresis sample buffer, boiled for 10 min, and the proteins separated by SDS-PAGE. The gels were fixed in 25% methanol/lO% acetic acid, washed with EN’HANCE (NEN Dupont) for 1 h, washed in

H20 for 1 h, dried, and exposed to autoradiographic film (XAR film; Kodak) with one enhancing screen a t

u-PA binding assays Single-chain human urokinase-type plasminogen ac-

tivator (scuPA; Abbott Laboratories, North Chicago, IL) was iodinated with Nalz5I (NEN Dupontj and 10- dogen (Pierce) according to the manufacturer’s instruc- tions. After purification by heparin-Sepharose chromatography (Andrade-Gordon and Strickland, 1986; Mignatti et al., 19911, the labeled protein was 97-99% trichloroacetic acid (TCA) precipitable and had a specific activity of 0.21 pciipmole.

Confluent BAE cells in 35 mm dishes were incubated in serum-free medium for 24 h and cultured an additional 18 h in serum-free medium in the presence or absence of bFGF (3 ng/ml) andlor MPA lop6 M. The cultures were first washed with PBS and acid washed with 50 mM glycine-HC1,O.l M NaC1, pH 3.0, for 3 min at room temperature with mild shaking in order to release endogenously produced u-PA from its receptor. After neutralization with half a volume of 0.5 M HEPES, 0.1 M NaC1, pH 7.5, the monolayers were washed three times with PBS containing BSA (1 mgiml) (PBS-BSA). The cells were incubated at 4°C for 1 h with 500 pl of binding medium consisting of a-MEM supplemented with 20 mM HEPES containing BSA (1 mgiml) and aprotinin (100 pg/ml) (Sigma) in the presence or absence of the [alaz01 mouse synthetic u-PA 13-33 peptide (1,343 pgiml), corresponding to the receptor-binding sequence (Appella et al., 1987). This peptide was added to determine the specificity of the u-PA binding. At the end of the incubation, 1251- s~~PA (0.2 pCi) was added to the dishes in 500 p1 of binding medium containing aprotinin (100 pgiml) and the incubation was continued for 1 h. After removal of the medium, the cells were washed four times with 1 ml/dish of PBS- BSA and the cell layer was lysed with 100 pl of 0.56 Triton X-100 in 0.1 M Tris-HC1, pH 8.1. The radioactivity associated with the binding medium, PBS-BSA washings, and the cell lysates was measured in a gamma counter.

RESULTS Angiostatic steroids inhibit PA activity in

BAE cells We first quantitated the amount of PA in BAE cells

incubated with different steroid preparations, including some which Crum et al. (1985) reported to be angiostatic in the CAM assay. At equal concentrations of the steroids the response of the BAE cells varied. As shown in Figure lA, conditioned medium from BAE cells incubated in the presence of M hydrocortisone, MPA, dexamethasone, or 17-alpha-hydroxyprogesterone demonstrated decreases of 95%,73%, 73%, and 27%, respectively, of PA levels compared to control cultures. Agents which were not angiostatic in the assays of Crum et al. (1985) (pregnenolone, progesterone) did not suppress PA levels to the same degree as other agents in this assay (Fig. 2). The degree of inhibition did not uniformly correlate with the reported efficacy of the compounds as angiostatic agents in the CAM assay. For

-70°C.

EFFECT OF ANGIOSTATIC STEROIDS ON PA AND PAI-I 571

Fig. 1. BAE cell PA production in the presence of steroids and bFGF. Confluent BAE cells were incubated with the indicated steroid preparations (10 M) in the absence (A) or presence (B) of bFGF (3 ngiml). Conditioned medium was collected and assayed for PA activity as described in Materials and Methods. Lane a: basal; lane b: hydrocortisone (hydro); lane c: MPA; lane d: dexamethasone (dexa); lane e: 17-alpha-hydroxyprogesterone (17-a-hydroxy).

Fig. 2. BAE cell PA production in the presence of different steroids. Confluent BAE cells were incubated with the indicated steroid preparations (lo-" M). Conditioned medium was collected and assayed for PA activity as described in Materials and Methods. Lane a: basal; lane b: hydrocortisone; lane c: MPA; lane d: progesterone; lane e: tetrahydro-S; lane E pregnenolone.

example, 17-alpha-hydroxyprogesterone, which was extremely angiostatic on the CAM (Crum et al., 1985), did not elicit a strong inhibition of PA levels. The expla-

Fig. 3. BAE cell PA production in the presence of steroids, heparin, and bFGF. Confluent BAE cells were incubated with MPA (lo-' M) in the presence or absence of heparin (10 pgiml) and/or bFGF (3 ngiml). Conditioned medium was collected and assayed for PA activity as described in Materials and Methods.

nation for this nonconcordance is not clear but may relate to differences in the assay systems.

When the angiogenic factor, bFGF, was added to BAE cultures, PA secretion by BAE cells was stimulated more than two-fold from the basal level. This stimulatory effect on PA production was inhibited by the presence of the steroids (Fig. 1B). When lo-" M steroid (hydrocortisone, MPA, dexamethasone, or 17- alpha-hydroxyprogesterone) was added simultaneously with bFGF, the PA levels in the conditioned medium were decreased by 74%, 67%, 67%, and 21%, respectively, of that observed with cells stimulated by bFGF alone. This result suggested that angiostatic steroids might act by inhibiting bFGF-stimulated increases in PA. Similar results were observed with cultures of microvascular (BCE) cells, i.e., decreases in the PA activity of steroid-treated cells compared to the control. Be- cause of the greater ease of culturing aortic endothelial cells, in our subsequent experiments we used only BAE cells.

Effect of heparin on BAE cell PA activity Since previous studies by others (Crum et al., 1985;

Gross et al., 1981) had demonstrated an angiostatic effect in the CAM only when steroids were coadministered with heparin, we tested the effect of heparin in our system. Control cells and heparin-treated cells produced comparable amounts of PA activity in the conditioned medium (Fig. 3) , although heparin by itself caused a slight inhibition. In cultures exposed to bFGF (3 ngiml), there was a greater than three-fold stimulation of PA. This was abrogated in the presence of MPA as noted above. The PA activity of BAE cells treated with heparin (10 pgiml) and MPA M) was similar to that noted when cells were incubated with MPA

M) alone. Similarly, PA production in cells incubated with bFGF plus steroid plus heparin was almost

BLEI ET AL. 572

equivalent to that noted when cells were exposed only to steroid and bFGF. Therefore, all subsequent experiments were performed in the absence of heparin unless noted.

Subsequent studies focused on the action of MPA since this was one of the more effective agents in sup- pressing PA activity. This drug has been shown to inhibit vascularization, collagenolysis, and tumor growth in vivo (Gross et al., 1981; Oikawa et al., 1988; Fuji- moto et a]., 1989a,b). Furthermore, the clinical efficacy of this agent has been studied for certain types of breast carcinoma (Jakobsen et al., 1986).

Dose-response effect of MPA on PA levels of BAE cells

The exquisite sensitivity of BAE cell PA production to MPA was demonstrated by decreases of PA activity at the following doses of MPA: lo-" M, lop7 M, M, and lo-' M, and this resulted in decreases of PA levels of 83%, 83%, 75%, and 39%, respectively, compared to the basal PA activity in the conditioned medium (data not shown). These data corroborate with what has previously been shown by Ashino-Fuse et al. (1989). Some variability existed among individual experiments with respect to basal BAE cell PA levels as well as the response of the cells to stirnulatory and inhibitory agents. This probably reflects subtle differences in the BAE cells and has been seen previously. However, the patterns of stimulation and inhibition were reproducible among experiments.

PA level of cell extracts We next compared possible differences in PA activity

in conditioned media and in cell extracts. As shown in Figure 4A, exposure of BAE cells to bFGF caused a threefold increase in PA activity in the culture medium and exposure to MPA (lo-' MI resulted in complete suppression of PA, as expected from our earlier experiments. TGF-P-treated cells were used as positive con- trols to demonstrate the decrease in PA both in the conditioned medium and the cell extract as previously described for this cytokine (Saksela et al., 1987; Sak- sela and Rifkin, 1988). These samples exhibited a decrease in PA activity in both the conditioned medium (Fig. 4A, column e) and the cell extract (Fig. 4B, column el. In contrast, the PA activity of bFGF-treated cells in the presence of 10 ' M MPA showed a striking differ- ence when the conditioned medium and the cell extracts were compared. The PA activity in the conditioned medium of these steroidibFGF-treated cells was suppressed to the basal level as described earlier. How- ever, the PA activity in cell extracts under these conditions was not decreased (Fig. 4B, column d) and was almost equivalent to the bFGF-treated sample (Fig. 4B, column b). These results suggested that the MPA-induced inhibition of bFGF-stimulated PA activity in the conditioned medium was the result of MPA directly blocking the effect of bFGF on PA synthesis (see Dis- cussion).

The samples used for the lz5I-fibrin plate assay (Fig. 4) were also assayed to characterize the type of PA (u-PA vs. tissue-type plasminogen activator [t-PA]) produced using plasminogen-containing overlay gels (see Methods). The zones of lysis (Fig. 5 ) appeared in a

Fig. 4. PA levels in cell extracts and conditioned medium. Confluent BAE cells were incubated with bFGF (3 ng/ml), MPA (10 ' M), bFGF plus MPA, or TGF-p (1 ngiml). Conditioned media (A) and cell extracts (B) were collected and assayed for PA activity as described in Materi- als and Methods. Lane a: basal; lane b: bFGF (3 ngiml); lane c: MPA

M); lane d: bFGF ( 3 ngiml) + MPA (lo-" M); lane e: TGF-p (1 ngiml).

position corresponding to a molecular weight (MW) of approximately 52 kDa, the known MW for bovine u-PA (Levin and Loskutoff, 19821, suggesting that the pre- dominant form of PA present in these cells is u-PA (Moscatelli, 1986). The intensity of the different bands (Fig. 5) also corroborated the PA activity patterns noted in the more quantitative fibrin plate assays; e.g., PA was increased in the presence of bFGF (Fig. 5c,h) and decreased in the presence of MPA (Fig. 5d,i). The observed bFGF-induced increase in fibrinolysis was

EFFECT OF ANGIOSTATIC STEROIDS ON PA AND PAI-1 573

Fig. 5. PA zymography. Confluent BAE cells were incubated with bFGF, TGF-p, MPA, or MPA plus bFgF a s in Figure 4. Aliquots of conditioned media (lanes a-e) and cell extracts (lanes f-j) were separated by SDS-PAGE. The gel was then placed atop a plasminogen- containing fibrin overlay gel and monitored a s described in Materials

blocked by the inclusion of MPA in the conditioned medium (Fig. 5, compare c and e) but not in the cell extracts (Fig. 5 compare h and j). TGF-P-treated cells demonstrated a decrease in PA levels in the conditioned medium (Fig. 5b) as well as in the cell extract (Fig. 5g) .

PAI-1 induction by MPA A possible explanation for the differences in PA ac-

tivities in conditioned medium vs. cell extracts in cells simultaneously exposed to bFGF and MPA was that MPA treatment stimulated the synthesis of the PA inhibitor PAI-1. An increase in PAI-1 production would inhibit PA in the culture medium but not bind to cell- associated scuPA, as only the active two-chain form of u-PA is able to complex with PAI-1 (Skriver et al., 1982; Wun et al., 1982; Andreasen et al., 1986). There- fore, we assayed PAI-1 levels by reverse zymography. This technique demonstrates the presence of PAI-1 by zones of inhibition of lysis due to PAI-1 in an indicator gel that has been otherwise completely lysed by plasmin formed by u-PA included in the gel. Figure 6 illus- trates results of the reverse zymography assay of the samples. The MW of the zones of inhibition corresponds to 45 kDa, the known MW of bovine PAI-1 (Erickson et al., 1984; Saksela e t al., 1987). TGF-P has been shown to induce PAI-1 synthesis in endothelial cells (Saksela and Rifkin, 1988) and, therefore, was used as a positive control (lanes c and g). Consistent with our hypothesis, a significant increase in PAI-1 was observed in the MPA-treated samples when compared to the control both in the conditioned medium (lane d vs. a) and in the cell extract (lane i vs. 0.

and Methods. Zones of lysis correspond to a MW of 52 kDa, the known MW of bovine u-PA. Lanes a,f: control; lanes b,g: TGF-(3 (1 ngirnl); l anes c ,h bFGF (3 ngiml); lanes d,i: MPA (10 ' MI; lanes e j : bFGF (3 ng/ml) plus MPA MI.

Immunoprecipitation of conditioned medium from 35S-labelled BAE cells with anti-PAI-1 antibodies is shown in Figure 7. The bands a t a position of 45 kDa represent PAI-1 as shown by Erickson et al. (1984, 1985) and Saksela et al. (1987). A very strong induction of PAI-1 was observed with TGF-P (lane b). Treatment of the cells with steroid ( lop6 M MPA) (lane d) resulted in increases in PAI-1 synthesis, although to a lesser degree than that seen with TGF-P. Interestingly, the amount of PAI-1 in the bFGF-treated sample was also increased (lane c) when compared to the control. This effect of bFGF on PA and PAI-1 production has been observed previously (Saksela et al., 1987; Pepper et al., 1990) and is also obvious in Figure 6 lane b vs. a (in Fig. 6). An additive induction of PAI-1 by steroids plus bFGF was seen (Fig. 7e), supporting the hypothesis that the steroids limit PA activity of BAE cells by increasing inhibitor levels. Lane f represents immunoprecipitation with nonimmune serum.

Effect of MPA on u-PA binding to BAE cells It has recently been shown that bFGF increases the

number of u-PA binding sites in bovine capillary endothelial cells and human umbilical vein endothelial cells (Mignatti et al., 1991). This effect is believed to increase pericellular proteolysis and potentiate the invasive capability of endothelial cells. Therefore, it was of interest to determine whether MPA affected the binding of scuPA to BAE cells. As shown in Figure 8, bFGF increased the binding of scuPA to BAE cells almost fourfold. The specificity of the binding was determined by competition with the Lala'"] mouse u-PA 13-33 pep-

574 BLEI ET AL.

Fig. 6. PAI-1 activity in control and steroid-treated cultures. Conflu- ent BAE cells were incubated with bFGF, TGF-b, MPA, or MPA + bFGF. Aliquots of conditioned media (lanes a-e) and cell extracts (lanes f-j) were separated by SDS-PAGE. This gel was then placed atop plasminogen and urokinase-containing fibrin overlay gels

tide which corresponds to the receptor-binding sequence of mouse u-PA. This peptide has been shown to compete with human scuPA binding to bovine endothelial cells (Mignatti et al., 1991). As shown in Figure 8, over 90% of scuPA binding to BAE cells was abolished by the mouse peptide, indicating that scuPA binding occurred through the cell membrane receptor for u-PA. MPA treatment of BAE cells did not affect the basal level of scuPA binding capacity nor did it abolish the increase in scuPA binding to bFGF-treated BAE cells. Thus MPA does not affect the bFGF-induced increase in u-PA binding sites.

DISCUSSION Recently several papers have described naturally oc-

curring inhibitors of angiogenesis, some of which are found in specific tissues. Proteins purified from cartilage (Lee and Langner, 1983; Moses et al., 1990) and ocular vitreous (Lutty et al., 1983), protamine (Taylor and Folkman, 1982), platelet factor 4 (Maione et al., 1990), antibiotics (Oikawa et al., 1989; Ingber et al., 1990; Tamargo et al., 1991), interferons (Tsurukoka et al., 1988; Sidky and Borden, 1987), and a variety of steroids in combination with heparin or heparin-like molecules (Crum et al., 1985; Folkman et al., 1989) have been shown to be antiangiogenic in several biological models. However, little is known as to the possible mechanisms of their actions.

In general, both microvascular (BCE) and large vessel (BAE) cells gave the same response. Although new vessel formation in vivo occurs in microvessels, both capillary endothelial cells and large vessel endothelial cells have been employed for in vitro studies. The data of Pepper et al. (1992) demonstrated that induction of PAI-1 expression was similar in wounded monolayers of bovine aortic and bovine capillary endothelial cells.

and allowed to develop as described in Materials and Methods. Zones of inhibition of lysis correspond to a MW of 45 kDa, the known MW of bovine PAI-1. L a n e a.fi control: lanes b.h: bFGF 3 (ndml): l anes c . ~ : TGF-P (1 ngiml); lanes d,i: MPA (10 MI; lanes e jybFGF (3 ngimi)

MPA (10-7 M).

The data of Ingber and Folkman (1989) demonstrated tube formation in bFGF-stimulated capillary cells, whereas Pepper et al. (1990) showed migration and tube formation in large vessel endothelial (calf pulmonary artery) cells in response to bFGF. The above studies show that cells derived from large and small vessels behave the same. If in vivo angiogenesis originates from microvasculature cells (Folkman, 1984), then these in vitro studies may duplicate the in vivo effects of these cells.

We have found that angiostatic steroids inhibit the fibrinolytic activity of endothelial cells exposed to the angiogenic factor bFGF. Ashino-Fuse et al. (1989) have shown that MPA inhibited PA activity in bovine capillary, aortic, and pulmonary artery endothelial cells. Our results demonstrate that in the case of MPA, this is related primarily to increased levels of PAI-1 production in the conditioned medium. Interestingly, whereas secreted PA levels are decreased in cells exposed to the combination of the angiogenic factor and the angiostatic steroid, cell-associated PA levels remain high. An explanation for this phenomenon is that the angiostatic steroids do not directly block the effect of bFGF but act indirectly and overcome the effect of increased protease production by inducing high levels of the PA inhibitor, PAI-1.

The primary type of PA in our cells was identified as u-PA by zymography. Thus, PA found in the conditioned medium, which is primarily two-chain u-PA, is inhibited by the PAI-1. Surface-bound two-chain u-PA, which is the product of activation of pro-u-PA is also inhibited by PAI-1 (Cubellis et al., 1990). However, most of the cell-associated PA is in the proenzyme form and is not inhibitable by PAI-1 (Skriver et al., 1982; Wun et al., 1982; Andreasen et al., 1986). Thus, this fraction of PA may represent the activity detected in

575 EFFECT OF ANGIOSTATIC STEROIDS ON PA AND PAI-1

t o l n l blnd lng

n o n i p e c l l l c blndlng

E P

m

Fig. 7. Immunoprecipitation of BAE cell conditioned medium by antibodies to PAI-1. BAE cells were labeled with "S-methionine in the presence or absence of TGF-P, bFGF, MPA, or bFGF plus MPA at the indicated concentrations. Conditioned medium was collected and im- munoprecipitated with polyclonal antibodies to bovine PAI-1 or nonimmune serum (NI) as described in Materials and Methods. Samples were separated by SDS-PAGE, and the dried gel was analyzed by autoradiography. The band a t 45 kDa corresponds to PAI-1. Lane a: control; lane b: TGF-P (1 ngiml); lane c: bFGF ( 3 ngiml); lane d: MPA

MI; lane E NI, nonimmune serum.

MI; lane e: bFGF (3 ng/ml) plus MPA

our assay system, since the pro-u-PA can be activated under conditions of the fibrin plate assay and its activity measured. These results strongly suggest that angiostatic steroids do not block the primary action of bFGF, but exert their effect by neutralizing the bFGF- induced increases in extracellular proteases. In angiogenesis, it is only the extracellular levels of PA and its inhibitors that are important. Although we do not have direct evidence that binding of u-PA to its receptor is stabilized by MPA, the indirect evidence of our data suggests no change in stability. Our results demonstrate that the steroids have no effect on u-PA binding to BAE cells (Fig. 8).

Interestingly both bFGF and MPA increase PAI-1 production in BAE cells. The amount of PAI-1 induced by bFGF appears to be greater than that induced by MPA. This result appears to be contradictory to observations made by measuring the PA activities of cells treated with these two agents. However, i t may be that the amount of active PAI-1 in the conditioned medium differs under the various conditions. PAI-1 is known to lose activity rapidly under tissue culture conditions. It is possible that MPA causes additional changes which stabilize the PAI-1 thereby enhancing the concentration of proteolytic inhibitor. The fact that the total

C o n t r o l MPR bFGF bFGF + MPR ( 1 0 - 6t-1) (3 n g / m ~ ~

Fig. 8. scuPA binding to BAE cells. Confluent BAE cells were incubated in the presence of MPA (10 ' M), bFGF (3 ngiml), or bFGF ( 3 nglml) + MPA (10 ' M). The monolayers were washed, acid treated to remove endogenous u-PA bound to the cells, and incubated in the presence or absence of the lala2'1 mouse u-PA 13-33 peptide, which corresponds to the receptor-binding region of uPA. '2"I-scuPA was added, and cell lysates were assayed for radioactivity. The scored bars represent specific binding of radiolabelled scuPA associated with the cell lysate for the conditions described. The shaded bars represent results of the same experiment done in the presence of excess lala"1 mouse u-PA 13-33 peptide.

amount of active PAI-1 in bFGF and bFGF plus MPA- treated samples appears similar after electrophoresis (Fig. 6) probably does not reflect the situation in the original culture medium since SDS-PAGE is known to activate inactive PAI-1.

The balance between proteolytic activity and inhibition is crucial. Whereas the induction of PAI-1 in the presence of the steroids alone does not appear to be as intense a s that with bFGF, in fact the total proteolytic balance is inhibitory, i.e., MPA plus bFGF is less proteolytic than bFGF itself. The overall effect of bFGF is an increase in PA production. Whereas bFGF induces both the inhibitor and PA, MPA induces only the inhibitor. It would be important to look a t other angiostatic steroids to see if they act in the same or different mechanisms. Given the variety, one may not expect them to act the same.

By interfering with the invasive step of the angiogenic response, one might successfully arrest patho- genic blood vessel formation. The prevention of endothelial cell basement membrane degradation could be achieved by the application of agents which would inhibit protease synthesis andlor secretion. Alterna- tively, the induction of increased levels of protease inhibitors might block invasion. Mignatti et al. (1989) have demonstrated that tissue inhibitor of metalloproteinase (TIMP), 1,lO-phenanthroline, as well as antibodies to collagenase and PA and inhibitors of plasmin inhibit endothelial cell invasion in an in vitro model of angiogenesis. Consistent with this result is the obser- vation of Moses et al. (1990) identifying the antiangiogenic molecule in shark cartilage as a TIMP-like an- ticollagenase. MPA has also been shown to block the collagenolytic activity associated with tumor growth

BLEI ET AL. 576

and vascularization (Gross e t al., 1981). It would be interesting to determine if angiostatic steroids also increased the production of TIMP as well as PAI-1.

Inhibitors of endothelial cell migration or proliferation would also function to delimit blood vessel growth. The interaction between endothelial cells and matrix- related molecules may contribute to these processes. Nicosia and Madri (1987) observed dynamic changes in microvascular extracellular matrix during angiogenesis. Tokida et al. (1990) have demonstrated that upon incubation with MPA (in the absence of heparin), BAE cells produce a laminin complex of a different structure from control cells. These investigators suggest that angiostatic steroids may influence intracellular process- ing of laminin subunits, resulting in a laminin complex associated with “quiescent” endothelial cells (i.e., nonangiogenic). This effect may represent an indepen- dent action of angiostatic steroids which blocks angiogenesis.

The modulation of u-PA activity by PAI-1 indirectly regulates both the local plasmin and collagenase activities. The major substrate for PA is plasminogen. Plas- minogen is present in both intravascular and extravas- cular sites including the extracellular matrix of endothelial cells (Mignatti et al., 1988). The active form of u-PA converts the proenzyme plasminogen to plasmin. Plasmin, also a serine protease, has a wide variety of substrates including several basement membrane components surrounding the cell (Dan# et al., 1985). However, since collagens are resistant to plasmin, active collagenases are required for the complete lysis of the basement membrane as well as the degradation of stromal collagens. Plasmin mediates the activation of procollagenase to active collagenase (Werb et al., 1977). Thus, a cascade of proteolytic activations, pro-u-PA to u-PA, plasminogen to plasmin, and prometalloprotein- ases to metalloproteinases must take place for the endothelial cells to breach the basement membrane.

Further experiments analyzing the regulation of PA and PAI-1 and assessing other spects of the angiogenic response in the presence of angiostatic steroids would be informative. Our data suggest that PA activity is suppressed by MPA indirectly by increased PAI-1 production. Van Zonneveld et al. (1988) have identified the promoter region of the PAI-1 gene and have shown PAI-1 induction in the presence of dexamethasone via a glucocorticoid responsive element within the PAI-1 gene. Similarly, Riccio et al. (1988) characterized the promoter region of the PAI-1 gene and demonstrated its regulation by glucocorticoids at the promoter level. It is likely that the induction of PAI-1 in our model system functions via this steroid-responsive sequence. How- ever, the steroid-responsive sequence may be more general than the glucocorticoid responsive element if steroids with no glucocorticoid properties exert their angiostatic effects via this mechanism.

Our results appear to be contradictory to those described by Ingber et al. (1986), who found that angiostatic steroids in the presence of heparin enhanced capillary basement membrane breakdown which may reflect increased, rather than decreased, proteolytic activity. It is noteworthy that the studies of Ingber et al. (1986) were different from those described in our work in that 1) the system of study involved chick CAM mi-

crovessels, whereas our studies thus far have been primarily on bovine aortic endothelial cells, and 2) angiostatic steroids induced avascular zones in growing vessels only when coadministered with heparin, whereas our studies evaluated effects of these agents on confluent monolayers of endothelial cells primarily without the addition of exogeneous heparin. Further experiments will be required to clarify these appar- ently different results.

It is important to note that the significance of heparin in endothelial cell physiology is somewhat para- doxical. The angiogenic proteins, acidic and basic fibroblast growth factor, both exhibit affinity for heparin which appears to modulate and stabilize the activity of these molecules (Vlodavsky et al., 1987; Sommer and Rifkin, 1989). On the other hand, heparin and heparin- like molecules appear to enhance the antiangiogenic response of certain steroids (Crum et al., 1985; Folk- man et al., 1989). However, our in vitro data do not indicate an increased inhibitory effect of angiostatic steroids on PA levels in BAE cell cultures in the presence of heparin compared to steroid alone.

In conclusion, a possible mechanism of action of an angiogenic steroid is described related to its influence on the endothelial cell PAiplasmin system. The angiostatic steroid MPA stimulates PAI-1 synthesis. Angio- genic factors (such as bFGF) stimulate endothelial cell PA production. Thus, the stimulation of PA activities by bFGF measured in the conditioned medium was abrogated by treatment with MPA. If angiogenesis is de- pendent upon an intact fibrinolytic system in endothelial cells, these effects of MPA may account for its ability to block angiogenesis.

ACKNOWLEDGMENTS The authors wish to thank Dr. D. Moscatelli for his

suggestions and advice during the course of this project. This work was supported by a training grant from the National Institute of General Medical Sciences 5 T32 GM 07552 (F.B.), NIH grants CA 49419 (E.L.W.), CA 34282 (D.B.R.), CA 20194 (D.B.R.), and The Benjamin Weiss Fund (E.L.W.).

LITERATURE CITED Andrade-Gordon, P., and Strickland, S. (1986) Interaction of heparin

with plasminogen activators and plasminogen: Effects on the activation of plasminogen. Biochemistry, 25r4033-4040.

Andreasen, P.A., Christensen, T.H., Huang, J.-Y., Nielsen, L.S., Wil- son, E.L., and Dan@, K. (1986) Hormonal regulation of extracellular plasminogen activators and Mr 54,000 plasminogen activator inhibitor in human neoplastic cell lines, studied with monoclonal antibodies. Mol. Cell. Endocrinol., 45:137-147.

Appella, E., Robinson, E.A., Ullrich, S.J., Stoppelli, M.P., Corti, A., Cassani, G., and Blasi, F. (1987) The receptor-binding sequence of urokinase. J. Biol. Chem., 262t44374440.

Ashino-Fuse, H., Takano, Y., Oikawa, T., Shimura, M., and Iwaguchi, T. (1989) Medroxyprogesterone acetate, an anti-cancer and antiangiogenic steroid, inhibits plasminogen activator in bovine endothelial cells. Int. J. Cancer, 44:859.

Ausprunk, D., and Folkman, J. (1977) Migration and proliferation of endothelial cells in pre-formed and newly formed blood vessels during tumor angiogenesis. Microvasc. Res., 14.53-65.

Crum, R., Szabo, S., and Folkman, J. (1985) A new class of steroids inhibits angiogenesis in the presence of heparin or a heparin frag- ment. Science, 230:1375-1378.

Cubellis, M.V., Wun, T-C., and Blasi, F. (1990) Receptor-mediated internalization and degradation of urokinase is caused by its specific inhibitor PAI-1. EMBO J., 9r1079-1085.

577 EFFECT OF ANGIOSTATIC STEROIDS ON PA AND PAI-1

Dane, K., Andreasen, P., Gr@ndahl-Hansen, J., Kristensen, P., Nielsen, L., and Skriver, L. (1985) Plasminogen activators and cancer. Adv. Cancer Res., 44:202-214.

Deutsch, D.G., and Mertz, E.T. (1970) Plasminogen purification from human plasma by affinity chromatography. Science, 170t1095- 1096.

Erickson, L.A., Lawrence, D.A., and Loskutoff, D.J. (1984) Reverse fibrin autoradiography: A method to detect and partially characterize protease inhibitors after d i u m dodecyl sulfate polyacrylamide gel electrophoresis. Anal. Biochem., 137:454-463.

Erickson, L.A., Hekman, C., and Loskdtoff, D.J. (1985) The primary plasminogen activator inhibitors in endothelial cells, platelets, and plasma are immunologically related. Proc. Natl. Acad. Sci., U.S.A., 82t8710-8714.

Folkman, J. (1984) Angiogenesis. In: Biology of Endothelial cells. E.A. Jaffe, ed. Martinus Nijhoff Publishers, Boston, MA, pp. 412-428.

Folkman, J. (1987) Angiogenesis. In: Thrombosis and Hemostasis 1987. M. Verstrate, J . Vermylen, H. Lijnen, and J . Arnout, eds. International Society on Thrombosis and Hemostasis and Leuven University Press, Leuven, Belgium, pp. 583-596.

Folkman, J., and Klagsbrun, M. (1987) Angiogenic factors. Science, 235r442-447.

Folkman, J., Haudenschild, C.C., and Zetter, B. (1979) Long-term culture of capillary endothelial cells. Proc. Natl. Acad. Sci., U.S.A., 76r5217-5221.

Folkman, J., Weisz, P., Joullie, M., Li, W., and Ewing, W. (1989) Control of angiogenesis with synthetic heparin substitutes. Science, 243.1490-1493.

Fujimoto, J., Hosoda, S., Fujita, H., and Okada, H. (1989a) Inhibition of tumor angiogenesis activity by medroxyprogesterone acetate in gynecologic malignant tumors. Invasion Metastasis, 9:269-277.

Fujimoto, J . , Hosoda, S., Fujita, H., and Okada, H. (1989133 Inhibition of tumor angiogenesis activity in C,H mouse mammary tumor by medroxyprogesterone acetate. Acta Obstet. Gynaecbl. Jpn., 41 :77- 82.

Goldie, I. (1970) The synovial microvasculature in rheumatoid arthritis and osteoarthritis. Acta Orthop. Scand., 40r751-764.

Granelli-Piperno, A., and Reich, E. (1978) A study of proteases and protease-inhibitor complexes in biological fluids. J . Exp. Med., 148.223-234.

Gross, J., Azizkhan, R.G., Biswas, C., Bruns, R.R., Hseih, D., and Folkman, J. (1981) Inhibition of tumor growth, vascularization, and collagenolysis in the rabbit cornea by medroxyprogesterone. Proc. Natl.Acad.Sci. U.S.A., 78:11761180.

Gross, J . , Moscatelli, D., Jaffe, E.A., andRifkin, D.B. (1982) Plasmino- gen activator and collagenase production by cultured capillary endothelial cells. J. Cell Biol., 95r974981.

Gross, J., Moscatelli, D., and Rifkin, D.B. (1983) Increased capillary endothelial cell protease activity in response to angiogenic stimuli in vitro. Proc. Natl. Acad. Sci. U.S.A., 80:2623-2627.

Ingber, D.E., and Folkman, J . (1989) Mechanochemical switching between growth and differentiation during fibroblast growth factor- stimulated angiogenesis in vitro: Role of extracellular matrix. J . Cell Biol., 109:317330.

Ingber, D.E., Madri, J.A., and Folkman, J. (1986) A possible mechanism for inhibition of angiogenesis by angiostatic steroids: Induc- tion of capillary basement membrane dissolution. Endocrinology, 119:176&1775.

Ingber, D., Fujita, T., Kishimoto, S., Sudo, K., Kanamaru, T., Brem, H., and Folkman, J. (19901 Synthetic analogues of fumagillin that inhibit angiogenesis and suppress tumor growth. Nature, 348:555- 557.

Jakobsen, A,, Frederiksen, P.L., Ml le r , K., Andersen, A.P., Brincker, H., Dombernowsky, P., Hansen, P.V., Hesselius, I., and Kjer, M. (1986) Medroxyprogesterone acetate and prednisone in advanced breast cancer. A randomized trial. Eur. J. Cancer Clin. Oncol., 22t1067-1072.

Lee, A., and Langner, E. (1983) Shark cartilage contains inhibitors of tumor angiogenesis. Science, 211 r1185-1187.

Levin, E.G., and Loskutoff, D.J. 11982) Cultured bovine endothelial cells produce both urokinase and tissue-type plasminogen activators. J. Cell Biol., 94t631-636.

Lutty, G.A., Thompson, D.C., Gallup, J.Y., Mello, R.J., Patz, A., and Fenselau, A. (1983) Vitreous: An inhibitor of retinal extract-induced neovascularization. Invest. Ophthalmol. Vis. Sci., 23:52-56.

Yaione, T.E., Gray, G.S., Petro, J., Hunt, Ad., Donner, A.L., Bauer, S.I., Carson, H.F., and Sharpe, R. (1990) Inhibition of angiogenesis by recombinant human platelet factor-4 and related peptides. Sci- ence, 247:77-79.

Mignatti, P., Welgus, H.G., and Rifkin, D.B. (1988) Role of degrada- tive enzymes in wound healing. In: The Molecular and Cellular Biology of Wound Repair. R.A.F. Clark, and P.M. Henson, eds. Ple- num Press, New York, pp. 497-523.

Mignatti, P., Tsuboi, R., Robbins, E., and Rifkin, D.B. (1989) In vitro angiogenesis on the human amniotic membrane: Requirement for basic fibroblast growth factor-induced proteinases. J. Cell Biol., 108:67 1-682.

Mignatti, P., Mazzieri, R., and Rifkin, D.B. (1991) Expression of the urokinase receptor in vascular endothelial cells is stimulated by basic fibroblast growth factor. J . Cell Biol., 113t1193-1201.

Moscatelli, D. (1986) Urokinase-type and tissue-type plasminogen activators have different distributions in cultured bovine capillary endothelial cells. J . Cell. Biochem., 30:19-29.

Moscatelli, D., and Rifkin, D.B. (1988) Membrane and matrix localiza- tion of proteinases: A common theme in tumor cell invasion and angiogenesis. Biochim. Biophys. Acta, 948:67-85.

Moscatelli, D., Presta, M., and Rifkin, D.B. (1986) Purification of a factor from human placenta that stimulates capillary endothelial cell protease production, DNA synthesis, and migration. Proc. Natl. Acad. Sci. U.S.A., 83:2091-2095.

Moses, M.A., Sudhalter, J., and Langer, R. (1990) Identification of an inhibitor of neovascularization from cartilage. Science, 248t1408- 1410.

Mulliken, J.B., and Young, A.E. (1988) Vascular Birthmarks: Heman- giomas and Malformations. W.B. Saunders Co., Philadelphia, PA.

Newell, F.W., ed. (1986) Ophthalmology Principles and Concepts. C.V. Mosby Co., St. Louis, MO.

Nicosia, R.F., and Madri, J.A. (1987) The microvascular extracellular matrix: Developmental changes during angiogenesis in the aortic ring plasma clot model. Am. J. Pathol., 128:7%90.

Oikawa, T., Hiragun, A,, Yoshida, Y., Ashino-Fuse, H., Tominaga, T., and Iwaguchi, T. (1988) Angiogenic activity of rat mammary carci- nomas induced by 7,12-dimethyl-benz(a)anthracene and its inhibition by medroxyprogesterone acetate: Possible involvement of antiangiogenic action of medroxyprogesterone acetate in its tumor growth inhibition. Cancer Lett., 43:85-92.

Oikawa, T., Hirotani, K., Shimamura, M., Ashino-Fuse, H., and Iwaguchi, T. (1989) Powerful antiangiogenic activity of Herbimycin A. J. Antibiot., 42r1202-1204.

Pepper, M.S., Belin, D., Montesano, R., Orci, L., and Vassalli, J.-D. (1990) Transforming growth factor-beta 1 modulates basic fibroblast growth factor-induced proteolytic and angiogenic properties of endothelial cells in vitro. J. Cell Biol., lllt743-755.

Pepper, M., Sappino, A.P., Montesano, R., Orci, L., and Vassalli, J.-D. (1992) Plasminogen activator inhibitor-1 is induced in migrating endothelial cells. J. Cell. Physiol., 153r129-139.

Riccio, A., Lund, L.R., Sartorio, R., Lania, A., Andreasen, P.A., Dano, K., and Blasi, F. (1988) The regulatory region ofthe human plasminogen activator inhibitor type-1 (PAI-1) gene. Nucleic Acids Res.,

Saksela, O., and Rifkin, D.B. (1988) Cell-associated plasminogen activation: Regulation and physiological functions. Annu. Rev. Cell Biol., 4:93-126.

Saksela, 0.. Moscatelli, D., and Rifkin, D.B. (1987) The opposing effects of basic fibroblast growth factor and transforming growth factor beta on the regulation of plasminogen activator activity in capillary endothelial cells. J. Cell Biol., 105r957-963.

Schwartz, S. (1978) Selection and characterization of bovine aortic endothelial cells. In Vitro, 14t966980.

Sidky, Y.A., and Borden, E.C. (1987) Inhibition of angiogenesis by interferons: Effects on tumor- and lymphocyte-induced vascular re- ponses. Cancer Res., 47:5155-5161.

Skriver, L., Nielsen, L.S. Stephens, R., and Dan@, K. (1982) Plasmino- gen activator released as inactive proenzyme from murine cells transformed by sarcoma virus. Eur. J . Biochem., 124:409414.

Sommer, A,, and Rifkin, D.B. (1989) Interaction of heparin with human basic fibroblast growth factor (FGF): Protection of the angiogenic protein from proteolytic degradation by a glycosaminoglycan. J. Cell. Physiol., 138:215-220.

Tamargo, R.J., Bok, R.A., and Brem, H. (1991) Angiogenesis inhibition by minocycline. Cancer Res., 51 :672-675.

Taylor, S., and Folkman, J. (1982) Protamine is an inhibitor of angiogenesis. Nature, 297:307-312.

Tokida, Y., Aratani, Y., Morita, A., and Kitagawa, Y. (1990) Produc- tion of two variant laminin forms by endothelial cells and shift of their relative levels by angiostatic steroids. J. Biol. Chem., 265t18123-18129.

Tsurukoka, N., Sugiyama, M., Tawaragi, Y., Tsujimoto, M., Nishi-

16t2805-2824.

578 BLEI ET AL.

hara, T., Goto, T., and Sato, N. (1988) Inhibition ofin vitro angiogenesis by lymphotoxin and interferon-gamma. Biochem. Biophys. Res. Commun.,155:429435.

Van Zonneveld, A.-J., Curriden, S.A., and Loskutoff, D.J. (1988) Type 1 plasminogen activator inhibitor gene: Functional analysis and glucocorticoid regulation of its promoter. Proc. Natl. Acad. Sci. U.S.A., 85t5525-5529.

Vlodavsky, I., Folkman, J., Sullivan, R., Fridman, R., Ishai-Michaeli, R., Sasse, J., and Klagsbrun, M. (1987) Endothelial cell-derived

basic fibroblast growth factor: Synthesis and deposition into suben- dothelial extracellular matrix. Proc. Natl. Acad. Sci. U.S.A., 84r2292-2296

Werb, Z., Mainardi, C., Vater, C., and Harris, E. (1977) Endogenous activation of latent collagenase by rheumatoid synovial cells. Evi- dence for a role of plasminogen activator. N. Eng. J. Med., 29fir1017-1023.

Wun, T-C., Ossowski, L., and Reich, E. (1982) A proenzyme form of urokinase. J. Biol. Chem., 257r7262-7268.

Mechanism of action of angiostatic steroids: Suppression of plasminogen activator activity via stimulation of plasminogen activator inhibitor synthesis

Documents