BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS Pages 633-638

Vol. 176, No. 2,1991 April30,1991

PLASMINOGEN ACTIVATOR INHIBITOR-1 IS INDUCED IN MICROVASCULAR ENDOTHELIAL CELLS BY A CHONDROCYTE-DERIVED TRANSFORMING GROWTH FACTOR-BETA M.S.Pepper, R.Montesano, L.Orci and J.-D.Vassalli Institute of Histology and Embryology, Department of Morphology, University of Geneva Medical Center, 1211 Geneva 4, Switzerland Received March 7,

1991

SUMMARY We have previously demonstrated that a chondrocyte-derived TGF-B inhibits spontaneous endothelial sprout formation in an in vitro model of angiogenesis (Pepper et al., J.Cell.Physiol. 146:170-179, 1991). We suggested that the inhibition might be mediated by an antiproteolytic effect. In this paper, we describe the induction of PAI-1 and PAI-1 mRNA in rnicrovascular endothelial cells by chondrocyte conditioned medium. This effect can be significantly reduced by the addition of anti-TGF-B antibodies to the conditioned medium, and can be mimicked by the addition of exogenous TGF-B1. Taken together with our previous observations, these results demonstrate that the inhibition of endothelial sprout formation occurs concomitantly with an increase in the production of PAI-1, a physiological plasminogen activator inhibitor. This suggests that TGF-B-induced antiproteolysis is responsible for the inhibition of sprout formation. ~ ~99~~aemi~ ~..... ~no.

Almost all tissues of the body are vascularized by an extensive capillary network. One notable exception in the adult is hyaline cartilage; during development, hyaline cartilage also appears transiently during the process of enchondral bone formation in the form of miniature anlagen of the future skeletal elements. In addition to this avascularity, cartilage explants or extracts as well as isolated chondrocytes have been shown to inhibit vascular invasion (1-5). Angiogenesis, the formation of new capillary blood vessels, consists of a number of sequential events which result in the formation of a capillary sprout. We and others have suggested that at least three elements of this process may be proteolytically mediated: degradation of the investing basement membrane, invasion of the interstitial extracellular matrix and capillary lumen formation (6). We have recently described an in vitro model of spontaneous endothelial sprout formation (7). When used in co-culture with isolated chondrocytes, sprout formation was markedly inhibited. On the basis of antibody inhibition studies, we identified a TGF-B as a mediator of this effect. We suggested that the TGF-Bmediated inhibition might have resulted from an antiproteolytic effect. The studies reported in this paper demonstrate that chondrocyte-conditioned medium induces PAI-1 in Abbrevi.ations used: TGF-B - transforming growth factor-B; PAI - plasminogen activator

inhibitor; BME - bovine microvascular endothelial. 0006-291X/91 $1.50

633

Copyright © 1991 by Academic Press, Inc. All rights of reproduction in any form reserved.

Vol. 176, No. 2, 1991

BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

microvascular endothelial cells, and that this induction can be significantly reduced by antiTGF-B antibodies.

MATERIALS AND METHODS

BME cells were cultured in gelatin-coated tissue culture flasks in complete BME medium consisting of minimal essential medium, alpha-modification (Gibco), 15% heat inactivated donor calf serum (Flow Laboratories, Ayrshire, Scotland), penicillin (500IU/ml) and streptomycin (100#g/ml). Chick embryo sternal chondrocytes and serumfree chondrocyte conditioned medium (CM) were prepared as previously described (7). Dialyzed and lyophylized CM was reconstituted in a volume of complete BME medium calculated to give an equivalent of 5x105 chondrocytes/ml and added to the cultures at 10fold dilutions in complete BME medium. Pre-immune rabbit gamma-globulins or anti-porcine platelet TGF-B1 antibodies (R. & D. Systems Inc., Minneapolis, MN) were added at a final concentration of 100#g/ml to reconstituted chondroc2cte conditioned medium containing Trasylol (200KIU/ml) and incubated at 4QC for 4 hours. (According to the manufacturer, the anti-TGF-g antibody neutralizes the activity of porcine and human TGF-B1 and porcine TGF-132, and shows no cross reactivity with acidic or basic fibroblast growth factors, platelet-derived growth factor or epidermal growth factor.) Confluent monolayers of BME cells in 24-well gelatin-coated plates were washed three times with PBS and 350#1 of reconstituted chondrocyte conditioned medium containing antibodies and Trasylol was added to duplicate wells. After a 12 hour incubation at 37-°C, cell extracts were prepared for the [125I] urokinase assay described below. Reverse zymography was as previously described (8). Human 33,000 M r urokinase was iodinated by the iodogen method, and the [125I]-labelled urokinase binding assay was as previously described (9). Immunoprecipitation was performed essentially as previously described (9). Briefly, samples were incubated for 2 hours at 4-°C either with normal rabbit serum, anti-PAI-1 or anti-PAI-2 antisera (provided by Drs. D.Loskutoff and E.Kruithof respectively). Fixed S.Aureus was added and the samples incubated at room temperature for a further 30 minutes. The samples were centrifuged in a microfuge for 5 minutes, the immunoprecipitate was discarded and the supernatant analysed by reverse zymography or the [125I]-labelled urokinase binding assay as described above. RNA preparation, Northern blots, in vitro transcription and hybridization were as previously described (8).

RESULTS Serum-free medium conditioned by chick embryo sternal chondrocytes grown in suspension culture was dialyzed, lyophilized and reconstituted in complete BME medium to an equivalent concentration of 5x105 chondrocytes/ml. When added in serial 10-fold dilutions to BME monolayers, conditioned medium induced the production of a 50kDa PAl in a dose-dependent manner, as determined by reverse zymography (Figure la). By immunoprecipitation, this inhibitor was shown to be PAI-1 (Figure lb). Similarly, when the cell extracts were analyzed by the [125I]-labelled 33kDa urokinase binding assay, PAI-1 was induced in a dose dependent manner (Figures 2a and 2b). This assay provided a quantitive estimate of PAI-1 production: gel slices corresponding to bound or free [125I]-labelled urokinase were excised and counted in a gamma-counter, and the relative levels of bound 634

V o l . 1 7 6 , N o . 2, 1991

BIOCHEMICAL A N D BIOPHYSICAL RESEARCH C O M M U N I C A T I O N S

BME

Platelet

50

C

1

2

3

4

50

5

N 2

1

N 2

1

Figure 1. Chondrocyte conditioned medium induces PAI-1 in BME cells, a) Reconstituted chondrocyte conditioned medium was added at 10-fold serial dilutions to BME cells ( C = control and 1,2,3,4 = 10000-, 1000-, 100- and 10-fold dilutions of 5, undiluted reconstituted medium conditioned by 5x105 chondrocytes/ml.) Cell extracts analysed by reverse zymography revealed the induction of a 50kDa PAl by undiluted (5) and 10-fold-diluted conditioned medium (4). b) The BME PAl induced by chondrocyte conditioned medium was immunoprecipitated by anti-PAI-1 antiserum (1), but not by anti-PAI-2 antiserum (2) or normal rabbit s e r u m (N). As a positive control, PAI-I from human platelet lysates was also immunoprecipitated by the anti-PAl-1 antiserum.

urokinase were normalized to the total (bound + flee) [125I]-labelled urokinase loaded per lane. This is illustrated in Figure 2a, which reveals a maximal 6,3-fold increase in bound [125I]-labelled urokinase in response to reconstituted conditioned m e d i u m f r o m 5x105 chondrocytes/ml. Chondrocyte-conditioned medium alone did not contain a P A l as determined by reverse zymography or the [125I]-labelled uroldnase binding assay (not shown). Hybridization of Northern blots of total cellular R N A from B M E cells exposed to

a

600 b

"=83 400

BME

Platelet

-",33

"" 8 3

a,. t..)

-,,=33

C

1

2

3

4

5

200 N 2 1 N 2 1

C

1

2

3

4

5

Fi_mare 2. Quantitation of PAI-1 induced by chondrocyte conditioned medium, a) Conditioned medium was added at 10-fold serial dilutions to BME cells (C ,1,2,3,4 and 5 are as described in Figure 1), and the cell extracts analysed by the 125I-labelled 33kDa urokinasebinding assay. Inset: a 50kDa PAI which shifted the [125I]-labelled 33kDa urokinase to an 83kDa enzyme-inhibitor complex was induced by undiluted (5) and 10-fold-diluted (4) • to bound and fr ee [1251]-labelled urokinase conditioned medium. Gel slices corresponding were excised and counted in a gamma-counter, and the relative levels of bound urokinase were normalized to the total (bound + free) uroldnase in each lane. Graphic representation reveals a maximal 6,3-fold increase in the [125I]-labelled urokinase-PAI-1 complex in response to reconstituted conditioned medium from 5x105 chondrocytes/ml, b) The BME PAI induced by chondrocyte conditioned medium was immunoprecipitated by anti-PAI-1 antiserum (1), but not by anti-PAI-2 antiserum (2) or normal rabbit serum (N) As a positive control, PAI-1 from human platelet lysates which also shifted pz~I]-labelled 33kDa urokinase to 83kDa was immunoprecipitated by the anti-PAI-1 antiserum. 635

Vol. 176, No. 2, 1991

BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

~28S

~18S

C

1

2

3 4 CM

5

1 2 TGF'B1

2 GAPDH

q)

®

C

NG

Mi

AT CM/NGCM/AT T/NG T/AT

Figure 3. Chondrocyte conditioned medium induces PAI-1 mRNA in BME cells. A northern blot of total BME cellular RNA (5#g/track) was hybridized with a [32p]-labe!led bovine PAI-1 cRNA probe. PAI-1 mRNA was induced by undiluted (5) and 10-fold-diluted (4) conditioned medium (CM). C,1,2,3,4 and 5 are as described in Figure 1. PAI-1 mRNA was :atso induced by 10pg/ml (1) and 100pg/ml (2) of TGF-B1. The same filter was hybridized with a [32p]-labelled chicken glyceraldehyde-3-phosphate dehydrogenase (GAPDH) cRNA probe, which reveled uniform loading of total BME cellular RNA. Figure 4. Anti-TGF-B antibodies reduce the PAI-t inductive properties of chondrocyte conditioned medium. Reconstituted conditioned medium from 5X10~ chondrocytes/ml (CM) was preincubated with anti-porcine platelet TGF-B! antibodies (AT) or normal rabbit gammaglobulins (NG) before addition to BME cells. When compared to CM preincubated with NG (CM/NG), the PAI-1 inductive properties of chondrocyte conditioned medium were significantly reduced by the anti.TGF-B antibodies (CM/AT) as assessed quantitatively by the [t25I]-labelled 33kDa urokinase binding assay. ~ne induction of PAI-1 by loopg/mi of human platelet-derived TGF-B1 was also significantly reduced by the anti-TGF-g antibodies (T/AT) when compared to TGF-B1 incubated with NG (T/NG). Values are mean + SEM; n = 3 for all conditions except T/NG and T/AT, where n = 2.

chondrocyte conditioned medium with a [32p]-labelled bovine PAI,1 cRNA probe (8), revealed a dose-dependent increase in bovine PAI-1 mRNA (Figure 3). Having previously identified a chondrocyte-derived TGF-13 as an inhibitor of endothelial sprout formation in vitro (7), and knowing that TGF-13 induces PAI-1 in endothelial cells, we preincubated chondrocyte-conditioned medium with anti-TGF-B antibodies before adding it to endothelial cell cultures to determine whether TGF-B might be the factor in conditioned medium which induces PAI-1 in BME cells. Anti-TGF-B antibodies reduced conditioned medium-induced PAI-1 by 61% (Figure 4). Preincubating human platelet-derived TGF-B1 (100pg/ml) with the antibodies under the same conditions as for conditioned medium reduced the PAI-1 inductive effect of TGF-131 by 74% (Figure 4). Finally, to determine the effective concentration of TGF-13 activity in chondrocyte conditioned medium, BME cells were incubated with human platelet-derived TGF-131 and levels of PAI-1 mRNA determined. Conditioned medium reconstituted to the equivalent of 5x105 chondrocytes/ml contained TGF-B activity equivalent to 100pg/ml of TGF-B1 (Figure 3). 636

Vol. 176, No. 2, 1991

BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

DISCUSSION

In an attempt to identify potential physiological inhibitors of angiogenesis, we have previously established a quantitative in vitro assay of capillary-like sprouting using BME cell aggregates embedded in three-dimensional collagen or fibrin gels (7). By co-culturing chick embryo sternal chondrocytes with the BME cells, we observed a marked inhibition of sprout formation. On the basis of antibody inhibition studies we demonstrated that a chondrocyte-derived TGF-13 was at least in part responsible for this inhibitory effect (7). Since TGF-13 induces the production of PAI-1 in rnicrovascular endothelial cells (8,10), we suggested that the antiproteolytic effect of PAL1 may play a role in the inhibition of sprout formation in our system. The studies in this paper were performed to determine whether chondrocytes might induce PAI-1 production by microvascular endothelial cells. We found that chondrocyte-conditioned medium, when added to microvascular endothelial monolayers, increases the production of PAI-1 and PAI-1 mRNA by these cells. In addition, this effect could be abrogated by t h e addition of anti-TGF-13 antibodies to the conditioned medium, and could be mimicked by the addition of TGF-gl. Although the anti-angiogenic properties of mature hyaline cartilage have been well described (11), the relevance of our in vitro findings to the inhibition of angiogenesis in vivo is not known. Mechanisms which have been implicated include: a) physical constraints imposed by the arrangement of cartilage-specific matrix macromolecules; b) inhibition of extracelltflar proteolytic activity, believed to be necessary for capillary endothelial cell invasion, by a variety of serine and metalloprotease inhibitors present in cartilage extracts or chondrocyte conditioned medium (see for example 12 and 13, and references therein); c) endogenous chondrocyte-derived factors present in cartilage matrix which specifically affect endothelial cells, the mechanism of inhibition likely to be most relevant to our observations. Our observations raise the possibility that in addition to endogenous protease inhibitors, a chondrocyte-mediated induction of protease inhibitor synthesis by endothelial cells themselves contributes to the anti-angiogenic properties of cartilage.

ACKNOWLEDGMENTS We are grateful to Drs. M.B.Furie and S.C.Silverstein for providing the BME cells, and Drs.D.Loskutoff and E.Kruitoff for providing the anti-PAI-1 and anti-PAI-2 antisera respectively. Excellent technical assistance was provided by M.Guisolan, and C.DiSanza, and photographic work was expertly done by P.,A.Ruttimann. This work was supported by grants trom the Swiss National Science Foundation (nos. 31-26625.89 and 31-9108.87) and The Sir Jules Thorn Charitable Trust.

REFERENCES .

2.

Brem, H. and Folkman, J. (1975) J. Exp. Med. 141, 427-439. Langer, R., Brem, H., Falterman, K., Klein, M. and Folkman, J. (1976) Science 193, 70-72. 637

Vol. 176, No. 2, 1991

BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

3. Kaminski, M., Kaminska, G., Jacobisiak, M. and Brezinski, W. (1977) Nature 268, 238-240. 4. Langer, R., Conn, H., Vacanti, H., Haudenschild, C. and Folkman, J. (1980) Proc. Natl. Acad. Sci. USA 77, 4331-4335. 5. Takigawa, M., Shirai, E., Enomoto, M., Pan, H.-O., Suzuki, F., Shiio, T. and Yugari, Y. (1987) Biochem. Int. 14, 357-363. 6. Pepper, M.S. and Montesano, R. (1991) Cell Diff. Dev. in press. 7. Pepper, M.S., Montesano, R., Vassalli, J.-D. and Orci, L. (1991) J. Cell. Physiol. 146, 170-179. 8. Pepper, M.S., Belin, D., Montesano, R., Orci, L. and Vassalli, J.-D. (1990) J. Cell Biol. 111, 743-755. 9. Vassalli, J.-D., Dayer, J.-M., Wohlwend, A. and Belin, D. (1984) J. Exp. Med. 159, 1653-1668. 10. Saksela, O., Moscatelli, D. and Rifldn, D.B. (1987) J. Cell Biol. 105, 957-963. 11. Kuettner, K.E. and Pauli, B.U. (1983) In Cartilage, Volume 1: Structure, Function and Biochemistry (B.K.Hall, Ed.) pp.281-312. Academic Press Inc., New York. 12. Yamada, H., Stephens, R.W., Nakagawa, T. and McNicol, D. (1988) J. Biochem. 104, 960-967. 13. Moses, M.A., Sudhalter, J. and Langer, R. (1990) Science 248, 1408-1410.

638

Plasminogen activator inhibitor-1 is induced in microvascular endothelial cells by a chondrocyte-derived transforming growth factor-beta.

We have previously demonstrated that a chondrocyte-derived TGF-beta inhibits spontaneous endothelial sprout formation in an in vitro model of angiogen...
762KB Sizes 0 Downloads 0 Views