EXPERIMENTAL

CELL

RESEARCH

203, 499-503 (1992)

SHORT NOTE Basic FGF and TGF-p Differentially Modulate lntegrin Expression of Human Microvascular Endothelial Cells JUDY ENENSTEIN,* NAHID S. WALEH,*T~ AND RANDALL H. KRAMER*$’ *Departments

of Anatomy San Francisco,

and Stomatology, and the $Cardiouascular California 94143; and TSRI International,

Basic fibroblast growth factor (bFGF) and transforming growth factor-B (TGF-8) are known to alter the migratory and proliferative capacity of endothelial cells in vitro and to stimulate angiogenesis in duo. One mechanism by which these cytokines induce their effects may be through the regulation of integrin adhesion receptor expression and activity. We examined the ability of these growth factors to modulate the expression of specific integrins in human microvascular endothelial cells (MEC). Immunoprecipitation of metabolically labeled MEC showed that bFGF upregulated the biosynthesis of (Ye, Q, &, and &. bFGF induced an increase in the levels of mRNA for (Yeand @,. TGF-fi increased synthesis of ax, a5, and &. These results suggest that bFGF and TGF-fi selectively alter integrin profiles and influence interactions of MEC with the extracellular matrix during neovascularization. In particular, the upregulation of the collagen/laminin receptor, a,fil, by bFGF may provide activated endothelial cells with an enhanced capacity to migrate through both their underlying basement membrane and the interstitial matrix. 8 1992 Academic Press, Inc.

INTRODUCTION Neovascularization following injury is dependent on a complex interplay of endothelial cells with cytokines and wound matrix. The cytokines basic fibroblast growth factor (bFGF) and transforming growth factor beta (TGF-P) are two of the best studied modulators of angiogenesis. Both bFGF and TGF-P are synthesized by large- and small-vessel endothelial cells [l, 21. bFGF,

1 To whom reprint requests should be addressed at University of California, Room HSW-604, Box 0512, San Francisco, CA 94143. 0512. Fax: (415) 476-4204. 499

Research Institute, University of California, Menlo Park, California 94025

which stimulates neovascularization in Go, is a very potent stimulator of proliferation and migration by endothelial cells in vitro [3]. High levels of endogenous bFGF correlate with the ability of endothelial cells to invade basement membranes in vitro [4]. TGF-fi has been found in sites of active blood vessel repair [5]. In duo, TGF-fl is thought to stimulate angiogenesis indirectly by inducing macrophages to produce cytokines [6]. In uitro, TGF-/3 antagonizes the angiogenie effects of bFGF by inhibiting the bFGF-induced stimulation of both proliferation and migration of aortic and capillary endothelial cells [4, 7, 81. The integrins are a family of heterodimeric glycoproteins which mediate cell-substratum and cell-cell interactions (for reviews, see [9, lo]). There are at least 8 classes of integrin /3 subunits, which interact with a much larger number of (Y subunits. The integrin LYsubunits 1 through 6 and v and p subunits 1, 3, and 5 have been identified on microvascular endothelial cells [ll, 121, and large vessels have at least a 2,3,5, and v and 8, and & [9, 131. Integrins are involved in several aspects of angiogenesis, including endothelial cell migration, proliferation, and formation of new vessels. Integrins show some specificity for different substrata. Therefore, a change of integrin pattern may be expected when endothelial cells leave the laminin/collagen IV-rich basement membrane to form sprouts in a matrix dominated by collagen I and fibronectin. Likewise, integrins that bind strongly to basement membranes would reappear as the homeostatic state is attained. To determine whether bFGF and TGF-P exert their effects on endothelial cell migration and proliferation in part by modulating integrin expression, we assessed the effects of these growth factors on integrin-specific protein and mRNA synthesis by microvascular endothelial cells (MEC) in vitro. Our studies, previously published in preliminary form [ 141, indicate that bFGF and TGF-fl both alter endothelial cell integrin expression. bFGF in 0014.4827/92 $5.00 Copyright 0 1992 by Academic Press, Inc. All rights of reproduction in any form reserved.

500

SHORT

particular upregulates several integrins which recognize ligands prevalent in the fibrin/fibronectin/collagen Irich matrix present after wounding. MATERIALS

AND

METHODS

Cell culture. Dermal MEC were isolated from human foreskin, as previously described [ll]. For analysis of integrin expression, cells were plated at a preconfluent density in medium containing 5 rig/ml bFGF. After 3 days, the medium was replaced, with or without bFGF or TGF-/3 at the indicated concentrations, which have previously been shown to be biologically active [3, 151. Under these conditions, toxic effects were not noted. Antibodies and probes. The specificity and sources of monoclonal antibodies against individual integrin subunits have been previously described [12]. cDNA fragments used as probes in Northern blot experiments were provided by Dr. Robert Pytela, San Francisco General Hospital. MEC (50-80% conMetabolic labeling and immunoprecipitation. fluent) were rinsed three times and incubated for 1 h in serum-free, methionine-free medium and treated with either 10 rig/ml bFGF or 6 rig/ml TGF-/3 for 18 h. Control cultures were not supplemented with growth factor. During the last 6 h of culture, cells were metabolically labeled with 40 &i/ml of [?S]methionine (New England Nuclear, spec act 29 TBq/mmol). Cells were washed twice in PBS, harvested with 2 mM EDTA, 0.05% BSA, lysed with N-octylglucoside, and prepared for immunoprecipitation as described [15]. Radioactivity was equalized for all lysates. Aliquots of lysates were reacted with excess anti-integrin antibodies overnight, treated with species-specific secondary antibody for at least 3 h, and immunoprecipitated with protein A-Sepharose beads for 2-3 h. Immune complexes were analyzed by SDS-PAGE followed by autoradiography [ll]. To ensure that the integrins were completely precipitated, cell lysates were subjected to double rounds of immunoprecipitation. Scanning densitometry. Autoradiograms were scanned with a video camera and the image was captured by the Apple image program. Images were analyzed using the scan analysis program. Matched backgrounds within a lane were subtracted from the values of individual bands. Northern blot analysis. Nearly confluent monolayers of MEC were treated for O-24 h with 10 rig/ml bFGF. Total cellular RNA was isolated by the guanidinium-cesium chloride method [16]. Samples of RNA (20 pg) were electrophoresed in 1% agarose gels containing 6% formaldehyde and were transferred by blotting to nylon membranes. Membranes were then probed with cDNA fragments of a2, 05, p,, and & integrin subunits labeled with [a-32P]dCTP by random oligonucleotide primer synthesis. Hybridizations were carried out at 60°C in 50% formamide. Filters were washed at room temperature and then at 65°C for 1 h.

RESULTS

AND

DISCUSSION

We examined the effects of bFGF and TGF-fl on the biosynthesis of 0, and & integrins by MEC. Immunoprecipitation with anti-& antibody indicated a significant increase in the synthesis of the /3, subunits and some (Y subunits after treatment with bFGF or TGF-P (Fig. 1). A predominant ct!chain in MEC is the q subunit [ll, 12, 161, which has an apparent molecular mass of - 145 kDa with or without reduction of disulfide bonds. The (~zsubunit in association with 0, was readily visible in reducing gels, where it was clearly increased in meta-

NOTE

bolically labeled cultures treated with bFGF (Fig. lB, lanes 7-9). Under reducing conditions, the CY~and (Ye subunits comigrate with & at -130 kDa. The effect of bFGF on the biosynthesis of c+ was pronounced and could be detected when cell lysates were immunoprecipitated with antibodies specific for cy2(Fig. 2). In four separate experiments, CQbiosynthesis following bFGF treatment increased an average of 2.8-fold over control (range 1.5 to 5.7-fold), as measured by scanning densitometry. Double immunoprecipitation showed that almost all the a2 came down in the first precipitation (Fig. 2). TGF-P treatment consistently stimulated (Yesynthesis, though to a lesser extent than bFGF (average 2.1fold, range 1.1-3.3). The increased level of a2 does not appear to be a result of an increase in precursor since the subunit we detected had the apparent molecular mass (145 kDa) of the mature form. Similarly, the increase in pi seems to consist almost entirely of the mature 0, subunit since only small amounts of precursor p, were detected after immunoprecipitation with anti-b, antibody (Fig. lA, lanes 7-9). bFGF, and to a lesser extent TGF-0, induced elevated synthesis of the (Yesubunit (Figs. 1A and lB, lanes 4-6). Biosynthesis of (Ye increased an average of 3.3-fold (range 1.3-5.5) by bFGF treatment and 2.1-fold (range l-4.7) by TGF-/3. In contrast, synthesis of the c+ subunit was noticeably decreased by bFGF (Fig. lA, lanes l-3). It is not clear if 01~levels are generally higher than &, but one explanation may be the presence of uncomplexed cy3. An increase in synthesis of &, averaging 1.5-fold (range 1.2-2.1), was detected by a complex-specific antibody (LM 609) in bFGF-treated but not consistently in TGF-P-treated cells (Figs. 1A and lB, lanes 11-13). Further studies are needed with cu,-specific antibodies to determine whether total 01, levels are altered by bFGF. In preliminary experiments, the integrin subunit ai, which is only weakly expressed in MEC [ 121, did not seem significantly affected by treatment with bFGF or TGF-P. Preliminary studies of a(6 suggest that both bFGF and TGF-P cause a significant decrease in its synthesis (not shown). The alterations in integrin synthesis by bFGF or TGF-P appear to be due to a combination of changes in the rate of synthesis of both the (Y and the fi subunits. For example, total &-associated cr subunits were increased along with the 6, subunit. Much of this increase appeared to be due to the increase in (Yeand c+ In addition, bFGF appears to increase the amount of metabolically labeled pi subunit associated with (Y* and c+,when immunoprecipitated by LYchain-specific antibodies (Fig. lA, lanes 4-6; Fig. 2). The average &lcuz ratio of radioactivity, as measured by densitometry, increased 2.3-fold (range 1.9-2.5; three experiments) following bFGF treatment of MEC. The &lc+, ratio in three experiments indicated a trend of increased & coprecipitation, but the

SHORT

501

NOTE

B

9

12345610 ------RI-

.

10

11 12 _

13

-116

CTFCTFCTFCTFN L-Ia3

a5

4

%4

FIG. 1. bFGF and TGF-fl alter integrin biosynthesis. Microvascular endothelial cells were treated for 18 h with bFGF (10 rig/ml) or TGF-fl (6 rig/ml) and then labeled with [%]methionine during the last 6 h. Cell lysates were irnmunoprecipitated with saturating amounts of monoclonal antibodies to a3 (PlB5), 01~(BlE5), /3, (AIIBS), or a,& (LM609), and analyzed by SDS-PAGE under nonreducing (A) or reducing (B) conditions. C, control, no treatment; T, TGF-P-treated cells; F, bFGF-treated cells; N, no primary antibody. These results represent at least four independent immunoprecipitation experiments for each antibody.

results were variable. The data suggest that bFGF may stimulate complex formation between (Yeor (Yeand Pi. Alternatively, bFGF may alter the turnover rate of the p1 subunit such that more newly labeled pi is incorporated into complexes. The effects of bFGF on integrin mRNA levels were assessed by Northern blot analysis. mRNA for the c+ subunit was noticeably increased after exposure of MEC to bFGF (Fig. 3A). The major (Yetranscript of w-8 kb was similar to the size previously reported [17]. The bFGF-induced increases in cyzmRNA paralleled the increases in (Yeseen by [35S]methionine labeling (Fig. 2). By contrast, (Lo translation was stimulated by bFGF without any change in levels of mRNA (Fig. 3B). Analysis of mRNA levels for the pi subunit indicated a significant increase after treatment of MEC with bFGF (Fig. 3C). As with the (YemRNA, 0, mRNA increased as nearly as 4 h after exposure to bFGF. The increase in /3, mRNA levels was marginal following bFGF treatment (Fig. 3D, lane c). Although some increase was seen by immunoprecipitation, the increase in a$i expression in response to bFGF is likely to have important consequences in the migratory and proliferative phases of angiogenesis as well as the reestablishment of cell-cell contacts in newly formed vessels. The a&, complex is the major cell-sur-

face receptor for interstitial collagen, as well as binding collagen IV, laminin, and possibly fibronectin [15, 1820]. Endothelial cells employ ~yJ?i both in focal contacts and at sites of cell-cell adhesion [ 211. The bFGF-stimulated increases in cu.& and LY,& synthesis would also facilitate endothelial cell migration through the provisional matrix, where there is increased deposition of fibronectin. One day after wounding in vivo, vessels stain weakly for fibronectin, and microvessels synthesize fibronectin by the second day [22, 231. Both c+,p, and cy,& bind to fibronectin [24, 161. In addition, (Y,& from endothelial cells binds to vitronectin, fibrinogen, and thrombospondin, all of which are components of the provisional matrix [25]. 1n viva, a,& is found in blood vessels of human granulation tissue [ 261, but p3 has been difficult to detect on resting small vessels. This suggests that & is specifically upregulated during wound healing and could facilitate the out-migration of endothelial cells during sprout formation. This study is the first report of bFGF altering integrin expression in endothelial cells. However, endothelial cell integrins have been found to respond to other cytokines [27]. And both bFGF and TGF-fl have been shown to be potent modulators of integrin expression in other cell types, either directly [14, 28-301 or by changing ligand synthesis. In osteosarcoma cells, crz& and a& in-

502

SHORT

NOTE

-a C

b

c

d D

-

-

a

b

c

d

a

b

c

d

FIG. 3. The effect of bFGF on integrin mRNA levels. Cells were grown for 4 or 24 h in the presence or absence of bFGF (10 rig/ml). Total cellular RNA was separated by formaldehyde agarose gel electrophoresis and transferred to a nitrocellulose filter. The membrane was probed with a3’P-labeled cDNA fragment of o(2(A), a5 (B), & (C), or & (D). Hours of exposure to bFGF: a, 0 h; b, 4 h; c, 24 h; d, 24 h no treatment. Note that in lane Dd; there was inefficient transfer of RNA. All of the mRNA results have been duplicated. The size of 28s and 18s ribosomal RNA is indicated by arrows.

FIG. 2. The effect of bFGF on 01~integrin biosynthesis as shown by exhaustive immunoprecipitation. Cell lysates were immunoprecipitated with antibodies to a2 (PlH5), then the supernatant was reprecipitated with anti-a, antibodies. Lanes 1-4 are nonreduced. Lanes 5-8 are reduced. C, control, no treatment; F, bFGF-treated cells; Cl and Fl, 1st precipitation; C2 and F2, 2nd precipitation.

Chapter; by AHA Grant 880763; and by NIH Grants CA-33834, CA51884, and DE-00242. REFERENCES 1.

creased in response to TGF-/3 while 01& decreased. This is comparable to the differential modulation we observed here with the endothelial cells. TGF-/3 also increases deposition of interstitial collagen and fibronectin [31, 51; such an increase in integrin ligand concentration may indirectly stimulate production of specific integrins. We have described some effects of bFGF and TGF-P on integrin expression that could have significance during angiogenesis. However, to understand the effects of bFGF and TGF-/3 on integrin expression in MEC in ho, we must know the temporal expression of these mediators during wounding, whether they are present in an active or inactive form, and whether bFGF and TGFp are themselves modulated by other cytokines. The significance of the modulation of integrins during angiogenesis in turn awaits complete understanding of integrin function not only in adhesion/migration but also in intracellular signaling that may regulate cellular activities such as protease synthesis and proliferation.

2.

3. 4. 5. 6.

7. 8. 9. 10. 11.

286. 12.

We thank Dr. William Carter, Dr. Virgil Wood, Dr. Caroline Damsky, and Dr. David Cheresh for antibodies and Dr. R. Pytela for cDNA fragments used as probes. We thank Evangeline Leash for her invaluable editorial assistance. This work was supported by a research fellowship from the American Heart Association, California Affiliate, 91-08, with funds contributed by the Alameda County

Schweigerer, L., Neufeld, G., Friedman, J., Abraham, J. A., Fiddes, J. C., and Gospodarowicz, D. (1987) Nature 325, 257259. Hannan, R. L., Kourembanas, S., Flanders, K. C., Rogelj, S. J., Roberts, A. B., Faller, D. V., and Klagsbrun, M. (1988) Growth Factors 1, 7-17. Folkman, J., and Klagsbrun, M. (1987) Science 235,442-447. Mignatti, P., Tsuboi, R., Robbins, E., and R&in, D. B. (1989) J. CellBiol. 108, 671-682. Madri, J., Reidy, M. A., Kocher, O., and Bell, L. (1989) Lab. Invest. 60, 755-765. Roberts, A. B., Sporn, M. B., Assoian, R. K., Smith, J. M., Roche, N. S., Wakefield, L. M., Heine, U. I., Liotta, L. A., Falanga, V., Kehrl, J. H., and Fauci, A. S. (1986) Proc. Natl. Acad. Sci. USA 83, 416774171. Frater-Schroder, M., Muller, G., Birchmeier, W., and Bohlen, P. (1986) Biochem. Biophys. Res. Commun. 137,295-302. Baird, A., and Durkin, T. (1986) Biochem. Biophys. Res. Commun. 138,476-482. Albelda, S. M., and Buck, C. (1990) FASEB J. 4, 2868-2880. Ruoslahti, E. (1991) J. Clin. Inuest. 87, l-5. Cheng, Y-F., and Kramer, R. (1989) J. Cell. Physiol. 139, 275-

13. 14. 15.

Kramer, R., Cheng, Y-F., and Clyman, R. (1990) J. Cell Biol. 111, 1233-1243. Basson, C. T., Knowles, W., Bell, L., Albelda, S., Castronovo, V., Liotta, L. A., and Madri, J. (1990) J. Cell Biol. 110, 789-801. Enenstein, J., and Kramer, R. H. (1990) J. Cell Biol. 111, 264a. Heino, J., Ignotz, R. A., Hemler, M. E., Crouse, C., and Massague, J. (1989) J. Biol. Chem. 264, 380-388.

SHORT 16. 17. 18. 19. 20. 21. 22.

23.

Cheng, Y-F., Clyman, R. I., Enenstein, J., Waleh, N., and Kramer, R. H. (1991) Exp. Cell Res. 194,69-77. Takada, Y., and Hemler, M. E. (1989) J. CellBiol. 109,397-407. Languino, L. R., Gehlsen, K. R., Wayner, E., Carter, W. G., Engvail, E., and Ruoslahti, E. (1989) J. Cell Biol. 109, 2455-2462. Elites, M. J., and Hemler, M. E. (1989) Proc. N&l. Acad. Sci. USA 86,9906-9910. Kirchhofer, D., Languino, L. R., Ruoslahti, E., and Pierschbather, M. D. (1990) J. Cell Biol. 111, 1245-1254. Lampugnani, M. G., Resnati, M., Dejana, E., and Marchisio, P. C. (1991) J. Cell Biol. 112,479-490. Clark, R. A. F., Dellepella, P., Manseau, E., Lanigan, J. M., Dvorak, H. F., and Colvin, R. B. (1982) J. Znuest. Dermatol. 79, 269-276. Clark, R. A. F., Quinn, J. H., Winn, H. J., Lanigan, J. M., Dellepella, P., and Calvin, R. B. (1982) J. Exp. Med. 156, 646-651.

Received March 25, 1992 Revised version received September

2, 1992

503

NOTE 24. 25.

Charo, I. F., Nannizzi, L., Smith, J. W., and Cheresh, (1990) J. Cell Biol. 111, 2795-2800. Clark, R. A. F. (1990) J. Znuest. Dermatol. 94, 128-134s.

D. A.

Rudolf, R., and Cheresh, D. (1990) Clin. Plast. Surg. 17, 457462. 27. Defilippi, P., Truffa, G., Stefanuto, G., Altruda, F., Silengo, L., and Tarone, G. (1991) J. Biol. Chem. 266, 7638-7645. 28. Bates, R. C., Rankin, L. M., Lucas, C. M., Scott, J. L., Krissansen, G. W., and Burns, G. F. (1991) J. Biol. Chem. 266, 1859318599. 29. Heino, J., and Massague, J. (1989) J. Biol. Chem. 264, 2180621811. 30. Ignotz, R. A., and Massague, J. (1986) J. Biol. Chem. 261,43374345. 31. Ignotz, R. A., Heino, J., and Massague, J. (1989) J. Biol. Chem. 264,389-392. 26.