JOURNAL OF CELLULAR PHYSIOLOGY 144:151-158 (1990)

Basic Fibroblast Growth Factor Expression in Human Omental Microvascular Endothelial Cells and the Effect of Phorbol Ester A. BIKFALVI, J. ALTERIO, A.L. INYANG, E. DUPUY, M. LAURENT, M.P. HARTMANN, 1. VIGNY, D. RAULAIS, Y, COURTOIS, A N D G. TOBELEM* lnstituf des vaissedu e l du sang (l.V.S) and lNSERM U I 5 0 and CNRS U A 334, Hopital Lariboisiere, 750 I0 (A.B., A.L.I., E.D., C.TI and INSERM U I 1 8, 75016 (].A,, M.L., M.P.H., L.V., D.R., Y.C.), Park, France The human omentum contains a potent, not yet identified angiogenic activity. The omentum is very vascularized. Therefore, we investigated whether human omental microvascular endothelial cells (HOME cells) express the angiogenic peptide basic fibroblast growth factor (bFGF). Cytosol prepared from HOME cells stimulated D N A synthesis in bovine epithelial lens cells (BEL cells). The mitogenic activity could be neutralized by an anti-bFGF antibody. Basic FGF-like material from the HOME cell cytosol was bound onto a heparin-Sepharose column at 0.6 M and was eluted at 3 M NaCI. The 3 M NaCl eluted material reacted with the specific anti-bFCF antibody in an ELlSA and stimulated D N A synthesis. It did not react with a specific anti-acidic fibroblast growth factor (aFGF) antibody. Western blotting experiments using the same bFGF antibody showed the presence of a major band of 17 Kd and a doublet of 20-22 Kd. Northern blotting of nonstimulated HOME cells using a specific 1.4 kb bFGF probe showed the presence of 5 molecular species of 6.6, 3.7, 2.2, 2.0, and 1.0 kb. N o aFGF mRNA was detected with a specific previously characterized 4.04 kb probe. 12-0-tetrad+ canoylphorbol 13-acetate (TPA) did not influence significantly the expression of bFGF at the protein and mRNA level in HOME cells. Thus, protein kinase C activation by TPA did not appear to modulate significantly the expression of bFGF for that cell type. Contrastingly, human umbilical vein endothelial cells (HUVE cells), which expressed no bFGF and aFGF mRNA at a basal level, were induced to express bFGF but not aFGF mRNA when stimulated by TPA. These results suggest that the described angiogenic activity could be the bFGF-like mitogen contained in HOME cells and that these cells are different from endothelial cells derived from large vessels (HUVE cells) regarding the expression of bFGF.

Basic fibroblast growth factor (bFGF) stimulates the growth and differentiation of a variety of normal and transformed cells (Gospodarowicz et al., 1987; Moscatelli et al., 1986a,b; Vlodavsky e t al., 1987a,b). It is mitogenic and chemotactic for vascular cells, such as endothelial cells (Gospodarowicz et al., 1987) and smooth muscle cells (Esch et al., 1985). Basic FGF induces angiogenesis in vitro (Montesano et al., 1986) and in vivo (Gospodarowicz et al., 1979; Hayek et al., 1987). It has been shown that bovine aortic and capillary endothelial cells synthetize bFGF that remains cell-associated and that has the potential to stimulate their own growth (Vlodavsky et al., 1987a,b; Schweigerer et al., 1987). Basic FGF does not possess a classical signal sequence and is not released into the culture medium (Fiddes et al., 1987). Nevertheless, bFGF is deposited in the extracellular matrix formed by these cells by a n unknown mechanism (Vlodavsky et al., 1987a,b; Baird and Ling, 1987). 12-0-Tetradecanoylphorbol 13-acetate (TPA) activates protein kinase C and modulates the expression of @

1990 WILEY-LISS, INC.

several oncogenes and growth factors such as the oncogene c-myc and c-fos (Rosengurt and Sinnet-Smith, 1987) and the platelet-derived growth factor (PDGF) B chain (Daniel et al., 1986). TPA inhibits bFGF activity and the cell surface binding to endothelial cells by down-regulating bFGF receptors (Doctrow and Folkman, 1988; Doctrow, 1989). The human omentum contains a potent angiogenic activity (Goldsmith et al., 1984, 1986). Neither the nature of this activity nor its cellular origin have yet been identified. The human omentum is highly vascularized and is used as a source for culturing microvascular endothelial cells (Kern et al., 1983). We have previously shown that human omental microvascular endothelial cells (HOME cells) exhibit high and low affinity receptors for bFGF and that bFGF is internalized and degraded by these cells (Bikfalvi et al., 1989).We have Received May 17, 1989; accepted May 15, 1990. "To whom reprint requestdcorrespondence should be addressed.

152

BIKFALVI ET AL

now extended our study by adressing the question of whether HOME cells, like bovine capillary endothelial cells, also contain a molecule similar or identical to bFGF and whether the stimulation by TPA modulates its expression. In this study, we demonstrate that HOME cells express bFGF but not acidic fibroblast growth factor (aFGF) and that TPA appears not to modulate significantly its expression. In this respect, home cells were found to be different from human umbilical vein endothelial cells (HUVE cells). The results suggest that the angiogenic activity contained in the omentum could be the bFGF-like mitogen expressed in HOME cells.

MATERIALS AND METHODS

Preparation of cytosol Confluent HOME cells (5 x lo7 cells) were deprived for 36 h r of serum in M 199 containing 1 pgiml insulin, 5 pgiml transferrin, and 0.5 mg1ml BSA, and incubated for 4.5 h r with 100 ngiml of TPA or not incubated. Then, the incubation medium was removed, and the cells were washed two times with PBS and harvested from the culture flasks by trypsinization (0.05 % trypsini0.025 % EDTA). Subsequently, the cell suspension was centrifuged a t 1,250 rpm for 10 min a t 4°C and resuspended in a small volume of PBS (0.5-1 ml). After 5 cycles of sonication (5 x 5 sec) the suspension was centrifuged at 100,OOOg for 2 h r at 4°C in a Beckman centrifuge (Beckman, Munchen, FRG). The clear supernatant was then collected and stored at 70°C until use. Protein was determined according to Bradford (1979) with immunoglobulin used as standard. ~

Materials Recombinant bFGF, 1251-protein-A (>1.16 Bq/mg) were purchased from Amersham (Les Ulis, France). All culture reagents, media, and limbro 12-multiwell plates were from Flow Laboratories (Mc Lean, VA). Twenty-four multiwell plates were obtained from Costar (Cambridge, CAI. Electrophoretic reagents including glycine, tris-(hydroymethyl) amino methane, sodium dodecyl sulfate (SDS) acrylamide, ammonium persulfate, bisacrylamide, 2-mercaptoethanol, Temed, and Coomassie blue R 250 were from Biorad (Richmond, CA). Nitrocellulose was also from Biorad (Richmond, CAI. Heparin-Sepharose CL-6B and Sephadex G-25 were from Pharmacia (Pharmacia Fine Chemicals, Uppsala, Sweden). EGF was obtained from Collaborative Research (Bedford, MA). Thrombin was a gift from M.C. Guillin (Hopital Beaujon, Paris, France). Anti-rabbit peroxidase coupled polyclonal antibodies were obtained from Biosys (Compiegne, France). All other materials and reagents were research grade. Cell cultures Human omental microvascular endothelial cells (HOME cells) and human umbilical vein endothelial cells (HUVE cells) were cultured and grown as described (Kern et al., 1983; Jaffe et al., 1973). The cells were grown in medium 199 containing 20% FCS, 2 mM glutamine, 50 U penicillin, 50 Fg/ml streptomycin, and 100 p,giml fungizone at 37°C in a humidified atmosphere with 5% CO,. After the first passage, the cells were grown in the same medium and 2% Ultroser G on gelatin coated flasks. For cytosol preparations and RNA extraction the cells were cultured on 175 cm2 flasks. For cytosol preparations and RNA extraction the cells were cultured on 175 cm2 flasks. The endothelial nature of both cell types was ascertained by a typical cobblestone appearance in confluent cultures and the presence of factor VIIUvon Willebrand factor antigen. Bovine epithelial lens cells (BEL cells) were cultured and grown as described (Plouet et al., 1984). The cells were cultured in minimal essential medium (MEM) containing 6% FCS, 2 mM glutamine, 50 U penicillin, 50 Fg/ml streptomycin, and 100 pg/ml fungizone a t 37°C in a humidified atmosphere with 5% C02. HOME cells were used between passages 2 and 4,HUVE cells between passages 1 and 3, and BEL cells between passages 20 and 30.

Preparation of polyclonal antibodies against bFGF and aFGF Anti-aFGF polyclonal antibodies were prepared in rabbits. Purified bovine brain aFGF was injected in PBS (100 Fgiml) with complete Freund adjuvant into the dorsal area of rabbits by multiple intradermal shots. Boosters were done 5 weeks later in incomplete Freund adjuvant. The animals were bled every week and the titers of the antisersa were measured by the ELISA method. Anti-bFGF polyclonal antibodies were raised by using synthetic peptides representing the Nterminal part (1-9 Tyr) of the protein coupled to tetanus toxin via glutaraldehyde but otherwise as indicated. The titers of the antisera were measured by ELISA. Maximal titers were obtained 15 days after the boost for anti-aFGF and 6 weeks for anti-bFGF. No cross-reactivity was observed with these antisera. The titer of anti-aFGF was estimated a t 112,000 and the titer of anti-bFGF a t 11500 when compared to non-immune serum. Anti-aFGF and anti-bFGF were able to recognize aFGF and bFGF in Western blots at dilutions of 11500 and 11200, respectively. These antibodies reacted with a similar titer with human and bovine acidic and basic FGF (data not shown). DNA synthesis assay 'H-thymidine incorporation on BEL cells was done as previously described (Plouet et al., 1984). Briefly, BEL cells were passaged and seeded a t a density of 30,000 cellsiwell in 24 multiwell plates in medium MEM containing 6% FCS. The cells were left without medium change for 6 days. The test samples were then added in t h a t medium. After 24 h r incubation, 'H-thymidine (1-2 $3) was added to the wells for 4 hr, and the experiment was achieved by washing the cells 3 times with PBS and solubilizing them with 0.2 M NaOH. The radioactivity was determined by scintillation counting. Western blot experiments Western blots were done as described (Burnette, 1981j. Cytosols a t indicated protein concentrations were solubilized in sodium dodecyl sulfate sample buffer, boiled, and run on a 18% SDS page gel according to Laemmli (1970) overnight. Subsequently, the

BASIC FGF EXPRESSION IN ENDOTIIELIAL CELLS

proteins were transferred from the gel to a nitrocellulose membrane and saturated after transfer with a tris buffer saline (TBS) containing 5% dried milk and 0.2% Tween 20 at pH 8 (Buffer A) overnight at room temperature. Then, the sheets were washed 2 times in TBS containing 0.5 % dried milk and 0.2% Tween at pH 8 (Buffer B). Subsequently, the nitrocellulose membranes were incubated with anti-bFGF antibody (11 100-11200) or non-immune serum at the same dilution and left for 2 h r at room temperature. Then, the sheets were washed 5 times in Buffer B during 1 h r and incubated with 1251-protein-A(30,000 cpmistrip) for 30 min. This procedure was followed by five washes in Buffer B during 1 hr, and the membranes were dried on a Wathman paper and exposed to autoradiography for 2-5 days at -70°C.

Heparin-Sepharose c h r o m a t o g r a p h y HOME cell cytosol was applied to a n HeparinSepharose column equilibrated with 0.6 M NaCl in PBS and recycled overnight. The column was then washed extensively with PBS containing 0.6 M NaCl and eluted with 3 M NaCl in PBS (flow rate 0.2 ml/ min, fractions of 1mlitube). The fractions were then screened in a direct ELISA using the specific anti bFGF and anti-aFGF antibodies and tested for biological activity in 'H-thymidine incorporation assay.

153

Probes The HBGF-IIlbFGF cDNA probe was a gift from J. Abraham and J. Fiddes (California Biotechnology, Inc., Mountain View, CA). This probe is a 1.4 kb bovine HBGF-2IbFGF cDNA clones into PBR 322 (Abraham et al., 1986). The cDNA insert contains the entire protein coding sequence of bovine HBGF-2ibFGF. The HBGF-llaFGF cDNA probe is a 4.04 kb bovine HBGFl/aFGF cDNA cloned into lambda gt 11 (Alterio et al., 1988).The cDNA insert contains the entire protein coding sequence of bovine HBGF-liaFGF. The glyceraldehyde-3-phophate dehydrogenase (GAPDH) cDNA probe was obtained from P. Fort (Department of Molecular Biology, Montpellier University, France). It is a 1.3 kb rat GAPDH cDNA cloned into PUC 18 (Fort et al., 1985). This probe was used as internal standard and utilized to estimate loading and transfer efficiency of RNA samples. TPA did not inf luence the level of transcription of GAPDH mRNA (Blanchard, personal communication).

N o r t h e r n blotting HOME cell and HUVE cell Poly (A+)-richRNA was electrophoresed on a 1%agarose gel containing 6% formaldehyde. The RNA was transferred onto a Genescreen plus membrane (Dupont de Nemours, Paris, France) in 10 x SSC by capillary transfer. DNA molecular weight markers were obtained by 32P-5'labeling of Direct ELISA a s s a y lambda DNA-Hind 111 restriction fragments. The probes were labeled with [al~ha-~'P]dCTP Ninety-four well ELISA plates were coated overnight (3,000 Ciimmole) by oligonucleotide priming using the at room temperature with bFGF, aFGF, or HeparinSepharose fractions or in BSA free-carbonate buffer priming Kit from Amersham (Les Ulis, France). The containing 0.05 M Na,CO, and NAHCO, at pH 9.6. The RNA blots were prehybridized at 42°C in buffer confollowing day, the plates were incubated with the same taining 50% formamide, 1% SDS, 1 M NaCl, and 10% buffer containing 2% BSA at 37°C for 2 hr. Then the dextran sulfate, and they were incubated in the same plates were washed 3 times with a buffer containing buffer at 42°C with the specific probes (lo6 cpm/ml). The blots were washed twice for 20 min in 0.1 x SSC PBS and 0.05% Tween (Wash Buffer) and incubated with anti-bFGF (11200) and aFGF antibodies (114,000) at 50°C and subjected to autoradiography. in the same buffer but containing 1% BSA (Buffer A) for 2 hr. After 5 washes in Wash Buffer the plates were RESULTS incubated in Buffer A with peroxidase conjugated antirabbit antibody (114000) and left for another 2 hr. FiHOME cells contain a mitogenic activity nally, the plates were washed again 5 times in Wash similar or identical to b F G F Buffer and incubated susequently in the dark with 0phenylene-diamine-dihydrochloride(OPDA) in citrate Cytosols prepared from HOME cells were tested for phosphate buffer (0.1 M citric acid, 0.2 M Na,PO,) con- their ability to stimulate DNA synthesis in BEL cells. taining H202 pH 5. The reaction was stopped after 20 BEL cells were chosen as a target for these experimin by adding 30 pl of 12.5% HzS04 to the wells. Ab- ments since they bind bFGF (Moenner et al., 1986) and sorbance was read at 495 nm. respond specifically to bFGF and aFGF (Plouet et al., 1984). Figure 1,b shows that the HOME cell-derived RNA isolation cytosol stimulated DNA synthesis and that the incuPreconfluent endothelial cells grown on 175 cm' bation with a polyclonal anti-bFGF antibody neutraflasks were washed once with M 199 and maintained lised this activity. The maximal stimulation of DNA for 36 h r in M 199 containing 5 pglml transferrin, 1 synthesis by the cytosol was reached at 100-200 pg/ pg/ml insulin, and 0.5 mgiml BSA. Then the cells were well (200-400 pg/ml) and the half-maximal stimulawashed 2 times in PBS. Total RNA was prepared from tion a t 30 pgiwell (60 pgiml). The inhibition of the cells lysed in a buffer containing 4 M quanidium cytosol stimulated DNA synthesis by the anti-bFGF isothiocyanate, 1M 2-mercaptoethanol, and 5% N-lau- antibody was of 75%, suggesting that the major fracrylsarcosine according to the cesium chlorid method as tion of the mitogenic activity in the HOME cell cytosol previously described (Alterio et al., 1988). Poly (A+)- was a bFGF-like activity. The data suggested also that rich RNA was prepared using Hybond mAP according HOME cells possibly also contained other not yet idento the instructions of the manufacturer (Amersham, tified growth-promoting activities. Les Ulis, France). To further characterize the mitogenic activity con-

154

BIKFALVI ET AL.

25000

I,

t

a

cYtosOl

T

bFGF

3 M NaCl

r'

0

a n t i - W G F antibody

o anti-aFGF a n t i b o d y

1

bFGF

+

. i "L

AB

I 1

10

40

50 100 150

ug/w.1

Fig. 1. a,b: Stimulation of DNA synthesis of BEL cells by HOME cell cytosol (Cs) and the effect of anti-bFGF antibody. Cytosols were prepared from 36 hr serum-deprived HOME cells and 3H-thymidine incorporation was run as indicated in Materials and Methods. In other experiments cytosol(100 pgiml) or recombinant bFGF (10ngiml) was added with or without bFGF antibody (final concentration 200 bgiml) to the cells. 3H-thymidine incorporation was performed as indicated. The figure is representative for three (a) and two (b) experiments run in triplicate (Results as meantSEM).

tained in HOME cells, heparin-Sepharose chromatography was performed. These experiments were carried out by using specific antibodies against the N-terminal portion of bFGF and the truncated form of aFGF. The cytosol was chromatographed on a heparin-Sepharose column equilibrated with 0.6 M NaCl. The material bound to the heparin-Sepharose was then eluted directly with 3 M XaC1 and tested in a direct ELISA with specific anti-aFGF and bFGF antibodies. It was shown that 3 fractions in the 3 M NaCl eluate were recognized by the specific polyclonal anti-bFGF antibody but not by a n antibody against aFGF (Fig. 2 ) . The material which was not bound on the heparinSepharose column did not react with the specific antibodies against bFGF and aFGF (data not shown). The mitogenic activity of the fractions which recognized the polyclonal anti-bFGF antibody was tested on BEL cells. Figure 3 shows t h a t the three fractions were mitogenic for BEL cells. Effect of TPA on the expression of the bFGF-like mitogenic activity Experiments were carried out in order to investigate whether TPA modifies the expression of bFGF for that cell type. In all experiments the cells were maintained for 36 h r in serum-free medium and subsequently stimulated with TPA. Longer starvation was generally avoided in order to minimize cell detachment and damage. Cytosols derived from 36 h r serum-deprived HOME cells showed a similar effect on DNA synthesis in BEL cells whether or not they were treated with TPA (data not shown). Similarly, Western blots of cytosols derived from 4.5 h r TPA-treated HOME cells using the specific polyclonal antibody did not show significant differences to untreated cells (Fig. 4). The Western blotting experiments demonstrated also that the specific polyclonal antibody against bFGF recognized a major 17 Kd protein but also a doublet of 20-22

50

45

55

60

65

Fraction Number

I

Fig. 2. Heparin-Sepharose chromatography of HOME cell-derived cytosol. HOME cell cytosol(19 mg protein) was extracted in 1 M NaCl and treated as indicated under Materials and Methods. After acidification for 1 hr (pH 5) the pH was adjusted to 7.2 and the cytosol was chromatographed as described. The fractions (1ml each) were tested in a ELISA with specific polyelonal anti-bFGF and aFGF antibodies.

Elor--

C

bFGF

47

48

49

Thr

EGF

Fig. 3. Effect of heparin-Sepharose fractions from HOME cell-derived cytosol on 3H-thymidine incorporation in EEL cells. The three fractions (fraction 47,48, and 49)recognized by the specific anti-bFGF antibody were diluted by half in test medium, arid 10 p1 of each fraction was tested for 3H-thymidine incorporation in BEL cells. Thrombin (Thr),EGF, and PDGF (not shown) a s a control did not stimulate ,5H-thymidineincorporation in BEL cells. The experiments were performed twice with similar results (Results as meant-SEMI.

Kd in the HOME cell cytosol (Fig. 4).These bands were not detected in control blots using non-immune serum instead of the specific antibodies (data not shown). Comparative Northern blots were performed in order to investigate whether TPA could also act a t the mRNA level. These experiments were conducted in comparison to HUVE cells since bFGF mRNA has been reported in HUVE cells (Hannan et al., 1988). Northern blotting experiments were done by using a specific 1.4 kb probe for bFGF and a specific 4.04kb probe for

aFGF. It was demonstrated that only bFGF was present in

155

BASIC FGF EXPRESSION IN ENDOTHELIAL CELLS

M.W

A

HOMEC

KDa

KDa

HUVEC 1 2 3

1 2 3 4 5

4 30

4 21.5

17-

9.5

9.5

6.6

6.6

4.3

4.3

2.2

2.2 2.0

2.0

4 14.3

1

2 3

4 5

Fig. 4. Western blot of cytosols of 4.5 hr TPA treated or untreated HOME cells. Cytosols of TPA treated (lane 4 = 150 pg and lane 5 = 300 pg) and untreated (lane2 :150 pg and lane 3 = 300 pg) HOME cells were solubilized in SDS sample buffer and run under reduced conditions on a 18% polyacrylamide gel. After transfer to a nitrocellulose membrane, incubation with anti-bFGF antibody and '251-protein A and autoradiography were done as described in Materials and Methods. Lane 1 = 1 kg rbFGF.

HOME cells and that the specific bFGF probe hybridized with five mRNA species of 6.) 6, 3.7, 2.2, 2.0, and 1.0 kb (Figs. 5A, 6). Basic FGF mRNA was still detectable after maintaining the cells for 68 hr in serum-free medium (these experiments were performed despite the risk of cell detachment) (reduction of 92% and 76% for the 6.6 kb and 3.7 kb band, respectively, in the 68 h deprived cells when compared to the 36 hr deprived cells) (Fig. 5A,B). Under the same experimental conditions (36 hr serum deprivation and absence of stimulation by TPA) no bFGF or aFGF mRNA could be detected in HUVE cells (Figs. 5A, 6). Northern blots from TPA-treated HOME cells showed some time-dependent modifications (Fig. 5A,B). The densitometric analysis of the 6.6 and 3.7 kb bands demonstrated a short and transitory increase of both bands at 1.5 hr (rel. area 145%and 180%,respectively) and a subsequent decrease. TPA was not capable of inducing aFGF mRNA in HOME cells (Fig. 6). Thus, treatment of HOME cells by TPA did not lead to an stable and significant increase of bFGF mRNA and did not induce aFGF mRNA. By contrast, treatment of HUVE cells with TPA resulted in a induction of the expression of bFGF but not aFGF mRNA (Figs. 5A, 6).

DISCUSSION Surgeons frequently use grafts from the omentum to improve the collateral circulation after cardiac and vascular surgery. It has been shown that an extract of the omentum contains a potent angiogenic activity (Goldsmith et al., 1984) and that this material increases the vascular perfusion after injection into the region of a defined ischemic area (Goldsmith et al., 1986). In support of the contention that the angiogenic activity described in the omentum is a bFGF-like mol-

GA PDHI

200

-

W

m

150

01

L

m

2

100

e

L

50

0

ns

i.sh

sh

6sh sd

Fig. 5. Basic FGF mRNA expression in HOME cells and HUVE cells and the effect of TPA. HOME cells or HUVE cells (6 x lo7 cellsitime point) were grown to preconfluence and serum-deprived for 36 hr. Thirty-six hours serum-deprived cells were used as nonstimulated control. After serum-deprivation, TPA (100 ngiml) was added for several time intervals and RNA was extracted and Northern blots were done as indicated in Materials and Methods. A Poly (A+)-richRNA (7 pg) from HOME cells, non-stimulated (lane 2), TPA treated for 1.5 h r (lane 3) and 8 hr (lane 4), and from 68 hr serum-deprived cells (lane 5). Poly (A+)-richRNA (4 pg! from HUVE cells, non-stimulated (lane 2) or TPA treated for 8 hr (lane 3). Poly (A+))-richRNA (3 pg) from bovine retina (lanes 1) was used as control. The sizes of 32P-labeledrestriction fragments of lambda DNA are indicated. After hybridization with the bFGF probe, the blots were washed and rehybridized with the GAPDH plasmid probe in order to determine the relative amounts of RNA loaded per lane. B: Densitometric analysis of bFGF mRNA expression in TPA treated and untreated HOME cells. With the autoradiogram shown above the silver grain densities of the two major bands (6.6 and 3.7 kb) were compared among treatment groups. The GAPDH autoradiographic signal was used as internal standard. The data are presented as percentages relative to the area of non-stimulated (ns) cells. Serumdeprived (sd).

156

BIKFALVI ET AL

HUVEC

mRNA. Thus, as has been demonstrated for bovine capillary endothelial cells (BCE cells) (Schweigerer et al., 1 2 3 1 2 3 1987) and bovine aortic endothelial cells (BAE cells) (Vlodavsky et al., 1987a,b1, human microvascular en9.5 dothelial cells contained a growth factor similar or identical to bFGF. However, there are differences for 6.6 HOME cell bFGF compared to the latter ones. The immunoblots showed the presence of higher molecular weight forms of bFGF. The existence of higher molec4.3 ular weight forms is not predictable from the nucleotide sequence, assuming that the initiation of transcription starts from the AUG codon at position -9. Nevertheless, higher molecular weight forms of bFGF 2.2 have been reported in human placenta (Moscatelli et al., 1986b; Sommer et al., 19871, in guinea pig brain 2 .o (Moscatelli et al., 19871, in rat brain (Presta et al., 19881, and in Kaposi sarcoma cells (Ensoli et al., 1989). Taken together, these data suggest that there are probably alternating initiation sites in the bFGF gene. This contention is reinforced by recent datas of Prats et al. (1989) and Florkiewicz et al. (19891, who have demonstrated that higher molecular weight forms are initiated by alternative CUG codons. It would be interesting to determine the exact sequence of the higher molecular weight alternative translational products of bFGF in HOME cells. GAPDH The second difference is that unlike the bovine endothelium, where two mRNA species were described Fig. 6 . Acidic FGF expression in HOME and HUVE cells and the effect of TPA. Cells were grown as described in Materials and Meth- (Schweigerer et al., 1987; Fiddes et al., 19871, HOME ods. Poly ( A +)-rich RNA from HOME cells, non-stimulated (lane2: 10 cells exhibits five mRNA species in our study. Kuropg) or stimulated for 8 h r by TPA (lane 3: 10 pg). Poly (A ' )-rich RNA gawa e t al. (1987) and Sternfeld et al. (1988) have also from HUVE cells, non-stimulated (lane2: 6 pg), or stimulated for 8 h r reported in human dermal fibroblast more than two by TPA (lane 3: 6 pg). Poly (A+)-richRNA from bovine retina (left panel, lane 1) and brain (right panel, lane 1) were used as controls. bFGF mRNA species of 7.0, 3.7, 2.2, and 1.5 kb. Similarly, i t has been shown that human astrocytoma cells After hybridization with the aFGF cDNA probe, the blots were washed and rehybridized with the GAPDH plasmid probe used as (Murphy et al., 19881 and Kaposi sarcoma cells (Ensoli positive control. et al., 1989) contain at least four bFGF mRNA species. Kurogawa et al. (1987) have suggested that this arose by different polyadenylation sites. Thus, the presence ecule contained in the omental microvessels, we have of several bFGF mRNA species is perhaps a general addressed the question of whether HOME cells synthe- feature for human cells. Among human endothelial cells other than HOME size an angiogenic activity similar or identical to cells, bFGF expression has been reported only in bFGF. We show herein that HOME cells express a mitogen HUVE cells. Hannan et al. (1988) have reported that which is similar if not identical to bFGF. HOME cell bFGF could be detected in HUVE cells at a protein and cytosol was capable of stimulating DNA synthesis in mRNA level. However, we were unable to detect bFGF BEL cells. BEL cells were chosen as a target because mRNA in non-stimulated, 36 h r serum-deprived they respond to both aFGF and bFGF, but not to PDGF, HUVE cells, suggesting that bFGF might be expressed EGF, and thrombin. The mitogenic activity of the in that cell type only a t low levels. This agrees with the HOME cell cytosol was inhibited by a polyclonal anti- results of Ensoli et al. (1989), who were also unable to bFGF antibody, suggesting t h a t the mitogenic activity find significant amounts of bFGF mRNA or protein in HUVE cells. is related to a molecule similar to bFGF. Activation of protein kinase C by TPA regulates the In order to further characterize the activity present in HOME cell-derived cytosol, Western blot experi- expression of several growth factors and oncogenes ments and heparin-Sepharose chromatography were (Daniel et al., 1986; Rosengurt and Sinnet-Smith, performed. The immunoblots not only showed the pres- 1987). However, TPA did not induces significant ence of a major molecular species of 17 Kd reacting changes in the expression of bFGF a t the protein level, with the specific anti-bFGF antibody in HOME cells but produced a short and transitory increase in the but also a doublet of 20 Kd-22 Kd, suggesting the pres- mRNA level in HOME cells. Nevertheless, protein kience of higher molecular weight products in that cell nase C activation by TPA in HOME cells induced retype. The ELISA of the 3 M NaCl eluted material using versible changes in the cellular phenotype and stimuspecific antibodies against aFGF and bFGF demon- lated the phophorylation of a 27 kD protein, but not strated clearly that bFGF was present in HOME cells DNA synthesis (Dupuy et al., 1989). Contrastingly, our but not aFGF. The 3 M NaC1-eluted material was also results show that TPA was capable of inducing bFGF mitogenic for BEL cells. In addition, Northern blotting mRNA in HUVE cells. Taken together, the results sugexperiments demonstrated bFGF mRNA but not aFGF gest that bFGF mRNAs are abundantly found in non-

HOMEC

BASIC FGF EXPRESSION IN ENDOTHELIAL CELLS

stimulated HOME cells and that protein kinase C may not influence significantly the transcription of bFGF for this cell type. However, Murphy et al. (1988) have recently reported that phorbol-12,13-dibutirate (PDBu) could induce a n increase in the transcription of bFGF mRNA in a human astrocytoma cell line. Thus, regulation of transcription of bFGF by PKC seems to be different for several bFGF producing cells. Feige and Baird (1989) have recently shown t h a t bFGF is phosphorylated in bovine capillary endothelial cells. Protein kinase C could therefore regulate bFGF in two different ways, modulating bFGF expression a t the gene level and phosphorylating bFGF at a protein level. Our results are apparently in contradiction with the data of Ohtaki et al. (1989), who have recently isolated aFGF from the bovine omentum. Another cell-type (i.e., fat cells, mesothelial cells), rather than microvascular endothelial cells contained in the omentum, is possibly responsible for the expression of aFGF. In conclusion, the present results indicate that microvascular endothelial cells derived from the human omentum express a mitogenic activity similar or identical to bFGF and that TPA does not induce a stable and prolonged increase in its expression. This molecule belongs possibly to the angiogenic activity described in the omentum.

ACKNOWLEDGMENTS The authors would like to thank Dr. J. Abraham and Dr. J. Fiddes (California Biotechnology inc., Mountain View, CAI for their gift of the 1.4 kb bFGF probe, and Dr. J.M. Blanchard (Montpellier university, France) for the gift of the GAPDH probe. The research was supported by the “Association de la Recherche Contre le Cancer (ARC).” LITERATURE CITED Alterio, J., Halley, C., Brou, C., Soussi, T., Courtois, Y., and Laurent, M. (1988) Characterization of a bovine acidic FGF cDNA clone and its expression in brain and retina. FEBS Lett., 242:41-46. Abraham, ?J., Mergia, A,, Whang, J.L., Tumolo, A,, Friedman, J., Hlerrild, C., Gospodarowicz, D., and Fiddes, J. (1986) Nucleotide sequence of a bovine clone encoding the angiogenic protein, basic fibroblast growth factor. Science, 233545-547. Baird, A., and Ling, N. (1987)Fibroblast growth factors are present in the extracellular matrix produced by endothelial cells in vitro: implications for a role of heparinase-like enzymes in the neovascular response. Biochem. Biophys. Res. Commun., 142:428-435. Bikfalvi, A., Dupuy, E., Inyang, A.L., Fayein, N., Leseche, G., Courtois, Y., and Tobelem, G. (1989) Binding, internalization and degradation of basic fibroblast growth factor in human microvascular endothelial cells. Exp. Cell. Res., 181r75-84. Bradford, M.M. (1979) A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein dye binding. Anal. Biochem., 72:248-254. Burnette, W.N. (1981) Western blotting: Electrophoretic transfert of proteins from sodium dodecyl sulfate-polyacrylamide gels to unmodified nitrocellulose and radiographic detection with antibody and radioiodinated protein A. Anal. Biochem., 112:195-203. Daniel, T.O., Gibbs, V.C., Milfay, D.F., Garoway, M.R., and Williams, L.T. (1986) Thrombin stimulates c-sis gene expression in microvascular endothelial cells. J. Biol. Chem., 261r9579-9582. Dupuy, E., Bikfalvi, A,, Rendu, F., Levi-Toledano, S., and Tobelem, G. (1989) Thrombin mitogenic responses and protein phosphorylations are different in endothelial cells derived from large and small vessels. Exp. Cell. Res., 185:363-372. Doctrow, S.R., and Folkman, J. (1988) Protein kinase C activators supress stimulation of capillary endothelial cell growth by angiogenic endothelial mitogens. J. Cell. Biol., L04r679-687. Doctrow, S.R. (1989) Protein kinase C-mediated modulation of bFGF

157

receptor in capillary endothelial cells. Abstract 1485, Am. SOC. Cell Biol., J . Cell. Biol, 107, Nb6. Ensoli, B., Nakrnura, S., Salahuddin, S.Z., Biberfeld, P., Larsson, L., Beaver, B., Wong-Stall, F., and Gallo, R.C. (1989) Aids Kaposi’s sarcoma derived cells express cytokines with autocrine and paracrine growth effects. Science, 243:223-226. Esch, F., Baird A,, Ling, N., Veno, N., Hill, F., Denoroy, L., Klepper, R., Gospodarowicz, D., Bohlen, P., and Guillemin, R. (1985) Primary structure of bovine pituitary bone fibroblast growth factor IFGF) and comparison with the amino terminal sequence of bovine brain acidic FGF. Roc. Natl. Acad. Sci. USA, 82:6507-6511. Feige, J.J., and Baird, A. (19891 Fibroblast growth factor is a substrate for protein phosphorylation and is phosphorylated in capillary endothelial cells in culture. Proc. Natl. Acad. Sci., USA, 86: 3174-3178. Fiddes J., Whang, J.L., Tumolo, A,, and Abraham, J. (1987) Genes for the angiogenic growth factors, basic and acidic fibroblast growth factor. In: Angiogenesis. D.B. Rifkin, M. Klagsbrun, eds. Cold Spring Harbor, New York, pp. 24-31. Fort, P., Marty, L., Piehaczyyk, M., Sabrouty, S., Dani, C., Jeanteur, P., and Blanchard, J.M. (1985) Various rat adult tissues express only one major mRNA species from thc glyceraldehyde3-phosphate-deshydrogenasemultigenic family. Nucl. Acids Res., 13:1431-1442. Gospodarowicz, D., Bialecki, L., and Trakal, T. (1979) The angiogenic activity of the fibroblast and epidermal growth factor. Exp. - Eye . Res., 28501-514. GosDodarowicz. D.. Neufeld. G.. and Schweigerer. L. (1987)Fibroblast growth factor: structural and biological properties. J . Cell. Physiol., (Suppl., 5):15-26. Goldsmith, H.S., Griffith, A.L., Kupferman, A,, and Catsimpoolas, N. (1986) Increased vascular perfusion after administration of a n omental lipid fraction. Surg. Gynecol. Obst., 162t579-583. Goldsmith, € 1 3 , Griffith, A.L., Kupferman, A., and Catsimpoolas, N. (1984) Lipid angiogenic factor from omentum. J. Am. Med. Assoc., 252:2034-2036. Halaban, R., Langdon, R., Birchall, N., Baird, A,, Scott, G., Moellmann, G., and McGuire, J. (1988) Basic fibroblast growth factor from human keratinocytes is a natural mitogen for melanocytes. J. Cell. Biol., 107:1611-1618. Hannan, R.L., Kourembanas, S., Flanders, K., Rogeli, S.J., Roberts, A.B., Faller, D., and Klagsbrun, M. (1988) Endothelial cells synthetize basic fibroblast growth factor and transforming growth factor beta. Growth Factors, I:?-17. Hayek, A., Culler, F., Beattie, G.M., Lopez, A.D., Cuevas, P., and Baird, A. (1987)An in vitro model for study of the angiogenic effect of basic fibroblast growth factor. Biochem. Biophys. Res. Commun., 147:876-880. Jaffe, E.A., Nachman, R.L., Becker, C.G., and Minick, C.V.R. (1973) Culture of human endothelial cells derived from umbilical veins. Identification by morphological and immunological criteria. J. Clin. Invest., 52r2745-1829. Kern, P., Knedler, A., and Eckel, R.H. (1983) Isolation and culture of human microvascular endothelial cells from adipose tissue. J . Clin. Invest., 71r1822-1829. Kurokawa, T., Sasada, R., Iwane, M., and Igarashi, K. (1987) Cloning and expression of cDNA encoding human basic fibroblast growth factor. FEBS Lett., 213:189-194. Laemmli, VK (1970) Cleavage of structual proteins during assembly of the head of bacteriophage T4. Nature, 227:680-685. Moenner, M., Chevallier, B., Badet, T. and Barrifault, D. (1986) Evidence and characterization of the receptor of eye-derived growth factor I, thc retinal form of bFGF, on bovine epithetial lens cells. Proc. Natl. Acad. Sci., USA 83.5024-5028, Moscatelli, D., Presta, M., Joseph-Silverstein, J., and Rifkin, D. (1986a) Both normal and tumor cells produce basic fibroblast growth factor. J. Cell. Physiol., 129273-276. Moscatelli, D., Presta, M., and Rifkin, D.B. (198613) Purification of a factor from human placenta that stimulates capillary endothelial cell protease production, DNA synthesis and migration. Proc. Natl. Acad. Sci., USA, 83.2091-20095, Moscatelli, D., Joseph-Silverstein, J., MNanejiass, R., and Rifkin, D.B. (1987) 25000 heparin-binding protein from guinea-pig brain is a high molecular weight form of basic fibroblast growth factor. Proc. Natl. Acad. Sci., USA, 84t5778-5782. Montesano, R., Vassalli, J.D., Baird, A., Guillemin, R., and Orci L. (1986) Basic fibroblast growth factor induces angiogenesis in vitro. Proc. Natl. Acad Sci. USA, 83:7297-7301. Murphy, P.R., Sato, Y., Sato, R., and Friesen, H. (1988) Regulation of multiple basic fibroblast growth factor messanger ribonucleic acid transcripts by protein kinase C. Mol. Endocrinol., 2t1196-1201.

158

BIKFALVI ET AL

Ohtaki, T., Wakamatsu, K., Ishiashi, Y., and Yasuhara, T. (1989) Purfication of acidic fibroblast growth factor from bovine omentum. Biochem. Biophys. Res. Commun., 161:169-175. Plouet, J., Courty, J., Olivi, M., Courtois, Y., and Barritault, D. (1984) High reliable and sensitive assay for the purification of cellular growth factors. J. Cell. Mol. Biol., 30:105-110. Presta, M., Rusnati, M., Maier, J.A.M.,and Ragnotti, G. (1988) Purification of basic fibroblast growth factor from rat brain: identification of a Mr. 22,000 immunoreactive form. Biochem. Biophys. Res. Commun., 155rl16 1-1 172. Rosengurt, E., Sinnet-Smith, J.W. (1987) Bombesin induction of c-fos and c-myc protooncogenes inswiss 3T3: significance in the mitogenic response. J. Cell. Physiol., 131t218-225. Schweigerer, L., Neufeld, G., Friedman, J., Abraham, J., Fiddes, J., and Gospodarowicz, D. (1987) Capillary endothelial cells express basic fibroblast growth factor, a mitogen that promotes their own growth. Nature, 325357-259.

Sommer, A., Brewer, M.T., Thompson, R.C., Moscatelli, D., Presta, M., and Rifkin, D.B. (1987) A form of basic FGF with a n extended amino terminus. Biochem. Biophys. Res. Commun., 144:543-550. Sternfeld, M.D., Hendrickson, J.E., Keeble, W., Rosenbaum, J., Robertson, J.E., Pittelkow, M.R., and Shipley, G.D. (1988) Differential expression of mRNA coding for heparin-binding growth factor type 2 in human cells. J. Cell. Physiol., 136:297-304. Vlodavsky, I., Fridman, R., Sullivan, R., Sasse, J.,and Klagsbrun, M. (1987a) Aortic endothelial cells synthetize basic fibroblast growth factor which remains cell associated and platelet derived growth factor-like protein which is secreted. J. Cell. Physiol., 131:402-408. Vlodavsky, I., Folkman, J., Sullivan. R., Fridman, R., Ishai-Michaeli, R., Sasse, J., and Klagsbrun, M. (198713) Endothelial cell derived basic fibroblast growth factor: synthesis and deposition into the extracellular matrix. Proc. Natl. Acad. Sci., USA, 54:2292-2296.