Exp. Eye Res. (1992) 55, 663-669
Preparation Localization
of Fluorescent Basic Fibroblast in Living Retinal Microvascular AILEEN
M. HEALY
AND
Growth Factor: Endothelial Cells
IRA M. HERMAN”
Program in Cell, Molecular and Developmental Biology, Tuft,s University Health Science Schools, 736 Harrison Avenue, Boston, MA 0217 I, U.S.A. (Received
Houston
2 December
1991 and accepted in revised form 7 I February
1992)
A biologically active fluorescent derivative of recombinant human basic fibroblast growth factor (bFGF) was prepared by immobilization on heparin-Sepharose 4B (HS) and derivatization with the fluorophore, Texas Red (TR). TR-bFGF was separated from free dye and carrier protein by elution from HS using 1.5 M NaCI. TR-bFGFcontained an average of two dye molecules bound per bFGF,retained its mitogenic activity and was visible using a fluorescence microscope equipped with silicon intensified target camera (SIT). TR-bFGF stimulated the growth of bovine aortic endothelial cells (BAEC), microvessel endothelial cells (MVEC) and BHK-2 1 cells grown in culture. BAEC,MVEC and BHK-21 cells treated with 20 ng ml-l (1 nM) TR-bFGFfor 72 hr were stimulated over serum controls by 8 7, 26 and 6 %, respectively. TR-bFGF stimulated EC growth was inhibited in a dose-dependentfashion when cells were coincubated with ,UM chloroquine. When EC were treated with TR-bFGF at 4°C and then monitored at 37% bright, focal, cytoplasmic spots were observed, which accumulated as punctate, perinuclear tluorescence. EC internalization of TR-bFGFwas inhibited 80% by the addition of loo-fold molar excessunlabeled bFGF or by maintaining cultures at 4°C. TR-bFGFcolocalized with an EC lysosomal marker, but TR-bFGFwas not detected in the nucleus. Results of these localization studies suggest that TR-bFGF stimulates EC proliferation without entering the nucleus. Key words: growth factors ; lysosomes: signal transduction.
1. Introduction
The controlled growth, development and remodeling of the adult vasculature is a complex processregulated, in part, by vascular cell interactions with soluble and matrix-bound growth factors. Vascular EC synthesize and store the heparin-binding growth factor, bFGF, within the subendothelial matrix (Vladovsky et al., 1987). bFGF mediates EC migration and proliferation in vitro and angiogenesis in vivo (for review see Burgess and Maciag, 1989 ; D’Amore and Thompson, 198 7). During angiogenesis, EC initially migrate through a hyaluronate-rich matrix and subsequently form cell-cell contacts within a matrix enriched in sulfated glycosaminoglycans (Ausprunk and Folkman, 19 77). Presumably, bFGFis stored within the vascular intima
in association
with
the glycosaminoglycan,
heparan sulfate proteoglycan (Moscatelli, 1988 ; Bashkin et al., 1989). Results of in vitro studies reveal that bFGF is released from matrix storage sites by proteolysis (Sakselaet al., 1990). Following its release from the subendothelial matrix in vivo, soluble bFGF may stimulate EC migration and proliferation, two key events during angiogenesis. Soluble bFGF may act locally to stimulate EC growth via high-affinity cellsurface FGF receptors. Low-affinity heparin binding sites may facilitate FGF receptor occupancy (Yayon et al., 1991). The bFGF receptors have been isolated and the genes encoding these receptors, cloned and
2. Materials and Methods Cell Culture
BAEC and retinal MVEC were isolated and characterized as previously described (Herman and
* For correspondence.
0014-4835/92/110663+07
sequenced (Neufeld and Gospodarowicz, 198 5 ; Lee et al., 19 8 9). FGF-receptor occupancy triggers receptormediated endocytosis followed by down regulation of cell-surface receptors (Moscatelli, 1988). The intracellular events that ensue, resulting in EC migration and proliferation, are undefined. The intracellular fate of endocytosed bFGF is unclear. Studies using radiolabeled bFGF suggest the FGF-receptor complex is endocytosed and the ligand degraded within lysosomes (Moscatelli, 1988 : Bikfalvi et al., 1989). However, the results of studies in which EC were treated with exogenous bFGF then flxed and prepared for immunofluorescence microscopy, suggest FGF is found in the nucleus (Baldin et al., 1990; Bouche et al., 1987). To determine the fate of bFGFin EC and to establish whether exogenously added bFGF transits to the nuclear compartment of living EC, we labeled bFGFwith the fluorophore, TR, and monitored its internalization in living ECin culture. In this report we describe the specific binding, internalization and Iysosomal targeting of a biologically active TR-bFGF. Our evidence suggests bFGF, which is exogenously added to target cell populations, stimulates cell proliferation without entering the nucleus.
$08.00/O
0 1992 Academic Press Limited
664
Castellot, 198 7 ; Herman and D’Amore, 1984). Cells were routinely cultured in Dulbecco’s modified Eagle’s medium (DMEM; Gibco Laboratories, Grand Island, NY) containing 10 % bovine calf serum (CS; Hyclone, Logan, UT), with penicillin, streptomycin and amphotericin (Gibco Laboratories, Grand Island, NY) and used between passagesfive and ten. BHK-2 1 cells (American Type Culture Collection, Rockville, MD) were cultured in DMEM/Ham’s F12 (Irvine Scientific, Santa Ana, CA) supplemented with 10 % CS and antibiotics. Proli,ferationAssays EC were plated in DMEM containing 10% CS at an initial density of 15 x lo4 cells per 2 cm2 well (Costar, Cambridge, MA). BHK-21 cells were treated similarly, except that cells were plated in DMEM/Ham’s F12 containing 10% CS. Twenty four hours after plating, culture media were removed and cells were washed twice with phosphate-buffered saline (PBS). To assay cell proliferation, test media containing 20 and 40 ng ml-’ of recombinant human bFGF (Takeda Chemical Ind., Osaka,Japan), with or without 0.5, 1.0 and 5.0 ,u~ chloroquine (Sigma Chemical Co., St. Louis, MO) in DMEM with 2 % CS,were added to each well. At the indicated days, cells from triplicate wells were trypsinized and electronically counted (Coulter Model ZF, Hialeigh, FL).
A.M.
HEALY
AND
I. M. HERMAN
BiochemicalAnalysis of TR-bFGF SDS-PAGE.TR-bFGF was electrophoresed on 15 % polyacrylamide gels (Laemmli, 1970) or 5-l 5 % polyacrylamide gradient gels under reducing conditions and detected by silver stain (Jesteret al., 1987). FluorescenceLocalization of TR-bFGF in Acrylamide Gels. TR-bFGF was electrophoresed in a 15 % acrylamide gel and viewed by UV irradiation using a longwave UV transilluminator (peak wavelength = 302 nm, UVP Inc., San Gabriel, CA) UV transilluminated gels containing TR-bFGF were photographed with Poloroid Instant film 667 (Poloroid Corp., Cambridge. MA). Staphylococcusaureus V-8 ProteaseDigestion oj TRbFGF and bFGF. TR-bFGF and bFGF were enzymatically digested with S. aureus V-8 protease (Sigma Chem. Co.) according to the procedure of Cleveland et al. (19 7 7). Briefly, TR-bFGFand bFGFwere dissolved in 0.5 y0 SDS, 0.125 M Tris-Cl pH 6.8. 10% glycerol, 0.001% bromphenol blue and containing O-2 Units S. aureus V-8 protease. bFGF and TR-bFGF were digested at 37’C for 1 hr using a 1: 1 weight ratio of growth factor to enzyme. The enzyme was inactivated by the addition of 2-mercaptoethanol and SDS to a final concentration of 10 and 2%. respectively. Samples were alkylated with 80 mM iodoacetamide (Sigma Chem. Co.) at 37°C for 30 min and electrophoresed on a 5-l 5% gradient gel. Peptide cleavage products were detected by silver stain or immunoblot (DeNofrio, Hoock and Herman, 1989).
bFGF Conjugation to Texas Red Texas Red-bFGFconjugation was performed while bFGF was adsorbed to heparin Sepharose 4B (HS; Pharmacia, Uppsala, Sweden). Typically, a 75 ~1 packed volume of HS was equilibrated in 250 ~1 of binding buffer (0.05 M NaCl, 0.0 1 M Tris-Cl, pH 7.5) containing 2.2 x 1O-5M ovalbumin (Sigma Chemical Co,, St Louis, MO), at room temperature for 30 min. HS was washed with 100 column volumes of binding buffer. 2.0 x 10e6M bFGF(300 ~1) was incubated with equilibrated HS in binding buffer at 4°C for 90 min. HS-bound bFGF was then re-equilibrated in 200 ~1 of conjugation buffer (0.01 M Na,CO,, 0.05 M NaCl, pH 8.8) containing 2.2 x 10e4M ovalbumin (100 ~1). 3.3 x 10M3M TR (Molecular Probes, Eugene, OR) dissolved in 50 ,ul conjugation buffer was added to HSbound bFGF. The final reaction volume equaled 3 50 pl. The reaction proceededon ice for 15 min with frequent agitation. To remove unbound dye and separate ovalbumin from TR-bFGF, HS was washed exhaustively with a 0.05, 0.5, 0.75, 1.0 and 1.25 M NaCl, pH 7.5. step gradient. TR-bFGF was eluted from HS with 1.5 M NaCl, pH 7.5. The labeling index or absorbance ratio of dye to protein was determined spectrophotometrically at 596 and 280 nm, respectively (Beckman Model DU 65, Fullerton, CA).
LocalizationExperiments EC and BHK-21 cells were plated in DMEM containing 10% CS onto gelatin-coated glass microscope coverslips and allowed to attach for 18 hr. The following morning, cells were washed twice with PBS and preincubated in DMEM containing 2 y0CSat 3 7°C for 2 hr. Cells were then cooled to 10°C and incubated with 20 or 40 ng ml-’ TR-bFGFin DMEM + 2 y0CSfor 2 hr. Cells were then washed twice with PBSand TRbFGF-containing media were replaced with fresh DMEM containing 2 % CSand cells were mounted in a temperature-regulated chamber slide. TR-bFGFlocalization in living EC or BHK-21 cells was monitored using either an IM35 fluorescence microscope or a Photomicroscope II (Zeiss, West Germany), both equipped with a x 63 objective (N.A. 1.4) and a Venus Two Stage Intensifier Video Detector, Model DV2 (Venus Sci. Incorp., Farmingdale, NY). Images were recorded on a Time-Lapse Video Recorder, Model AG 6050 (Panasonic, Seacaucus, NJ). Data shown were collected with the Photomicroscope II. Colocalization of TR-bFGF and Lysosomes
EC were cultured as described for localization experiments, then washed briefly in PBS and fixed in
BFGF
IN RETINAL
ENDOTHELIUM
methanol at 4°C for 5 min. Following three, 5 min PBS washes, cells were sequentially incubated with rabbit and anti-lgp120, an immunoglobulin directed against a 120 000 Da lysosomal membrane glycoprotein (I. Mellman, Yale University, New Haven, CT) and rhodamine-labeled goat anti-rabbit immunoglobulin (Cooper Biomedical, Malvern, PA).
665
20 ng ml-’ an optimal concentration. A lysosomotropic base, which disrupts the endocytic pathway, was employed to assesswhether alkalinization of the lysosomal compartment perturbs bFGF and TR-bFGF (A)
(B)
3. Results Texas Red Labeled bFGF
We directly labeled bFGF using the fluorescent dye TR. The conjugation reaction was conducted with bFGF bound to HS, which produced TR-bFGF labeled with varying numbers of dye molecules. Electrophoresis and silver staining of recombinant bFGF [Fig. l(A), lane 11and TR-bFGFrevealed a labeled molecule with a relative mobility of 19,250 Da, which corresponds to a labeling index (i.e. 596/280 nm) of 0.5 [Fig. l(A), lane 21. This represents two TR molecules per bFGF. Electrophoresed TR-bFGF was also visualized following exposure to a 302 nm UV source [Fig. l(A), lane 31. TR-bFGF preparations did not contain contaminating free dye, ovalbumin or unlabeled bFGF [Fig. l(A)]. Peptide maps were generated to determine whether covalent modifications of bFGF with TR significantly altered its secondary structure. To this end, bFGF and TR-bFGFwere treated with S. aureusprotease V-8 and electrophoresed through gradient gels. Following silver staining, five low-molecular-weight polypeptide cleavage products were observed [Fig. l(B), lanes 2 and 4). All the bFGF and TR-bFGF fragments were identical, with relative mobilities ranging from 11000-5 000 Da. bFGF,TR-bFGFand the resulting V8 cleavage peptides were also immunoreactive with anti-bFGF IgG by Western blot analysis (not shown).
TR-bFGF Promotes
EC Growth
To determine if bFGFmaintained growth promoting activity following labeling with TR, target cell populations with known sensitivity to bFGF were assayedfor growth promotion. BAEC,MVEC and BHK21 cells were grown in culture in the presence or absence of either bFGF or TR-bFGF as described. Proliferation assay results are summarized in Table I. As indicated, after 72 hr in culture bFGF stimulated MVEC growth 2.3 times as compared with serum controls, and TR-bFGF stimulated MVEC growth 1.4 times above serum controls. BAEC grown in the presence of bFGF were stimulated 3.6 times above serum controls, while TR-bFGF stimulated BAEC growth 1.9 times. BHK-21 cells are not markedly stimulated by TK-bFGF. By 96 hr, cells under all growth conditions attained confluence. bFGF and TRbFGFstimulation of ECgrowth is dose-dependent,with
I2
I
3
2
3
4
Electrophoreticanalysisof bFGF and TRQFGF. (A) Silver stain of a 15% acrylamidegel containing bFGF(lane FIG. 1.
1) and TR-bFGF (lane 2). Fluorescent TR-bFGF is detected in unfixed gels following UV irradiation (lane 3). (B) Silver stained, 5-15 % gradient gel containing bFGF (lane 1). bFGF plus S. aureus V-8 protease (lane 2). TR-bFGF (lane 3) and TR-bFGF plus S. aureus V-8 protease (lane 4). Molecular weight markers are indicated in kDa.
TABLE
I
l’R-bFGF stimulates cell prohjeration TR-bFGF/ Serum
bFGF/ Serum Chloroquine
0
~/AM
.~-.
MVEC BAEC BHK
2.3t 3.6 1.4
0
5j.N
1 ,uM
5 ,UM
~~
1.9*** 3.3*** 1.3**
1.3*** 3.5*** 1.2*
1.4 1.3** 1.0*** 1.9 2.0** 2.3** 1.0 1.1* 1.0*
t The numbers represent the fold increase in cell number from triplicate wells cultured for 3 days then counted. bFGF or TR-bFGF stimulated cells (20 ng ml-‘)/control cells (serum alone). SEM < 13 % for all groups. P values represent analysis of chloroquine-treated vs. control cells (no chloroquine treatment) in FGF and TR-bFGF stimulated cells: *P > 0.05. ** P < 0.05 and *** P < 0901.
TABLE
11
Texasred-bFGFlabeling index Labeling index (596/280 nm) 0.20-0.49 0.50-0.79 0.80-2.50
“/o Stimulation* 27 (n =: 31 61 (n = 4) 21 (n =: 3)
* xstimulation = lOO(i- x)/(y -x). x = controlcells. ECcultured for three days in media containing 2% cs. y = cells treated with 20 ng ml-’ bFCF. and 2 = cells treated with 20 ng ml-’ TR-bFGF. EEK55
666
A.M.
HEALY
AND
I. M. HERMAN
FIG. 2. TR-bFGFLocalization in Living EC. Phase contrast (A and C) and fluorescent (B and D) images of BAEC in culture. EC treated with 20 ng ml-’ TR-bFGF for 2 hr at 4-10°C prior to warming to 37°C for approximately 1 hr before viewing (A and B). EC cultured for 24 hr after the addition of 20 ng ml-’ TR-bFGF(C and D). Arrowheads indicate phase dense granules that fluoresce upon LJVexcitation.
stimulated growth. EC and BHK-21 cells were grown in the presence or absence of chloroquine for 3 days (Table I). bFGF and TR-bFGF stimuiated EC growth was inhibited 6-20X in the presence of 1 ,UM chloroquine and 647% with 5 ,UM chloroquine. Higher concentrations of chIoroquine were toxic to cells during the 72 hr incubation, while lower concentrations did not inhibit either bFGF. TR-bFGF stimulated growth or growth in serum alone. BHK-2 1 cells were unresponsive to chloroquine treatment. The extent of derivatization of FGF by TR affected its growth promoting activity. Diminished biological activity was observed with underconjugated (i.e. 596/280 nm = 0.20-049) or overconjugated molecules (i.e. 596/280 nm = 0.8 or greater) (Table II). Optimal biological activity was attained when the absorbance ratio was 0.50-0.79. We chose the
derivatized FGF with optimal growth promoting activity to perform our localization studies. TR-bFGF Localization in EC We localized TR-bFGF in living cells to map its intracellular fate. BAEC were incubated with TR-bFGF at 4°C and monitored at 37°C for approximately 1 hr [Figs 2(A) and (B)]. Using fluorescence microscopy, TR-bFGF was localized in the cytoplasm and, by phase contrast optics. we determined the signal to be contained within a subset of perinuclear, phase dense granules (Figs 2 and 3). When cultures were monitored at 10°C eliminating the 37°C warm-up and thereby preventing internalization, we were unable to detect TR-bFGF at the cel1 surface. When TR-bFGF treated EC were monitored for up to 24 hr [Fig. 2(C)
BFGF
IN RETINAL
ENDOTHELIUM
667
co-localized within EC lysosomes (Fig. 3). The lysosomes appeared to accumulate around the nucleus of EC and were absent from the leading edge of the cell shown (Fig. 3). Whereas most of the phase dense granules stained with the lysosomal marker, TR-bFGF was detected only in a subset of these lysosomes. Localization of TR-bFGF in BAEC incubated with 50 ,LM chloroquine for 2 hr resulted in TR-bFGF internalization, but the fluorescence remained diffuse and did not concentrate in a perinuclear pattern (not shown). 4. Discussion
FIG. 3. Colocalization of TR-bFGF and EC lysosomes.TRbFGF vital staining (A, arrows) corresponds with dark cytoplasmic granules observed under phase contrast (B, arrows) and lgp 120 (a golgi-specific marker) immunolocalization detectedin fixed EC.An etched coverslip (seen as a line through B) was used to mark the location of the cell following vital staining.
and (D)], the number of phase dense granules decreased with time [Fig. 2(D)]. However, the perinuclear pattern of TR-bFGF fluorescence persisted throughout this time course and fluorescence was not detected in the nucleus. The cytoplasmic fluorescence was abolished by denaturing or boiling TR-bFGFor by competition of TR-bFGF with excess unlabeled bFGF. ECincubated with 40 ng ml-’ TR-bFGFin the presence of loo-fold molar excessof unlabeled bFGF, led to an 80 % reduction in the number of TR-bFGFpositive cells (not shown). Experiments carried out using MVEC gave similar results. To ascertain the cytoplasmic
compartment
con-
taining TR-bFGF, we co-localized TR-bFGF fluorescence with an antibody against a lysosomal membrane glycoprotein, anti-lgp 120, and found TR-bFGF
A fluorescent derivative of bFGF (TR-bFGF) was prepared, which retained its ability to bind HS and stimulate vascular EC growth in vitro. Treatment of living EC with TR-bFGF results in its internalization and cytoplasmic localization within EC lysosomes. Furthermore, addition of the lysosomotropic base, chloroquine. disrupted the lysosomal targeting of TRbFGF and inhibited TR-bFGF stimulated EC growth. TR-bFGFcould not be detected within the EC nucleus up to 24 hr after its addition. The initial protocol for generating TR-bFGFdid not employ FGF immobilization onto HS. bFGF labeling with TR in solution resulted in a fluorescent molecule with a high labeling index but virtually all growth promoting activity was destroyed. Binding bFGFto HS prior to derivatization, on the other hand, led to a labeled molecule that maintains mitogenic activity. Of interest is the observation that ‘under ’ conjugated and ‘over’ conjugated TR-bFGF were not as effective in stimulating cell growth when compared to those bFGF molecules with an intermediate number of TR molecules covalently attached. It seemsplausible that the degree of derivatization influences the bFGF’s tertiary structure, which may alter mitogenic activity. This is comparable to conjugation of antibodies with fluorescent dyes since underconjugated IgGs (lessthan 2 per IgG) cannot be seen, overconjugated IgGs (greater than 5 per IgG) bind non-specifically and those possessing2-5 dye molecules per IgG stain cells specifically (Cebra and Goldstein, 196 5 ). The heparinbinding domain and high affinity cell surface receptorbinding domain of bFGF are reported to be distinct (Kurokawa, Doctrow and Klagsbrun, 1989). However, the bFGF-HS interactions exploited in these studies presumably restricted the number of sites available for covalent modification and thereby protected domains responsible for receptor occupancy and subsequent growth stimulation. TR-bFGF maintained biochemical and biological properties similar to unlabeled bFGF. TR-bFGF, like unlabeled bFGF. was eluted from HS with 1.5 M NaCl. Staphylococcus aureus V-8 protease digestion of bFGF and TR-bFGF [Fig. l(B)] resulted in cleavage products with identical relative mobilities. Proliferation studies demonstrated TR-bFGFstimulated EC growth 1.4-1.9
A. M. HEALY
668
times above serum controls. This growth stimulation was not due to contaminating unlabeled bFGF [Fig. 1(A)]. Oxidative dimers of TR-bFGF detected upon electrophoresis, which could not be removed by interdisulfide reduction, indicate intermolecular actions occured during the conjugation reaction. TRbFGF dimers may be biologically inactive thus contributing to the decreased growth promoting activity observed in labeled versus unlabeled bFGF. Following receptor-mediated endocytosis, bFGF is reportedly degraded within lysosomes (Bikfalvi et al.. 1989 ; Moscatelli, 1988). In our studies, perturbation of the lysosomal compartment with 11~ chloroquine marginally decreased bFGF and TR-bFGF stimulated EC growth. The alkalinization of the lysosomal compartment may interrupt formation of some intermediary in the stimulatory pathway, or perhaps FGF-receptor turnover may be affected. Interestingly, BAEC were more readily stimulated by exogenous bFGF and TR-bFGF and less affected by chloroquine treatment than MVEC. This may be explained by differences in bFGF cell surface receptor numbers and/or endogenous bFGF levels. Tsuboi. Sato and R&in (1990), studying MVEC subclones, reported that an inverse relationship exists between bFGF receptor number and endogenous bFGF levels. If MVEC synthesize relatively more endogenous bFGF than BAEC and express fewer bFGF receptors, then MVEC might be less responsive to exogenous bFGF stimulation and more susceptible to chloroquine disrupted receptor function. In further support of this correlation, BHK-2 1 cells reportedly have ten-times more bFGF receptors than EC (Neufeld et al., 1988), do not synthesize endogenous bFGF (Healy and Herman, unpubl. obs.) and had a minimal response to chloroquine treatment (Table I). Our findings indicate TR-bFGF is internalized, presumably via receptor-mediated endocytosis, and then is targeted to lysosomes. Internalization occured within minutes after warming cells to 37°C and the TR-bFGF signal was detected in a subset of phase dense granules that were confirmed to be lysosomes using antibody anti-lgp-120 co-localization. Interestingly, TR-bFGF fluorescence could not be detected at the EC surface perhaps because receptor copy number is low. TR-bFGF was also not readily detected in the extracellular matrix, which has previously been reported to be a site for bFGF binding (Rogelj et al.. 1989; Moscatelli. 1988). TR-bFGF localization was most readily detected after the ligand was sequestered or, in effect, concentrated in putative endosomes and lysosomes. TR-bFGF fluorescence was not detected within the EC nucleus. This result differs from the recent immunocytochemical studies by Baldin et al. (1990) in which anti-bFGF antibodies were used to localize exogenously added bFGF within the EC nucleolus. Two major differences between these two studies is our use of live cells, thus eliminating the possibility of fixation
AND
I. M. HERMAN
artifacts, and our use of a directly labeled FGF. Recent anti-bFGF localization studies indicate that bFGF nuclear fluorescence is absent from methanol-fixed, but can be seen in formaldehyde-fixed and detergenttreated EC (Healy and Herman, 1992). However, the likelihood remains that low levels of TR-bFGF or a non-fluorescent degradation product of TR-bFGF is present in the nucleus, but below our limits of detection. Using P5-bFGF and subcellular fractionation, Baldin et al. (1990) demonstrated that the full length form of bFGF accumulated in the nucleus. It is also possible that the intact bFGF non-specifically associates with the polyanionic nuclear compartment during cell fractionation. Recent evidence indicates that endogenous bFGF may be isolated biochemically from cell membranes, the cytoplasm and nuclei of the human hepatoma cell line SK Hep-1 (Brigstock, Sasse and Klagsbrun, 199 1). In normal or transformed cells, it remains unclear how bFGF might enter the nucleus. The consecutive basic amino acid sequence that constitutes a nuclear pore recognition sequence or a zinc finger motif have not been identified for bFGF. however, Imamura and co-workers ( 1990) identified a putative nuclear translocation sequence in aFGF and generated deletion mutants of this region (amino acid residues 21-27). Results of studies by Imamura et al. indicate the FGF truncation mutants were able to induce tyrosine phosphorylation, but not DNA replication ; however, cytoplasmic or nuclear localization was never ascertained. Current studies offer an additional explanation for the differences observed in FGF localization. bFGF and the FGF-related growth factor, int-2, both contain at least two in-frame translation start codons. One, an AUG codon, may be responsible for extracellular and cytoplasmic targeting and another, CUG-initiated form, is found in the nucleus (Acland et al., 1990: Florkiewicz, Baird and Gonzalez, 1991 ; Brigstock. Sasse and Klagsbrun, 1991). Collectively, these findings suggest exogenous and endogenous bFGF may be processed differentially and that two distinct mechanisms may exist for bFGF cell stimulation. The evidence reported here does not support the hypothesis of bFGF nuclear translocation. In the classical sense of polypeptide growth factor stimulated cell proliferation, bFGF internalization into the endothelial cytoplasm may be secondary to its receptor interaction, which is required for stimulating cell growth via second messengers. Acknowledgements The authors would like to thank Dr J. Folkman and the Takeda Chem. Ind.. Osaka. Japan, for the generous supply of growth factor. We are also grateful to Dr D. A. Albertini for the use of his imaging station and Dr I. Mellman for anti-lgp 120 antibody. This work was supported, in part. by a Sigma Xi Grant-In-Aid of Research award to A.M.H. and National Institutes of Health grants H.L. 35570 and 34739 to I.M.H.
BFGF
IN RETINAL
ENDOTHELIUM
References Acland. P., Dixon, M.. Peters, G. and Dickson, C. (1990). Subcellular fate of the Int-2 oncoprotein is determined by choice of initiation codon. Nature 343. 662-S. Ausprunk. D. H. and Folkman, J. (1977). Migration and proliferation of endothelia1 cells in preformed and newly formed blood vessels during tumor angiogenesis. Microvasc. Res. 14, 53-65. Baldin. V., Roman, A.-M., Bose-Bierne, I., Amalric. F. and Rouche, G. ( 19 90). Translocation of bFGF to the nucleus is C, phase cell cycle specific in bovine aortic endothelial cells. EMBO 1. Y(5). 151 l-17. Bashkin, P.. Doctrow. S.. Klagsbrun. M., Svahn. C. M.. Folkman. J. and Vlodavsky, I. (1989). Basic fibroblast growth factor binds to subendothelial extracellular matrix and is released by heparanases and heparin-like molecules. Biochemistry 28. 173-43. 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 Rrs. 181. 75-84. Bouche. G.. Gas, N., Prats, H., Baldin, V., Tauber, J-P., Tiessier. J. and Amalric. F. (1987). Basic fibroblast growth factor enters the nucleolus and stimulates the transcription of ribosomal genes in ABAE cells undergoing G,,-G, transition. Prof. Natl. Acad. Sci. USA 84. 6770-4. Brigstock, D. R., Sasse. J. and Klagsbrun. M. (199 1). Subcellular distribution of basic fibroblast growth factor in hepatoma cells. Growth Factors 4, 189-96. Burgess, W. H. and Maciag. T. (1989). The heparin-binding (fibroblast) growth factor family of proteins. Ann. Rev. Hiochem. 58, 575-606. Cebra. J. J. and Goldstein, G. (1965). Chromatographic purification of tetramethylrhodamine-immune globulin conjugates and their use in the cellular localization of rabbit gamma globulin peptide chains. J. Immunol. 95. 230-45. Cleveland. I). W.. Fisher, S. T., Kirschner, M. W. and Laemmli. U. K. (1977). Peptide mapping by limited proteolysis in sodium dodecyl sulfate and analysis by gel electrophoresis. J. Biol. Chem. 252(3), 1102-6. D’Amore, P. A. and Thompson, R. W. (1987). Mechanisms of angiogenesis. Ann. Rev. Physiol. 49, 453-64. DeNofrio. D.. Hoock, T. and Herman, I. M. (1989). Functional sorting of actin isoforms in microvascular pericytes. 1. Cell. Rio!. 109, 19 l-202. Florkiewicz, R. Z.. Baird, A. and Gonzalez, A.-M. (1991). Multiple forms of bFGF: differential nuclear and cell surface localization. Growth Factors 4, 265-75. Gillespie, I,. I... Paterno. G. D. and Slack, J. M. W. (1989 ). Analysis of competence: receptors for fibroblast growth factor in early Xenopus embryos. Developnzent 106. 20 3-8. Healy. A. M. and Herman, I. M. (1992). Density dependent accumulation of basic fibroblast growth factor in the subendothelial matrix. Eur. j. Cell Biol. (in press). Herman, I. M. and Castellot. J. J. (1987). Regulation of smooth muscle growth by endothelial-synthesized matrices. Arteriosclrrosis 7. 463-9.
669
Herman, I. M. and D’Amore, P. A. (19841. Capillary endothelial chemotaxis : loss of stress fibers in response to retina-derived growth factors. I. Muscle Res. Cell Motil. 5, 679-709. Imamura, T.. Engleka, K., Zhan, X.. Jackson, A., Maier. J. A. and Maciag, T. (1990). Recovery of mitogenic activity of a growth factor mutant with a nuclear translocation sequence. Science 249, 1567-70. Jester, J. V., Rodrigues. M. M. and Herman, 1. M. (19 87). Characterization of avascular cornea1 wound healing fibroblasts: new insights into the myofibroblast. Am. 1. Pathol. 127. 140-8. Klagsbrun, M. and Edelman, E. R. (1989). Biological and biochemical properties of fibroblast growth factors. (Review). Arteriosclerosis 9, 269-78. Kurokawa, M.. Doctrow. S. R. and Klagsbrun, M. (1989). Neutralizing antibodies inhibit the binding of basic fibroblast growth factor to its receptor but not to heparin. J. Biol. Chem. 264( 13), 7686-9 1. Lee, P. L., Johnson, D. E., Cousens. L. S.. Fried, V. A. and Williams, L. T. ( 19 89 ). Purification and complementary DNA cloning of a receptor for basic libroblast growth factor. Science 245, 57-60. Laemmli, U. K. (1970). Cleavage of structural proteins during the assembly of the head of Bacteriophage T4. Nature 227, 680-5. Moscatelli, D. (19 88). Metabolism of receptor-bound and matrix-bound basic fibroblast growth factor by bovine capillary endothelial cells. I. Cell. Bio. 107, i53-9. Neufeld. G. and Gospodarowicz. D. (19 8 5). The identification and partial characterization of the fibroblast growth factor receptor of baby hamster kidney cells. J. Cell. Bio. 260, 13860-8. Neufeld, G.. Mitchell, R., Ponte, P. and Gospodarowicz, D. (I 988). Expression of human basic fibroblast growth factor cDNA in baby hamster kidney derived cells results in autonomous cell growth. 1. 011. Riol. 106, 1385-94. Kogelj. S., Klagsbrun, M.. Atzmon, R.. Kurokawa, M., Haimovitz. A., Fuks, Z. and Vlodavsky, I. (1989). Basic fibroblast growth factor is an extracellular matrix component required for supporting the proliferation of vascular endothelial cells and the differentiation of PC12 cells. J. Cell. BioI. 109, 823-9. Saksela, 0.. Moscatelii, D., Sommer, A. and R&in. D. B. (19 90). Endothelial cell-derived heparan sulfate binds fibroblast growth factor and protects it from proteolytic degradation. 1. Cell. Biol. 107. 743-5 I. Tsuboi. R.. Sato, Y. and R&in. D. B. (1990). Correlation of cell migration, cell invasion, receptor number, proteinase production, and basic fibroblast growth factor levels in endothelial cells. 1. Cell. Bio. 110, 5 1l-l 7. Vlodavsky, I., Folkman. J., Sullivan. R.. Fridman. R., IshaiMichaeli. R., Sasse, J. and Klagsbrun, M. (1987). Endothelial cell-derived basic fibroblast growth factor; synthesis and deposition into subendothelial extracellular matrix. Proc. Natl. Acad. Sci. USA 84, 2292-6. Yayon. A., Klagsbrun, M., Esko, J. D.. Leder, P. and Ornitz, D. M. (1991). Cell surface, heparin-like molecules are required for binding of basic fibroblast growth factor to its high affinity receptor. Cell 64. 84 1-X.
Preparation Localization
of Fluorescent Basic Fibroblast in Living Retinal Microvascular AILEEN
M. HEALY
AND
Growth Factor: Endothelial Cells
IRA M. HERMAN”
Program in Cell, Molecular and Developmental Biology, Tuft,s University Health Science Schools, 736 Harrison Avenue, Boston, MA 0217 I, U.S.A. (Received
Houston
2 December
1991 and accepted in revised form 7 I February
1992)
A biologically active fluorescent derivative of recombinant human basic fibroblast growth factor (bFGF) was prepared by immobilization on heparin-Sepharose 4B (HS) and derivatization with the fluorophore, Texas Red (TR). TR-bFGF was separated from free dye and carrier protein by elution from HS using 1.5 M NaCI. TR-bFGFcontained an average of two dye molecules bound per bFGF,retained its mitogenic activity and was visible using a fluorescence microscope equipped with silicon intensified target camera (SIT). TR-bFGF stimulated the growth of bovine aortic endothelial cells (BAEC), microvessel endothelial cells (MVEC) and BHK-2 1 cells grown in culture. BAEC,MVEC and BHK-21 cells treated with 20 ng ml-l (1 nM) TR-bFGFfor 72 hr were stimulated over serum controls by 8 7, 26 and 6 %, respectively. TR-bFGF stimulated EC growth was inhibited in a dose-dependentfashion when cells were coincubated with ,UM chloroquine. When EC were treated with TR-bFGF at 4°C and then monitored at 37% bright, focal, cytoplasmic spots were observed, which accumulated as punctate, perinuclear tluorescence. EC internalization of TR-bFGFwas inhibited 80% by the addition of loo-fold molar excessunlabeled bFGF or by maintaining cultures at 4°C. TR-bFGFcolocalized with an EC lysosomal marker, but TR-bFGFwas not detected in the nucleus. Results of these localization studies suggest that TR-bFGF stimulates EC proliferation without entering the nucleus. Key words: growth factors ; lysosomes: signal transduction.
1. Introduction
The controlled growth, development and remodeling of the adult vasculature is a complex processregulated, in part, by vascular cell interactions with soluble and matrix-bound growth factors. Vascular EC synthesize and store the heparin-binding growth factor, bFGF, within the subendothelial matrix (Vladovsky et al., 1987). bFGF mediates EC migration and proliferation in vitro and angiogenesis in vivo (for review see Burgess and Maciag, 1989 ; D’Amore and Thompson, 198 7). During angiogenesis, EC initially migrate through a hyaluronate-rich matrix and subsequently form cell-cell contacts within a matrix enriched in sulfated glycosaminoglycans (Ausprunk and Folkman, 19 77). Presumably, bFGFis stored within the vascular intima
in association
with
the glycosaminoglycan,
heparan sulfate proteoglycan (Moscatelli, 1988 ; Bashkin et al., 1989). Results of in vitro studies reveal that bFGF is released from matrix storage sites by proteolysis (Sakselaet al., 1990). Following its release from the subendothelial matrix in vivo, soluble bFGF may stimulate EC migration and proliferation, two key events during angiogenesis. Soluble bFGF may act locally to stimulate EC growth via high-affinity cellsurface FGF receptors. Low-affinity heparin binding sites may facilitate FGF receptor occupancy (Yayon et al., 1991). The bFGF receptors have been isolated and the genes encoding these receptors, cloned and
2. Materials and Methods Cell Culture
BAEC and retinal MVEC were isolated and characterized as previously described (Herman and
* For correspondence.
0014-4835/92/110663+07
sequenced (Neufeld and Gospodarowicz, 198 5 ; Lee et al., 19 8 9). FGF-receptor occupancy triggers receptormediated endocytosis followed by down regulation of cell-surface receptors (Moscatelli, 1988). The intracellular events that ensue, resulting in EC migration and proliferation, are undefined. The intracellular fate of endocytosed bFGF is unclear. Studies using radiolabeled bFGF suggest the FGF-receptor complex is endocytosed and the ligand degraded within lysosomes (Moscatelli, 1988 : Bikfalvi et al., 1989). However, the results of studies in which EC were treated with exogenous bFGF then flxed and prepared for immunofluorescence microscopy, suggest FGF is found in the nucleus (Baldin et al., 1990; Bouche et al., 1987). To determine the fate of bFGFin EC and to establish whether exogenously added bFGF transits to the nuclear compartment of living EC, we labeled bFGFwith the fluorophore, TR, and monitored its internalization in living ECin culture. In this report we describe the specific binding, internalization and Iysosomal targeting of a biologically active TR-bFGF. Our evidence suggests bFGF, which is exogenously added to target cell populations, stimulates cell proliferation without entering the nucleus.
$08.00/O
0 1992 Academic Press Limited
664
Castellot, 198 7 ; Herman and D’Amore, 1984). Cells were routinely cultured in Dulbecco’s modified Eagle’s medium (DMEM; Gibco Laboratories, Grand Island, NY) containing 10 % bovine calf serum (CS; Hyclone, Logan, UT), with penicillin, streptomycin and amphotericin (Gibco Laboratories, Grand Island, NY) and used between passagesfive and ten. BHK-2 1 cells (American Type Culture Collection, Rockville, MD) were cultured in DMEM/Ham’s F12 (Irvine Scientific, Santa Ana, CA) supplemented with 10 % CS and antibiotics. Proli,ferationAssays EC were plated in DMEM containing 10% CS at an initial density of 15 x lo4 cells per 2 cm2 well (Costar, Cambridge, MA). BHK-21 cells were treated similarly, except that cells were plated in DMEM/Ham’s F12 containing 10% CS. Twenty four hours after plating, culture media were removed and cells were washed twice with phosphate-buffered saline (PBS). To assay cell proliferation, test media containing 20 and 40 ng ml-’ of recombinant human bFGF (Takeda Chemical Ind., Osaka,Japan), with or without 0.5, 1.0 and 5.0 ,u~ chloroquine (Sigma Chemical Co., St. Louis, MO) in DMEM with 2 % CS,were added to each well. At the indicated days, cells from triplicate wells were trypsinized and electronically counted (Coulter Model ZF, Hialeigh, FL).
A.M.
HEALY
AND
I. M. HERMAN
BiochemicalAnalysis of TR-bFGF SDS-PAGE.TR-bFGF was electrophoresed on 15 % polyacrylamide gels (Laemmli, 1970) or 5-l 5 % polyacrylamide gradient gels under reducing conditions and detected by silver stain (Jesteret al., 1987). FluorescenceLocalization of TR-bFGF in Acrylamide Gels. TR-bFGF was electrophoresed in a 15 % acrylamide gel and viewed by UV irradiation using a longwave UV transilluminator (peak wavelength = 302 nm, UVP Inc., San Gabriel, CA) UV transilluminated gels containing TR-bFGF were photographed with Poloroid Instant film 667 (Poloroid Corp., Cambridge. MA). Staphylococcusaureus V-8 ProteaseDigestion oj TRbFGF and bFGF. TR-bFGF and bFGF were enzymatically digested with S. aureus V-8 protease (Sigma Chem. Co.) according to the procedure of Cleveland et al. (19 7 7). Briefly, TR-bFGFand bFGFwere dissolved in 0.5 y0 SDS, 0.125 M Tris-Cl pH 6.8. 10% glycerol, 0.001% bromphenol blue and containing O-2 Units S. aureus V-8 protease. bFGF and TR-bFGF were digested at 37’C for 1 hr using a 1: 1 weight ratio of growth factor to enzyme. The enzyme was inactivated by the addition of 2-mercaptoethanol and SDS to a final concentration of 10 and 2%. respectively. Samples were alkylated with 80 mM iodoacetamide (Sigma Chem. Co.) at 37°C for 30 min and electrophoresed on a 5-l 5% gradient gel. Peptide cleavage products were detected by silver stain or immunoblot (DeNofrio, Hoock and Herman, 1989).
bFGF Conjugation to Texas Red Texas Red-bFGFconjugation was performed while bFGF was adsorbed to heparin Sepharose 4B (HS; Pharmacia, Uppsala, Sweden). Typically, a 75 ~1 packed volume of HS was equilibrated in 250 ~1 of binding buffer (0.05 M NaCl, 0.0 1 M Tris-Cl, pH 7.5) containing 2.2 x 1O-5M ovalbumin (Sigma Chemical Co,, St Louis, MO), at room temperature for 30 min. HS was washed with 100 column volumes of binding buffer. 2.0 x 10e6M bFGF(300 ~1) was incubated with equilibrated HS in binding buffer at 4°C for 90 min. HS-bound bFGF was then re-equilibrated in 200 ~1 of conjugation buffer (0.01 M Na,CO,, 0.05 M NaCl, pH 8.8) containing 2.2 x 10e4M ovalbumin (100 ~1). 3.3 x 10M3M TR (Molecular Probes, Eugene, OR) dissolved in 50 ,ul conjugation buffer was added to HSbound bFGF. The final reaction volume equaled 3 50 pl. The reaction proceededon ice for 15 min with frequent agitation. To remove unbound dye and separate ovalbumin from TR-bFGF, HS was washed exhaustively with a 0.05, 0.5, 0.75, 1.0 and 1.25 M NaCl, pH 7.5. step gradient. TR-bFGF was eluted from HS with 1.5 M NaCl, pH 7.5. The labeling index or absorbance ratio of dye to protein was determined spectrophotometrically at 596 and 280 nm, respectively (Beckman Model DU 65, Fullerton, CA).
LocalizationExperiments EC and BHK-21 cells were plated in DMEM containing 10% CS onto gelatin-coated glass microscope coverslips and allowed to attach for 18 hr. The following morning, cells were washed twice with PBS and preincubated in DMEM containing 2 y0CSat 3 7°C for 2 hr. Cells were then cooled to 10°C and incubated with 20 or 40 ng ml-’ TR-bFGFin DMEM + 2 y0CSfor 2 hr. Cells were then washed twice with PBSand TRbFGF-containing media were replaced with fresh DMEM containing 2 % CSand cells were mounted in a temperature-regulated chamber slide. TR-bFGFlocalization in living EC or BHK-21 cells was monitored using either an IM35 fluorescence microscope or a Photomicroscope II (Zeiss, West Germany), both equipped with a x 63 objective (N.A. 1.4) and a Venus Two Stage Intensifier Video Detector, Model DV2 (Venus Sci. Incorp., Farmingdale, NY). Images were recorded on a Time-Lapse Video Recorder, Model AG 6050 (Panasonic, Seacaucus, NJ). Data shown were collected with the Photomicroscope II. Colocalization of TR-bFGF and Lysosomes
EC were cultured as described for localization experiments, then washed briefly in PBS and fixed in
BFGF
IN RETINAL
ENDOTHELIUM
methanol at 4°C for 5 min. Following three, 5 min PBS washes, cells were sequentially incubated with rabbit and anti-lgp120, an immunoglobulin directed against a 120 000 Da lysosomal membrane glycoprotein (I. Mellman, Yale University, New Haven, CT) and rhodamine-labeled goat anti-rabbit immunoglobulin (Cooper Biomedical, Malvern, PA).
665
20 ng ml-’ an optimal concentration. A lysosomotropic base, which disrupts the endocytic pathway, was employed to assesswhether alkalinization of the lysosomal compartment perturbs bFGF and TR-bFGF (A)
(B)
3. Results Texas Red Labeled bFGF
We directly labeled bFGF using the fluorescent dye TR. The conjugation reaction was conducted with bFGF bound to HS, which produced TR-bFGF labeled with varying numbers of dye molecules. Electrophoresis and silver staining of recombinant bFGF [Fig. l(A), lane 11and TR-bFGFrevealed a labeled molecule with a relative mobility of 19,250 Da, which corresponds to a labeling index (i.e. 596/280 nm) of 0.5 [Fig. l(A), lane 21. This represents two TR molecules per bFGF. Electrophoresed TR-bFGF was also visualized following exposure to a 302 nm UV source [Fig. l(A), lane 31. TR-bFGF preparations did not contain contaminating free dye, ovalbumin or unlabeled bFGF [Fig. l(A)]. Peptide maps were generated to determine whether covalent modifications of bFGF with TR significantly altered its secondary structure. To this end, bFGF and TR-bFGFwere treated with S. aureusprotease V-8 and electrophoresed through gradient gels. Following silver staining, five low-molecular-weight polypeptide cleavage products were observed [Fig. l(B), lanes 2 and 4). All the bFGF and TR-bFGF fragments were identical, with relative mobilities ranging from 11000-5 000 Da. bFGF,TR-bFGFand the resulting V8 cleavage peptides were also immunoreactive with anti-bFGF IgG by Western blot analysis (not shown).
TR-bFGF Promotes
EC Growth
To determine if bFGFmaintained growth promoting activity following labeling with TR, target cell populations with known sensitivity to bFGF were assayedfor growth promotion. BAEC,MVEC and BHK21 cells were grown in culture in the presence or absence of either bFGF or TR-bFGF as described. Proliferation assay results are summarized in Table I. As indicated, after 72 hr in culture bFGF stimulated MVEC growth 2.3 times as compared with serum controls, and TR-bFGF stimulated MVEC growth 1.4 times above serum controls. BAEC grown in the presence of bFGF were stimulated 3.6 times above serum controls, while TR-bFGF stimulated BAEC growth 1.9 times. BHK-21 cells are not markedly stimulated by TK-bFGF. By 96 hr, cells under all growth conditions attained confluence. bFGF and TRbFGFstimulation of ECgrowth is dose-dependent,with
I2
I
3
2
3
4
Electrophoreticanalysisof bFGF and TRQFGF. (A) Silver stain of a 15% acrylamidegel containing bFGF(lane FIG. 1.
1) and TR-bFGF (lane 2). Fluorescent TR-bFGF is detected in unfixed gels following UV irradiation (lane 3). (B) Silver stained, 5-15 % gradient gel containing bFGF (lane 1). bFGF plus S. aureus V-8 protease (lane 2). TR-bFGF (lane 3) and TR-bFGF plus S. aureus V-8 protease (lane 4). Molecular weight markers are indicated in kDa.
TABLE
I
l’R-bFGF stimulates cell prohjeration TR-bFGF/ Serum
bFGF/ Serum Chloroquine
0
~/AM
.~-.
MVEC BAEC BHK
2.3t 3.6 1.4
0
5j.N
1 ,uM
5 ,UM
~~
1.9*** 3.3*** 1.3**
1.3*** 3.5*** 1.2*
1.4 1.3** 1.0*** 1.9 2.0** 2.3** 1.0 1.1* 1.0*
t The numbers represent the fold increase in cell number from triplicate wells cultured for 3 days then counted. bFGF or TR-bFGF stimulated cells (20 ng ml-‘)/control cells (serum alone). SEM < 13 % for all groups. P values represent analysis of chloroquine-treated vs. control cells (no chloroquine treatment) in FGF and TR-bFGF stimulated cells: *P > 0.05. ** P < 0.05 and *** P < 0901.
TABLE
11
Texasred-bFGFlabeling index Labeling index (596/280 nm) 0.20-0.49 0.50-0.79 0.80-2.50
“/o Stimulation* 27 (n =: 31 61 (n = 4) 21 (n =: 3)
* xstimulation = lOO(i- x)/(y -x). x = controlcells. ECcultured for three days in media containing 2% cs. y = cells treated with 20 ng ml-’ bFCF. and 2 = cells treated with 20 ng ml-’ TR-bFGF. EEK55
666
A.M.
HEALY
AND
I. M. HERMAN
FIG. 2. TR-bFGFLocalization in Living EC. Phase contrast (A and C) and fluorescent (B and D) images of BAEC in culture. EC treated with 20 ng ml-’ TR-bFGF for 2 hr at 4-10°C prior to warming to 37°C for approximately 1 hr before viewing (A and B). EC cultured for 24 hr after the addition of 20 ng ml-’ TR-bFGF(C and D). Arrowheads indicate phase dense granules that fluoresce upon LJVexcitation.
stimulated growth. EC and BHK-21 cells were grown in the presence or absence of chloroquine for 3 days (Table I). bFGF and TR-bFGF stimuiated EC growth was inhibited 6-20X in the presence of 1 ,UM chloroquine and 647% with 5 ,UM chloroquine. Higher concentrations of chIoroquine were toxic to cells during the 72 hr incubation, while lower concentrations did not inhibit either bFGF. TR-bFGF stimulated growth or growth in serum alone. BHK-2 1 cells were unresponsive to chloroquine treatment. The extent of derivatization of FGF by TR affected its growth promoting activity. Diminished biological activity was observed with underconjugated (i.e. 596/280 nm = 0.20-049) or overconjugated molecules (i.e. 596/280 nm = 0.8 or greater) (Table II). Optimal biological activity was attained when the absorbance ratio was 0.50-0.79. We chose the
derivatized FGF with optimal growth promoting activity to perform our localization studies. TR-bFGF Localization in EC We localized TR-bFGF in living cells to map its intracellular fate. BAEC were incubated with TR-bFGF at 4°C and monitored at 37°C for approximately 1 hr [Figs 2(A) and (B)]. Using fluorescence microscopy, TR-bFGF was localized in the cytoplasm and, by phase contrast optics. we determined the signal to be contained within a subset of perinuclear, phase dense granules (Figs 2 and 3). When cultures were monitored at 10°C eliminating the 37°C warm-up and thereby preventing internalization, we were unable to detect TR-bFGF at the cel1 surface. When TR-bFGF treated EC were monitored for up to 24 hr [Fig. 2(C)
BFGF
IN RETINAL
ENDOTHELIUM
667
co-localized within EC lysosomes (Fig. 3). The lysosomes appeared to accumulate around the nucleus of EC and were absent from the leading edge of the cell shown (Fig. 3). Whereas most of the phase dense granules stained with the lysosomal marker, TR-bFGF was detected only in a subset of these lysosomes. Localization of TR-bFGF in BAEC incubated with 50 ,LM chloroquine for 2 hr resulted in TR-bFGF internalization, but the fluorescence remained diffuse and did not concentrate in a perinuclear pattern (not shown). 4. Discussion
FIG. 3. Colocalization of TR-bFGF and EC lysosomes.TRbFGF vital staining (A, arrows) corresponds with dark cytoplasmic granules observed under phase contrast (B, arrows) and lgp 120 (a golgi-specific marker) immunolocalization detectedin fixed EC.An etched coverslip (seen as a line through B) was used to mark the location of the cell following vital staining.
and (D)], the number of phase dense granules decreased with time [Fig. 2(D)]. However, the perinuclear pattern of TR-bFGF fluorescence persisted throughout this time course and fluorescence was not detected in the nucleus. The cytoplasmic fluorescence was abolished by denaturing or boiling TR-bFGFor by competition of TR-bFGF with excess unlabeled bFGF. ECincubated with 40 ng ml-’ TR-bFGFin the presence of loo-fold molar excessof unlabeled bFGF, led to an 80 % reduction in the number of TR-bFGFpositive cells (not shown). Experiments carried out using MVEC gave similar results. To ascertain the cytoplasmic
compartment
con-
taining TR-bFGF, we co-localized TR-bFGF fluorescence with an antibody against a lysosomal membrane glycoprotein, anti-lgp 120, and found TR-bFGF
A fluorescent derivative of bFGF (TR-bFGF) was prepared, which retained its ability to bind HS and stimulate vascular EC growth in vitro. Treatment of living EC with TR-bFGF results in its internalization and cytoplasmic localization within EC lysosomes. Furthermore, addition of the lysosomotropic base, chloroquine. disrupted the lysosomal targeting of TRbFGF and inhibited TR-bFGF stimulated EC growth. TR-bFGFcould not be detected within the EC nucleus up to 24 hr after its addition. The initial protocol for generating TR-bFGFdid not employ FGF immobilization onto HS. bFGF labeling with TR in solution resulted in a fluorescent molecule with a high labeling index but virtually all growth promoting activity was destroyed. Binding bFGFto HS prior to derivatization, on the other hand, led to a labeled molecule that maintains mitogenic activity. Of interest is the observation that ‘under ’ conjugated and ‘over’ conjugated TR-bFGF were not as effective in stimulating cell growth when compared to those bFGF molecules with an intermediate number of TR molecules covalently attached. It seemsplausible that the degree of derivatization influences the bFGF’s tertiary structure, which may alter mitogenic activity. This is comparable to conjugation of antibodies with fluorescent dyes since underconjugated IgGs (lessthan 2 per IgG) cannot be seen, overconjugated IgGs (greater than 5 per IgG) bind non-specifically and those possessing2-5 dye molecules per IgG stain cells specifically (Cebra and Goldstein, 196 5 ). The heparinbinding domain and high affinity cell surface receptorbinding domain of bFGF are reported to be distinct (Kurokawa, Doctrow and Klagsbrun, 1989). However, the bFGF-HS interactions exploited in these studies presumably restricted the number of sites available for covalent modification and thereby protected domains responsible for receptor occupancy and subsequent growth stimulation. TR-bFGF maintained biochemical and biological properties similar to unlabeled bFGF. TR-bFGF, like unlabeled bFGF. was eluted from HS with 1.5 M NaCl. Staphylococcus aureus V-8 protease digestion of bFGF and TR-bFGF [Fig. l(B)] resulted in cleavage products with identical relative mobilities. Proliferation studies demonstrated TR-bFGFstimulated EC growth 1.4-1.9
A. M. HEALY
668
times above serum controls. This growth stimulation was not due to contaminating unlabeled bFGF [Fig. 1(A)]. Oxidative dimers of TR-bFGF detected upon electrophoresis, which could not be removed by interdisulfide reduction, indicate intermolecular actions occured during the conjugation reaction. TRbFGF dimers may be biologically inactive thus contributing to the decreased growth promoting activity observed in labeled versus unlabeled bFGF. Following receptor-mediated endocytosis, bFGF is reportedly degraded within lysosomes (Bikfalvi et al.. 1989 ; Moscatelli, 1988). In our studies, perturbation of the lysosomal compartment with 11~ chloroquine marginally decreased bFGF and TR-bFGF stimulated EC growth. The alkalinization of the lysosomal compartment may interrupt formation of some intermediary in the stimulatory pathway, or perhaps FGF-receptor turnover may be affected. Interestingly, BAEC were more readily stimulated by exogenous bFGF and TR-bFGF and less affected by chloroquine treatment than MVEC. This may be explained by differences in bFGF cell surface receptor numbers and/or endogenous bFGF levels. Tsuboi. Sato and R&in (1990), studying MVEC subclones, reported that an inverse relationship exists between bFGF receptor number and endogenous bFGF levels. If MVEC synthesize relatively more endogenous bFGF than BAEC and express fewer bFGF receptors, then MVEC might be less responsive to exogenous bFGF stimulation and more susceptible to chloroquine disrupted receptor function. In further support of this correlation, BHK-2 1 cells reportedly have ten-times more bFGF receptors than EC (Neufeld et al., 1988), do not synthesize endogenous bFGF (Healy and Herman, unpubl. obs.) and had a minimal response to chloroquine treatment (Table I). Our findings indicate TR-bFGF is internalized, presumably via receptor-mediated endocytosis, and then is targeted to lysosomes. Internalization occured within minutes after warming cells to 37°C and the TR-bFGF signal was detected in a subset of phase dense granules that were confirmed to be lysosomes using antibody anti-lgp-120 co-localization. Interestingly, TR-bFGF fluorescence could not be detected at the EC surface perhaps because receptor copy number is low. TR-bFGF was also not readily detected in the extracellular matrix, which has previously been reported to be a site for bFGF binding (Rogelj et al.. 1989; Moscatelli. 1988). TR-bFGF localization was most readily detected after the ligand was sequestered or, in effect, concentrated in putative endosomes and lysosomes. TR-bFGF fluorescence was not detected within the EC nucleus. This result differs from the recent immunocytochemical studies by Baldin et al. (1990) in which anti-bFGF antibodies were used to localize exogenously added bFGF within the EC nucleolus. Two major differences between these two studies is our use of live cells, thus eliminating the possibility of fixation
AND
I. M. HERMAN
artifacts, and our use of a directly labeled FGF. Recent anti-bFGF localization studies indicate that bFGF nuclear fluorescence is absent from methanol-fixed, but can be seen in formaldehyde-fixed and detergenttreated EC (Healy and Herman, 1992). However, the likelihood remains that low levels of TR-bFGF or a non-fluorescent degradation product of TR-bFGF is present in the nucleus, but below our limits of detection. Using P5-bFGF and subcellular fractionation, Baldin et al. (1990) demonstrated that the full length form of bFGF accumulated in the nucleus. It is also possible that the intact bFGF non-specifically associates with the polyanionic nuclear compartment during cell fractionation. Recent evidence indicates that endogenous bFGF may be isolated biochemically from cell membranes, the cytoplasm and nuclei of the human hepatoma cell line SK Hep-1 (Brigstock, Sasse and Klagsbrun, 199 1). In normal or transformed cells, it remains unclear how bFGF might enter the nucleus. The consecutive basic amino acid sequence that constitutes a nuclear pore recognition sequence or a zinc finger motif have not been identified for bFGF. however, Imamura and co-workers ( 1990) identified a putative nuclear translocation sequence in aFGF and generated deletion mutants of this region (amino acid residues 21-27). Results of studies by Imamura et al. indicate the FGF truncation mutants were able to induce tyrosine phosphorylation, but not DNA replication ; however, cytoplasmic or nuclear localization was never ascertained. Current studies offer an additional explanation for the differences observed in FGF localization. bFGF and the FGF-related growth factor, int-2, both contain at least two in-frame translation start codons. One, an AUG codon, may be responsible for extracellular and cytoplasmic targeting and another, CUG-initiated form, is found in the nucleus (Acland et al., 1990: Florkiewicz, Baird and Gonzalez, 1991 ; Brigstock. Sasse and Klagsbrun, 1991). Collectively, these findings suggest exogenous and endogenous bFGF may be processed differentially and that two distinct mechanisms may exist for bFGF cell stimulation. The evidence reported here does not support the hypothesis of bFGF nuclear translocation. In the classical sense of polypeptide growth factor stimulated cell proliferation, bFGF internalization into the endothelial cytoplasm may be secondary to its receptor interaction, which is required for stimulating cell growth via second messengers. Acknowledgements The authors would like to thank Dr J. Folkman and the Takeda Chem. Ind.. Osaka. Japan, for the generous supply of growth factor. We are also grateful to Dr D. A. Albertini for the use of his imaging station and Dr I. Mellman for anti-lgp 120 antibody. This work was supported, in part. by a Sigma Xi Grant-In-Aid of Research award to A.M.H. and National Institutes of Health grants H.L. 35570 and 34739 to I.M.H.
BFGF
IN RETINAL
ENDOTHELIUM
References Acland. P., Dixon, M.. Peters, G. and Dickson, C. (1990). Subcellular fate of the Int-2 oncoprotein is determined by choice of initiation codon. Nature 343. 662-S. Ausprunk. D. H. and Folkman, J. (1977). Migration and proliferation of endothelia1 cells in preformed and newly formed blood vessels during tumor angiogenesis. Microvasc. Res. 14, 53-65. Baldin. V., Roman, A.-M., Bose-Bierne, I., Amalric. F. and Rouche, G. ( 19 90). Translocation of bFGF to the nucleus is C, phase cell cycle specific in bovine aortic endothelial cells. EMBO 1. Y(5). 151 l-17. Bashkin, P.. Doctrow. S.. Klagsbrun. M., Svahn. C. M.. Folkman. J. and Vlodavsky, I. (1989). Basic fibroblast growth factor binds to subendothelial extracellular matrix and is released by heparanases and heparin-like molecules. Biochemistry 28. 173-43. 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 Rrs. 181. 75-84. Bouche. G.. Gas, N., Prats, H., Baldin, V., Tauber, J-P., Tiessier. J. and Amalric. F. (1987). Basic fibroblast growth factor enters the nucleolus and stimulates the transcription of ribosomal genes in ABAE cells undergoing G,,-G, transition. Prof. Natl. Acad. Sci. USA 84. 6770-4. Brigstock, D. R., Sasse. J. and Klagsbrun. M. (199 1). Subcellular distribution of basic fibroblast growth factor in hepatoma cells. Growth Factors 4, 189-96. Burgess, W. H. and Maciag. T. (1989). The heparin-binding (fibroblast) growth factor family of proteins. Ann. Rev. Hiochem. 58, 575-606. Cebra. J. J. and Goldstein, G. (1965). Chromatographic purification of tetramethylrhodamine-immune globulin conjugates and their use in the cellular localization of rabbit gamma globulin peptide chains. J. Immunol. 95. 230-45. Cleveland. I). W.. Fisher, S. T., Kirschner, M. W. and Laemmli. U. K. (1977). Peptide mapping by limited proteolysis in sodium dodecyl sulfate and analysis by gel electrophoresis. J. Biol. Chem. 252(3), 1102-6. D’Amore, P. A. and Thompson, R. W. (1987). Mechanisms of angiogenesis. Ann. Rev. Physiol. 49, 453-64. DeNofrio. D.. Hoock, T. and Herman, I. M. (1989). Functional sorting of actin isoforms in microvascular pericytes. 1. Cell. Rio!. 109, 19 l-202. Florkiewicz, R. Z.. Baird, A. and Gonzalez, A.-M. (1991). Multiple forms of bFGF: differential nuclear and cell surface localization. Growth Factors 4, 265-75. Gillespie, I,. I... Paterno. G. D. and Slack, J. M. W. (1989 ). Analysis of competence: receptors for fibroblast growth factor in early Xenopus embryos. Developnzent 106. 20 3-8. Healy. A. M. and Herman, I. M. (1992). Density dependent accumulation of basic fibroblast growth factor in the subendothelial matrix. Eur. j. Cell Biol. (in press). Herman, I. M. and Castellot. J. J. (1987). Regulation of smooth muscle growth by endothelial-synthesized matrices. Arteriosclrrosis 7. 463-9.
669
Herman, I. M. and D’Amore, P. A. (19841. Capillary endothelial chemotaxis : loss of stress fibers in response to retina-derived growth factors. I. Muscle Res. Cell Motil. 5, 679-709. Imamura, T.. Engleka, K., Zhan, X.. Jackson, A., Maier. J. A. and Maciag, T. (1990). Recovery of mitogenic activity of a growth factor mutant with a nuclear translocation sequence. Science 249, 1567-70. Jester, J. V., Rodrigues. M. M. and Herman, 1. M. (19 87). Characterization of avascular cornea1 wound healing fibroblasts: new insights into the myofibroblast. Am. 1. Pathol. 127. 140-8. Klagsbrun, M. and Edelman, E. R. (1989). Biological and biochemical properties of fibroblast growth factors. (Review). Arteriosclerosis 9, 269-78. Kurokawa, M.. Doctrow. S. R. and Klagsbrun, M. (1989). Neutralizing antibodies inhibit the binding of basic fibroblast growth factor to its receptor but not to heparin. J. Biol. Chem. 264( 13), 7686-9 1. Lee, P. L., Johnson, D. E., Cousens. L. S.. Fried, V. A. and Williams, L. T. ( 19 89 ). Purification and complementary DNA cloning of a receptor for basic libroblast growth factor. Science 245, 57-60. Laemmli, U. K. (1970). Cleavage of structural proteins during the assembly of the head of Bacteriophage T4. Nature 227, 680-5. Moscatelli, D. (19 88). Metabolism of receptor-bound and matrix-bound basic fibroblast growth factor by bovine capillary endothelial cells. I. Cell. Bio. 107, i53-9. Neufeld. G. and Gospodarowicz. D. (19 8 5). The identification and partial characterization of the fibroblast growth factor receptor of baby hamster kidney cells. J. Cell. Bio. 260, 13860-8. Neufeld, G.. Mitchell, R., Ponte, P. and Gospodarowicz, D. (I 988). Expression of human basic fibroblast growth factor cDNA in baby hamster kidney derived cells results in autonomous cell growth. 1. 011. Riol. 106, 1385-94. Kogelj. S., Klagsbrun, M.. Atzmon, R.. Kurokawa, M., Haimovitz. A., Fuks, Z. and Vlodavsky, I. (1989). Basic fibroblast growth factor is an extracellular matrix component required for supporting the proliferation of vascular endothelial cells and the differentiation of PC12 cells. J. Cell. BioI. 109, 823-9. Saksela, 0.. Moscatelii, D., Sommer, A. and R&in. D. B. (19 90). Endothelial cell-derived heparan sulfate binds fibroblast growth factor and protects it from proteolytic degradation. 1. Cell. Biol. 107. 743-5 I. Tsuboi. R.. Sato, Y. and R&in. D. B. (1990). Correlation of cell migration, cell invasion, receptor number, proteinase production, and basic fibroblast growth factor levels in endothelial cells. 1. Cell. Bio. 110, 5 1l-l 7. Vlodavsky, I., Folkman. J., Sullivan. R.. Fridman. R., IshaiMichaeli. R., Sasse, J. and Klagsbrun, M. (1987). Endothelial cell-derived basic fibroblast growth factor; synthesis and deposition into subendothelial extracellular matrix. Proc. Natl. Acad. Sci. USA 84, 2292-6. Yayon. A., Klagsbrun, M., Esko, J. D.. Leder, P. and Ornitz, D. M. (1991). Cell surface, heparin-like molecules are required for binding of basic fibroblast growth factor to its high affinity receptor. Cell 64. 84 1-X.
