Crorc1th Fnctors, 1992, Vol. 7, pp. 215-231 Repfints available directly from the publisher Photocopying permitted by license only
0 1992 Harwood Academic Publishers GinbH Printed in the United Kingdom
Cell Cycle-Dependent Localization of Irnmunoreactive Basic Fibroblast Growth Factor to Cytoplasm and Nucleus of Isolated Ovine Fetal Growth Plate Chondrocytes
Growth Factors Downloaded from informahealthcare.com by UB Heidelberg on 11/14/14 For personal use only.
(Receiwd August 19 1991, Accepted April 30 1992) Basic fibroblast growth factor (basic FGF) is a potent mitogen for chondrocytes iri vitro and is present in developing cartilage iri vivo. Studies of intracellular basic FGF localization in other cell types have revealed a transient nuclear presence. We have examined ovine fetal growth plate chondrocytes for the presence of intracellular basic FGF by immunocytochemistry. Chondrocytes were isolated from the proximal tibia1 growth plate of lamb fetuses between 75 and 80 days’ gestation using collagenase, and were cultured in monolayer before use between passages 3 and 6. In non-synchronized cell cultures 58?6% of cells (mean+s.d., r1=3) demonstrated cytoplasmic staining for immunoreactive basic FGF. Of these cells, 18+3% also exhibited strong nuclear staining. Chondrocytes were growth-restricted and restarted into the cell cycle with 2% (vol/vol) fetal calf serum. The timing of S phase was followed by nuclear labelling of nuclei with [‘HI thymidine followed by autoradiogaphy, or by the incorporation of [‘HI thymidine into trichloroacetic acid-precipitable DNA i n parallel cultures. A cytoplasmic presence of immunoreactive basic FGF did not appear, following immunocytochemistry, until the second half of G I with 97% of cells immunopositive 2 hr prior to S phase. Nuclear staining for basic FGF appeared 2 hr before peak nuclear labelling index, and 56% of cell nuclei were immunopositive. Following entry into S phase cytoplasmic and nuclear basic FGF immunostaining rapidly disappeared. When these experiments were repeated with or without the presence of anti-basic FGF antibody or heparin, the presence of the antibody significantly reduced peak [’HI thymidine incorporation into DNA during S phase while exposure to heparin increased this. However, the proportion of cells demonstrating cytoplasmic or nuclear staining for immunoreactive basic FGF, and the time of onset of staining, were unaltered. Incubation of cells with suramin blocked subsequent DNA synthesis and no intracellular basic FGF was visualized. Cellconditioned culture medium, extracellular matrix and cytoplasm from synchronized cultures of chondrocytes were taken at time points throughout the cell cycle and assessed for basic FGF content by radioimniunoassay. Basic FGF was detectable in each compartment and steadily rose throughout the second half of G, to reach maximum values around the S phase. The accumulation of basic FGF in medium, matrix and c to lasm was blocked by the presence of cycloheximide. When cells were labelled with [Y HI leucine P . and newly-synthesized basic FGF immunoprecipitated from cytoplasmic or nuclear fractions, peak values of peptide were seen in the cytoplasm 2 hr prior to its transient appearance in the nucleus which was immediately before entry into S phase.
Correspondence: Dr David J. Hill, Depts of Medicine and I’hysiology, Lawson Research Institute, St. Joseph’s Health Centre, 268 Grosvenor Street, London, Ontario, N6A 4V2, Canada. Tel: (519) 439-3271 x4716. Fax: (519) 438-3525.
21.5
HILL and LOGAN
216
The amount of immunoprecipitable, tritiated basic FGF recovered from cytoplasm and nucleus was unaffected by incubation with antibody against basic FGF. Separation of immunoprecipitated, tritiated basic FGF by gel electrophoresis revealed two major forms with molecular sizes 18 kDa and 23 kDa in both the cytoplasm and nucleus. Only the former was present in conditioned culture medium. The results suggest that chondrocytes synthesize and release basic FGF during the second half of G I in response to hormonal signals contained in fetal calf serum. This basic FGF contributes to the subsequent DNA synthetic rate of the cells. During late GI a brief translocation of basic FGF from cytoplasm to nucleus occurs, but this is not directly related to alterations in DNA synthesis.
Growth Factors Downloaded from informahealthcare.com by UB Heidelberg on 11/14/14 For personal use only.
KEYWORDS: fibroblast growth factor, fetus, chondrocyte, immunocytochemistry, nucleus
INTRODUCTION
MATERIALS AND METHODS
Basic fibroblast growth factor (basic FGF) is widely expressed in embryonic tissues and persists until late gestation in rat (Gonzalez et al., 1990; Hebert et al., 1990). Radioimmuno- and bioassayable basic FGF have been detected in the developing limb buds in the chick (Munaim, Klagsbrun and Toole, 1988; Seed, Olwin and Hauschka, 1988) while immunoreactive basic FGF was detected in growth plate chondrocytes of the rat fetus (Gonzalez et al., 1990). Since basic FGF is a potent mitogen for chondrocytes isolated from the epiphyseal growth plates of the rabbit, rat and chick (Makower, Wroblewski and Pawlowski, 1989; Crabb et al., 1990; Kato and Iwamato, 1990) it is possible that it acts as an autocrine or paracrine growth factor for skeletal tissue during early life. In support of this hypothesis we recently showed that isolated growth plate chondrocytes from the ovine fetus contained and released immunoreactive and bioactive basic FGF (Hill et al., 1992). Unlike other isolated cell types, almost half of the immunoreactive basic FGF released was recovered from conditioned culture medium in soluble form, the remainder being recovered from the extracellular matrix and cell membranes, and the cytoplasm. It has been demonstrated previously for other cell types which synthesize basic FGF that a cell cycle-dependent translocation of basic FGF occurs between cytoplasm and nucleus, although the relevance of this to its biological actions is unknown (Bladin et al., 1990; Renko et al., 1990). We have examined ovine fetal growth plate chondrocytes to determine whether such an intracellular translocation occurs in this cell type, and if so, what is its relationship to the autocrine/paracrine role of basic FGF as a mitogen.
Recombinant human basic FGF was kindly provided by Dr P. Barr, Chiron Corporation, Emeryville, CA; and a rabbit polyclonal antibody (Ab 773) raised against the 1-24 synthetic fragment of bovine basic FGF provided by Dr A. Baird, Whittier Institute for Diabetes and Medicine, La Jolla, CA. A purified IgG was obtained by precipitation of the antiserum with 30% (wt/vol) ammonium sulphate and passage over a protein A-Sepharose affinity column. The antibody demonstrated less than 1% cross-reactivity with acidic FGF and no crossreactivity with hst/ks or FGF-5. Cross-reactivity with FGF-6 is not known. However, FGF-6 shows low sequence homology with the fragment of basic FGF used as the antigen (Gonzalez et al., 1990). The same antibody was found to block the mitogenic actions of exogenous basic FGF on vascular endothelial cells in ziifro (Dr A. Baird, Whittier Institute, La Jolla, CA; personal communication). Biotinylated anti-rabbit IgG and biotin peroxidase complex used during immunocytochemistry were obtained from Vector Laboratories, Burlingame, CA. Plastic tissue culture flasks (75 cm2 area, NUNC), Hank’s buffered salts solution (HBSS), Dulbecco’s modified Eagle’s medium (DMEM), fetal calf serum (FCS) and trypsin-EDTA, were each purchased from Gibco Ltd., Burlington, Ont., Canada. A glucose-free formulation of DMEM was obtained from Imperial Laboratories Ltd., Salisbury, England; Dulbecco’s phosphate buffered saline (PBS), fatty acid-free bovine serum albumin, fraction V (BSA), cycloheximide, trasylol (aprotinin), phenylmethylsulfonyl fluoride, protein A-Sepharose and chloramine T from Sigma Chemical Co., St. Louis, MO; and %well
CELL CYCLE-DEPENDENT bFGF LOCALIZATION
Growth Factors Downloaded from informahealthcare.com by UB Heidelberg on 11/14/14 For personal use only.
Lab-tek chamber slides from Miles Laboratories Inc., Naperville, IL. Collagenase were obtained from Boehringer Ingelheim (Canada) Ltd., Dorval, Quebec, Canada. Hoeschst fluorochrome 33258 and 3-3'-diaminobenzidine were purchased from Aldrich Chemical Co., Milwaukee, WI; euthanyl from MCT Pharmaceuticals, Mississauga, Canada; and [3H] thymidine, ['HI leucine and ['2sI]Na from Amersham International, Mississauga, Ontario, Canada. Centricon micro-concentrators were obtained from Amicon, Danvers, MA; donkey anti-rabbit immunoglobulin-coated cellulose (Sac-Cell) from IDS Ltd, Washington, Tyne and Wear, England; and suramin from FBA Pharmaceuticals Inc, New York, NY. Heparin was purchased from Organon Teknica, Toronto, Ontario, Canada. Isolation of Chondrocytes Five pregnant sheep of mixed breed with single insemination dates were used in this study. Animals were killed with sodium pentobarbitone (Euthanyl) on days 75-80 of gestation (term 145 days), a laparotomy performed, and the fetuses delivered. Epiphyseal growth plate cartilage was dissected from the proximal tibia and chondrocytes isolated by incubation with 0.2% (wt/vol) collagenase as described in detail previously (Hill and De Sousa, 1990). Isolated chondrocytes were plated in 75cm' plastic culture flasks in DMEM supplemented with 10%' (vol/vol) FCS, penicillin (100 pU/ml), streptomycin (100 puglrnl) and fungizone (2.5 pg/ml), and allowed to grow to confluency over 2-3 days. Cultures were maintained a t 37°C in a humidified atmosphere of 5% C02-95% air. Chondrocytes were passaged with trypsin-EDTA three times before use in experiments between passages 3 and 6. We previously showed that cells continued to synthesize type I1 collagen over this passage number (Hill and De Sousa, 1990). Assessment of Cell Viability Chondrocyte viability throughout the culture period was monitored by several methods. Firstly, cells were visually inspected to ensure that cells were attached and had maintained a flattened morphology. Cultures which showed any evidence of rounded or free-floating cells were discarded. Secondly, the possible release of
217
lactate dehydrogenase (LDH) from damaged cells into the culture medium was assessed following 12, 24 or 48 hr of culture. The presence of LDH in cell-conditioned medium was detected by a colorimetric technique using a kit from Sigma Chemical Co. Thirdly, representative chondrocyte cultures were exposed to trypan blue at the end of incubation and the presence of positively-staining cells detected by light microscopy. Immunocytochemical Localization of Basic FGF in Chondrocytes and Estimation of DNA Synthesis Chondrocytes were plated at a density of 0 . 2 ~ 1 0 ~ cells per well in 8-well glass chamber slides and maintained in 200 p1 glucose-free DMEM supplemented with 8.7mM glucose and 10% (vol/vol) FCS for 24 hr, by which time cells were approximately 50% confluent. Cells were washed twice with PBS (pH 7.4) and growth-restricted by exposure to serum-free DMEM (8.7 mM glucose) containing 0.1% (vol/vol) FCS for 48 hr. After a further wash with PBS, this medium was replaced with DMEM (8.7mM glucose) supplemented with 2% (vol/vol) FCS to restart cells into the cell cycle, with or without 25 ,ug/ml antibasic FGF antibody, an equivalent protein concentration of protein A-Sepharose-extracted normal rabbit serum, 70pg/ml heparin or 100 p g l m l suramin. Chondrocytes were cultured for between 2 and 44 hr, with the addition of 1 pCi/ml ['HI thymidine in lop1 DMEM for the final 2 hr of incubation. Medium was removed and the cells from the upper four chambers washed three times with PBS before fixation with 2% (wt/vol) paraformaldehyde and 2% (vol/vol) glutaraldehyde in 0.1 M PBS (pH 7.5) at 4°C for 30min. Cells in the lower four chambers were used for the quantitation of DNA content a s described subsequently. Following removal of the chamber superstructure, cells which had been fixed were dehydrated by exposure of the glass slides to an ascending ethanol series (79, 90 and 100%) and air-dried. Slides were stored a t -20°C prior to immunocytochemistry. Immunocytocheniistry was performed using the avidin-biotin peroxidase technique (Hsu et al., 1981). Cells were re-hydrated in a descending ethanol series (100, 90 and 70%) into PBS, and incubated in 1% (vol/vol) hydrogen peroxide in
Growth Factors Downloaded from informahealthcare.com by UB Heidelberg on 11/14/14 For personal use only.
218
HILL and LOGAN
PBS for 15 min to eliminate endogenous peroxidase activity, followed by 5% (wt/vol) BSA in PBS to block non-specific binding. Slides were incubated with 2.5 p g l m l anti-basic FGF antibody (Ab773) diluted in 0.01 M PBS containing 2% (wt/vol) BSA, and 0.01% (wt/vol) sodium azide pH 7.5 (100 p1 per slide) for 48 hr at 4°C in a humidified chamber. In some experiments 0.3% (vol/vol) Triton was included in the buffer. This was not associated with any change in the staining pattern or intensity. Following incubation, slides were washed with PBS buffer and incubated with biotinylated goat anti-rabbit immunoglobulin (1:500 dilution; 100 pl) for 1.5 hr at 23°C. Following a further wash, the slides were incubated with avidin and biotinylated peroxidase (1OOpl) for 1 hr at 23"C, washed with PBS and then with 50 mM Tris buffer (pH 7.5). The basic FGF was visualized using freshly prepared 3,3'diaminobenzidine tetrahydrochloride (1.89 mM containing 0.03% (vol/vol) of 30% (vol/vol) hydrogen peroxidase). The reaction was quenched by washing in excess 50mM Tris buffer and cells were lightly counterstained with Carazzi's haematoxylin, dehydrated in an ascending ethanol series and xylene, and mounted under glass coverslips. Chondrocytes were examined by light microscopy and the percent cells showing positive immunoreactive staining in either the cytoplasm or the nucleus was calculated by point counting in defined fields. Specificity of staining was verified by several criteria. Immunostaining was absent when (a) non-immune, protein ASepharose-purified rabbit IgG was substituted for the primary antibody; (b) the biotinylated second antibody was omitted; and (c) the primary antibody was pre-incubated for 24 hr at 4°C with 240 nM basic FGF in a siliconized, polypropylene tube before incubation on the slide. Nuclear labelling was performed on the same cells used for immunocytochemistry. To estimate nuclear labelling index, coverslips were removed with xylene and 100% ethanol, and the slides dipped in Kodak NTB-2 liquid autoradiographic emulsion for 4 weeks at 4°C before processing with Kodak D19 developer and rapid fixer. Cells were remounted under glass coverslips and the percent nuclear labelling index estimated by light microscopy. In some experiments, DNA synthesis was estimated from the incorporation of ['HI thymidine
into solubilized DNA. These studies used the lower four chambers of each slide while the upper four were used for histology. Following culture and incubation with ['HI thymidine as described above, cells were washed with PBS to remove unincorporated isotope, the DNA and protein precipitated with 1 ml ice-cold trichloroacetic acid (10% wt/vol) and scraped free of the slide and solubilized by overnight incubation at 37°C with 1 M sodium hydroxide (500 pl per well). Isotope incorporation was measured by liquid scintillation counting and expressed as disintegrations per min (d.p.m.1 per pg cell DNA. The DNA content of each of the lower four chambers was measured for all experiments by fluorimetry, using Hoeschst fluorochrome 33258, as described previously (Hill and De Sousa, 1990). Some experiments were performed with chondrocytes which had not been growth restricted in serum-free medium prior to immunocytochemical staining and analysis of nuclear labelling index. These cells were grown in chamber slides until approximately 50% confluent, the medium changed to DMEM+2% (vol/vol) FCS, and incubated for 18 hr. ['HI thymidine (1 pCi/ml) was added for the final 2 hr of incubation before cells were washed and fixed as decribed earlier. This allowed the calculation of basic FGF presence and nuclear labelling index in unsynchronized, free-cycling cells.
Sample Preparation and Basic FGF Radioimmunoassay To estimate basic FGF synthesis by chondrocytes the cells were first grown in 75 cm' plastic flasks until approximately 80% confluent. Cells were then washed with PBS and the medium replaced with DMEM+O.l% FCS (vol/vol) for 48 hr, in order to deplete cells of any basic FGF accumulated from the previous growth medium. Chondrocytes were washed again with PBS and the medium replaced with 10 ml glucose-free DMEM supplemented with 8.7mM glucose, and 2% (vol/vol) FCS prior to incubation for between 16 and 30 hr. ['HI thymidine (1 pCi/ml) was added for the final 2 hr of culture. Some incubations were performed in the presence of 5 0 p g l m l cycloheximide. Culture medium was removed, centrifuged at 3000xg for 15 min to remove particulates, and stored at -70°C. The cells were washed with PBS and removed from the culture
Growth Factors Downloaded from informahealthcare.com by UB Heidelberg on 11/14/14 For personal use only.
CELL CYCLE-DEPENDENT bFGF LOCALIZATION
flask with 4 ml trypsin-EDTA (0.25% vol/vol) during a 2 min incubation a t 37°C. Aprotiniii (1.7 ml, 250 p g l m l ) was then added immediately and the cell suspension removed a n d placed on ice. Chondrocytes were sedimented by centrifugation at l000xg for 10 min a t 4°C and the supernatant aspirated and retained. This was considered to contain any basic FGF liberated from cell membranes or extracellular matrix by trypsin a n d is referred to as the matrix fraction. Cells were washed in PBS, re-pelleted by centrifugation at 1OOOxg for 1 0 m i n a t 4"C, a n d 250pg aprotinin added in 1 ml PBS supplemented with 1 M NaCl and 0.3% (vol/vol) Triton (pH 7.5). Chondrocytes were lysed by centrifugation at 10,OOOxg for 20 min a n d the supernatant collected for basic FGF radioimmunoassay. This fraction was considered to represent the intracellular content of basic FGF, and is referred to as the cytoplasmic fraction. The residual particulate cell pellet was used to calculate the incorporation of ['HI thymidine into DNA. Each pellet was resuspended in loop1 PBS a n d precipitated with 10% (wt/vol) trichloroacetic acid. Following centrifugation a n d aspiration of the supernatant the pellet was washed twice more with trichloroacetic acid before solubilization in 1 M NaOH a n d liquid scintillation counting. Medium, matrix a n d cytoplasmic fractions were each de-salted by centrifugation at 1SOOxg for 1 hr over Centricon micro concentrators with 10 kDa molecular weight cut-off membranes. Once volumes had been reduced to approximately 200 pl the samples were dried in a Savant rotary vacuum centrifuge, a n d resuspended i n 1OOpl PBS immediately before assay. The basic FGF content of each fraction was estimated hy radioimmunoassay, a s described below. In some experiments, a n alternate method was used to prepare t 11 e matrix fraction. Follow i n g rem o va 1 of medium and washing with PBS the chondrocytes were quickly washed once with 1 ml of 2 M NaCl in 20 mM HEPES buffer ( p H 7.4) followed by 2 waslies with PBS. The 2 M NaCl wash was retained for basic FGF radioimmunoassay. This method was previously shown to be effective in removing basic FGF from heparin-like binding sites in the extracellular matrix of endothelial cell cultures (Flaumenhaft et al., 1989). Both exposure to salt and incubation with dilute trypsin yielded similar recoveries of basic FGF. Following washing in PBS cells were removed from the culture
21')
dish with trypsin EDTA and the cytoplasmic fractions prepared as described above. Basic FGF was iodinated by a modification of the Cliloramine T method of Greenwood, Hunter a n d Gover (1963). One pg of basic FGF in 10 p l of PBS was added to 2 pl of 1.5 M potassium phosphate (pH 7.6) and 0.5 mCi (5pl) ["iIl-labelled sodium iodide in a siliconized, polypropylene conical tube. Chloramine T (3pl of a 0.1 mg/ml concentration in PBS) was added for 2 m i n and the reaction tube gently agitated. A further 3p1 chloramine T solution was then added for an incubation of 1.5 min, a n d a third addition of 3 pl m a d e before a final incubation of 1 min. The reaction was terminated by the addition of 5pl SO mM tyrosine hydrochloride and the vessel incubated for a further 2 min before addition of 200 pl 60 mM potassium iodide. All incubations were performed at room temperature. Iodinated basic FGF was separated from the reaction mixture by gel filtration on a Sephadex G-50 column ( 3 0 ~ 0 . cm) 7 eluted with 4 mM HCl containing 75 mM NaCl a n d 0.1% (wt/vol) BSA. Fractions of 1 ml were collected a n d stored in siliconized polypropylene tubes at -20°C until assay. All samples were measured with iodinated basic FGF from a single iodination reaction of specific activity 85 pCi/pg. Radioimmunoassay was performed using a final volume of 25Opl in polypropylene assay tubes. The reaction mixture consisted of 125pl assay buffer (0.01 M phosphate buffer containing 0.1% (wt/vol) sodium azide a n d 0.01 M NaCl, p H 7.5),50 p1 diluted primary antiserum (Ah 773, final concentration 625 p g / m l ) and 50 pl test sample or basic FGF standard preparation. A standard curve ranging from 7.8 fM/tube to 1000 fM/tube was used. After mixing the reagents were incubated for 24 hr at 4°C before the addition of ['"I] basic FGF (-
0 1992 Harwood Academic Publishers GinbH Printed in the United Kingdom
Cell Cycle-Dependent Localization of Irnmunoreactive Basic Fibroblast Growth Factor to Cytoplasm and Nucleus of Isolated Ovine Fetal Growth Plate Chondrocytes
Growth Factors Downloaded from informahealthcare.com by UB Heidelberg on 11/14/14 For personal use only.
(Receiwd August 19 1991, Accepted April 30 1992) Basic fibroblast growth factor (basic FGF) is a potent mitogen for chondrocytes iri vitro and is present in developing cartilage iri vivo. Studies of intracellular basic FGF localization in other cell types have revealed a transient nuclear presence. We have examined ovine fetal growth plate chondrocytes for the presence of intracellular basic FGF by immunocytochemistry. Chondrocytes were isolated from the proximal tibia1 growth plate of lamb fetuses between 75 and 80 days’ gestation using collagenase, and were cultured in monolayer before use between passages 3 and 6. In non-synchronized cell cultures 58?6% of cells (mean+s.d., r1=3) demonstrated cytoplasmic staining for immunoreactive basic FGF. Of these cells, 18+3% also exhibited strong nuclear staining. Chondrocytes were growth-restricted and restarted into the cell cycle with 2% (vol/vol) fetal calf serum. The timing of S phase was followed by nuclear labelling of nuclei with [‘HI thymidine followed by autoradiogaphy, or by the incorporation of [‘HI thymidine into trichloroacetic acid-precipitable DNA i n parallel cultures. A cytoplasmic presence of immunoreactive basic FGF did not appear, following immunocytochemistry, until the second half of G I with 97% of cells immunopositive 2 hr prior to S phase. Nuclear staining for basic FGF appeared 2 hr before peak nuclear labelling index, and 56% of cell nuclei were immunopositive. Following entry into S phase cytoplasmic and nuclear basic FGF immunostaining rapidly disappeared. When these experiments were repeated with or without the presence of anti-basic FGF antibody or heparin, the presence of the antibody significantly reduced peak [’HI thymidine incorporation into DNA during S phase while exposure to heparin increased this. However, the proportion of cells demonstrating cytoplasmic or nuclear staining for immunoreactive basic FGF, and the time of onset of staining, were unaltered. Incubation of cells with suramin blocked subsequent DNA synthesis and no intracellular basic FGF was visualized. Cellconditioned culture medium, extracellular matrix and cytoplasm from synchronized cultures of chondrocytes were taken at time points throughout the cell cycle and assessed for basic FGF content by radioimniunoassay. Basic FGF was detectable in each compartment and steadily rose throughout the second half of G, to reach maximum values around the S phase. The accumulation of basic FGF in medium, matrix and c to lasm was blocked by the presence of cycloheximide. When cells were labelled with [Y HI leucine P . and newly-synthesized basic FGF immunoprecipitated from cytoplasmic or nuclear fractions, peak values of peptide were seen in the cytoplasm 2 hr prior to its transient appearance in the nucleus which was immediately before entry into S phase.
Correspondence: Dr David J. Hill, Depts of Medicine and I’hysiology, Lawson Research Institute, St. Joseph’s Health Centre, 268 Grosvenor Street, London, Ontario, N6A 4V2, Canada. Tel: (519) 439-3271 x4716. Fax: (519) 438-3525.
21.5
HILL and LOGAN
216
The amount of immunoprecipitable, tritiated basic FGF recovered from cytoplasm and nucleus was unaffected by incubation with antibody against basic FGF. Separation of immunoprecipitated, tritiated basic FGF by gel electrophoresis revealed two major forms with molecular sizes 18 kDa and 23 kDa in both the cytoplasm and nucleus. Only the former was present in conditioned culture medium. The results suggest that chondrocytes synthesize and release basic FGF during the second half of G I in response to hormonal signals contained in fetal calf serum. This basic FGF contributes to the subsequent DNA synthetic rate of the cells. During late GI a brief translocation of basic FGF from cytoplasm to nucleus occurs, but this is not directly related to alterations in DNA synthesis.
Growth Factors Downloaded from informahealthcare.com by UB Heidelberg on 11/14/14 For personal use only.
KEYWORDS: fibroblast growth factor, fetus, chondrocyte, immunocytochemistry, nucleus
INTRODUCTION
MATERIALS AND METHODS
Basic fibroblast growth factor (basic FGF) is widely expressed in embryonic tissues and persists until late gestation in rat (Gonzalez et al., 1990; Hebert et al., 1990). Radioimmuno- and bioassayable basic FGF have been detected in the developing limb buds in the chick (Munaim, Klagsbrun and Toole, 1988; Seed, Olwin and Hauschka, 1988) while immunoreactive basic FGF was detected in growth plate chondrocytes of the rat fetus (Gonzalez et al., 1990). Since basic FGF is a potent mitogen for chondrocytes isolated from the epiphyseal growth plates of the rabbit, rat and chick (Makower, Wroblewski and Pawlowski, 1989; Crabb et al., 1990; Kato and Iwamato, 1990) it is possible that it acts as an autocrine or paracrine growth factor for skeletal tissue during early life. In support of this hypothesis we recently showed that isolated growth plate chondrocytes from the ovine fetus contained and released immunoreactive and bioactive basic FGF (Hill et al., 1992). Unlike other isolated cell types, almost half of the immunoreactive basic FGF released was recovered from conditioned culture medium in soluble form, the remainder being recovered from the extracellular matrix and cell membranes, and the cytoplasm. It has been demonstrated previously for other cell types which synthesize basic FGF that a cell cycle-dependent translocation of basic FGF occurs between cytoplasm and nucleus, although the relevance of this to its biological actions is unknown (Bladin et al., 1990; Renko et al., 1990). We have examined ovine fetal growth plate chondrocytes to determine whether such an intracellular translocation occurs in this cell type, and if so, what is its relationship to the autocrine/paracrine role of basic FGF as a mitogen.
Recombinant human basic FGF was kindly provided by Dr P. Barr, Chiron Corporation, Emeryville, CA; and a rabbit polyclonal antibody (Ab 773) raised against the 1-24 synthetic fragment of bovine basic FGF provided by Dr A. Baird, Whittier Institute for Diabetes and Medicine, La Jolla, CA. A purified IgG was obtained by precipitation of the antiserum with 30% (wt/vol) ammonium sulphate and passage over a protein A-Sepharose affinity column. The antibody demonstrated less than 1% cross-reactivity with acidic FGF and no crossreactivity with hst/ks or FGF-5. Cross-reactivity with FGF-6 is not known. However, FGF-6 shows low sequence homology with the fragment of basic FGF used as the antigen (Gonzalez et al., 1990). The same antibody was found to block the mitogenic actions of exogenous basic FGF on vascular endothelial cells in ziifro (Dr A. Baird, Whittier Institute, La Jolla, CA; personal communication). Biotinylated anti-rabbit IgG and biotin peroxidase complex used during immunocytochemistry were obtained from Vector Laboratories, Burlingame, CA. Plastic tissue culture flasks (75 cm2 area, NUNC), Hank’s buffered salts solution (HBSS), Dulbecco’s modified Eagle’s medium (DMEM), fetal calf serum (FCS) and trypsin-EDTA, were each purchased from Gibco Ltd., Burlington, Ont., Canada. A glucose-free formulation of DMEM was obtained from Imperial Laboratories Ltd., Salisbury, England; Dulbecco’s phosphate buffered saline (PBS), fatty acid-free bovine serum albumin, fraction V (BSA), cycloheximide, trasylol (aprotinin), phenylmethylsulfonyl fluoride, protein A-Sepharose and chloramine T from Sigma Chemical Co., St. Louis, MO; and %well
CELL CYCLE-DEPENDENT bFGF LOCALIZATION
Growth Factors Downloaded from informahealthcare.com by UB Heidelberg on 11/14/14 For personal use only.
Lab-tek chamber slides from Miles Laboratories Inc., Naperville, IL. Collagenase were obtained from Boehringer Ingelheim (Canada) Ltd., Dorval, Quebec, Canada. Hoeschst fluorochrome 33258 and 3-3'-diaminobenzidine were purchased from Aldrich Chemical Co., Milwaukee, WI; euthanyl from MCT Pharmaceuticals, Mississauga, Canada; and [3H] thymidine, ['HI leucine and ['2sI]Na from Amersham International, Mississauga, Ontario, Canada. Centricon micro-concentrators were obtained from Amicon, Danvers, MA; donkey anti-rabbit immunoglobulin-coated cellulose (Sac-Cell) from IDS Ltd, Washington, Tyne and Wear, England; and suramin from FBA Pharmaceuticals Inc, New York, NY. Heparin was purchased from Organon Teknica, Toronto, Ontario, Canada. Isolation of Chondrocytes Five pregnant sheep of mixed breed with single insemination dates were used in this study. Animals were killed with sodium pentobarbitone (Euthanyl) on days 75-80 of gestation (term 145 days), a laparotomy performed, and the fetuses delivered. Epiphyseal growth plate cartilage was dissected from the proximal tibia and chondrocytes isolated by incubation with 0.2% (wt/vol) collagenase as described in detail previously (Hill and De Sousa, 1990). Isolated chondrocytes were plated in 75cm' plastic culture flasks in DMEM supplemented with 10%' (vol/vol) FCS, penicillin (100 pU/ml), streptomycin (100 puglrnl) and fungizone (2.5 pg/ml), and allowed to grow to confluency over 2-3 days. Cultures were maintained a t 37°C in a humidified atmosphere of 5% C02-95% air. Chondrocytes were passaged with trypsin-EDTA three times before use in experiments between passages 3 and 6. We previously showed that cells continued to synthesize type I1 collagen over this passage number (Hill and De Sousa, 1990). Assessment of Cell Viability Chondrocyte viability throughout the culture period was monitored by several methods. Firstly, cells were visually inspected to ensure that cells were attached and had maintained a flattened morphology. Cultures which showed any evidence of rounded or free-floating cells were discarded. Secondly, the possible release of
217
lactate dehydrogenase (LDH) from damaged cells into the culture medium was assessed following 12, 24 or 48 hr of culture. The presence of LDH in cell-conditioned medium was detected by a colorimetric technique using a kit from Sigma Chemical Co. Thirdly, representative chondrocyte cultures were exposed to trypan blue at the end of incubation and the presence of positively-staining cells detected by light microscopy. Immunocytochemical Localization of Basic FGF in Chondrocytes and Estimation of DNA Synthesis Chondrocytes were plated at a density of 0 . 2 ~ 1 0 ~ cells per well in 8-well glass chamber slides and maintained in 200 p1 glucose-free DMEM supplemented with 8.7mM glucose and 10% (vol/vol) FCS for 24 hr, by which time cells were approximately 50% confluent. Cells were washed twice with PBS (pH 7.4) and growth-restricted by exposure to serum-free DMEM (8.7 mM glucose) containing 0.1% (vol/vol) FCS for 48 hr. After a further wash with PBS, this medium was replaced with DMEM (8.7mM glucose) supplemented with 2% (vol/vol) FCS to restart cells into the cell cycle, with or without 25 ,ug/ml antibasic FGF antibody, an equivalent protein concentration of protein A-Sepharose-extracted normal rabbit serum, 70pg/ml heparin or 100 p g l m l suramin. Chondrocytes were cultured for between 2 and 44 hr, with the addition of 1 pCi/ml ['HI thymidine in lop1 DMEM for the final 2 hr of incubation. Medium was removed and the cells from the upper four chambers washed three times with PBS before fixation with 2% (wt/vol) paraformaldehyde and 2% (vol/vol) glutaraldehyde in 0.1 M PBS (pH 7.5) at 4°C for 30min. Cells in the lower four chambers were used for the quantitation of DNA content a s described subsequently. Following removal of the chamber superstructure, cells which had been fixed were dehydrated by exposure of the glass slides to an ascending ethanol series (79, 90 and 100%) and air-dried. Slides were stored a t -20°C prior to immunocytochemistry. Immunocytocheniistry was performed using the avidin-biotin peroxidase technique (Hsu et al., 1981). Cells were re-hydrated in a descending ethanol series (100, 90 and 70%) into PBS, and incubated in 1% (vol/vol) hydrogen peroxide in
Growth Factors Downloaded from informahealthcare.com by UB Heidelberg on 11/14/14 For personal use only.
218
HILL and LOGAN
PBS for 15 min to eliminate endogenous peroxidase activity, followed by 5% (wt/vol) BSA in PBS to block non-specific binding. Slides were incubated with 2.5 p g l m l anti-basic FGF antibody (Ab773) diluted in 0.01 M PBS containing 2% (wt/vol) BSA, and 0.01% (wt/vol) sodium azide pH 7.5 (100 p1 per slide) for 48 hr at 4°C in a humidified chamber. In some experiments 0.3% (vol/vol) Triton was included in the buffer. This was not associated with any change in the staining pattern or intensity. Following incubation, slides were washed with PBS buffer and incubated with biotinylated goat anti-rabbit immunoglobulin (1:500 dilution; 100 pl) for 1.5 hr at 23°C. Following a further wash, the slides were incubated with avidin and biotinylated peroxidase (1OOpl) for 1 hr at 23"C, washed with PBS and then with 50 mM Tris buffer (pH 7.5). The basic FGF was visualized using freshly prepared 3,3'diaminobenzidine tetrahydrochloride (1.89 mM containing 0.03% (vol/vol) of 30% (vol/vol) hydrogen peroxidase). The reaction was quenched by washing in excess 50mM Tris buffer and cells were lightly counterstained with Carazzi's haematoxylin, dehydrated in an ascending ethanol series and xylene, and mounted under glass coverslips. Chondrocytes were examined by light microscopy and the percent cells showing positive immunoreactive staining in either the cytoplasm or the nucleus was calculated by point counting in defined fields. Specificity of staining was verified by several criteria. Immunostaining was absent when (a) non-immune, protein ASepharose-purified rabbit IgG was substituted for the primary antibody; (b) the biotinylated second antibody was omitted; and (c) the primary antibody was pre-incubated for 24 hr at 4°C with 240 nM basic FGF in a siliconized, polypropylene tube before incubation on the slide. Nuclear labelling was performed on the same cells used for immunocytochemistry. To estimate nuclear labelling index, coverslips were removed with xylene and 100% ethanol, and the slides dipped in Kodak NTB-2 liquid autoradiographic emulsion for 4 weeks at 4°C before processing with Kodak D19 developer and rapid fixer. Cells were remounted under glass coverslips and the percent nuclear labelling index estimated by light microscopy. In some experiments, DNA synthesis was estimated from the incorporation of ['HI thymidine
into solubilized DNA. These studies used the lower four chambers of each slide while the upper four were used for histology. Following culture and incubation with ['HI thymidine as described above, cells were washed with PBS to remove unincorporated isotope, the DNA and protein precipitated with 1 ml ice-cold trichloroacetic acid (10% wt/vol) and scraped free of the slide and solubilized by overnight incubation at 37°C with 1 M sodium hydroxide (500 pl per well). Isotope incorporation was measured by liquid scintillation counting and expressed as disintegrations per min (d.p.m.1 per pg cell DNA. The DNA content of each of the lower four chambers was measured for all experiments by fluorimetry, using Hoeschst fluorochrome 33258, as described previously (Hill and De Sousa, 1990). Some experiments were performed with chondrocytes which had not been growth restricted in serum-free medium prior to immunocytochemical staining and analysis of nuclear labelling index. These cells were grown in chamber slides until approximately 50% confluent, the medium changed to DMEM+2% (vol/vol) FCS, and incubated for 18 hr. ['HI thymidine (1 pCi/ml) was added for the final 2 hr of incubation before cells were washed and fixed as decribed earlier. This allowed the calculation of basic FGF presence and nuclear labelling index in unsynchronized, free-cycling cells.
Sample Preparation and Basic FGF Radioimmunoassay To estimate basic FGF synthesis by chondrocytes the cells were first grown in 75 cm' plastic flasks until approximately 80% confluent. Cells were then washed with PBS and the medium replaced with DMEM+O.l% FCS (vol/vol) for 48 hr, in order to deplete cells of any basic FGF accumulated from the previous growth medium. Chondrocytes were washed again with PBS and the medium replaced with 10 ml glucose-free DMEM supplemented with 8.7mM glucose, and 2% (vol/vol) FCS prior to incubation for between 16 and 30 hr. ['HI thymidine (1 pCi/ml) was added for the final 2 hr of culture. Some incubations were performed in the presence of 5 0 p g l m l cycloheximide. Culture medium was removed, centrifuged at 3000xg for 15 min to remove particulates, and stored at -70°C. The cells were washed with PBS and removed from the culture
Growth Factors Downloaded from informahealthcare.com by UB Heidelberg on 11/14/14 For personal use only.
CELL CYCLE-DEPENDENT bFGF LOCALIZATION
flask with 4 ml trypsin-EDTA (0.25% vol/vol) during a 2 min incubation a t 37°C. Aprotiniii (1.7 ml, 250 p g l m l ) was then added immediately and the cell suspension removed a n d placed on ice. Chondrocytes were sedimented by centrifugation at l000xg for 10 min a t 4°C and the supernatant aspirated and retained. This was considered to contain any basic FGF liberated from cell membranes or extracellular matrix by trypsin a n d is referred to as the matrix fraction. Cells were washed in PBS, re-pelleted by centrifugation at 1OOOxg for 1 0 m i n a t 4"C, a n d 250pg aprotinin added in 1 ml PBS supplemented with 1 M NaCl and 0.3% (vol/vol) Triton (pH 7.5). Chondrocytes were lysed by centrifugation at 10,OOOxg for 20 min a n d the supernatant collected for basic FGF radioimmunoassay. This fraction was considered to represent the intracellular content of basic FGF, and is referred to as the cytoplasmic fraction. The residual particulate cell pellet was used to calculate the incorporation of ['HI thymidine into DNA. Each pellet was resuspended in loop1 PBS a n d precipitated with 10% (wt/vol) trichloroacetic acid. Following centrifugation a n d aspiration of the supernatant the pellet was washed twice more with trichloroacetic acid before solubilization in 1 M NaOH a n d liquid scintillation counting. Medium, matrix a n d cytoplasmic fractions were each de-salted by centrifugation at 1SOOxg for 1 hr over Centricon micro concentrators with 10 kDa molecular weight cut-off membranes. Once volumes had been reduced to approximately 200 pl the samples were dried in a Savant rotary vacuum centrifuge, a n d resuspended i n 1OOpl PBS immediately before assay. The basic FGF content of each fraction was estimated hy radioimmunoassay, a s described below. In some experiments, a n alternate method was used to prepare t 11 e matrix fraction. Follow i n g rem o va 1 of medium and washing with PBS the chondrocytes were quickly washed once with 1 ml of 2 M NaCl in 20 mM HEPES buffer ( p H 7.4) followed by 2 waslies with PBS. The 2 M NaCl wash was retained for basic FGF radioimmunoassay. This method was previously shown to be effective in removing basic FGF from heparin-like binding sites in the extracellular matrix of endothelial cell cultures (Flaumenhaft et al., 1989). Both exposure to salt and incubation with dilute trypsin yielded similar recoveries of basic FGF. Following washing in PBS cells were removed from the culture
21')
dish with trypsin EDTA and the cytoplasmic fractions prepared as described above. Basic FGF was iodinated by a modification of the Cliloramine T method of Greenwood, Hunter a n d Gover (1963). One pg of basic FGF in 10 p l of PBS was added to 2 pl of 1.5 M potassium phosphate (pH 7.6) and 0.5 mCi (5pl) ["iIl-labelled sodium iodide in a siliconized, polypropylene conical tube. Chloramine T (3pl of a 0.1 mg/ml concentration in PBS) was added for 2 m i n and the reaction tube gently agitated. A further 3p1 chloramine T solution was then added for an incubation of 1.5 min, a n d a third addition of 3 pl m a d e before a final incubation of 1 min. The reaction was terminated by the addition of 5pl SO mM tyrosine hydrochloride and the vessel incubated for a further 2 min before addition of 200 pl 60 mM potassium iodide. All incubations were performed at room temperature. Iodinated basic FGF was separated from the reaction mixture by gel filtration on a Sephadex G-50 column ( 3 0 ~ 0 . cm) 7 eluted with 4 mM HCl containing 75 mM NaCl a n d 0.1% (wt/vol) BSA. Fractions of 1 ml were collected a n d stored in siliconized polypropylene tubes at -20°C until assay. All samples were measured with iodinated basic FGF from a single iodination reaction of specific activity 85 pCi/pg. Radioimmunoassay was performed using a final volume of 25Opl in polypropylene assay tubes. The reaction mixture consisted of 125pl assay buffer (0.01 M phosphate buffer containing 0.1% (wt/vol) sodium azide a n d 0.01 M NaCl, p H 7.5),50 p1 diluted primary antiserum (Ah 773, final concentration 625 p g / m l ) and 50 pl test sample or basic FGF standard preparation. A standard curve ranging from 7.8 fM/tube to 1000 fM/tube was used. After mixing the reagents were incubated for 24 hr at 4°C before the addition of ['"I] basic FGF (-
