CELL STRUCTURE AND FUNCTION 17:

9-17

(1992)

© 1992 by Japan Society for Cell Biology

Roles of Various Growth Factors in Growth of HumanOsteosarcoma Cells Which Can Grow in Protein-Free Medium Kenji Yamada*1, Yoshino Yoshitake1, Hiromichi Norimatsu2 and Katsuzo Nishikawa**1 ^Department of Biochemistry, Kanazawa Medical University, Uchinada, Ishikawa 920-02, and ^Department of Orthopaedic Surgery, Kagawa Medical School, 1750-1 Ikenobe, Kita-gun, Kagawa 761-01, Japan Key Words: osteosarcoma cells/protein-free

culture/FGF/EGF/TGF-a

ABSTRACT.The human osteosarcoma cell line (OST-l-PF) can grow in protein-free Coon's modified Ham's F12 medium. Growth of the cells in protein-free medium was partially density-dependent and partially depressed by mediumchange. An extract and conditioned mediumof OST-l-PF cells contained high mitogenic activity for BALB/c3T3cells. The growth factor in the cells was purified and identified as a basic fibroblast growth factor (bFGF)-like factor on the basis of its elution profile on heparin-affinity chromatography and the result of immunoblotting. An unidentified factor in a conditioned mediumeliciting most of the DNAsynthesis-stimulating activity showed a weak affinity for heparin. Various additions, including serum and growth factors, stimulated the growth of OST-l-PF cells in protein-free medium. Of these factors, epidermal growth factor (EGF), transforming growth factor-a (TGF-a), acidic fibroblast growth factor (aFGF) and bFGF were the most potent mitogens. High-affinity receptors of EGFand FGFwere found on the surface of these cells. These results indicate that autonomous growth of OST-l-PF cells in protein-free medium is mainly controlled by an intracellular mechanism.

There have been manystudies on the roles of growth factors in the growth of cultured osteosarcoma cells. A

cells,

shownto have a receptor for epidermal growth factor (EGF) and to respond to treatment with EGFby increased prostaglandin production, but not DNAsynthe-

Members of the fibroblast growth factor (FGF) family are potent mitogens for manytypes of cells of mesodermal and neuroectodermal origin, as are angiogenic factors in vivo (reviewed in 4, 8, 12, 27, 36). They are

clonal

line of human osteosarcoma

sis (30). Platelet-derived stimulated prostaglandin

cells,

G-292,

was

growth factor (PDGF) also production by this cell line

(31). On the other hand, the human U-2 OS osteosarcoma cell line was found to secrete a growth factor structurally related to a homodimer of PDGFA-chains (3, 16).

Another

human osteosarcoma

cell

* On leave of absence from the Department

Surgery, Kagawa Medical School. ** To whomcorrespondence sed.

and reprint

line,

MG-63

of Orthopaedic

request should be addres-

Abbreviations: aFGF, acidic fibroblast growth factor; bFGF, basic fibroblast growth factor; BSA, bovine serum albumin; CHAPS,3[(3-cholamidopropyl)dimethylammonio]-l-propane

calf serum; DME,Dulbecco's

cell

line

(Osteogenic

growth factor;

CS,

modified Eagle's medium; EGF, epider-

mal growth factor; FBS, fetal mance liquid chromatography;

IGF-II, insulin-like

sulfonate;

bovine IGF-I,

growth factor-II;

serum; HPLC, high-perforinsulin-like growth factor-I;

OST-1 , A human osteosarcoma

Sarcoma Takase);

PF-C-F12, protein-free

PDGF,

platelet-derived

Coon's modified Ham's F12

medium; PBS, phosphate-buffered saline; RIA, radioimmunoassay; RPMI-1640, Rosewell Park Memorial Institute tissue culture medium-1640; TGF-a, transforming growth factor-a; TGF-/3,

transforming growth factor-/3.

was reported

to have functional

receptors

for

PDGF (13, 39) and for insulin-like growth factor I (IGFI), which is a potent mitogen for the cells in vitro (26).

structurally related proteins and include acidic FGF, basic FGF, and the proto-oncogene products desig-

nated as int-2, hst-l/ks, hst-2 (FGF-5), and FGF-6. One of these proteins, bFGF, is widely distributed in various normal tissues and tumors, but its production and functional roles in osteosarcoma cells have not yet been reported.

It was first isolated

as a single-chain

protein

composed of 146 amino acids (7), but later, forms that weretruncated or extended at the aminoterminus were isolated (ll, 19, 32, 38). As reviewed by Barnes and Sato (2), the use of serumfree media has obvious advantages for investigating the effects of growth factors, hormones and other defined materials in the pathology and etiology of cancer, including autocrine, paracrine and endocrine mechanisms (33). In this study we used a humanosteosarcoma cell line, OST-l-PF, that can grow in protein-free medium. The present paper reports that this cell line produced bFGFlike factor and expressed both EGF and FGF receptors. These cells showedincreased replication in response to

K. Yamada et al.

(24). Protein concentrations

exogenous bFGF, aFGF, EGFor transforming growth factor-a (TGF-a), but the response to bFGF did not seem to include an autocrine mechanism. MATERIALS

Cells and culture.

Biocinchonic

cell

line

(OST-1, Osteogenic Sarcoma Takase), which was established by Dr. Yamazaki (President of Kanazawa Medical University) in 1964, was a generous gift from Dr. Tomita (Department of Orthopaedic Surgery, Kanazawa University School of Medicine) (37). These cells were maintained in a mediumconsisting of Rosewell Park Memorial Institute tissue culture medium1640 (RPMI-1640) supplemented with lOO U/ml of penicillin and 100//g/ml of streptomycin and 15% fetal bovine serum (FBS). The cell line OST-1-PF was selected

Rock ford, IL) with

BSAas a standard. A nalysis ofDNAsynthesis-stimulating activities in a cell extract (OST-1-PF-CE) and conditioned medium (OST-l-PFCM)ofOST-l-PF cells. For analysis of the mitogenic activities in OST-1-PF-CE and 0ST-1-PF-CM, their elution profiles on heparin affinity chromatography were examined as described previously (24). Con fluent OST-l-PF cells grown in CF12 medium containing \% CS in 850-cm2 Falcon roller bottles were washed with PF-C-F12 medium for 1 day, and then the cells were collected. A crude extract of 7 x 107 cells in 7 ml of 10 mMTris-HCl buffer (pH 7.5), prepared by freezing and then thawing the cells, was applied to a TSKHeparin 5PW HPLC column (7.5 x75 mm) which had been equilibrated with the same buffer containing 0.5 M NaCl and 0.1% 3-[(3cholamidopropyl)dimethylammonio] - 1 -propane sulfonate (CHAPS).The column was washed with 16ml of the same buffer and developed with a linear gradient of 0.5-2.0 M NaCl in the same buffer for 1 hr at a flow rate of 0.8 ml/min. Fractions of 1.2 ml were collected. Conditioned medium of OSTl-PF cells (OST-1-PF-CM) was collected to avoid contamination by the lysate of dead cells as follows. Con fluent OST-lPF cells grown in 10 ml of PF-C-F12 in 100-mm dishes were washed with PF-C-F12 medium for 1 day and then maintained in 10 ml of the same medium for 4 days. Then the conditioned medium was collected and centrifuged and the resulting supernatant (50 ml) was applied to the heparin-Sepharose column (1.4 x 5 cm) (Pharmacia, Uppsala) that had been equilibrated with the same buffer containing 0.65 M NaCl and 0.1% CHAPS.The column was washed with 20ml of the buffer and developed with 120 ml of a linear gradient of 0.652.0 MNaClin the same buffer. The flow rate was adjusted to 80 ml/hr and the eluate was collected in ll-ml fractions. Aliquots of the fractions from both columns were assayed for stimulation of DNAsynthesis in BALB/c3T3-3Kcells, as de-

AND METHODS

The human osteosarcoma

of FGFs were determined with a

Acid (BCA) kit (Pierce,

from OST-1 cells

that could grow in protein-free Coon's modified Ham's F12 medium supplemented with 15 mMHEPES (pH7.3), 100 U/ml penicillin and 100 //g/ml streptomycin (PF-C-F12) by continuous culture of the original cells in this protein-free medium. The stock culture was maintained in PF-C-F12 medium

in dishes that had been coated with type IV collagen (Sigma, St. Louis, MO).These cells were subcultured by treatment with 0.25% trypsin/0.025% EDTAin phosphate-buffered saline (PBS) and then with 0.1% soybean trypsin inhibitor (Sigma, St. Louis, MO) in PBS. The BALB/c3T3-3Kcells were used for tests on the abilities

of various growth factors to stimulate DNAsynthesis (25, 40).

The cells were cultured in Dulbecco's modified Eagle's me-

dium (DME) containing 10% calf serum (CS) and all the other

supplements described above. These cells were cultured at 37°C in a humidified atmosphere of 5% CO2 in air. Assay ofDNAsynthesis. The abilities of various growth factors, including bFGF, to stimulate DNAsynthesis of cultured BALB/c3T3-3K cells were assayed as described previously (25, 40). One unit of activity was defined as the amount with the same ability as 1 mg of CS protein to stimulate the incorporation of [3H]-thymidine into DNA. Growth experiments. OST-1-PF cells were plated at a density of2x 104 in 2ml of PF-C-F12 medium in 35-mm Falcon dishes that had been coated with type IV collagen. Samples were added at the time of inoculation. After 7 days, the cells were harvested by trypsinization and counted in a Coulter counter. Values are shown as averages of triplicate experiments.

scribed above.

Purification of bFGF-likefactorfrom an extract of OST-lPF cells {OST-1-PF-CE). A bFGF-like factor from OST-1PF-CE was prepared by the methods of Schweigerer et al. (28) and Klagsbrun et al. (19) with some modifications. Briefly, 8.9 x 108 OST-l-PF cells were harvested from monolayer cultures in roller bottles by trypsinization. The cells were suspended in 100ml of 1 M NaCl/lOmM Tris-HCl buffer (pH 7.5) containing 1 //g/ml of leupeptin and 15juM pepstatin (both from the Peptide Institute Inc., Osaka), 1 mMphenylmethylsulfonyl fluoride (Sigma), and 1 mMN-ethylmaleimide and 1 mMEDTA(both from Wako, Osaka). The cells were homogenized in a Potter-Elvehjehm homogenizer, and the homogenate was sonicated for 3 min and centrifuged at 27,000 g for 10 min. The resulting supernatant was dialyzed overnight against distilled water to lower the concentration of NaCl to below 0.15 Mand then centrifuged at 40,000g for 1 hr. The supernatant was adjusted to pH 6.0 and applied to a CM-Sephadex C-50 column (5 x 5 cm) (Pharmacia, Uppsala) that

Clonal growth was assayed as described previously (21).

Briefly, OST-l-PF cells were plated at a density of 2 x 104 cells in 5 ml of the same mediumas described above in Falcon dishes (60 mmwith 2-mmgrid) that had been coated with type IV collagen. Cell numbers in four 4-mm2 fields were counted for 5 days by microscopy and the cell numbers per dish were calculated.

Purifications of bFGF and aFGF. bFGF and aFGFwere purified from bovine brain by the method of Gospodarowicz et al. (10) with slight modifications as described previously 10

Growth of Osteosarcoma Cells in Protein-Free Medium had been equilibrated with 0.1 Msodium phosphate buffer (pH 6.0). The material was then eluted stepwise with 0.1 Msodium phosphate buffer containing 0.15 M and 0.65 MNaCl. The material eluted with 0.65 MNaCl was applied to a heparin-Sepharose column (1.4 x 5 cm) that had been equilibrated with 0.65 M NaCl/lOmM Tris-HCl (pH7.2). The column was washed with about 5 column volumes of the same solution, and the adsorbed material was then eluted with 120 ml of a linear gradient of 0.65-2.0 M NaCl in 10 mMTris-HCl (pH 7.5) at a flow rate of60 ml/h. Fractions of 10 ml were col-

lected and tested for DNAsynthesis-stimulating

activity.

(18),

and then the radioactivity

was counted as described

above. For competitive binding assays, the experimental conditions were as described above except that the binding medium consisted of 0.4ml of C-F12/0.1% BSA, 0.05 ml of 26 ng/ml 125I-hEGF (45,000-48,000 cpm/well) or 10-30 ng/ml 125I-aFGF (62,000-68,000 cpm/well), 0.05 ml of unlabeled hEGF, human recombinant transforming growth factor-a

(hTGF-tf) and aFGFor bFGFat appropriate concentrations in 0.1% BSA/PBS/0.02% NaN3. The value for nonspecific binding of radioactivity (usually about 0.2 or 1.0% of the total tracer added) in the presence of excess hEGF (1 //g/ml) or aFGF (100 ng/ml), respectively, was subtracted. All measurements were done at least in duplicate, and average values

The

active fractions were combined, dialyzed against distilled water and lyophilized. The powder was dissolved in distilled water and used for immunoblotting. All operations were performed at about 4°C. Immunoblotting. Proteins were separated by electrophoresis on SDS-12.5% polyacrylamide gel and then transferred electrophoretically to a nitrocellulose sheet. The sheet wasin-

are shown. Other materials.

Monoclonal

antibodies

against

bFGF

(bFM-1 and bFM-2) and hEGF (HA) were obtained and purified as described previously (23, 41). Recombinant hEGF and hTGF-a were obtained from Wakunaga, Osaka. Porcine TGF-^1 was obtained from R&DSystems, Inc. (Minneapolis, MN). PDGF(AB) was a generous gift from Dr. Kaji, (Tokyo Metropolitan Institute of Gerontology, Tokyo). C-F12 medium was a product of Hazleton Inc., (St. Lenexa, KS). FBS was from Flow Lab. Inc. (North Ryde, CA). CS was from Gibco Lab., (Grand Island, NY). Na 125I and 3H-thymidine were obtained from the Radiochemical Centre (Amersham).

cubated first in blocking buffer (5% BSA/PBS) and then for 30 min with monoclonal antibody against bFGF (bFM-2, 5 jug/m\) (23) diluted with washing buffer (0.1% Tween 20/ PBS). The sheet was washed 4 times for 15 min each time with washing buffer and incubated with 500 ng/ml of biotinylated rabbit anti-mouse IgG (Vector Laboratories, Inc. Burlingam, CA) in washing buffer for 30 min. It was then washed again with washing buffer and incubated with horseradish peroxidase-avidin conjugate according to the manufacturer's instructions (Vector Laboratories, Inc.). The sheet was then washed extensively with washing buffer and the peroxidase activity was located with diaminobenzidine tetrahydrochloride.

Type IV collagen,

human transferrin,

BSA, bovine insulin

and heparin were from Sigma. RESULTS

Cellgrowth inprotein-free medium. The OST-l-PF cell line was obtained from the OST-1 cell line by maintenance of the latter in PF-C-F12 medium, which is use-

125I-Labeled EGF and 125I-labeled aFGF binding to the cells. The receptors for EGF and FGF on OST-l-PF cells were surveyed by binding assays with 125I-labeled ligands as described previously (25, 41), with some modifications. Human

recombinant EGF (hEGF) (Wakunaga, Hiroshima) and puri-

fied bovine aFGF were labeled with 125I by the chloramine-T method, as described previously (23, 25). OST-l-PF cells were inoculated at a density of 1 x 105 cells in 2ml of C-F12 me-

dium containing 1% CSinto 24-well multiwell plates (Costar, Broadway Cambridge, 2 cm2) that had been coated with type

IV collagen. The cells were grown to con fluency, and then the mediumwas changed to PF-C-F12 mediumand the cells were cultured for another day. The medium was then changed to binding medium (0.5 ml) consisting of 0.45ml of C-F12/ 0.1% BSA, 0.05 ml of 0.2-10.0 ng/ml 125I-labeled hEGF (125IhEGF) (4,400-177,000 cpm/well) or 0.2-4.0 ng/ml 125I-labeled aFGF (125I-aFGF) (4,500-89,500 cpm/well) in 0.1% BSA/PBS/0.02% NaN3, and the wells were incubated for 3 hr at 20°C. For 125I-hEGF binding assay, the cells were washed three times with 1 ml of cold PBS and solubilized in 0.5 ml of 0.5 N NaOH, and then the radioactivity bound to the cells was counted in an Aloka-autowell gamma system (ARC-300). For 125I-aFGF binding assay, the cells were washed four times with 0.1% BSA/PBS/0.02% NaN3 and once with 250^g/ml heparin in PBS to reduce the nonspecific binding of 125I-aFGF

10A

Fig. 1. Growth of OST-l-PF cells in protein-free mediumin dishes coated with type IV collagen. Clonal growth was assayed under the experimental

conditions

described

in the

text.

A. Effect

of CS.

Triangles, PF-C-F12 medium; Circles, C-F12 medium containing 1% CS. B. Effect of cell density. The cells were plated at a density of 2 x 104 or 2x 103 in PF-C-F12 medium. C. Effect of medium change.

Circles, no treatment; squares, sham treatment every 24 hr; crosses, mediumchange to fresh mediumevery 24 hr. ll

K. Yamada et al.

growth of these cells only partially. Characterization of growth factors produced by OST-l-PF cells. To characterize the activity in an extract of OST-l-PF cells to stimulate DNAsynthesis in BALB/c3T3cells, we applied the extract to a heparin5PWHPLCcolumn and examined the elution profile with a linear gradient of NaCl (Fig. 2A). Almost all the activity was bound to the column and eluted with 1.31.5 M NaCl with about 100% recovery, indicating that the activity was due to a bFGF-like factor (24, 29). The activity corresponded to 620 units, or 12 ng, of bovine bFGFper 107 cells, calculated on the basis of the specific activity of purified bovine bFGF (5 x 107 units/mg protein). To confirm the identity of the activity as that of bFGF, we purified this growth factor. By the purification steps described in Materials and Methods, 370 ng of bFGF-like factor was obtained from an extract of 8.9x 108 cells with about 65% recovery. For immuno-

ful for protein-free culture of various cell lines (22, 24, 34). The cells grew in protein-free medium in type IV collagen-coated dishes with a doubling time of 48 hr after a short lag period (Fig. 1A). The cells grew in medium containing 1% CS with a doubling time of 24 hr, which was less than that in protein-free medium. Concentrations of CS higher than 1% had no further effect on the growth (data not shown). For cell growth, type IV collagen-coated dishes were preferable to uncoated, poly-D-lysine-coated (data not shown).

or type I collagen-coated

dishes

To determine whether the factor(s) that is produced and secreted into the protein-free mediumcontributes to growth of these cells, we examined the cell density dependence of growth and also the effect of medium change on the growth. The cells were plated at a density of 2x 104 or 2x 103 in 5 ml of protein-free medium in 60-mm dishes coated with type IV collagen (Fig. IB). The doubling time at low density (68 hr) was slightly more than that at high density (44 hr). When the cells were inoculated into protein-free mediumand the medium was replaced by fresh medium every 24hr, the growth rate was slightly decreased after mediumchange (Fig. 1C). Shamtreatment by removing the medium and returning it resulted in a growth rate similar to that of untreated cells. These results indicate that OST-l-PF cells grow autonomously in protein-free medium and

blot

analysis

of the bFGF-like

factor,

the bFM-2 ob-

tained as described previously (23), which recognized de-

natured bFGFfrom various species, was used as a probe. The cross-reactivity of the protein with bFM-2 was the same as that of bovine bFGF, and the antibody did not cross-react with aFGF. The bFGF-like factor from the cell extract migrated as a single, main band with an apparent Mr of 16,500, similar to bovine bFGF (Fig. 2B). For analysis of the activity to stimulate DNAsynthesis in BALB/c3T3 cells in the conditioned medium of B

A

d

that an exogenous factor(s) or secretory factor(s) which is liberated from the cells into the medium affects

Mr 0.10

2.0

x 10~3 - 97.4 - 66.2

1*

-42.7

- 31.0 *£

- 21.5 - 14.4 0.0

0.00

'0.0

BP (32ml)

20

30 Fraction

number

40 (1.2ml/fraction)

50

60

aFGF

bFGF

CE

Fig. 2. A. Heparin-5PW HPLCof the growth factor from an extract of OST-l-PF cells. Experimental conditions were as described in the text. BP, flow through fraction. B. Immunoblot analysis of bFGF-like growth factor using MAbagainst bFGF (bFM-2). Experimental conditions were as described in the text. Lane a, bovine aFGF (100 ng); lane b, bovine bFGF (100 ng); lane c, bFGF-like factor purified from an extract of OST-l-PF cells (24 ng). The following Mr markers from Bio-Rad (Richmond, CA) were used: phosphorylase B (Mr, 97,400), BSA (Mr, 66,200), ovalbumin (Mr, 45,000), carbonic anhydrase (Mr, 31,000), soybean trypsin inhibitor (Mr, 21,500) and lysozyme (Mr, 14,400). 12

Growth of Osteosarcoma Cells in Protein-Free Medium Table I. Effects

of various additions

CELLS IN PROTEIN-FREE

C-F12

on growth of OST-l-PF MEDIUM.a) ^ of co nt ro l ce ll n um be r

A d d i ti o n (ju g / m l) N o n e b)

(10 0 )

I n su lin (5 )

28 1

T r a n s f e r r in ( 5 0 )

230

BSA (200)

230

Insulin+Transform+BSA (5, 50, 200)

487

h E G F ( 0 .0 1 )

496

b F G F (0 .0 1 )

45 1

PDGF (0.01)

191

T G F - /3 1 (0 .0 1 )

202

M o n o cl o n a l a n t ib o d y ag a i n st b F GF ( 1 0)

158

1%

452

5

c

å>

蝣 ^ 1.0 r > 01 z

I8 O.Q 5

Fraction

10

number (llml/fraction)

C S

.^ ,,7+ V,

Fig. 3. Heparin-Sepharose affinity chromatography of the growth factor from conditioned mediumof OST-l-PF cells. Experimental conditions were as described in the text. BP, flow through fraction;

a) Cell numbers after 7 days were counted in a Coulter counter. Experimental conditions were as described in the text. b> The cells

W, washing fraction.

OST-l-PF cells,

the medium was applied to a heparin-

Sepharose column (Fig. 3). In contrast to the activity in the cell extract, most of activity in the conditioned medium was recovered in the unabsorbed fractions with less than 0.65 MNaCl, and minor fractions were eluted with about 0.9M and 1.3 MNaCl. A similar elution profile was obtained whena small volumeof the conditioned medium was subjected to heparin-5PW HPLC (data not shown). Although the major activity in the unabsorbed fraction was not identified as PDGF, EGF or TGF-a by the respective

radioimmunoassays

were plated

at a density

of 2xlO4/dish.

The cell

number after 7 days in the control culture (7.0x 104-l.l cells/dish) without any addition was taken as 100%.

300

x 105

1-

O1%cs

(RIAs),

some of the activity was shown to be due to insulin-like

growth factor-II (IGF-II) according to its RIA (data not shown). Further major activity.

study is required

Effects of various additions cells.

We studied

the effects

to characterize

the