MFO

CDU

R

fcucntUcc(rdIdwtct

tlcflIts5Isfsoas

Biochemical and Biophysical Research Communications 263, 786–789 (1999)

Article ID bbrc.1999.1451, available online at http://www.idealibrary.com on

0CA

itogenic and Antiapoptotic Effects of Insulin-like Growthactor Binding Protein-6 in the Human Osteoblasticsteosarcoma Cell Line Saos-2/B-10

hristoph Schmid, Claudia Keller, Martina Gosteli-Peter, and Jurgen Zapfivision of Endocrinology and Diabetes, Department of Internal Medicine,niversity Hospital, Ramistrasse 100, CH-8091 Zurich, Switzerland

eceived August 27, 1999

MATERIALS AND METHODS

Gwi9caccnAeCfit

iaCw

mDsiFbgnDclmwc2u4a

asi

Insulin-like growth factor (IGF) I is a potent mitogenor human osteosarcoma cells such as the Saos-2/B-10ell line. IGF binding proteins (IGFBPs) prevent stim-lation of DNA synthesis by IGFs. In contrast to re-ombinant human (rh) IGFBP-2, -3, -4, and -5, 10–100M rhIGFBP-6 stimulated [3H]thymidine incorpora-ion into DNA and multiplication of Saos-2/B-10 cells.pon withdrawal of serum, 30 nM IGFBP-6 also de-reased apoptosis (within 4 h) and increased proteinontent and sodium-dependent phosphate uptakewithin 24 h), but less potently than IGF I. 125I-labeledhIGFBP-6 did not bind to the cells, and cold IGFBP-6id not affect 125I-labeled IGF I binding. Production ofGF I, IGF II, and IGFBP-6 by the cells or significantegradation of rhIGFBP-6 could not be detectedithin 24 h of incubation. Thus, among the rhIGFBPs

ested, rhIGFBP-6 is unique in stimulating osteosar-oma cell growth. Furthermore, it has an antiapop-otic effect. © 1999 Academic Press

Six different high affinity insulin-like growth fac-or binding proteins (IGFBP-1 to IGFBP-6) of mo-ecular mass 22–32 kDa have been identified andloned (1). IGFBP-6 was isolated from cerebrospinaluid (2) and also from the medium of fibroblasts (3).GFBP-6 binds IGF II with a 100-fold higher affinityhan IGF I (4) and is abundant in adult human serum;erum concentrations are in the range of 5–10 nM (1,). IGF I is a potent mitogen for Saos-2/B-10 cells (6);GF II is 2- to 3-fold less potent. While testing thepecificity of the biological activity of rhIGFBPs, weound that IGFBP-6 but not IGFBP-5 stimulated DNAynthesis in these cells (7). Since stimulatory effectsf an IGFBP in a cell line which expresses little IGF Ind IGF II were unexpected, we undertook the presenttudy.

786006-291X/99 $30.00opyright © 1999 by Academic Pressll rights of reproduction in any form reserved.

Cell culture. Saos-2/B-10 cells (kindly provided by Drs. S. B. and. A. Rodan) (8) were grown and passaged in medium supplementedith fetal calf serum (FCS, 10%) as described (6, 9). Cells were plated

n Falcon multiwell tissue culture dishes at a density of 2 3 105 per.6-cm2 surface area (35 mm diameter), kept for 3 days in 5% FCS-ontaining growth medium, then rinsed with serum-free medium,nd the medium was replaced by serum-free Ham’s F12 mediumontaining gentamycin (50 mg/ml), glutamine (2 mmol/liter) andharcoal-treated bovine serum albumin (BSA) at 1 g/liter. Recombi-ant human (rh) IGF I and rhIGF II were from former Ciba–GeigyG, Basel. RhIGFBP-2 was a gift from Dr. J. Shuster, Chiron (Em-ryville, CA), rhIGFBP-3 from Dr. A. Sommer, Celtrix (Santa Clara,A). RhIGFBP-4 and rhIGFBP-6 were expressed in yeast and puri-ed as described (4). Aliquots of test agents were added directly tohe cultures.

[3H]Thymidine incorporation into DNA. Confluent cultures werencubated in 1 ml serum-free F12 medium containing 1 g/liter BSAnd test agents for 18 h, pulsed with [3H]thymidine (Amersham, 85i/mmol; 1 mCi per dish) and further incubated for 3 h at 37°C. DNAas precipitated and incorporated radioactivity was measured (6, 9).

Photometric enzyme immunoassay for determination of cytoplas-ic histone-associated DNA fragments. An ELISA kit (Cell Deathetection ELISAPLUS from Boehringer-Mannheim) was used to mea-

ure cytosolic oligonucleosome-bound DNA. 42,000 cells were seededn 24 multiwell (2-cm2 surface area) dishes and grown for 3 days inCS-containing medium, rinsed with serum-free medium and incu-ated for 4 h in 200 ml serum-free Ham’s F12 medium containing 1/l BSA and test agents. The ELISA monitors enrichment of oligo-ucleosomes in the cells by monoclonal antibodies directed againstNA and histones. To avoid potential effects caused by differences in

ell adherence and recovery, both the cell supernatant and the cellayer lysate were analyzed. The media were combined with 2 3 400l phosphate-buffered saline (PBS), pH 7.3, which was used forashing the cell layers (to harvest detached and loosely adhering

ells). After centrifugation for 10 min at room temperature at 200g,00 ml of lysis buffer was added to the pellet, and a 20-ml aliquot wassed for the assay. The cell layer on the dishes was lysed directly into00 ml lysis buffer, centrifuged for 10 min at 200g at room temper-ture, and 20 ml was used for analysis as described by the supplier.

Phosphate transport studies. Cells were grown on 9.6-cm2 surfacerea dishes for 3 days in 5% FCS-containing medium, rinsed witherum-free medium and exposed to test medium. 32PO4 uptake stud-es were performed in buffer containing 140 mM NaCl or choline

chloride instead of NaCl (for measuring Na1-independent phosphateumUtp

2tirr

pmtfa

Toc

R

ldIweIccI2Is(ttblas

tftIDacI

tms

I(iwmI

a2cttId2tw

TABLE 1

CI

CI

IIIII

ro[weiierfI

Vol. 263, No. 3, 1999 BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

ptake) during 10 min at room temperature in the presence of 0.1M cold KH2PO4 and 1 mCi 32PO4 (200 mCi/mmol, from Amersham,K). Na-dependent phosphate transport was calculated by subtrac-

ion of the Na-independent component from the total uptake in theresence of Na as described (9).

Blot analysis for IGFBPs. Conditioned media were collected after4 h, centrifuged at 250g for 10 min (to remove detached cells) andhe supernatants were stored at 220°C. 125I-IGF II ligand blot andmmunoblot analysis were performed using polyclonal antiseraaised in rabbits to check for the recovery of added rhIGFBP-6 andhIGFBP-3 as described (10).

Determination of cell number, protein content, and alkaline phos-hatase activity. After 4 days of culture (the last 24 h in testedium), cells were rinsed with PBS, detached from the dishes with

rypsin and counted in a hemocytometer. Parallel dishes were usedor the determination of protein content and of alkaline phosphatasectivity (9).

Statistical analysis. The results are presented as mean 6 SEM.he difference between means was evaluated by the Wilcoxon rank-rder test. A probability level of a random difference of p , 0.05 isonsidered statistically significant.

ESULTS

DNA synthesis, cell replication. IGFBP-6 stimu-ated [3H]thymidine incorporation into DNA in a dose-ependent manner (Table 1). Since we noticed thatGFBP-6 stimulated DNA synthesis, we checkedhether IGFBP-6 modified DNA synthesis in the pres-

nce of a maximally active concentration (5 nM) of IGF. In these experiments, IGFBP-6 alone tended to in-rease DNA synthesis at 10 nM and exerted a signifi-ant stimulatory effect at 30 nM (Table 1a). 30 nMGFBP-6 also increased cell number (Table 1b, Table). When tested in combination with IGF I, 30 nMGFBP-6 neither attenuated nor further enhanced thetimulatory effects of 5 nM IGF I on DNA synthesisTable 1a) and on the number of the cells attached tohe culture dishes (Table 1b). IGFBP-2, -3, -4, and -5 byhemselves had no effect on DNA synthesis (Table 1c),ut all these preparations reliably blocked the stimu-atory effects of rhIGFs on DNA synthesis when testedt 10-fold molar excess over IGF I and IGF II (nothown).

Apoptosis. Microscopic examination of the cell cul-ures revealed that a large number of cells detachedrom the dishes after serum withdrawal, i.e., duringhe test period. Not only 5% FCS but also IGF I andGF II prevented cells from detaching. Gel analysis ofNA extracted from detaching cells harvested after 8 hnd after 24 h revealed marked DNA fragmentation inontrol cells but not in cells treated by 5 nM IGF I, IGFI or 5% FCS (not shown).

To further investigate apoptosis with a more sensi-ive assay, we used an ELISA to detect DNA frag-ents. Much less DNA was degraded to oligonucleo-

omes in cell cultures treated with 5 nM IGF I, 5 nM

787

GF II, or 5% FCS, as estimated from the absorbanceA405 nm 2 A490 nm) values for both detaching and adher-ng cells. IGFBP-6 at 30 nM also inhibited apoptosisithin 4 h, but 5 nM IGF I or IGF II or 5% FCS wereore potent (Table 2). Nevertheless, the effect of 30 nM

GFBP-6 was consistent and statistically significant.

Phosphate transport, alkaline phosphatase activitynd protein content. Phosphate transport of the Saos-/B-10 cell line has been previously characterized. Theurrent studies were conducted at a phosphate concen-ration of 0.1 mM, i.e., at a concentration which is closeo the KM of the transport system (9). Not only 5 nMGF I but also 30 nM IGFBP-6 stimulated Na-ependent phosphate (but not alanine) uptake by Saos-/B-10 cells (Table 3). Again, IGF I was more potenthan IGFBP-6. IGF I is effective within 30 min (9),hereas stimulation of phosphate transport by

Effects of rhIGFBP-6 on Proliferation of Saos-2/B-10 Cells

Control IGFBP-6 (10 nM) IGFBP-6 (30 nM)

(a) [3H]Thymidine incorporation

ontrol 6,373 6 673 7,102 6 699 11,832 6 1,315b

GF I(5 nM)

32,367 6 3,794a 32,378 6 4,096a 41,550 6 4,829a

(b) Cell number

ontrol 258,500 6 11,655 284,400 6 14,259 309,275 6 13,991b

GF I(5 nM)

383,200 6 19,901a 383,350 6 9,606a 399,150 6 31,160a

(c) [3H]Thymidine incorporation

IGFBP concentration

3 nM 10 nM 30 nM 100 nM

GFBP-2 0.99 6 0.03GFBP-3 0.98 6 0.03 1.01 6 0.06GFBP-4 0.94 6 0.07GFBP-5 0.93 6 0.07c

GFBP-6 1.07 6 0.02 1.27 6 0.06b 1.75 6 0.14b 3.56 6 0.16b

Note. (a) Effects of IGF I and IGFBP-6 on 3[H]thymidine incorpo-ation into DNA (cpm/dish 3 3 h). (b) Effects of IGF I and IGFBP-6n cell number per dish. (c) Effects of rhIGFBP-2, -3, -4, -5, and -6 on3H]thymidine incorporation into DNA (ratio treated/control). Cellsere pulsed with [3H]thymidine (18–21 h) or counted (24 h) afterxposure to test media. In (a), the values of [3H]thymidine cpmncorporated per 3 h and dish from 5 experiments (3 in triplicates, 2n quadruplicates) are given; in (b), the cell numbers per dish from 5xperiments in duplicates (n 5 10). In (c), values are expressed asatios of treated/control; data from 2 experiments in quadruplicatesor IGFBP-2 and -4 and from 3 experiments in quadruplicates forGFBP-3 and -6 are given.

a p , 0.05 IGF I-treated vs controls.b p , 0.05 IGFBP-6-treated vs control.c Taken from Ref. 7.

IaInndc

nmwTiw

ctpIbralylfibcdb

cell culture media, IGFBP-6 production was undetect-ah

D

cscitscatn

2aminIidtpIdap

b

CII

I

I

I

e

TABLE 2

iE

CI

I

I

I

F

clwttf

Vol. 263, No. 3, 1999 BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

GFBP-6 was not significant after 2 and 6, but onlyfter 24 h. The effects of 5 nM IGF I and 10 or 30 nMGFBP-6 on Na-dependent phosphate transport wereot additive. 5 nM IGF I and, to a smaller extent, 30M IGFBP-6 increased the protein content per cultureish. IGFBP-6 did not affect alkaline phosphatase spe-ific activity (Table 3).

IGFBP-6 recovery. Saos-2/B-10 cells express a highumber of 125I-IGF I specific binding sites; half maxi-al displacement is observed by 0.3 nM cold IGF I orith 2–3-fold higher concentrations of cold IGF II.ype 1 IGF receptors appear to mediate all the biolog-

cal effects of IGF I and IGF II. Preincubation of cellsith 10 or 30 nM IGFBP-6 for 24 h did not alter

125I-IGF I binding. Binding sites for 125I-rhIGFBP-6ould not be detected (not shown). Since high concen-rations of IGFBP-6 were required to stimulate cellroliferation, we also considered the possibility thatGFBP-6 may be degraded and that fragments maye responsible for the observed effects. However,hIGFBP-6 was completely recovered from the mediumfter 24 h of incubation, as assessed by 125I-IGF IIigand blot (not shown) and IGFBP-6 immunoblot anal-sis (Fig. 1). Thus, rhIGFBP-6 did not attach to the cellayer nor to the plastic of the culture dishes. Thisnding is distinct from findings with rhIGFBP-3 whichinds to the cell layer and to plastic (not shown), de-reases over time and is degraded to fragments asetected by IGFBP-3 immunoblot (Fig. 1). As assessedy Northern and Western blot analysis of concentrated

Detection of Nucleosomes (Arbitrary Absorbance Values)n Cytoplasmic Fractions of Saos-2/B-10 Cells after 4 h ofxposure to Test Media

Treatment

A405 nm 2 A490 nm

detaching cells(medium)

A405 nm 2 A490 nm

adhering cells(cell layer)

Cell number(3105/dish)

ontrol 0.65 6 0.09 0.39 6 0.06 2.0 6 0.2GFBP-6

(10 nM)0.55 6 0.09 0.33 6 0.06a 2.3 6 0.2

GFBP-6(30 nM)

0.38 6 0.07a 0.20 6 0.02a 2.7 6 0.2a

GF I(5 nM)

0.04 6 0.01a 0.02 6 0.01a 3.1 6 0.3a

GF II(5 nM)

0.06 6 0.01a 0.03 6 0.01a 3.4 6 0.3a

CS (5%) 0.13 6 0.02a 0.08 6 0.01a 3.5 6 0.3a

Note. Cells were exposed to test media for 4 h. Lysate aliquots ofells detaching from and adhering to the culture dishes were ana-yzed as described under Materials and Methods. Parallel dishesere used for the determination of cell number after 24 h of incuba-

ion in the same test media. Values are from three experiments inriplicate (assayed in duplicate) for absorbance (A405 nm 2 A490 nm) androm four experiments in duplicate for cell number per dish.

a p , 0.05 treated vs control.

788

ble in Saos-2/B-10 cells but readily detected in normaluman fibroblasts.

ISCUSSION

IGF I is a potent mitogen for human osteosarcomaells. IGF II is 2- to 3-fold less potent than IGF I intimulating cell growth and, correspondingly, alsoompetes less potently with 125I-IGF I for receptor bind-ng. IGFBP-6 sequesters IGF II, and, much less effec-ively, IGF I (4, 10). Nevertheless, IGFBP-6 per setimulated DNA synthesis and cell number signifi-antly at concentrations greater than 10 nM (Tables 1nd 2). Thus, IGFBP-6 is unique among all the IGFBPsested in this cell line, since IGFBP-2, -3, -4, and -5 hado stimulatory effects (Table 1c, Ref. 7).Serum and IGF withdrawal causes apoptosis of Saos-

/B-10 cells. IGF I and IGF II were potent inhibitors ofpoptosis as judged by a marked decrease in cytoplas-ic histone-associated DNA fragments during 4 h of

ncubation (Table 2). Prevention of apoptosis of malig-ant as well as of normal cells is a typical feature ofGFs (11–13). However, at 30 nM IGFBP-6 also inhib-ted apoptosis, although less than maximally effectiveoses (5 nM) of IGF I or II. Furthermore, it increasedhe protein content per dish and Na-dependent phos-hate transport within 24 h, again less than 5 nM IGF(Table 3). In contrast to 5 nM IGF I, 30 nM IGFBP-6id not affect alkaline phosphatase specific activity,nd it had no effect on Na-dependent phosphate trans-ort in short-time experiments (up to 6 h).IGF I and IGF II expression by Saos-2/B-10 cells was

elow the limit of detection by Northern blot of total

TABLE 3

Protein Content, Alkaline Phosphatase (ALP) Activity, andNa-Dependent Phosphate Transport

TreatmentProtein

(mg/dish)

ALP activity(mmol/h 3

mg protein)

NadPi transport(pmol/mg protein

3 10 min)

ontrol 73 6 8 109 6 8 725 6 74GF I (5 nM) 136 6 9a 84 6 6a 1845 6 297a

GFBP-6(10 nM)

79 6 7b 108 6 7 790 6 91

GFBP-6(10 nM) 1 IGF I

140 6 10a 91 6 3a 1736 6 311a

GFBP-6(30 nM)

87 6 7b 106 6 6 975 6 121b

GFBP-6(30 nM) 1 IGF I

144 6 8a 92 6 5a 1515 6 236a

Note. Cells were exposed to test media for 24 h. Values from sixxperiments (four in triplicates, two in quadruplicates) are shown.

a p , 0.05 IGF I-treated vs corresponding controls.b p , 0.05 IGFBP-6-treated vs corresponding control.

RdnpfbrctaItbctilchdmr

CtaI

weight fragments (Fig. 1). An autocrine role cannot becdom

rotnsbs

A

sytfi

R

1

1

1

1

1

1

fte3fMml

Vol. 263, No. 3, 1999 BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

NA and by radioimmunoassays of concentrated me-ia, in agreement with published results (14): whereasormal human osteoblast-like cells and fibroblasts ex-ress significant levels of both IGF I and IGF II, trans-ormed human osteosarcoma cells do not. A high num-er of high affinity type I IGF receptors explains theemarkable sensitivity of Saos-2/B-10 cells to low con-entrations of IGF I and IGF II. Therefore, the failureo detect IGF production does not definitely exclude anutocrine/paracrine role of IGFs. However, since addedGFBP-3 which effectively sequesters IGFs and blocksheir stimulatory effects (6, 10), did not decrease theasal rate of DNA synthesis (Table 1, Refs. 6 and 7), aontribution of locally released IGFs to the growth ofhese cells is unlikely. Similarly, the fact that IGFBP-6nhibited apoptosis already within 4 h renders it un-ikely that this effect is due to IGFs produced by theells. Whereas Saos-2/B-10 osteosarcoma cells expressigh affinity binding sites for IGF I, we were unable toetect binding of 125I-IGFBP-6 to the cells. Further-ore, 10 or 30 nM IGFBP-6 had no effect at all on

eceptor binding of 125I-IGF I (not shown).Synthetic heparin-binding peptides derived from theOOH-terminal third of IGFBP-6 have been reported

o stimulate glucose uptake by endothelial cells at 15nd 30 mM (15). However, IGFBP-6, in contrast toGFBP-3, was not degraded into small molecular

FIG. 1. IGFBP-6 and IGFBP-3 immunoblot analysis of serum-ree rhIGFBP-6- and rhIGFBP-3-containing medium before and af-er exposure to Saos-2/B-10 cells for 24 h. Cells were grown andxposed to serum-free test medium containing 30 nM rhIGFBP-6 or0 nM rhIGFBP-3. The supernatant was collected, centrifuged androzen. Twenty microliter-aliquots were used for blot analysis (see

aterials and Methods). Numbers left to the blot indicate molecularasses of markers. On top, the letter “b” designates before and the

etter “a” after incubation with cells.

789

onsidered for IGFBP-6 since this protein is not pro-uced by Saos-2/B-10 cells. However, a paracrine rolef IGFBP-6 produced by neighboring fibroblasts in tu-ors might be envisaged.IGFBP-6 did not stimulate DNA synthesis in normal

at osteoblastic cells (10). While the biological effectsn the Saos-2/B-10 cell line are unique for IGFBP-6,he mechanism of its action is unclear because we haveot been able to identify IGFBP-6 receptors. One pos-ibility to explain the described biological effects mighte a cooperative action together with an unidentifiedubstance produced by the tumor cells.

CKNOWLEDGMENTS

We thank C. Veldman for help with the phosphate transporttudies, C. Hauri and C. Zwimpfer for performing IGFBP blot anal-sis and IGF radioimmunoassays, M. Salman for the preparation ofhe manuscript, and the Swiss National Science Foundation fornancial support (Grant 32-46808.96).

EFERENCES

1. Jones, J. I., and Clemmons, D. R. (1995) Endocr. Rev. 16, 3–34.2. Roghani, M., Hossenlopp, P., Lepage, P., Balland, A., and

Binoux, M. (1989) FEBS Lett. 255, 253–258.3. Martin, J. L., Willetts, K. E., and Baxter, R. C. (1990) J. Biol.

Chem. 265, 4124–4130.4. Kiefer, M. C., Schmid, C., Waldvogel, M., Schlapfer, I., Futo, E.,

Masiarz, F. R., Green, K., Barr, P. J., and Zapf, J. (1992) J. Biol.Chem. 267, 12692–12699.

5. Baxter, R. C., and Saunders, H. (1992) J. Endocrinol. 134, 133–139.

6. Schmid, C., Rutishauser, J., Schlapfer, I., Froesch, E. R., andZapf, J. (1991) Biochem. Biophys. Res. Commun. 179, 579–585.

7. Schmid, C., Schlapfer, I., Gosteli-Peter, M. A., Froesch, E. R., andZapf, J. (1995) Prog. Growth Factor Res. 6, 167–173.

8. Rodan, S. B., Wesolowski, G., Ianacone, J., Thiede, M. A., andRodan, G. A. (1989) J. Endocrinol. 122, 219–227.

9. Veldman, C. M., and Schmid, C. (1998) Growth Horm. IGF Res.8, 55–63.

0. Schmid, C., Schlapfer, I., Keller, A., Waldvogel, M., Froesch,E. R., and Zapf, J. (1995) Biochem. Biophys. Res. Commun. 212,242–248.

1. Hill, P. A., Tumber, A., and Meikle, M. C. (1997) Endocrinology138, 3849–3858.

2. Parrizas, M., Saltiel, A. R., and LeRoith, D. (1997) J. Biol. Chem.272, 154–161.

3. Stewart, C. E. H., and Rotwein, P. (1996) Physiol. Rev. 76,1005–10026.

4. Okazaki, R., Conover, C. A., Harris, S. A., Spelsberg, T. C., andRiggs, B. L. (1995) J. Bone Miner. Res. 10, 788–795.

5. Booth, B. A., Boes, M., Andress, D. L., Dake, B. L., Kiefer, M. C.,Maack, C., Linhardt, R. J., Bar, K., Caldwell, E. E. O., Weiler, J.,and Bar, R. S. (1995) Growth Regul. 5, 1–17.