THE JOURNAL OF BIOLOGICAL CHEMISTRY 0 1988 by The American Society for Biochemistry and Molecular Biology, Inc.

Vol. 263, No. 27, Issue of September 25, pp. 13916-13921,1988 Printed in U. S. A.

Transcriptional Regulation of Osteopontin Production in Rat Osteosarcoma Cells by Type ,6 Transforming Growth Factor*

(Received for publication, November 16, 1987)

Masaki NodaS, Kyonggeun YoonS, Charles W. Princes, William T. Butlerll, and Gideon A. Rodan$ From the $Department of Bone Biology and Osteoporosis Research, Merck Sharp and Dohme Research Laboratories, West Point, Pennsylvania 19486, the §Department of Nutrition Sciences, University of Alabama at Birmingham, Birmingham, Alabama 35294, and the 7Department of Biological Chemistry, Dental Branch, University of Texas Health Sciences Center at Houston, Houston, Texas 77225

Type B transforming growth factor (TGFB) was shown to regulate the production of several extracel- lular matrix proteins. Osteopontin (OP) is a recently discovered bone matrix protein which was shown to promote the attachment of osteoblastic rat osteosar- coma ROS 17/23 cells to their substrate. We examined the effects of TGFB on OP production and OP mRNA in ROS 17/2.8 cells. Four-day treatment with 4 ng/ml TGFBl increased substantially the level of osteopontin in the cell culture media, as estimated by immunoblot- ting. Metabolic labeling showed that this effect was associated with a 3-4-fold increase in OP biosynthesis. TGFBl also increased, in a dose-dependent manner starting at 0.4 ng/ml, the steady-state level of OP mRNA. The increase in OP mRNA was first detected 48 h after the addition of TGFBl and lasted at least until 120 h. The half-life of OP mRNA, estimated in the presence of 5,6-dichloro-l-~-~-ribofuranosylben- zimidazole, was about 10 h and was not altered by TGFB1. On the other hand, the increase in OP mRNA was blocked by actinomycin D. Nuclear run-on assays indicated that TGFBl increased the rate of transcrip- tion of the OP gene. Examination of hormonal inter- actions showed that TGFBl opposed or compensated for the reduction in OP mRNA produced by dexameth- asone and that TGFBl did not further augment OP mRNA levels which had been increased by 1,25-dihy- droxyvitamin DB treatment. TGFB2 had similar effects on OP gene expression as TGFB1. In conclusion, TGFB promotes the production of osteopontin in the osteo- blastic osteosarcoma cells through a pathway which is at least in part mediated by transcriptional events.

Osteopontin (bone sialoprotein I) is a 44-kDa glycoprotein which is rich in aspartic and glutamic acid and contains 1 phosphothreonine and 12 phosphoserine residues (1). Using immunolocalization, osteopontin was detected in bone matrix, osteoid, osteoblasts, osteocytes, and preosteoblast-like fibro- blasts (2), as well as in certain neurons and neurosensory cells (3). In osteoblasts, the protein wes shown by immunoelectron microscopy to be localized in the Golgi apparatus (4).

High levels of OP’ mRNA were found in rat calvariae, rat

* The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “aduertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

I The abbreviations used are: OP, osteopontin; TGFP, type (3 trans- forming growth factor; HEPES, 4-(2-hydroxyethyl)-l-piperazineeth- anesulfonic acid; FBS, fetal bovine serum; SDS-PAGE, sodium do- decyl sulfate-polyacrylamide gel electrophoresis.

osteosarcoma cells, and rat kidney, and barely detectable levels were present in other tissues (5). In rat embryonic calvariae, the expression of OP mRNA, relative to total RNA, started to rise at day 15 of gestation and peaked in the newborn (5). This pattern paralleled that of osteocalcin mRNA. Although the exact role of osteopontin in osteogenesis is unclear, its amino acid sequence, deduced from the cDNA, contains the residues Gly-Arg-Gly-Asp-Ser (6), a sequence identical to the cell-binding domain of fibronectin and several other cell adhesion molecules. Osteopontin promoted the at- tachment of rat osteosarcoma cells (6) and human gingival fibroblasts (7) to plastic dishes, and this effect was inhibited by Arg-Gly-Asp-containing peptides (6).

Very little is known about the regulation of this recently discovered protein. We have observed that the level of osteo- pontin secreted by rat osteoblastic osteosarcoma cells (ROS 17/2.8) was increased 1.5-%fold by low concentrations of 1,25- dihydroxyvitamin D3, and this increase was inhibited by ac- tinomycin D (8). The OP mRNA levels were increased by 1,25-dihydroxyvitamin D3 treatment and were decreased by dexamethasone treatment (5).

Type P transforming growth factor (9, 10) is a potential autocrine regulator of bone formation since it is produced by osteoblasts ( l l ) , is abundant in bone matrix (12,13), regulates bone cell proliferation (14), and was shown to promote the differentiation of muscle-derived mesenchymal cells into chondrocytes (12, 15). In addition, TGFP enhanced the expression of the bone-related proteins, alkaline phosphatase, type I collagen, and osteonectin in ROS 17/2.8 cells through a mechanism likely to involve transcription (16). In other tissues and cell types, TGFP has been shown to promote the synthesis of extracellular matrix proteins such as fibronectin and collagen (9, 17-19), to reduce the production of metallo- proteinases (20), and to enhance the production of tissue inhibitor of metalloproteinases (20) and plasminogen activa- tor inhibitor (21), all of which are conducive to an anabolic effect on connective tissue. TGFP indeed is produced in wound tissues (22); stimulates total protein, collagen, and DNA content in wound-healing chambers (23); and promotes the rate of wound repair in vivo (24).

The object of this study was to examine the effect of TGFP on OP synthesis and its mode of action in ROS 17/2.8 cells. This cell line has been shown to possess many characteristics of osteoblasts (25) including formation of calcified matrix in vivo (26), high alkaline phosphatase (27), type I collagen and osteocalcin production (28, 29), and response to parathyroid hormone (27). This report shows that low concentrations of TGFPl and TGFB2 stimulate the production of osteopontin by a mechanism which involves de novo synthesis and tran- scriptional control.

13916

TGFP Regulation of Osteopontin Synthesis 13917

MATERIALS AND METHODS

Porcine platelet-derived TGFPl and TGFP2 were purchased from R & D Systems Inc. (Minneapolis, MN). [a-32P]dCTP (3000 Ci/ mmol), [a-32P]UTP (800 Ci/mmol), and [14C]serine (171 mCi/mmol) were purchased from Amersham Corp. Antibodies to osteopontin were raised in rabbits and affinity-purified as described (2). The cDNAs for rat osteopontin and alkaline phosphatase were obtained as previously described (5, 30). The cDNAs for type I procollagen (alRz), fibronectin (prlf-1), tubulin (RPT3), and actin were the gen- erous gifts of Drs. Rowe, Kream (University of Connecticut Health Center, Farmington, CT), Hynes (MIT, Center for Cancer Research, Cambridge, MA), and Farmer (Boston University School of Medicine) (see Refs. 31-33).

Cell Culture-Rat osteosarcoma ROS 17/23 cells were maintained as described previously in Ham's F-12 medium containing 28 mM HEPES, 2.5 mM L-glutamine, and 1.1 mM CaC12 and supplemented with 100 pg/ml kanamycin and 5% FBS (34). The cells were routinely subcultured once a week. For experiments, cells were plated at 10,000- 20,000 cells/cm2 in 9.5-cm2 wells or in 150- or 500-cmZ dishes in 5% FBS unless described otherwise. MC3T3E1 cells were kindly provided by Dr. Kodama (Tohoku University, Fukushima, Japan) and were maintained as described previously in a-minimal essential medium supplemented with 60 pg/ml kanamycin and 10% FBS (35). The cells were routinely subcultured every 3 days.

Zmmunoblot (Western) Analysis-ROS 17/23 cells grown to con- fluence in 150-cm2 dishes were rinsed three times with serum-free media and were subsequently cultured in 30 ml of serum-free media with or without TGFpl or TGFP2 at 4 ng/ml for 96 h. At that time, the media were collected, and the cells were trypsinized and counted in a Coulter Counter. The cell number in TGFP-treated cultures was 15-20% lower than that in control cultures. The media were spun at 3000 X g for 5 min to remove cellular debris; phenylmethylsulfonyl fluoride (0.3 mM) and EDTA (20 mM) were added to the supernatant. The media were concentrated 24-fold in a Centricon-10 (Amicon Corp., Danvers, MA) at 4 "C according to the manufacturer's protocol. Aliquots corresponding to equal numbers of cells were incubated with 5% P-mercaptoethanol in the electrophoresis buffer and were ana- lyzed by 10% sodium dodecyl sulfate-polyacrylamide gel electropho- resis (SDS-PAGE) (36). The proteins were then transferred to nitro- cellulose filters (37) which were incubated with anti-osteopontin immunoaffinity-purified antibodies in buffer containing 10 mM Tris- HC1, pH 7.5, 0.05% Tween 20 for 1 h at 23 "C after pretreatment of filters for 1 h with a nonfat dry milk powder (Carnation, Los Angeles, CA) solution. To visualize the antibodies, the filters were incubated for 1 h with anti-rabbit IgG goat antibody conjugated with horseradish peroxidase (Bethesda Researyh Laboratories), and color development was carried out using 4-chloro-1-naphthol as substrate according to the manufacturer's instructions (Bethesda Research Laboratories).

Metabolic Labeling of Osteopontin-At confluence, ROS 17/2.8 cells cultured in 9.5-cm2 wells were rinsed three times with phosphate- buffered saline and were incubated in serum-free media with or without TGFpl or TGFP2 (4 ng/ml) for 48 h. The cells were then incubated with 1 ml of fresh serum-free media containing 10 pCi of ~-[U-"C]serine for 4 or 48 h. Cells and media were harvested sepa- rately in the presence of proteinase inhibitors as described above. Some cells in parallel cultures wereotrypsinized and counted in a Coulter Counter. Media were centrifuged at 3000 X g for 5 min to remove cellular debris. Aliquots corresponding to equal numbers of cells were treated with 5% P-mercaptoethanol in loading buffer and were analyzed on 10% SDS-PAGE. The gels were treated with En- lightening (Du Pont-New England Nuclear) and fluorographed.

Estimation of Osteopontin mRNA (Northern Ana1ysis)"Cells were plated into 150- or 500-cmZ dishes, and RNA was prepared as de- scribed by Greenberg and Ziff (38). Briefly, the cells were rinsed three times with ice-cold phosphate-buffered saline, scraped in phosphate- buffered saline, and pelleted at 500 X g for 5 min. The cell pellet (3- 5 X 10' cells) was resuspended in 1 ml of Nonidet P-40 lysis buffer (10 mM Tris-HCI, pH 7.4, 10 mM NaC1, 3 mM MgC12, 0.5% (v/v) Nonidet P-40), incubated for 5 min on ice, and centrifuged at 500 X g for 5 min. The supernatant was removed and mixed with 1/20 volume of 0.2 M vanadyl ribonucleoside complex and centrifuged at 10,000 X g for 10 min at 4 "C. The supernatant was then mixed with an equal volume of digestion buffer (0.2 M Tris-HC1, pH 7.5, 0.44 M NaCl, 2% SDS, 25 mM EDTA) containing 200 pg/ml proteinase K and incubated at 37 "C for 60 min. After phenol/chloroform extrac- tion and precipitation by ethanol, 10 pg of RNA was fractionated on 1% formaldehyde (0.44 M)-agarose gel and transferred to nylon filters

(Hybond N, Amersham Corp.) by electroblotting (39). Filters were prehybridized overnight at 42 "C in hybridization buffer containing 50% formamide, 5 X SSC, 200 pg/ml sonicated salmon sperm DNA, and 5 X Denhardt's solution. Hybridization was carried out in hy- bridization buffer plus 4 X lo6 cpm of cDNA/5 ml. The cDNA inserts were labeled with [a-32P]dCTP by using random primers and Klenow fragments as described by Feinberg and Vogelstein (40) to a specific activity greater than 10' cpm/pg. Filters were washed at 65 "C in 0.1 X SSC, 0.1% SDS and were exposed to x-ray film at -70 "C using an intensifying screen.

Nuclear Run-on Assay-Isolation of nuclei, in vitro transcription, and subsequent determination of the rates of transcription and hy- bridization were done according to published methods (41-43) with minor modifications. Nuclei (2-3 X 10') were incubated in a 200-111 reaction mixture containing 50 mM Tris-HC1, pH 7.5, 100 mM am- monium sulfate, 1.8 mM dithiothreitol, 1.8 mM MnCL, 80 units of RNasin, 0.3 mM each ATP, GTP, and CTP, and 100 pCi of [a-"PI UTP. Radiolabeled RNA was extracted and hybridized to cDNAs immobilized on nitrocellulose or nylon filters. For each cDNA, 2 pg of linearized and alkaline-denatured plasmid DNA was immobilized on filters either by baking at 80 "C for 2 h or by exposure to ultraviolet light for 5 min. Filters were prehybridized at 42 "C overnight in 50% formamide, 2 mM Tris-HC1, pH 7.6, 10 mM EDTA, 0.1% SDS, 5 X SSC, 5 X Denhardt's solution, 10 pg/ml sonicated salmon sperm DNA, 10 pg/ml polyadenylic acid, 20 pg/ml tRNA. RNA probes at 1- 2 X lo6 cpm in 0.5 ml were hybridized at 42 "C for 3 days in hybridization buffer. Filters were then washed in 2 X SSC, 0.1% SDS three times at room temperature for 5 min each, followed by 10 min of washing in 0.1 X SSC, 0.1% SDS at 60 "C. Filters were exposed to x-ray films at -70 "C using intensifying screens.

RESULTS

Fig. 1 (upper) shows the effect of TGFP1, and TGFP2 on osteopontin levels in conditioned media of ROS 17/22 cells as estimated by Western blots. Migration of OP in SDS- PAGE was reported to depend on the concentration of poly- acrylamide: 45 kDa in 15% and 75 kDa in 5-15% gradient gels. In 10% gels, purified OP migrated as a BO-kDa band. The OP band was substantially enhanced follbwing TGFP treatment. Both TGFPl and TGFP2 showed similar potency. The specificity of the immunoblotting reaction was demon- strated by the disappearance of the OP band when the anti- body was preincubated with 40 pg/ml pure osteopontin (Fig. 1, lower). Using the same method, we detected no signal in cell extracts (data not shown), suggesting that most of the osteopontin is secreted soon after production.

Fig. 2 presents SDS-PAGE fluorographs of media from metabolically labeled ROS 17/2.8 cells cultured in the pres- ence and absence of aTGFP1. It appears that osteopontin, (60 kDa) is the most prominent band present in the media follow- ing [14C]serine labeling, except for high molecular mass bands which are probably collagen (data not shown). The OP band was enhanced about 3-4-fold (estimated by scanning densi- tometry) following TGFP treatment, indicating stimulation of de novo synthesis. A labeling period of 4 h following 48 h of treatment with TGFP (Fig. 2, upper) was sufficient to demonstrate a rise in radiolabeled OP in the medium. This observation is also consistent with rapid release of newly synthesized OP. TGFP2 had similar effects as TGFPl in this assay (Fig. 2, lower). Similar results were obtained when cells were metabolically labeled in the presence of 5% serum (data not shown).

Seventy-two hours of treatment with TGFPl or TGFp2 increased the steady-state level of the 1.5-kilobase OP mRNA in ROS 17/2.8 cells by about 3-fold (Fig., 3, upper). Concom- itantly, there was a substantial decrease in the steady-state level of fibronectin mRNA estimated in the same filters. As seen in Fig. 3 (center), the TGFpl effect on osteopontin mRNA levels was dose-dependent, starting at 0.4 ng/ml (16 pM). The specificity of this effect was shown by the similar abundance of actin mRNA in the same filters (Fig. 3, lower).

13918 TGFP Regulation of Osteopontin Synthesis

1 2 3 . '",

97 - 68 -

OP

43 -

1 2 7. 97-

68 - I - 0 P

43 - FIG. 1. Effect of TGFB on osteopontin production in culture

media of ROS 17/2.8 cells. Confluent cells in 150-cm2 dishes were treated with 4 ng/ml TGFPl or TGFP2 for 96 h in serum-free media. Media were concentrated in a Centricon-10 as described under "Ma- terials and Methods." Media aliquots corresponding to 420,000 cells were examined by Western immunoblot analysis as described under "Materials and Methods." Molecular masses in kilodaltons are indi- cated. Authentic osteopontin (data not shown) migrated as a band with an apparent molecular mass of 60,000 Da on 10% SDS-PAGE, indicated as OP (upper). Lane I , control; lane 2, TGFp1; lane 3, TGFB2. Lower lane I , 40 pg/ml purified osteopontin was present during preincubation with the anti-osteopontin antibody as described under "Materials and Methods"; lane 2, same as lane I , except for the addition of 40 pg/ml bovine serum albumin instead of purified osteopontin. The figures represent the results from one of two similar experiments.

To examine further the level of mRNA regulation, mRNA synthesis was blocked using 5,6-dichloro-l-~-~-ribofuranosyl- benzimidazole after ROS 17/2.8 cells were cultured in the presence or absence of TGFB1 for 48 h and mRNA half-life was estimated. As shown in Fig. 4, the half-life was about 10 h and was not altered by TGFBl treatment. On the other hand, the RNA polymerase inhibitor actinomycin D blocked the TGFD effect on mRNA abundance (Fig. 5). Nuclear run- on assays were conducted to estimate directly the TGFB effect on the transcriptional rate of the osteopontin gene. As shown in Fig. 6, TGFPl increased the synthesis of radiolabeled transcripts of the OP gene. This was a selective effect since no significant difference in the transcription rate of the actin gene was observed.

Pilot experiments indicated that the increase in OP mRNA levels in response to TGFB was first detected a t 48 h; the above experiments were therefore conducted at 48 and 72 h. However, in previous studies: we found that the effects of TGFj3 on collagen and alkaline phosphatase mRNA were clearly detectable a t 24 h. To reconcile these observations, we followed the time course of the TGFP-dependent changes in the mRNA levels for fibronectin, type I collagen, and OP in the same RNA preparations for 120 h. Equal amounts of RNA from TGFB1-treated cells and the corresponding controls were hybridized with the three probes. As shown in Fig. 7, the rise in collagen message in response to TGFPl is clearly seen

M. Noda, unpublished data.

1 2

1 2 3

26 - FIG. 2. Effect of TGFB on metabolically labeled proteins in

culture media of ROS 17/2.8 cells. Confluent cells in 9.5-cm2 wells were treated with 4 ng/ml TGFPl or TGFp2 for 48 h in serum- free media. Then, media were changed to 1 ml of fresh serum-free media containing 10 pCi/ml ~-["C]serine for 4 (upper) or 48 (lower) h. Proteins in media aliquots corresponding to 110,000 cells were fractionated on 10% SDS-PAGE. Gels were treated with Enlightening and fluorographed. X-ray films were exposed for 3 days (upper) and 1 day (lower). Lane I , control; lane 2, TGFS1; lane 3, TGFB2. The figures represent the results from one of two similar experiments.

at 24 h and is sustained thereafter. However, the rise in OP mRNA levels is first apparent a t 48 h, and the reduction in fibronectin mRNA levels is obvious only after 72 h. The collagen and OP mRNA basal and stimulated levels also fluctuated with media changes: collagen mRNA increased 24 h and OP mRNA 48 h after exposure to fresh media. These experiments were performed three times with similar results. Moreover, TGFB had a similar effect on OP mRNA in serum- free media (data not shown).

Dexamethasone was known to decrease and 1,25-dihy- droxyvitamin D3 to increase O P mRNA levels (5). The com- bined effects of TGFj31 and each of these modulators on OP mRNA accumulation are shown in Fig. 8. In the presence of saturating concentrations of dexamethasone (100 nM) and TGFBl (4 ng/ml), there was no net change in OP mRNA levels relative to nontreated controls. On the other hand, 1,25- dihydroxyvitamin D3 (10 nM) was a more potent stimulator of O P mRNA accumulation than was TGFb1 (4 ng/ml), and no further effect was observed when cells were treated with both agents. In the same filters, fibronectin mRNA accumu- lation was reduced by TGFPl regardless of the presence of dexamethasone or 1,25-&hydroxyvitamin Ds.

To test the generality of the TGFP effect on osteopontin

TGFP Regulation of Osteopontin Synthesis 13919

TGFPl (ng/ml)

0 .4 .8 4

18S1

TGFp1 (nglml)

o Control TGFD 1

i W I P a

0 0 12 24

HOUR

FIG. 4. Effect of TGFB on half-life of osteopontin mRNA. ROS 17/2.8 cells were cultured for 48 h in the presence or absence of 4 ng/ml TGF@l. 5,6-Dichloro-l-j3-~-ribofuranosylbenzimidazole (25 pg/ml) was then added to culture media, and cytoplasmic RNA was isolated a t 0, 12, and 24 h after the addition of the drug. The cells maintained normal morphology and attachment throughout the drug treatment period. The amounts of osteopontin mRNA were analyzed by Northern blot analysis as described under "Materials and Meth- ods." Density of bands on autoradiograms was measured by densiti- metry. Data are expressed as mean f S.D. from two dishes and represent one of two similar experiments.

C A T T A - ..

0 .4 .8 4 - . . .

28s - 18s *

18SD - FIG. 3. Effect of TGFB on steady-state levels of osteopontin

mRNA. ROS 17/2.8 cells were cultured in the presence or absence of TGFB for 72 h. Northern blots of total RNA .(lo pgllane) for osteopontin cDNA were conducted as described under "Materials and Methods." Upper: 1st lane, control (C) , 2nd and 3rd lanes, 4 ng/ml TGFj3l and TGFB2, respectively. FN, fibronectin. Center, dose de- pendence of TGFj3l effects on osteopontin mRNA levels. TGF@l concentrations are indicated above the respective lanes. Lower, the same filter (center) was probed with rat @-actin cDNA. The figures represent the results from one of three similar experiments.

mRNA in osteoblast-like cells, we examined its effect on the cell line MC3T3E1, which is a nontransformed osteoblast- like cell line derived from neonatal mouse calvariae (35). As shown in Fig. 9, osteopontin mRNA was substantially in- creased in MC3T3E1 cells by TGFBl treatment.

DISCUSSION

The findings presented above show that TGFP stimulates the production of the bone extracellular matrix protein osteo- pontin. TGFB also modulates the production of other extra- cellular matrix proteins, such as collagen and fibronectin (9, 17-19), which participate in cell attachment and may have effects on cellular differentiation. In other mesenchymal cells, TGFP stimulates fibronectin synthesis (9, 17, 19, 44). It was shown, for example, that the effects of TGFP on anchorage-

FIG. 5. Effect of actinomycin D on TGFB-induced accumu- lation of osteopontin mRNA levels. ROS 17/2.8 cells were treated with vehicle (lane C), 4 ng/ml TGFj31 (lane 'I"), 200 ng/ml actinomycin D (lane A ) , or both (lane TA). After 48 h, cytoplasmic RNA was isolated and subjected to Northern blot analysis (10 pgllane) and autoradiography to detect osteopontin mRNA as described under "Materials and Methods." The figure represents the results from one of three similar experiments.

C T

FIG. 6. TGFBl effects on the transcription rate of osteopon- tin gene (nuclear run-on assay). After grown to subconfluency, ROS 17/2.8 cells were rinsed three times with phosphate-buffered saline and subsequently cultured in serum-free medium in the absence (lane C) or presence (lane 7') of 4 ng/ml TGFj3l for 96 h with one change of media a t 48 h. The nuclear run-on assay was conducted as described under "Materials and Methods." Arrowhead I, osteopontin; arrowhead 2, rat actin; arrowhead 3, PBS+ (plasmid vector). The figure represents the results from one of three similar experiments.

13920 TGFP Regulation of Osteopontin Synthesis 0 12 24 48 72 96 120 “””

C T C T C T C T C T C T

FIG. 7. Time course for effects of TGFj3l osteopontin and other extracellular matrix protein mRNAs in ROS 17/2.8 cells. Cells were cultured for the indicated periods of time (in hours) in the absence (lanes C ) or presence (lanes 7‘) of 4 ng/ml TGFfi1. Media were replaced every 48 h. Northern blots of cytoplasmic RNA (10 pgllane) were hybridized to oligo-labeled rat cDNA probes. FN, fibronectin (prlf-1); COL, type I procollagen (a1R2); OP, osteopontin. The figure represents the results from one of three similar experi- ments.

+ 1,25 + DEX (OH),D,

TGFP - + - + - + “ FN c & * - ‘ r n - -

. ... -%- ””%m

COL . 28s *

18s c OP -I

FIG. 8. Effects of dexamethasone (100 nM) and 1,25-dihy- droxyvitamin Ds (10 nM) on TCFBl(4 ng/ml)-induced changes in osteopontin and fibronectin mRNA accumulations. ROS 17/2.8 cells were treated for 72 h with the combination of hormones and TGFfi1 as indicated. Total RNA (10 pgllane) was analyzed by Northern blotting as described under ”Materials and Methods.” The figure represents the results from one of two similar experiments. DEX, dexamethasone; lr25-(0H)&, 1,25-dihydroxyvitamin D3.

18s =

FIG. 9. Effect of TGFj31 on osteopontin mRNA accumula- tion in MC3T3E1 cells. Cells were grown to subconfluence in 10% FBS and then were subsequently cultured in the absence (lane C) or presence (lane 7‘) of TGFPl (2 ng/ml) in 2% FBS for 72 h. Total RNA (10 pg/lane) was analyzed by Northern blotting as described under “Materials and Methods.” The figure represents the results from one of two similar experiments.

independent growth of NRK-49F cells could be neutralized by the Gly-Arg-Gly-Asp-Ser-Pro peptide, which interferes with cellular attachment to fibronectin (17). Interestingly in

this context is the fact that TGFP reduces the fibronectin mRNA in ROS 17/2.8 cells. One could speculate that the attachment protein osteopontin, which contains the Arg-Gly- Asp-Ser sequence, may replace fibronectin and possibly con- vey to the cells a more “bone-specific” attachment. It is also of interest that the change in osteopontin mRNA levels is first detected at 48 h, whereas the increase in collagen message is clearly detectable a t 24 h. Assuming similar detectability of these mRNAs in Northern blots, these differences suggest different mechanisms for the regulation of the respective mRNAs. The fibronectin mRNA actually increased somewhat a t 24 h in response to TGFP in repeated experiments; but due to the limited precision of these estimates, this observation requires further validation. If these observed kinetic differ- ences occur in vivo, they raise interesting questions regarding tissue-specific regulation of gene expression, which could be further investigated in this system.

Stimulation of the expression of osteopontin by TGFP in osteoblast-like cells in concert with increases in the mRNA levels of alkaline phosphatase, type I collagen, and osteonectin (16) support, on one hand, the relation of this new protein to the osteoblastic phenotype and, on the other hand, broaden the phenotype-related effects of TGFj3 in these cells. TGFP also increased the steady-state mRNA level of osteopontin in mouse calvaria-derived osteoblast-like cells (MC3T3E1), sug- gesting that the observations on osteopontin gene expression made in ROS 17/28 cells may apply to osteoblast-like cells in general.

Glucocorticoids and 1,25-dihydroxyvitamin D3 have pro- nounced effects on the shape of ROS 17/2.8 cells, presumably through cytoskeletal involvement. Glucocorticoids cause cell spreading and actin cable formation (45) along with the decrease in osteopontin mRNA (5). On the other hand, 1,25- dihydroxyvitamin DS causes stellar shape changes (46) and increases osteopontin mRNA (5). Glucocorticoid treatment failed to reduce osteopontin mRNA levels when cells were also treated with TGFP1. The molecular basis for this opposite regulation of the osteopontin gene remains to be determined. 1,25-Dihydroxyvitamin D3 (10 nM) increased osteopontin mRNA levels more than TGFPl (4 ng/ml), and the latter had no further effect in the presence of the former. The molecular basis for the similar but nonadditive effect of the two agents on OP mRNA also remains to be elucidated. The reduction of fibronectin mRNA produced by TGFP was not affected by the presence of dexamethasone or 1,25-dihydroxyvitamin D3,

suggesting separate control of the fibronectin gene and point- ing to complex regulation in this multihormonal system.

Very little is known about the mechanism of action of TGFP. It was shown that TGFP increases the abundance of steady-state mRNA levels of actin, fibronectin, procollagen, and tissue inhibitor of metalloproteinases in fibroblasts or fetal rat calvariae (14, 19, 20, 47). We have recently shown that the enhancement of alkaline phosphatase mRNA by TGFP in ROS 17/2.8 cells is most likely due to transcriptional control (16). Our findings indicate that TGFB enhancement of osteopontin mRNA is regulated at least in part through transcriptional control. In analogy to the a2(I) collagen gene, where nuclear factor I-binding sites were found to be activated by TGFj3, similar regulatory sequences may be present in the osteopontin gene. On the other hand, TGFP was reported to increase mRNA levels of collagen and fibronectin in human fibroblasts without affecting the rate of transcription (19). The diverse effects of TGFj3 on growth and differentiation in various cell types suggest multiple, possibly indirect mecha- nisms of action. The time lag between TGFB addition and osteopontin mRNA changes observed in these experiments

TGFP Regulation of Osteopontin Synthesis 13921

and the opposite effects of TGFP on osteopontin and fibro- nectin mRNAs are consistent with indirect action involving additional mediators (transacting factors?). This possibility and other aspects of gene regulation in osteoblastic cells can be further investigated in this cell line. Finally, it remains to be seen to what extent these observations relate to the effects of TGFB on bone metabolism in uiuo.

Acknowledgments-We wish to thank Drs. Harrison and Kream for their advice regarding nuclear transcription experiments. We thank Dr. David Rowe for rat procollagen cDNA clones and Dr. Hynes for a rat fibronectin cDNA clone.

REFERENCES 1. Prince, C. W., Oosawa, T., Butler, W. T., Tomana, M., Bhown,

A. S., Bhown, M., and Schrohenloher, R. E. (1987) J. Biol. Chem. 262,2900-2907

2. Mark, M. P., Prince, C. W., Oosawa, T., Gay, S., Bronckers, A. L. J. J., and Butler, W. T. (1987) J. Histochem. Cytochem. 3 5 , 707-716

3. Mark, M. P., Prince, C. W., Gay, S., Austin, R. L., and Butler, W. T. (1988) Cell Tissue Res. 251,23-30

4. Mark, M. P., Prince, C. W., Gay, S., Austin, R. L., Bhown, M., Finkelman, R. D., and Butler, W. T. (1987) J. Bone Miner. Res.

5. Yoon, K., Buenaga, R., and Rodan, G. A. (1987) Biochem. Bio-

6. Oldberg, A., Franzen, A., and Heinegard, D. (1986) Proc. Natl.

7. Somerman, M. J., Prince, C. W., Sauk, J. J., Foster, R. A., and

8. Prince, C. W., and Butler, W. T. (1987) Collagen Relat. Res. 7 ,

9. Sporn, M. B., Roberts, A. B., Wakefield, L. M., and Assoian, R. K. (1986) Science 233, 532-534

10. Sporn, M. B., Roberts, A. B., Wakefield, L. M., and de Crom- brugghe, B. (1987) J. Cell Biol. 105 , 1039-1045

11. Robey, P. G., Young, M. F., Flanders, K. C., Roche, N. S., Kondaiah, P., Reddi, A. H., Termine, J. D., Sporn, M. B., and Roberts, A. B. (1987) J. Cell Biol. 105,457-463

12. Seyedin, S. M., Thomas, T. C., Thompson, A. Y., Rosen, D. M., and Piez, K. A. (1985) Proc. Natl. Acad. Sci. U. S. A. 82,2267- 2271

13. Hauschka, P. V., Maurakos, A. E., Lafrati, M. D., Doleman, S. E., and Klagsbrun, M. (1986) J. Biol. Chem. 261,12665-12674

14. Centrella, M., McCarthy, T. L., and Canalis, E. (1987) J. Bwl. Chem. 262,2869-2874

15. Seyedin, S. M., Thompson, A. Y., Bentz, H., Rosen, D. M., McPherson, J. M., Conti, A., Siegel, N. R., Gallupi, G. R., and Piez, K. A. (1986) J. Biol. Chem. 2 6 1 , 5693-5695

16. Noda, M., and Rodan, G. A. (1987) J. Cell. Physiol. 133 , 426- 437

17. Ignotz, R. A., and Massagui., J. (1986) J. Bwl. Chem. 261,4337- 4345

18. Roberts, A. B., Sporn, M. B., Assoian, R. K., Smith, J. M., Roche, N. S., Wakefield, L. M., Heine, U. I., Liotta, L. A., Falanga, V.,

2,337-346

phys. Res. Commun. 148, 1129-1136

Acad. Sci. U. S. A. 83,8819-8823

Butler, W. T. (1987) J. Bone Miner. Res. 2,259-265

305-313

Kehrl, J. H., and Fauci, A. S. (1986) Proc. Natl. Acad. Sci. U.

19. Raghow, R., Postlethwaite, A. E., Keski-Oja, J., Moses, H. G., and Kang, A. H. (1987) J. Clin. Znuest. 79, 1285-1288

20. Edwards, D. R., Murphy, G., Reynolds, J. J., Whitham, S. E., Docherty, A. J. P., Angel, P., and Heath, J. K. (1987) EMBO

21. Laiho, M., Saksela, O., Andreasen, P. A., and Keski-Oja, J. (1986) J. Cell Biol. 103 , 2403-2410

22. Cromack, D. T., Sporn, M. B., Roberts, A. B., Merino, M. J., Dart, L. L., and Norton, J. A. (1987) J. Surg. Res. 42,622-628

23. Sporn, M. B., Roberts, A. B., Shull, J. H., Smith, J. M., Ward, J. M., and Sodek, J. (1983) Science 2 1 9 , 1329-1331

24. Mutose, T. A., Pierce, G. F., Thomason, A., Gramates, P., Sporn, M. B., and Devel, T. F. (1987) Science 2 3 7 , 1333-1336

25. Rodan, G. A,, and Rodan, S. B. (1984) in Bone and Mineral Research (Peck, W. A., ed) Vol. 2, pp. 244-285, Elsevier/North- Holland Biomedical Press, Amsterdam

26. Shteyer, A., Gazit, D., Passi-Even, L., Bab, I., Majeska, R. J.,

Znt. 3 9 , 49-54 Gronowicz, G., Lurie, A., and Rodan, G. A. (1986) Calcif. Tissue

27. Majeska, R. J., Rodan, S. B., and Rodan, G. A. (1980) Endocri-

28. Kream, B. E., Row, D., Smith, M. D., Maher, V., and Majeska, R. J. (1986) Endocrinology 199 , 1922-1928

29. Price, P. A., and Baukol, S. A. (1980) J. Biol. Chem. 2 5 5 , 11660- 11663

30. Noda, M., Yoon, K., Thiede, M., Buenaga, B., Weiss, M., Hen- thorn. P.. Harris. H.. and Rodan. G. A. (1987) J. Bone Miner.

S. A. 83,4167-4171

J. 6,1899-1904

nology 107 , 1494-1503

, I

Res. 2,161-164 . .

31. Genovese. C.. Rowe. D.. and Kream. B. (1984) Biochemistrv 23, , I

6210-6216’ , . . ”

32. Schwartzbauer. J. E.. Tamkun. J. W.. Lemischka. I. R., and Hynes, R. 0.’(1983)Cell35,421-431

Biol. 4, 1313-1319 33. Bond, J. F., Robinson, G. S., and Farmer, S. R. (1984) Mol. Cell.

34. Majeska, R. J., Nair, B. C., and Rodan, G. A. (1985) Endocrinology

35. Sudo. H.. Kodama, H.. Amaeai. Y., Yamamoto, S., and Kasai, S. 116,170-179

(1983) J . Cell BiA. 96,19i- i98 . .

36. Laemmli, U. K. (1970) Nature 227,680-685 37. Burnette, W. N. (1981) Anal. Biochem. 112 , 195-203 38. Greenberg, M. E., and Ziff, E. B. (1984) Nature 311,433-438 39. Thomas, P. (1980) Proc. Natl. Acad. Sci. U. S. A. 77, 5201-5205 40. Feinberg, A. P., and Vogelstein, B. (1984) Anal. Biochem. 137 ,

41. Groudine, M., Peretz, M., and Weintraub, H. (1981) Mol. Cell.

42. McKnight, G. S., and Palmiter, R. D. (1979) J. Biol. Chem. 2 5 4 ,

43. Harrison, J., Peterson, D., Lichter, A., Rowe, D., and Kream, B.

44. Ignotz, R. A., Endo, T., and Massagui., J. (1987) J. Biol. Chem.

45. Wiren, K. M., and Rodan, G. A. (1985) in The Chemistry and Biology of Mineralized Tissues (Butler, W. T., ed) pp. 320-325, Ebsco Media Inc., Birmingham, AL

46. Gronowicz. G.. Eean. J. J.. and Rodan. G. A. (1986) J. Bone

266-267

Biol. 1,281-288

9050-9058

(1987) J. Bone Miner. Res. 2, 51, 118 (abstr.)

262,6443-6446

Miner. Res. 1,441-455 ’

Cell. Physiol. 127, 83-88

. .

47. Leof, E. B., Proper, J. A., Getz, M. J., and Moses, H. L. (1986) J.