Altered Cbfa1 expression and biomineralization in an osteosarcoma cell line

ELSEVIER

Journal of Orthopaedic Research

Journal of Orthopaedic Research 22 (2004) 404-410

www.elsevier.com/locate/orthres

Altered Cbfa 1 expression and biomineralization in an osteosarcoma cell line H. Perinpanayagam

a,b,e,*

, G. Schneider a,c, K. Holtman C. Stanford a,cid

a,

R. Zaharias

a,

Doii7.s1nstitute.fbr Dmicrl Rcsrurcli, Collqe of Diwistrj,, Unicwsitjv o f Ioivu, l o i v r i City, I A 52242, U S A Diyrrtnient qf'Enclodontics, Co1lcg.e of Dmitistr.v, University of l o i i w , I n ~ City, a I A 52242, USA Diyxrrtnicnt of' Prosthotlontics, Colli~gcof Dcntislrjb, University of loit~a.Ion,u City, I A 52242, USA " Dc~purtnieiitnf Orllroprtlic Surgery, Colliye of Medicine, Unirersity of' l o i t ~ t r ,Ioiiu City. I A 52242, U S A D i p i r t n i m t of' Pcriotlonlic.? untl Enilotloiitics, School qf Dental Medicini,. St~rleUnirersity of Nen' York ut Buffillo, Buflulo. N Y 14214-3008, USA

'I

'

Received 23 July 2002; accepted 30 July 2003

Abstract

Osteoblast differentiation and expression are regulated by Cbfal transcription factors. Recent evidence suggests that Cbfal may also regulate bone mineralization. The purpose of this study was to characterize Cbfal expression in relation to mineralization in rat UMR 106-01 osteoblastic cell cultures. UMR106-01 BSP cultures consistently form bone-like mineral, whereas the UI subclone mineralize gradually. BSP and UI cultures were grown for 48 h and then treated with P-glycerophosphate. BSP cultures had alizarin red stained calcifications and mineral-like deposits within 24 h of phosphate. Atomic absorption spectroscopy measured significantly ( P < 0,0001) more calcium in the phosphate-treated BSP cultures than in the UI. Cbfal message was detected in the BSP and UI cultures, but the Cbfal N-terminal isoform was deficient in UI and appeared to be up-regulated in the phosphate-treated BSP cultures. Cbfal protein levels were also reduced in the UI. D N A sequence from the RT-PCR products was utilized to design Taqman Real-time RT-PCR reagents. Quantitative Real-time RT-PCR analysis showed that Cbfal mRNA levels relative to endogenous 18 s rRNA were lower in the slower mineralizing UI cultures. Furthermore, the Cbfal N-terminal isoform mRNA levels were significantly ( P < 0.001) lower in the slower mineralizing cultures. Transfection with Cbfal or isoform antisense caused a significant ( P < 0,001) reduction in mineralization. Therefore, Cbfal expression may be associated with bone-like mineral formation in rat UMR106-01 osteoblastic cell cultures. 0 2003 Orthopaedic Research Society. Published by Elsevier Ltd. All rights reserved. K c ~ i r o d s :UMR106-01; Osteoblast; Cbfal; Mineralization: Real-time PCR

Introduction

In vitro bone formation can be studied in the rat osteosarcoma-derived UMR106-01 BSP cell line (which expresses elevated levels of bone sialoprotein) [I 3,151. UMR 106-01 cells have been characterized as differentiated osteoblasts which express the major osteoblast proteins: alkaline phosphatase, osteocalcin, osteopontin, bone sialoprotein and type 1 collagen [ 15,181. Further* Corresponding author. Address: Department of Periodontics and Endodontics, School of Dental Medicine. State University of' New York at Buffalo. Buffalo, NY 14214-3008, USA. Tel.: +I-716-8293847; fax: +I-716-837-7623. E-niiril uclrlims: hp7(ujbuffalo.edu (H. Perinpanayagam).

more, confluent UMR106-01 BSP cultures rapidly initiate bone-like mineral formation within 24 h of phosphate treatment [ 151. This rapid osteoblast-mediated mineralization phenomenon provides a convenient in vitro model to study the processes involved in the initiation of bone mineral formation. In contrast to the rapid mineralization of BSP cultures, a variant subclone has been isolated which has reduced rates of mineral formation relative to the parental cell line [14]. These UI cells (isolated at The University of Iowa) resemble osteoblasts in cell morphology and expression of alkaline phosphatase, osteocalcin, bone sialoprotein, osteopontin and type I collagen [14]. However unlike the rapidly mineralizing parental BSP cell line, UI cultures are slower to mineralize.

0736-0266/$ - see front matter 0 2003 Orthopaedic Research Society. Published by Elsevier Ltd. All rights reserved. doi: 10.1016/S0736-0266(03)00196-7

If. Perinpcinuytiguiii et ul. I Joirrncrl

of OrrliopucJclicRescwcli

Osteoblast differentiation and extracellular matrix synthesis are regulated by the Cbfal transcription factor [6], and osteoblast-mediated bone mineral formation may also be influenced by Cbfal [7]. Furthermore, Cbfa 1 antisense oligonucleotides were found to inhibit the formation of mineral nodules in primary rat osteoblast culture [l]. The purpose of this study was to compare and contrast UMR106-01 BSP and UI cells with regard to Cbfal expression and the initiation of mineralization. The specific aims were to determine if steady state levels of Cbfal mRNA are elevated in the rapidly mineralizing BSP cultures compared to the slower minerdizing UI subclone, and if inhibition of Cbfal expression reduces mineralization. This study found that UMR 106-01 cells provide a useful model to study the events leading to osteoblast-mediated initiation of mineralization. The BSP cells express Cbfal and readily mineralize, whereas the UI cells consistently express 100-fold less of the Cbfal N-terminal isoform and only gradually mineralize. Furthermore, the inhibition of Cbfal expression in BSP cells decreased their mineralization. Therefore, it was concluded that Cbfal expression may facilitate the initiation of matrix mineralization in rat UMR106-01 osteoblastic cell cultures.

Methods

405

N-terminal isoforin were normalized lo 18 s rRNA and calculated by the comparative method of relative quantitation in multiplex reactions (Perkin Elmer). Statistical analysis was by one-way ANOVA ( P < 0.05) and Tukey‘s multiple comparison test. Chfir I

UI1 ti.srns1,

1rili1.sfec1ions

Antisense oligonucleotides to Cbfal [I]. to the isoforin [8] and a scrambled control sequence were chosen (Fig. 5). BSP cells were transiently transfecied with either phosphothioate-modified Cbfa 1, isoform or scrambled control antisense oligonucleotides, by using the Lipofectamine Plus Reagent kit (RRL, Gaithersburg. MD). Each trdnsfect was plated in triplicate cultures, grown for 48 h and treated with phosphate for 24 h. as previously described. The 72 h c~iltiires were analyzed by alizarin red stain and atomic absorption spectroscopy for calcium.

Results

Replicate groups of BSP and UI osteoblasts were grown in triplicate microdot cultures for 48 h and then incubated with ( 5 mM) or without (0 mM) P-glycerophosphate (BGP) for an additional 24 h. Phase contrast light microscopy demonstrated an abundance of mineral-like deposits in the phosphate-treated BSP cultures (Fig. 1A). Mineral deposits were absent from the phosphate-treated UI cultures and from the untreated

B

BG P

A

BGP

5

0

Cell l.ullur<’ UMR106-OI BSP and UI cells were suspended in EMEM and plated as I0 p1 droplets (5 x lo4 cells) in separate tissue culture wells. After 1 h of cell attachment, the wells were flooded with growth media and at 48 h the cultures were treated with either 0 or 5 mM j3-glycerophosphate (BGP) for 24 h. Triplicate 72 h cultures were examined by light microscopy, stained with 2% alizarin r e d 3 and analyzed by atomic absorption spectroscopy for calcium.

22 (2004) 404-410

5

0

BSP

BSP

UI

UI

E.vpression of’ Chjul

Conventional RT-PCR primers were designed to detect total Cbfal message (200 bp) and the N-terminal isoforin (688 bp), and glyceraldehyde 3-phosphate dehydrogenase (GAPDH 345 bp) as a control. Triplicate 72 h cultures were harvested and total cell RNA extracted with the RNeasy Mini Kit (Qiagen, Valencia, CA). The extracts were analyzed by conventional RT-PCR with the advantage one-step RTPCR Kit (Clontech, Palo Alto, CA). The Cbfal amplicons were recovered with the QIAquick Kit (Qiagen) and DNA sequenced. Whole cell lysates were harvested from replicate 72 h cultures. Polyclonal antibody to Cbfal (Alpha Diagnostic International, San Antonio, Tx) was utilized to prepare immunoprecipitates for Western blot analysis.

c

6o

-e

50

0 OBGP

-

SBGP

€ 3 40-

.2 i

z%

30-

=*

20-

0 0

E

10 -

0-

-

BSP

Rid-tinie quun fitution of’ Chjul esprcssion Real-time RT-PCR primers and a probe were designed to the junction of exons 3 and 4 to measure total Cbfal message, and to the junction of exons I and 2 to measure the N-terminal isoform, with Primer Express software (Perkin Elmer). TaqMan ribosomal RNA control reagents (Perkin Elmer) were used to measure 18 s rRNA as a control. Each RNA extract was analyzed in triplicate by quantitative Real-time multiplex RT-PCR with the TaqMan Gold RT-PCR Kit (Perkin Elmer) and an ABI Prism 7700 Sequence Detection System (Perkin Elmer). Relative steady state mRNA levels of Cbfal and the

n-

UI

Fig. I . UMR106-01 BSP cultures rapidly mineralize. (A) Light microscopy showed mineral-like deposits in only the phosphate-treated ( 5 BGP) BSP cultures. Magnification. 4x. (B) Alizarin red staining showed calcifications in only the phosphate-treated (5 BGP) BSP cultures. (C) Atomic absorption spectroscopy demonstrated appreciable quantities of calcium (56.3 2 4.8 nmollculture) in the phosphatetreated (5 BGP) BSP cultures. that were significantly (P < 0.0001) higher than in the phosphate-treated (5 BGP) U1 or untreated (0 BGP) BSP and UI cultures.

BSP and UI cultures. Similarly, alizarin red-S staining demonstrated an abundance of calcifications in the phosphate-treated BSP cultures, that were absent from the phosphate-treated U1 cultures, and the untreated BSP and UI cultures (Fig. 1B). Atomic absorption spectroscopy found appreciable quantities of calcium (56.3 k 4.8 nmol/culture) in the phosphate-treated BSP cultures, that were significantly more ( P < 0.0001) than the background levels that were measured in the phosphate-treated U1 cultures and in the untreated BSP and UI (Fig. IC). Extended culture conditions (7 days) showed that the UI did eventually mineralize, albeit more slowly and to a lesser degree than the rapidly mineralizing BSP cultures (data not shown). BSP und (/I osteohlu.rts e.xpress Chjicl

The conventional RT-PCR analysis of RNA extracts from triplicate phosphate-treated and untreated BSP and UI cultures consistently demonstrated the expression of Cbfal mRNA (Fig. 2A). Cbfal mRNA expression appeared to be similar in the phosphate-treated and untreated BSP and U1 cultures. However, when the RNA extracts were analyzed by conventional RT-PCR for the Cbfal N-terminal iso-

A

BSP

UI

BGP:

5

0

0

BSP UI

Cbfal IgG

65kD

Fig. 3. UMR106-01 UI cells express less Cbfal protein. Western blot analysis of immunoprecipitated Cbfal demonstrated Cbfal protein in BSP and U I cell extracts. The Cbfal protein levels were lower in the UI.

form, differences in expression were detected (Fig. 2B). The conventional RT-PCR analyses were consistently unable to detect appreciable quantities of the isoform mRNA in the phosphate-treated and untreated UI cultures. The isoform was detected in the BSP cultures and appeared to be up-regulated in the phosphate-treated and mineralizing BSP cultures. When these RT-PCR products were purified and sequenced, their Cbfal identities were confirmed. These sequences were found to confirm the partial rat sequence that is known [20] and were highly homologous to the known mouse CBFAl gene [S]. Likewise, Western blot analysis with commercially available polyclonal antibody (Alpha Diagnostic International, San Antonio, Tx) to Cbfal identified Cbfal protein in BSP and in UI whole cell extracts which had been enriched for Cbfal through immunoprecipitation (Fig. 3 ) . The Cbfal protein levels were higher in BSP cells than in UI.

5

BSP e.xpress more Cbful isofortn tlzan UI osteohlusts 345bp

GAPDH

Cbfal

200bp

B

UI BGP:

0

BSP 5

0

5

lsoform

688bp

GAPDH

345bp

Fig. 2. UMR106-01 BSP and U I cells express Cbfal. ( A ) Conventional RT-PCR analysis demonstrated Cbfal m R N A (200 bp) in all of the phosphate-treated ( 5 BGP) and untreated (0 BGP) BSP and U I cultures. Glyceraldehyde 3-phosphate dehydrogenase (345 bp) served as a positive control for the RT-PCR reactions. (B) Conventional RT-PCR analysis demonstrated isoform mRNA (688 bp) in only the phosphatetreated ( 5 BGP) and untreated (0 BGP) BSP cultures.

The rat Cbfal mRNA sequences that were obtained were utilized to design Taqman Real-time RT-PCR primers and probes (Fig. 4) [19,20]. The primers to detect Cbfal mRNA were designed from exons 3 and 4, and the primers to detect the Cbfal N-terminal isoform were from exons 1 and 2 (Fig. 4). The Cbfal primers were designed to span the junction of exons 3 and 4, and to generate a 59 bp PCR product. The isoform primers were designed to span the junction of exons 1 and 2, and to generate an 80 bp PCR product. The probes were designed to overlay the exon junctions, so that they would be unlikely to hybridize to genomic DNA that may have been present. This ensured that if genomic DNA contaminants were present. their PCR amplification was non-contributory to the Real-time analysis. R N A extracts from the replicate groups of phosphate-treated and untreated BSP and UI cultures were analyzed for steady state levels of Cbfal mRNA by quantitative Real-time RT-PCR. The relative steady state levels of Cbfal mRNA were higher in the BSP cultures than in the UI, but these differences were not statistically significant (Fig. 5A). The Cbfal mRNA levels in the phosphate-treated BSP and UI cultures were higher than the untreated BSP and UI cultures,

H

Piwiiprinti~ rr,ytiiii c’t

ti/

I Jour i z [ d of Ortho/m~tlriR~wrrri h 22 12004) 404 410

407

1 ATGCTTCATTCGCCTCACACAACCACAGAACCACAAGTGCGGTGC~CTTTCTCCAGGAGGACAGC~GGAGGCCCTGGTGTTTAAATGGTTAATCT

u

Exon 1 2 101 CTGCAGGTCACTACCAGCCACCGAGACCAACCGAGTCAGTGAGTGCTCTGAGCACAGTCCATGCAGTAATATTTAAGGCTGCAAGCAGTATTTACAACAG I s o f o r m F o r w a r d RT I s o f o r m Probe I s o f o r m Reverse Exon 2 8 3 201 AGGGCACAAGTTCTATCTGGAAAAAAAAGGAGGGGACTATGGCGTC~CAGCCTCTTCAGTGCAGTGACACCGTGTCAGCAAAACTTCTTTTGGGATCCG RT I s o f o m Antisense 301 AGCACCAGCCGGCGCTTCAGCCCCCCCTCCAGCAGCCTGCAGCCCGGCAAGATGAGCGACGTGAGCCCGGTAGTGGCTGCTCAGCAGCAGCAGCAGCAGC

401 AGCAGCAGCAGCAGCAGCAACAGCAGCAACAGCAGCAGCAGCAGCAGCAGCAGCAGCAGCAGCAGCAGCAGGAGGC~GCCGCAGCAGCAGCAGCGGCAGCGGC

501 GGCGGCAGCAGCGGCGGCAGCCGTGCCCCGGTTGAGGCCGCCGCACGACAACCGCACCATGGTGGAGATCATCGCGGACCACCCGGCCGAACTGGTCCGC

u

Exon 3 4 601 ACCGACAGCCCCAACTTCCTGTGCTCCGTGCTGCCCTCGCACTGGCGGTGC~CAAGACCCTGCCCGTGGCCTTCAAGGTTGTAGCCCTCGGAGAGGTAC C b f a l Antisense C b f a l Forward RT C b f a l Probe Cbfal 701 CAGATGGGACTGTGGTTACCGTCATGGCCGGGAATGATGAGAACTACTCTGCCGAGCTACG~TGCCTCCGCTGTTATGAAAAACCAAGTAGCCAGGTT Reverse RT Exon 4 5 801 CAACGATCTGAGATTTGTAGGCCGGAGCGGACGAGGCAAGAGTTTCACTTTGACCATAACGGTCTTCAC~ATCCTCCCCAAGTGGCCACTTACCACAGA

u

u

Exon 5 6 901 GCTATTAAAGTGACAGTGGACGGTCCCCGGGAACCAAGAAGGCACAGACA

Fig. 4. Molecular reagents were designed. Rat Cbfal mRNA sequence was obtained from the sequencing of RT-PCR products. Real-time RT-PCR primers and probes were designed to span the junction of exons 3 and 4 to identify Cbfal. and to the junction of exons 1 and 2 to identify the CbFal N-terminal isoform (isoform). Antisense oligonucleotides to Cbfa I and to the N-terminal isoforin were chosen.

respectively, but these differences were also statistically insignificant. In contrast, the relative levels of the Cbfal N-terminal isoform mRNA in the phosphate-treated and untreated BSP cultures were significantly ( P < 0.001) higher (100-fold) than in the phosphate-treated and untreated UI cultures (Fig. 5B). The isoform mRNA levels in the phosphate-treated BSP and UI cultures were higher than in the untreated BSP and UI cultures, respectively, but these differences were also statistically insignificant.

with the control oligonucleotide and fewer calcifications in the cultures which had been transfected with the Cbfal antisense and the isoform antisense (Fig. 6A). Atomic absorption spectroscopy measured significantly ( P < 0.001) less (23-27%) calcium in the cultures that had been transfected with antisense, than in those transfected with the control oligonucleotide (Fig. 6B). However, both the antisense and control oligonucleotide transfected cultures had appreciable amounts of calcium that were significantly ( P < 0.001) more than in the cultures that did not receive phosphate treatment.

Innhibition of Chful and n?inerolixtion in BSP osteohlusts

Discussion In order to evaluate the role of Cbfal in regulating the mineralization phenotype, transient transfection studies were performed. BSP cells were transfected with either antisense or scrambled control oligonucleotides (Fig. 4). These were grown in replicate groups of triplicate microdot cultures for 48 h and then incubated with P-glycerophosphate (5 mM) for an additional 24 h. Alizarin red staining demonstrated an abundance of calcifications in the cultures which had been transfected

Cellular regulation of the steps leading to the initiation of mineral formation is a key hallmark of the temporo-spatial regulation of mineralization in biological systems. In this study, UMR106-01 BSP cultures rapidly initiated mineralization. Likewise, prior studies in our laboratory found that UMR106-01 BSP cultures treated with organic or inorganic phosphates formed bone-like mineral deposits in the extracellular matrix

408

8

5BGP

2 z

A € lo! 6 OBGP T

2a

0

UI

BSP

B

2 K E

-z

1000-

00 BGP 5 BGP

I -

100-

E

10-

1-

r=5

UI

ESP

Fig. 5. UMR106-01 BSP cells over-express the Cbfal isoform. ( A ) Real-time RT-PCR analysis measured similar levels of Cbfal mRNA in all of the phosphate-treated ( 5 BGP) and untreated (0 BGP) BSP and UI cultures. (B) Real-time RT-PCR analysis measured significantly ( P < 0.001) higher (100-fold) levels of isoform mRNA in the phosphate-treated (5 BGP) and untreated (0 BGP) BSP cultures, than in the phosphate-treated ( 5 BGP) and untreated (0 BGP) U I cultures.

A Control

Cbfal

lsoform

Control

Cbfal

lsoform

Fig. 6. Inhibition of Cbfal expression reduces mineralization. (A) Alizarin red htaining showed that BSP cultures transfected with scrambled oligonucleotides had abundant calcifications, whereas those transfected with Cbfal or isoform antisense were less calcified. (B) Atomic absorption spectroscopy found significantly ( P < 0.001) less calcium in cultures transfected with Cbfal (23%) or isoforin (27%) antisense. than in those transfected with scrambled oligonucleotides.

within 24 h [15]. We found that mineralization was dependent on dose and time, and on the activity of alkaline phosphatase, the transcription of RNA, and the syn-

thesis, processing and secretion of bone-related proteins. These were small needle-like crystals associated with fibrillar extracellular matrix deposits and intracellular spherical structures, with an X-ray diffraction pattern resembling human bone [15]. This pattern resembled the early stages in mineral formation in primary osteoblast cultures [2,4,9] and normal rat bone sections [3], and their chemical composition was bio-apatitic in nature, resembling normal bone in crystal size and calcium phosphate ratio [15]. Furthermore, when UMR cells were experimentally transplanted in animals they formed mineralized bone trabeculae at ectopic sites [I 11. In contrast to the rapidly mineralizing BSP cultures, we found that the UMR106-01 UI cultures were slow to initiate mineralization. Whereas the BSP cultures initiated mineralization within 72 h of phosphate treatment, the UI cultures took an additional 48 h (data not shown). Likewise, prior studies found that the UMR106-01 UI subclone were slower to mineralize than their parental BSP cell line [14]. However, the UI cells did resemble BSP osteoblasts and expressed an osteoblast-like phenotype. They expressed similar levels of collagen and osteocalcin, but 10-fold less bone sialoprotein than the BSP cells. Their level of alkaline phosphatase was lower than in the BSP cells, but was similar to the levels in rat calvarial osteoblasts, and they did completely hydrolyze organo-phosphates within 24 h like the BSP. There may be differences between the BSP and UI cells that account for their dissimilar rates of mineralization. Prior studies have found that the BSP cells synthesize large amounts of a highly sulfated form of bone sialoprotein [12] compared to the UI and other gradually mineralizing osteoblasts. This may at least partially account for their rapid initiation of mineralization, since bone sialoprotein can serve as a protein nucleator of hydroxyapatite crystal formation [lo]. This overexpression of bone sialoprotein may be due to various factors including several regulatory elements that have been identified in the bone sialoprotein promoter [lo]. These include consensus sites for the Cbfal transcription factor, a master regulator of osteoblasts [8]. Indeed, fibroblasts that were forced to express Cbfal, activated their expression of bone sialoprotein and osteocalcin [8]. When Cbfal was specifically inhibited in postnatal and fully differentiated osteoblasts in transgenic mice, there was a decrease in bone formation despite adequate osteoblast numbers [7]. Therefore, Cbfal was proposed as a transcriptional activator of bone formation. Our study also found that UMR106-01 BSP and UI cells consistently expressed Cbfa 1. Conventional RTPCR analysis identified Cbfal mRNA in the phosphatetreated and untreated BSP and UI microdot cultures. Since Cbfal expression is known to be restricted to the mineralized tissue-forming cells that include, chondrocytes, osteoblasts and odontoblasts [ 5 ] , Cbfal expression in UMR106-01 cells is consistent with their bone-like

mineral formation. In fact, expression of the Cbfal transcription factor, the extracellular matrix protein osteocalcin and the formation of mineral are key characteristics of the osteoblast phenotype. Therefore the characterization of Cbfal expression and mineralization in this study in conjunction with the analysis of osteocalcin expression and mineralization in prior studies [ 14,151, serves to confirm the osteoblast phenotype of BSP and U1 cells. Although Cbfal mRNA levels appeared to be similar in BSP and U I cultures, we hypothesized that other aspects of Cbfal expression may differ between the rapidly mineralizing BSP cultures and the slowly mineralizing U1. Recent studies have identified novel Nterminal sequence and several different Cbfa I isoforms [20]. Some studies found that the Cbfal N-terminal sequences may modulate transactivation and heterodimerization, and up-regulate osteoblast genes, whereas others have also claimed that the N-terminal sequence may not be crucial for transactivation [8,16,19]. In this study we found that conventional RT-PCR was able to detect the N-terminal transcript in BSP extracts, but not in the UI. Furthermore, there appeared to be an up-regulation of the N-terminal isoform in phosphatetreated cultures. These results were extended by quantitative Real-time analysis. Message levels for the Cbfal N-terminal isoform in the phosphate-treated and untreated BSP cultures were significantly ( P < 0.001) higher (100-fold) than in the phosphate-treated and untreated UI. These Real-time measurements appear to be the first quantitative data demonstrating the relative abundance of Cbfal mRNA, and the first evidence of differences in Cbfal message levels between osteoblast cell lines that suggests transcriptional regulation. Transcriptional regulation of Cbfa 1 expression in BSP and UI osteoblasts suggests that changes in expression may be observed during culture. In this study, phosphate treatment caused an increase in Cbfal levels. Quantitative Real-time RT-PCR analysis showed that isoform mRNA levels in the phosphate-treated BSP and UI cultures were approximately twice the levels in the untreated BSP and UI cultures, respectively. However the differences between mean values were found to be statistically insignificant. Complimentary to the observation of Cbfal expression and mineral formation in UMR106-01 cultures, is the approach of active intervention. Specific genes can be targeted with complementary antisense oligonucleotides that duplex with mRNA, and cause their degradation, impaired nuclear processing or blocked protein translation [ 171. In this study, antisense oligonucleotides to Cbfal caused a decrease in mineralization. BSP cultures that were transiently transfected with antisense had fewer calcifications and significantly ( P < 0.001) less calcium than those transfected with scrambled control oligonucleotides. When the same antisense oligonucleotide

to the Cbfal isoform were transfected into primary rat osteoblasts (ROS17/2.8), there was a dramatic decrease in collagen, osteocalcin and osteopontin [8]. Likewise, the Cbfal antisense sequence in primary rat osteoblasts (ROS17/2.8) caused a dose response decrease in alkaIine phosphatase, osteocalcin and mineral nodules [I]. Together these studies show that the inhibition of Cbfal expression in osteoblasts leads to decreased expression of osteoblast associated proteins and reduced mineral formation. In summary, the rat osteosarcoma-derived UMR 10601 UI cells express Cbfal and gradually mineralize, whereas the UMR106-01 BSP cells consistently express 100-fold more of the Cbfal N-terminal isoform and rapidly mineralize. Furthermore, the inhibition of Cbfal expression in BSP cells decreases their mineralization. These findings suggest that altered expression of Cbfal may partially account for the differences in biomineralization in UMR106-01 BSP and UI subclones.

Acknowledgements Thanks to Dr. Kevin Knudtson and the University of Iowa DNA Facility for guidance with the Real-time PCR analyses, and Xian-Jin Xie for guidance with the statistical analyses. This work was supported by the National Institutes of Health: K16-DE00175-14 (HP), P30-DE I0 126-10 (GS), P60-DE 13076 (GS); University of Iowa: Roy J. Carver Charitable Trust (CS); and State University of New York at Buffalo (HP).

References [I] Banerjee C, McCabe LR, Choi JY, Hiebert SW, Stein JL, Stein GS. et al. Runt homology domain proteins in osteoblast differentiation: AML3KBFA I is a major component of a bone-specific complex. J Cell Biochem 1997;66:1-8. "4 Bellows CG, Aubin JE. Heersche J N , Antosz ME. Mineralized bone nodules formed in vitro from enzymatically released rat calvaria cell populations. Calc Tissue Int 1986;38:143-54. [3] Bianco P, Riminucci M. Silvestrini G, Bonucci E, Termine JD, Fisher LW, et al. Localization of bone sialoprotein (BSP) to golgi and post-golgi secretory structures in osteoblasts and to discrete sites in early bone matrix. J Histochem Cytochem 1993:41:193203. [4] Casser-Bette M, Murray AB. Closs El. Erfle V, Schmidt J. Bone formation by osteoblast-like cells in a three-dimensional cell culture. Calc Tissue Int 1990;46:46 56. [5] D'Souza R N , Aberg T. Gaikwad J, Cavender A, Owen M, Karsenty G. et al. Cbfal is required for epithelial-mesenchymal interactions regulating tooth development in mice. Development 1999;126:2911-20. [6] Ducy P. CBFAI: a molecular switch in osteoblast biology [in process citation]. Dev Dyn 2000:219:461-7 1. [7] Ducy P. Starbuck M. Priemel M, Shen J , Pinero G, Geoffroy V, et al. A Cbfal -dependent genetic pathway controls bone formation beyond embryonic development. Genes Dev 1999:13: 1025-36.

[8] Ducy P, Zhang R, Geoffroy V, Ridall AL, Karsenty G. Osf21 Cbfal: a transcriptional activator of osteoblast dilferentiation. Cell 1997:89:747- 54. [9] Ecarot-Charrier B, Shepard N, Charette G , Grynpas M, Glorieux FH. Mineralization in osteoblast cultures: a light and electron microscopic study. Bone 1988:9:147-54. [lo] Ganss B, Kim RH, Sodek J. Bone sialoprotein. Crit Rev Oral Biol Med 1999:10:79-98. [Ill Martin TJ, Ingleton PM, Coulton LA, Melick RA. Metabolic properties of hormonally responsive osteogenic sarcoma cells. Clin Orthop 1979247-54. [I21 Midura RJ, McQuillan DJ. Benham KJ, Fisher LW, Hascall VC. A rat osteogenic cell line (UMR 106-01) synthesizes a highly sulfated form of bone sialoprotein. J Biol Chem 1990:265:5285-91. [I31 Schneider GB, Zaharias R, Stanford C. Osteoblast integrin adhesion and signaling regulate mineralization. J Dent Res 2001;80:15404. [I41 Stanford C. Holtman K, Martin J, Stevens J, Midura RJ. Loss of BSP expression correlates with loss of biomineralization. J Dent Res 1996;75:910.

[IS] Stanford CM, Jacobson PA, Eanes ED, Lenibke LA, Midura RJ. Rapidly forming apatitic mineral in an osteoblastic cell line (UMR 106-01 BSP). J Biol Chem 1995;270:9420-8. [I61 Thirunavukkarasu K. Mahajan M, McLarren KW, Stifani S, Karsenty G . Two domains unique to osteoblast-specific transcription factor Osf2/Cbfal contribute to its transactivation function and its inability to heterodimerize with Cbfbeta. Mol Cell Biol 1998; 18:4197-208. [I71 van der Krol AR, Mol JN, Stuitje AR. Modulation of eukaryotic gene expression by complementary RNA or DNA sequences. Biotechniques 1988;6:958-76. [I81 Wang A, Martin JA. Lembke LA. Midura RJ. Reversible suppression of in vitro biomineralization by activation of protein kinase A. J Biol Chem 2000:275: I 1082-9 1 . [I91 Xiao ZS, Hinson TK, Quarles LD. Cbfal isoform overexpression upregulates osteocalcin gene expression in non-osteoblastic and pre-osteoblastic cells. J Cell Biochem 1999:74:596- 605. [20] Xiao Z S . Thomas R, Hinson TK. Quarks LD. Genomic structure and isoform expression of the mouse, rat and human CbfaI/Osf2 transcription factor. Gene 1998;214187-97.