145

Vol. 30 No. 3

The Journal of Toxicological Sciences,Vol.30, No.3, 145-156, 2005

PHENYTOIN STIMULATES CHONDROGENIC DIFFERENTIATION INMOUSE CLONAL CHONDROGENIC EC CELLS, ATDC5

Akinobu OKADA*, Teruaki SHIOMI*, Yoshinobu AOKI and Michio FUJIWARA

Safety Research Laboratories, Yamanouchi Pharmaceutical Co., Ltd.,

1-1-8 Azusawa, Itabashi-ku, Tokyo 174-8511, Japan

(Received December 21, 2004; Accepted March 13, 2005)

ABSTRACT — Phenytoin (DPH) is known to affect bone formation. However, the mechanism of thiseffect has not been well understood. In this study, we evaluated the effects of DPH on cartilage formationin a model system using ATDC5 cells, a clonal murine chondrogenic cell line. Alcian blue staining forcartilage nodules and real-time reverse-transcription polymerase chain reaction for the expression of genesencoding type II collagen, aggrecan, transforming growth factor (TGF)-β1, bone morphogenetic protein(BMP)-4, parathyroid hormone-related peptide (PTHrP), indian hedgehog (Ihh), and patched (Ptc) wereperformed in ATDC5 cells cultured with DPH. The ATDC5 cells demonstrated enhanced cartilage for-mation in cultures with DPH. During promoted chondrogenic differentiation, it was observed that DPHincreased the mRNA expression of TGF-β1, BMP-4, Ihh, and Ptc, in a dose-dependent manner on Days5 to 15. In contrast, other antiepileptic drugs, phenobarbital and valproic acid had no effects on chondro-genesis in ATDC5 cells and osteogenesis in MC3T3-E1 cells. Our results provide fundamental evidencethat DPH has a direct stimulatory effect on cartilage formation by regulating TGF-β and hedgehog sig-naling molecules, and that DPH effect on bone formation, including chondrogenesis and osteogenesis, isdistinct from other antiepileptic drugs as suggested in clinical settings.

KEY WORDS: Antiepileptic drugs, ATDC5, Chondrogenesis, EC cells, Phenytoin

INTRODUCTION

Phenytoin (diphenylhydantoin, DPH) is a widelyused antiepileptic drug, but has been found to affectbone metabolism during long-term use. In patientsreceiving chronic DPH therapy, increased thicknessand density of maxillary and calvarial bones has beenshown (Kattan, 1970; Lefebvre et al., 1972; Johnson,1984). Sasaki et al. (1999) reported a clinical case ofexcessive exostosis that had formed in a patient chron-ically treated with DPH. In addition, there have beenmany in vivo and in vitro studies which suggest thatDPH has direct positive effects on osteogenesis. In rab-bits, the healing of experimentally produced fracturesof mandibular bone was accelerated by DPH (Sklans etal., 1967). In human culture cells, DPH has osteoblasticeffects at micromolar concentrations (Lau et al., 1995).Moreover, DPH increased the osteoblast number andosteoid thickness, as well as the volume of bone tissuesin rats (Ohta et al., 1995). However, bone formation is

a complex process including osteogenesis and chon-drogenesis, and the effect of DPH on chondrogenesishas not yet been elucidated. The evaluation of DPHeffects on chondrogenesis provides fundamental infor-mation for understanding the effect of DPH on boneformation in clinical settings.

Chondrocytes undergo coordinated proliferation,differentiation and apoptosis to produce a cartilagescaffold that is mineralized during new bone formation.The process of chondrocyte differentiation is regulatedby glucocorticoids (GCs) and thyroid hormones (THs).In growing rats treated with corticosterone, the growthplate was reduced, which was attributed to impairedchondrocyte proliferation and increased hypertrophicchondrocyte apoptosis (Silvestrini et al., 2000). It wasreported that hypothyroid growth arrest results fromdisorganization of a growth plate in which there is a rel-ative failure of hypertrophic chondrocyte differentia-tion as well as production of an abnormal cartilagematrix (Stevens et al., 2000). These findings clearly

Correspondence: Michio FUJIWARA

* Two of the authors contributed equally to this work.

Present address: Drug Safety Research Laboratories, Astellas Pharma Inc., 2-1-6 Kashima, Yodogawa-ku,Osaka 532-8514, Japan

146

A. OKADA et al.

Vol. 30 No. 3

demonstrate that GCs and THs have opposite actionson chondrogenic proliferation and differentiation invivo. In addition, DPH is known to have teratogenicpotency inducing cleft palate in humans and laboratoryanimals (Dansky and Finnell, 1991; Finnell and Dan-sky, 1991), as GCs do. (Goldman et al., 1978). Kat-sumata et al. (1982) demonstrated that DPH and GCsshare a common receptor which is responsible for theteratogenic effects of these chemicals. This DPHmechanism has been supported by later experiments(Goldman, 1984; Katsuyama et al., 1985; Kay et al.,1990). The significant interaction of DPH with THreceptor (TR) was also reported previously (Franklynet al., 1985). These findings imply, in terms of themolecular mechanism, that DPH may act in two waysto inhibit or stimulate chondrocyte differentiation,through GC receptor (GR) or TR signaling, respec-tively. Thus it is of great interest to identify whetherDPH has stimulatory or inhibitory effects on chondro-genesis.

In this study, we employed the clonal murine cellline ATDC5. This line of embryonal carcinoma (EC)cells is isolated from the feeder-independent teratocar-cinoma stem cell line AT805 (Atsumi et al., 1990), andis well established as an in vitro model for the differen-tiation of chondrocytes (Shukunami et al., 1997). Inthe presence of insulin, there appeared to be areas ofthe condensation in the culture from which proliferat-ing chondrocytes were generated to form cartilagenodules. The progressive expression of type II collagenand aggrecan mRNA was initiated along with conden-sation and the subsequent growth of the nodules(Shukunami et al., 1997). The differentiation ofATDC5 cells is coordinately regulated by the mole-cules fundamental for the normal chondrogenesis invivo. Previous studies demonstrated that oppositeinhibitory and stimulatory effects were exhibited inATDC5 cells treated with GCs and THs, respectively,as in vivo (Miura et al., 2002; Siebler et al., 2002), sug-gesting that the evaluation for determining the chon-drogenic effects of DPH in this cell line was accurate.Moreover, in these cells, the supplementation of Indianhedgehog (Ihh) stimulated chondrogenic differentia-tion (Akiyama et al., 1999; Enomoto-Iwamoto et al.,2000), and parathyroid hormone-related peptide(PTHrP) inhibited cellular condensation and the subse-quent formation of cartilage nodules (Shukunami etal., 1996). The progression of chondrogenic differenti-ation of ATDC5 cells also involves autocrine signalingby transforming growth factor (TGF)-β and bone mor-phogenetic protein (BMP)-4 (Kawai et al., 1999;

Shukunami et al., 2000).Firstly, we evaluated the direct effects of DPH on

chondrogenic differentiation in ATDC5 cells. Next, tounderstand the molecular mechanism of DPH-affectedchondrogenesis, the expression of TGF-β1, BMP-4,Ihh, patched (Ptc, a Ihh receptor), and PTHrP wereinvestigated by quantitative real-time reverse-tran-scription polymerase chain reaction (qRT-PCR). Addi-tionally, effects of other widely used antiepilepticdrugs, phenobarbital (PBT) and valproic acid (VPA),on chondrogenesis and osteogenesis were examined onATDC5 and MC3T3-E1 cells, respectively. Since PBTand VPA have toxicological and teratological effectssimilar to DPH, the comparison of the effect on boneformation between these chemicals may be critical tounderstanding the effects of DPH and its mechanism inbone formation. The results obtained herein show thatDPH directly promoted not only osteogenic differenti-ation, but also chondrogenic differentiation. PBT andVPA had no effect on either the osteogenic or the chon-drogenic differentiation processes in vitro, confirmingthe distinctness of the effect of DPH on bone formationin clinical settings.

MATERIALS AND METHODS

ReagentsAlpha-modified minimum essential medium (α-

MEM), a 1:1 mixture of Dulbecco’s modified Eagle’sand Ham’s F12 medium (DMEM/F12), and fetalbovine serum (FBS) were purchased from InvitrogenCorp. (Carlsbad, CA). β-Glycerophosphate, bovineinsulin (I), human transferrin (T), and sodium selenite(S) were obtained form Sigma Chemical Co. (St.Louis, MO). 5,5-Diphenylhydantoin sodium salt(DPH) was from Sigma. Both phenobarbital sodium(PBT) and sodium valproate (VPA) were purchasedfrom Wako Pure Chemical Industries, Ltd. (Osaka,Japan).

Cell culture and treatment with antiepilepticsATDC5 and MC3T3-E1 cells were purchased

from RIKEN cell bank (Ibaraki, Japan). Stock cultureswere grown in growth media of DMEM/F12 contain-ing 5% FBS or α-MEM containing 10% FBS with 50µg/mL ascorbic acid for ATDC5 or MC3T3-E1 cells,respectively, at 37°C in a humidified atmosphere of 5%CO2 in air. After purchasing, cells were not usedbeyond passage 10, and were passed every 3 days.They were plated at an initial density of 1.0 × 104 cells/cm2 in culture dishes. For ATDC5, after a 48 hr precul-

Phenytoin in cartilage formation.

147

Vol. 30 No. 3

ture period, the medium was changed to the growthmedium which included ITS (10 µg/mL insulin, 10 µg/mL transferrin and 3 × 10−8 M sodium selenite) with orwithout each concentration of antiepileptics (cultureday 0). While for MC3T3-E1, after a 48-hr precultureperiod, the medium was changed to the growthmedium with 10 mM β-glycerophosphate and eachconcentration of antiepileptics (culture day 0). Themedium was replaced every other day. Duplicate ortriplicate cultures were used for each experiment. Theselected concentrations for the antiepileptics were 1,10, and 100 µM of DPH, 10 and 100 µM of PBT, and100 and 1,000 µM of VPA, which showed low cytotox-icity in these cells (data not shown).

Alcian blue and Alizarin red S stainingsATDC5 and MC3T3-E1 cells were plated in 6-

multiwell plates (Corning Incorporated, Corning, NY)and cultured. At each time point, cells were rinsed withphosphate-buffered saline, and fixed on ice with 99.5%methanol for 2 min or 70% ethanol for 15 min for alu-cian blue or alizarin red stainings, respectively. FixedATDC5 and MC3T3-E1 cells were then stained with0.1% alcian blue 8 GX (Sigma) for 2 hr and 1%alizarin red S (Kanto Kagaku Co., Inc., Tokyo, Japan)for 5 min, respectively, at room temperature. The plateswere washed with tap water, air-dried, and photo-graphed.

Assay for Alkaline Phosphatase (ALPase) activityMC3T3-E1 cells were seeded in 24-multiwell

plates (Asahi Techno Glass Corp., Tokyo, Japan) andcultured. At each time point, the activity of ALPasewas measured by using p-nitrophenyl phosphate(pNPP) as a substrate. Cell layers were homogenizedon ice in 0.9% NaCl containing 0.2% Triton X-100,and then centrifuged. The supernatant was assayed in a

reaction mixture of 0.5 mM pNPP and 0.5 mM MgCl2.The reaction was carried out for 15 min at 37°C, andthe concentration of p-nitrophenol generated was mea-sured at 415 nm using a SPECTRAmax 250 Micro-plate Spectrophotometer (Molecular Devices Corp.,Sunnyvale, CA). The serial dilutions of p-nitrophenol(Sigma) were also measured as the standard. Proteincontent in the cell layer was determined by the Brad-ford method using a Protein Assay Kit (Bio-Rad Labo-ratories, Hercules, CA). Values of ALPase activitywere standardized by protein content, and expressed asnmol/µg protein/hour.

Total RNA preparation and qRT-PCRATDC5 cells were cultured in 6 cm dishes (Corn-

ing). Procedures for total RNA preparation and qRT-PCR were described previously (Okada et al., 2002).Primer sequences used in this study are shown in Table1. Quantitative real-time PCR was carried out in anABI Prism 7900 Sequence Detector (Applied Biosys-tems, Foster, CA). The number of template copiespresent at the start of the reaction was determined bycomparison to a standard scale prepared from mousegenomic DNA. The expression level of each targetgene was calculated by standardizing the target genecopy number with the glyceraldehyde-3-phosphate-dehydrogenase (GAPDH) copy number in a sample.The analysis of the results is based on triplicate sam-ples from two independent experiments.

Statistical analysisStatistical analysis was carried out using Dun-

can’s multiple comparison test for qRT-PCR as well asALPase activity studies. Data are reported as the mean± SD, and are considered significantly different atp<0.05.

Table 1. Primer sequences used in this study.PositionGenBankaReverse primer (5’ - 3’)Forward primer (5’ - 3’)Target

31294-31488M65161ACTGCGGTTGGAAAGTGTTTGAAGACCGTCATCGAGTACCGAType II collagen6160-6270L07049GGTGCCCTTTTTACACGTGAAAACTTCTTTGCCACCGGAGAAggrecan1223-1326M13177GGTCCCAGACAGAAGTTGGCAAGAACTGCTGTGTGCGGCTGF-β1521-621X56848GCCGGTAAAGATCCCTCATGGGACTTCGAGGCGACACTTCBmp-4297-388M60056TTCCCGGGTGTCTTGAGTGCGGTTTGGGTCAGACGATGPTHrP2106-2207BC046984CAAAGGCTCAGGAGGCTGGATGGACTCATTGCCTCCCAGAIhh3168-3270U46155AAAGGCCAAAGCCACGTGGATGGGCCTCATTGGGATCPtc43-299M32599ATTTGATGTTAGTGGGGTCTCGCAAAATGGTGAAGGTCGGTGTGGAPDH

a: GenBank accession number.

148

A. OKADA et al.

Vol. 30 No. 3

RESULTS

Effects of DPH on cartilage formation in ATDC5cells

To evaluate chondrogenic potency of DPH,ATDC5 cells were cultured both with and withoutDPH. Two days after inoculation, ATDC5 cells rapidlyproliferated to form a confluent monolayer. At thispoint, they remained undifferentiated with a fibroblas-tic morphology. After reaching confluency, insulin andeach concentration of DPH were added to the culturemedium. As shown in Photo 1, without DPH, ATDC5cells showed chondrogenic differentiation through thecondensation stage to form cartilage nodules. The sizeand number of cartilage nodules increased with eachday of culture. DPH stimulated cell condensation andincreased the formation of cartilage nodules in a dose-dependent manner. Changes in cartilage nodule forma-

tion were verified by staining ATDC5 cells with alcianblue, as shown in Photo 2. No alcian blue staining wasseen on Day 6 for any culture group. In the absence ofDPH on Day 12, spotty staining of cartilage nodulesappeared in the center area of the culture. The degreeof staining then spread out through the dish over timeand reached its highest intensity on Day 20 (Photo 2).On Day 12, DPH had no influence on staining com-pared to the control (not shown). However, on Day 20,an apparent increase in the number of stained spots wasnoted in ATDC5 cells cultured with 100 µM DPH(Photo 2). Thus, DPH at 100 µM stimulates chondro-genic differentiation to form cartilage nodules inATDC5 cells.

Effects of DPH on chondrogenic marker genes inATDC5 cells

A number of extracellular matrix (ECM) marker

Photo 1. Phase contrast micrographs of ATDC5 cells with and without DPH. In the absence of DPH (Control),ATDC5 cells are at the undifferentiated condensation stage on Day 5, and typical cartilage nodules areformed by Day 15. The bar represents 200 µm.

Phenytoin in cartilage formation.

149

Vol. 30 No. 3

molecules, including type II collagen and aggrecan, arewell characterized during chondrogenesis in vivo andin ATDC5 cells (Shukunami et al., 1997). To supportthe stimulatory effect of DPH on the chondrogenic dif-ferentiation, the expressions of type II collagen andaggrecan mRNAs were measured using qRT-PCR inATDC5 cells cultured with DPH at 1, 10, and 100 µMon Days 5, 10, and 15. Both mRNAs were expressed onDay 5, and expression increased with each culture day(Fig. 1). On Day 5, expression levels of type II collagentreated with DPH were comparable to the control at 1and 10 µM, but significantly higher at 100 µM. In theabsence of DPH, type II collagen was markedlyincreased at Day 10 and slightly decreased at Day 15.DPH induced the higher expression of type II collagendose-dependently on Days 10 and 15 at 10 and 100µM. While, in the control cells, aggrecan mRNAincreased continuously from Days 5 to 15. At 10 and100 µM, an increase in aggrecan was detected whencompared to the control on Days 10 and 15 (Fig. 1).These results taken together indicate that DPH inducestype II collagen and aggrecan expressions in ATDC5cells in a dose-dependent manner. Thus, changes in the

expression of ECM molecules strongly confirm theDPH effects in promoting the chondrogenic differenti-ation of ATDC5 cells.

Effects of DPH on the expression of TGF-ββββ1, BMP-4, PTHrP, Ihh, and Ptc mRNAs during the chon-drogenic differentiation of ATDC5 cells

In order to understand the mechanism of DPHeffects on chondrogenesis, we next examined theexpression of signaling molecules important for chon-drogenesis in ATDC5 cells. TGF-β and hedgehog sig-naling cascades were selected to examine the effect ofDPH in ATDC5 cells. These cascades are fundamentalfor normal chondrogenesis both in vivo and in ATDC5cells, and, therefore, their alteration in ATDC5 cellsmay be clear evidence for demonstrating DPH effectsin clinical settings. In control cells, TGF-β1 mRNAwas detected on Day 5 and increased on Day10, mean-while, BMP-4 was expressed on Day 5 and maintainedto Day 15. In the presence of DPH, significantly higherexpressions of TGF-β1 were exhibited throughoutDays 5 to 15 in ATDC5 cells exposed to 100 µM con-centrations (Fig. 2). Similarly, the expression of BMP-

Photo 2. Effects of DPH on the formation of cartilage nod-ules. ATDC5 cells cultured in 6-multiwell plateswith and without DPH were stained with 0.1% alcianblue on Days 12 (control) and 20 (control and DPH).The figure represents a typical well at each timepoint and concentration.

150

A. OKADA et al.

Vol. 30 No. 3

4 was induced by treatment with 10 and 100 µM ofDPH on Day 10 (Fig. 2). According to hedgehog sig-naling, an elevation of Ihh and Ptc occurred on cultureDays 5 to 15 in cells cultured without DPH. At 10 and/or 100 µM, increased expressions in both Ihh and Ptcwere observed on Day 10 or 15 (Fig. 3). These findingsindicate that DPH promotes chondrogenic differentia-tion of ATDC5 cells by stimulating TGF-β and hedge-hog signaling molecules. There was no marked changein PTHrP expression in cells exposed to any concentra-tion of DPH (Fig. 3).

Effects of VPA and PBT on cartilage formation andmineralization

PBT and VPA are other antiepileptic drugs widely

used in clinical settings. In addition to their pharmaco-logical potency, these chemicals are known to cause avariety of toxicological and teratological effects similarto DPH in humans and laboratory animals (Dansky andFinnell, 1991; Finnell and Dansky, 1991). Since theosteogenic and chondrogenic potencies of PBT andVPA have not yet been reported in vivo and in vitro, theevaluation of the effect of these chemicals may provideinformation important for understanding the similaritiesor differences in the effect mechanisms of antiepilepticdrugs on adult and fetal bone formations. Thus thealcian blue staining and qRT-PCR for type II collagenand aggrecan were performed in ATDC5 cells cultured

Fig. 1. Effects of DPH on the expression of type II collagenand aggrecan. The level of mRNA expression for typeII collagen (upper) and aggrecan (lower) was deter-mined for ATDC5 cells cultured with DPH on Days5, 10, and 15. The data represents the mean ± SDfrom triplicate samples in 2 independent experiments.ap<0.05 and bp<0.01 vs. control.

Fig. 2. Effects of DPH on the expression of TGF-β1 andBMP-4. The level of mRNA expression for TGF-β1(upper) and BMP-4 (lower) was determined forATDC5 cells cultured with DPH on Days 5, 10, and15. The data represents the mean ± SD from triplicatesamples in 2 independent experiments. ap<0.05 andbp<0.01 vs. control.

Phenytoin in cartilage formation.

151

Vol. 30 No. 3

with 10 and 100 µM PBT, or 100 and 1,000 µM VPA.Moreover, to compare the effects on mineralization withthat of DPH, alizarin red S staining for mineralized nod-ules and the measurement of ALPase activity were per-formed on MC3T3-E1 cells treated with and withoutDPH, PBT, or VPA. The clonal murine calvarial cell lineMC3T3-E1 has been well used as an in vitro model forosteoblast differentiation and maturation in the presenceof β-glycerophosphate (Kodama et al., 1981; Sudo etal., 1983).

In ATDC5 cells, no differences in the formationof cartilage nodules or their staining with alcian bluewas observed in the presence or absence of PBT orVPA at any concentration (Fig. 4). Both type II col-lagen as well as aggrecan mRNA expression werecomparable in ATDC5 cells cultured with PBT or VPA

Fig. 3. Effects of DPH on the expression of PTHrP, Ihh, andPtc. The level of mRNA expression for PTHrP(upper), Ihh (middle), and Ptc (lower) was deter-mined for ATDC5 cells cultured with DPH on Days5, 10, and 15. The data represents the mean ± SDfrom triplicate samples in 2 independent experiments.ap<0.05 and bp<0.01 vs. control.

Fig. 4. Effects of PBT and VPA on the chondrogenesis ofATDC5 cells. ATDC5 cells cultured with or withoutPBT (10 and 100 µM) or VPA (100 and 1,000 µM)were stained with 0.1% alcian blue on Day 20(upper). The figure represents a typical well at eachtime point. The lower columns show the level ofmRNA expression for type II collagen and aggrecanin ATDC5 cells cultured with PBT or VPA on Day15. The data represent the mean ± SD from triplicatesamples in 2 independent experiments. bp<0.01 vs.control.

152

A. OKADA et al.

Vol. 30 No. 3

as compared to the control on Day 15 (Fig. 4), confirm-ing that PBT and VPA have no direct effect on thechondrogenic differentiation of ATDC5 cells. Mineral-ized nodules stained with alizarin red S were absent inall groups including the control on Day 10, and thenappeared moderately on Day 20 in the control for the

MC3T3-E1 cells (Photo 3). However, DPH at 100 µMmarkedly increased both the number and the area of thestained spots in the dish in which the MC3T3-E1 cellshad been cultured for 20 days (Photo 3). Treatmentwith DPH at 10 and 100 µM also induced ALPaseactivity in a dose-dependent manner on Day 7, when

Photo 3. Effects of DPH, PBT, and VPA on the formation ofmineralized nodules. MC3T3-E1 cells cultured withDPH, PBT, or VPA were stained with 1% alizarin redS on Days 10 (control) and 20 (control and others).The figure represents a typical well at each time pointand concentration.

Phenytoin in cartilage formation.

153

Vol. 30 No. 3

enzyme activity began to increase in the untreated-cul-ture (Fig. 5, Kodama et al., 1981). In contrast, noeffects of PBT and VPA were observed on the forma-tion of mineralized nodules and ALPase activity inMC3T3-E1 cells (Photo 3 and Fig. 5). These findingsclearly demonstrate that the stimulatory effects of DPHon chondrogenesis and osteogenesis are distinct fromthat of PBT and VPA in the systems used in this study.

DISCUSSION

Much in vivo and vitro evidence indicates thatDPH has a promoting effect on bone formation, espe-cially osteogenesis, in clinical and laboratory settings.However, the effect of DPH on chondrogenesis has notbeen well elucidated, and the molecules responsiblefor the DPH effect remain unknown. This study is aninitial report to demonstrate that DPH has a stimula-tory effect on chondrogenic differentiation in ATDC5cells as well as osteogenic differentiation in MC3T3-E1 cells. This study also demonstrates that the effect ofDPH on bone formation is distinct from that of PBTand VPA. Importantly, our findings show that theenhancement of the expression of signaling molecules

such as TGF-β1, BMP-4, Ihh, and Ptc occurs duringchondrogenesis promoted by DPH. Thus this studyprovides fundamental information for understandingthe effects of DPH treatment on bone formation in aclinical setting.

ATDC5 and MC3T3-E1 cells are well character-ized in in vitro models and have been used for investi-gating the molecular mechanisms of chondrogenic andosteogenic differentiation, respectively. In this study,we showed that DPH stimulated the osteogenic differ-entiation of MC3T3-E1 cells at concentrations of 10-100 µM. This effect is in agreement with previousstudies in rats treated intraperitoneally at 1-5 mg/kg/day (serum DPH levels were not determined) (Ohta etal,, 1995), fetal rat calvaria cells at 12.5-200 µM(Ikedo et al., 1999), and human osteoblasts at 5-10 µM(Lau et al., 1995). Thus it has been confirmed thatDPH is an apparent osteogenic compound in vivo andin vitro at micromolar concentrations. It is of interestthat the induction of chondrogenic differentiation hasbeen shown here in ATDC5 cells cultured with DPH atconcentrations of 10-100 µM ATDC5 showed anincreased-formation of cartilage nodule in a dose-dependent manner with DPH on Day 20, as shown inPhotos 1 and 2. However, the expression of chondro-genic markers was a more sensitive indicator than thealcian blue staining for cartilage nodules, as expressionincreased for both type II collagen and aggrecanmRNA, starting from Days 5 and 10, respectively.Later effects on aggrecan expression reflected a normalsequential expression of these markers, suggesting thatDPH accelerates the differentiation process in ATDC5cells. Following these stages, ATDC5 cells transited tothe mineralized stage by changing the CO2 concentra-tion to 3% on Day 21 (Shukunami et al., 1997). How-ever, we focused on the effects of DPH on the earlyphases of the ATDC5 cells, and the later effect on min-eralization was studied in MC3T3-E1 cells. Takentogether, these findings clearly show that DPH at 10-100 µM directly stimulates both the osteogenic andchondrogenic processes during bone formation invitro. The recommended therapeutic range of serumDPH concentration for treating epileptic patients isbetween 10 and 20 µg/ml (equivalent to 36.5-73 µM)(Levy, 1980), indicating that the in vitro effectobserved in this study could be clinically relevant.

Concerning the chondrogenic signaling molecule,the stimulatory effect of DPH was detected in theexpression of TGF-β1, BMP-4, Ihh, and Ptc duringchondrogenic differentiation in ATDC5 cells. Previousstudies demonstrated that autocrine signaling by TGF-β

Fig. 5. Effects of DPH, PBT, and VPA on ALPase activityduring osteogenesis in MC3T3-E1 cells. MC3T3-E1cells were cultured with DPH, PBT, or VPA. Theactivity of ALPase was measured on Day 7. The datarepresents the mean ± SD from triplicate samples in 3independent experiments. ap<0.05 and bp<0.01 vs.control.

154

A. OKADA et al.

Vol. 30 No. 3

and BMP-4 are involved in the progression of earlychondrogenic differentiation, including initiation andconversion of cell condensation, in ATDC5 cells (Kawaiet al., 1999; Shukunami et al., 2000). Thus the inductionof TGF-β1 and BMP-4 may be responsible for the earlydifferentiation promoted by DPH treatment. EspeciallyBMP-4 is expressed highly in early culture stages (Days3 to 14) in ATDC5 cells and declines thereafter(Akiyama et al., 2000), and its signaling was criticallyrequired for the conversion of condensing undifferenti-ated cells into chondrocytes (Shukunami et al., 2000).Thus in this study, the transient increase in BMP-4expression noted on Day 10 of culture with DPH mayreflect the critical requirement of BMP-4 in this stage ofATDC5 cell differentiation. In addition, DPH stimulatedthe increase in Ihh expression as well as Ptc expressionas shown in Fig. 3. That the supplementation of Ihhstimulated the late-phase chondrogenic differentiationof ATDC5 cells (Akiyama et al., 1999; Enomoto-Iwa-moto et al., 2000) suggests that DPH-increased Ihhcould accelerate the late-phase chondrogenesis evenmore. Moreover, the induction of Ptc observed in theculture with DPH was possibly due to the increased Ihh,since the recombinant of Ihh induced Ptc expression(Akiyama et al., 1999). In contrast, PTHrP is a potentinhibitor of chondrogenic differentiation of chondro-cytes (Lee et al., 1996; Schipani and Provot, 2003),inhibited cellular condensation and the subsequent for-mation of cartilage nodules in ATDC5 cells (Shukunamiet al., 1996), and is known to be induced by Ihh in pre-chondrocytes (Vortkamp et al., 1996). In this study,although changes in PTHrP were not evident in ATDC5cells cultured with DPH during chondrogenic differenti-ation by Day 15, its expression may increase sequen-tially in the later termination stages of chondrogenesis.Taken together, these findings suggest that the expres-sion of molecules involved in TGF-β and hedgehog sig-naling is coordinately induced by DPH treatment andhighly responsible for DPH-accelerated chondrogenesisin ATDC5 cells.

DPH is known to have the teratogenic potency toinduce cleft palate and axial skeletal malformations inhumans and laboratory animals (Dansky and Finnell,1991; Finnell and Dansky, 1991). The involvement ofGR in DPH-induced cleft palate has been suggestedsince DPH and GCs share a common receptor(Katsumata et al, 1982; Goldman 1984; Katsuyama etal., 1985; Kay et al., 1990). Therefore, it is speculatedthat DPH inhibits chondrogenic differentiation ofATDC5 cells via GR. A previous study by Shibata etal. (1996) detected the inhibitory effects of the terato-

genic compound YM9429, which induces cleft palatein rats, in the chondrogenic differentiation of ATDC5cells. However, our ATDC5 results show an oppositeeffect, suggesting a lesser possibility of DPH actionthrough GR. In contrast, the pattern of skeletal abnor-malities closely resembles that observed in miceexposed to retinoic acid (RA), and in phenotypes ofmice lacking the genes encoding homeobox (Hox) pro-teins (Ross et al., 2000). In MC3T3-E1 cells, retinoidsinduced their osteogenic differentiation (Park, 1997),and the expression of Hox genes were regulated byTGF-β superfamily members (Kloen et al., 1997). Inaddition, the expression of Hox genes was induced inthe BMP-implanted rat tissue (Iimura et al., 1994).These findings demonstrate that there may be an asso-ciation between RA, TGF-β, and Hox signaling inosteogenic differentiation. Although, in this study,DPH effects on Hox and RA signaling molecules werenot determined, the increased expression of RA recep-tors in mouse fetuses exposed to DPH (Waes, 1999)has been reported. Thus DPH may possibly alter theexpression of Hox genes and RA signaling moleculesdirectly or indirectly via members of the TGF-β super-family in tissues of adults and embryos. However, weshow no change here in any parameters for chondro-genic or osteogenic differentiations in ATDC5 andMC3T3-E1 cells cultured with the other antiepileptics,PBT and VPA, which have teratogenic effects similarto DPH on fetal skeletons. Therefore, further evalua-tions to clarify the molecular target of antiepileptics onskeletogenesis for adults and embryos are necessary. Inconclusion, this study demonstrates for the first timethat DPH could have a positive chondrogenic effect onbone formation, and that DPH-promoted chondrogenicdifferentiation could be in part mediated by Ihh, Ptc,TGF-β1, and BMP-4 in ATDC5 cells.

ACKNOWLEDGMENT

The authors would l ike thank Dr. ChisaShukunami, Kyoto University, for her valuable techni-cal assistance on this manuscript. The authors wouldalso like acknowledge members of the Safety ResearchLaboratories of Yamanouchi Pharmaceutical Co., Ltd.for their helpful support and comments.

REFERENCES

Akiyama, H., Shigeno, C., Iyama, K., Ito, H., Hiraki,Y., Konishi, J. and Nakamura, T. (1999): Indianhedgehog in the late-phase differentiation in

Phenytoin in cartilage formation.

155

Vol. 30 No. 3

mouse chondrogenic EC cells, ATDC5: Upregu-lation of type X collagen and osteoprotegerinligand mRNA. Biochem. Biophys. Res. Com-mun., 257, 814-820.

Akiyama, H., Shukunami, C., Nakamura, T. andHiraki, Y. (2000): Differential expressions ofBMP family genes during chondrogenic differ-entiation of mouse ATDC5 cells. Cell Struct.Funct., 25, 195-204.

Atsumi, T., Miwa, Y., Kimata, K. and Ikawa, Y.(1990): A chondrogenic cell line derived from adifferentiating culture of AT805 teratocarci-noma cells. Cell Differ. Dev., 30, 109-116.

Dansky, L.V. and Finnell, R.H. (1991): Parental epi-lepsy, anticonvulsant drugs, and reproductiveoutcome: Epideminologic and experimentalfindings spanning three decades; 2: human stud-ies. Reprod. Toxicol., 5, 301-335.

Enomoto-Iwamoto, M., Nakamura, T., Aikawa, T.,Higuchi, Y., Yuasa, T., Yamaguchi, A., Nohno,T., Noji, S., Matsuya, T., Kurisu, K., Koyama,E., Pacifici, M. and Iwamoto, M. (2000):Hedgehog proteins stimulate chondrogenic celldifferentiation and cartilage formation. J. BoneMiner. Res., 15, 1659-1668.

Finnell, R.H. and Dansky, L.V. (1991): Parental epi-lepsy, anticonvulsant drugs, and reproductiveoutcome: Epideminologic and experimentalfindings spanning three decades; 1: animal stud-ies. Reprod. Toxicol., 5, 281-299.

Franklyn, J.A., Davis, J.R., Ramsden, D.B. andSheppard, M.C. (1985): Phenytoin and thyroidhormone action. J. Endocrinol., 104, 201-204.

Goldman, A.S., Sharpior, B.H. and Katsumata, M.(1978): Human foetal palatal corticoid receptorsand teratogens for cleft palate. Nature, 272,464-466.

Goldman, A.S. (1984): Biochemical mechanism ofglucocorticoid and phenytoin-induced cleft pal-ate. Curr. Top. Dev. Biol., 19, 217-239.

Iimura, T., Oida, S., Takeda, K., Maruoka, Y. and Sasaki,S. (1994): Changes in homeobox-containing geneexpression during ectopic bone formation inducedby bone morphogenetic protein. Biochem. Bio-phys. Res. Commun., 201, 980-987.

Ikedo, D., Ohishi, K., Yamauchi, N., Kataoka, M.,Kido, J. and Nagata, T. (1999): Stimulatoryeffects of phenytoin on osteoblastic differentia-tion of fetal calvaria cells in culture. Bone, 25,653-660.

Ito, H., Akiyama, H., Shigeno, C. and Nakamura, T.

(2000): Parathyroid hormone-related peptideinhibits the expression of bone morphogeneticprotein-4 mRNA through a cyclic AMP/proteinkinase A pathway in mouse clonal chondrogenicEC cells, ATDC5. Biochim. Biophys. Acta,1497, 237-243.

Johnson, J.P. (1984): Acquired craniofacial featuresassociated with chronic phenytoin therapy. Clin.Pediatr., 23, 671-674.

Katsumata, M., Gupta, C., Baker, M.K., Sussdorf,C.E. and Goldman, A.S. (1982): Diphenylhy-dantoin: An alternative ligand of a glucocorti-coid receptor affecting prostaglandin generationin A/J mice. Science, 218, 1313-1315.

Katsuyama, M., Gupta, C. and Goldman, A.S. (1985):Glucocorticoid receptor IB: Mediator of anti-inflammatory and teratogenic functions of bothglucocorticoids and phenytoin. Arch. Biochem.Biophys., 243, 385-395.

Kattan, K.R. (1970): Calvarial thickening after dilan-tin medication. Am. J. Roentgenol. RadiumTher. Nuclear Med., 110, 102-106.

Kawai, J., Akiyama, H., Shigeno, C., Ito, H., Konishi,J. and Nakamura, T. (1999): Effects of trans-forming growth factor-β signaling on chondro-genesis in mouse chondrogenic EC cells,ATDC5. Eur. J. Cell Biol., 78, 707-714.

Kay, E.D., Goldman, A.S. and Daniel, J.C. (1990):Common biochemical pathway of dysmorpho-genesis in murine embryos: Use of the glucocor-ticoid pathway by phenytoin. TeratogenesisCarcinog. Mutagen., 10, 31-39.

Kloen, P., Visker, M.H.P.W., Olijve, W., van Zoelen,E.J.J. and Boersma, C.J.C. (1997): Cell-type-specific modulation of hox gene expression bymembers of the TGF-β superfamily: A compari-son between human osteosarcoma and neuro-blastoma cell lines. Biochem. Biophys. Res.Commun., 233, 365-369.

Kodama, H., Amagai, Y., Sudo, H., Kasai, S. andYamamoto, S. (1981): Establishment of a clonalosteogenic cell line from newborn mouse calva-ria. Jap. J. Oral Biol., 23, 899-901.

Lau, K-H.W., Nakade, O., Barr, B., Taylor, A.K.,Houchin, K. and Baylink, D.J. (1995): Pheny-toin increases markers of osteogenesis for thehuman species in vitro and in vivo. J. Clin.Endocrinol. Metab., 80, 2347-2353.

Lee, K., Lanske, B., Karaplis, A.C., Deeds, J.D.,Kohno, H., Nissenson, R.A., Kronenberg, H.M.and Segre, G.V. (1996): Parathyroid hormone-

156

A. OKADA et al.

Vol. 30 No. 3

related peptide delays terminal differentiation ofchondrocytes during endochondral bone devel-opment. Endocrinology, 137, 5109-5118.

Lefebvre, E.B., Haining, R.G. and Labbe, R.F. (1972):Coarse facies, calvarial thickening and hyper-phosphatasia associated with long-term anticon-vulsant therapy. N. Engl. J. Med., 286, 1301-1302.

Levy, R.H. (1980): Phenytoin: Biopharmacology. Adv.Neurol., 27, 315-321.

Miura, M., Tanaka, K., Komatsu, Y., Suda, M.,Yasoda, A., Sakuma, Y., Ozawa, A. and Nakao,K. (2002): Thyroid hormones promote chondro-cyte differentiation in mouse ATDC5 cells andstimulate endochondral ossification in fetalmouse tibias through iodothyronine deiodinasesin the growth plate. J. Bone Miner. Res., 17,443-454.

Ohta, T., Wergedal, J.E., Gruber, H.E., Baylink, D.J.and Lau, K-H.W. (1995): Low dose phenytoin isan osteogenic agent in the rat. Calsified TissueInt., 56, 42-48.

Okada, A., Ohta, Y., Buchanan, D.L., Sato, T. andIguchi, T. (2002): Effect of estrogens on ontoge-netic expression of progesterone receptor in thefetal female rat reproductive tract. Mol. Cell.Endocrinol., 195, 55-64.

Park, C.K., Ishimi, Y., Ohmura, M., Yamaguchi, M. andIkegami, S. (1997): Vitamin A and carotenoidsstimulate differentiation of mouse osteoblasticcells. J. Nutr. Sci. Vitaminol., 43, 281-296.

Ross, S.A., McCaffery, P.J., Drager, U.C. and Luca,L.M. (2000): Retinoids in embryonal develop-ment. Physiol. Rev., 80, 1021-1054.

Sasaki, H., Ikedo, D., Kataoka, M., Kido, J., Kitamura,S. and Nagata, T. (1999): Pronounced palataland mandibular tori observed in a patient withchronic phenytoin therapy: A case report. J.Periodontol., 70, 445-448.

Schipani, E. and Provot, S. (2003): PTHrP, PTH, andthe PTH/PTFrP receptor in endochondral bonedevelopment. Birth Defects Res. Part C, 69,352-362.

Shibata, M., Fujino, E., Sato, Y., Numazawa, S.,Yoshida, T. and Kuroiwa, Y. (1996): Inhibitionof insulin-induced chondrogenic differentiationof embryonic carcinoma cells by a teratogeniccompound YM9429. Biochem. Biophys. Res.Commun., 225, 593-599.

Shukunami, C., Shigeno, C., Atsumi, T., Ishizeki, K.,Suzuki, F. and Hiraki, Y. (1996): Chondrogenic

differentiation of clonal mouse embryonic cellline ATDC5 in vitro: Differentiation-dependentgene expression of parathyroid hormone (PTH)/PTH-related peptide receptor. J. Cell Biol., 133,457-468.

Shukunami, C., Ishizeki, K., Atsumi, T., Ohta, Y.,Suzuki, F. and Hiraki, Y. (1997): Cellular hyper-trophy and calcification of embryonal carci-noma-derived chondrogenic cell line ATDC5 invitro. J. Bone Miner. Res., 12, 1174-1188.

Shukunami, C., Akiyama, H., Nakamura, T. andHiraki, Y. (2000): Requirement of autocrine sig-naling by bone morphogenetic protein-4 forchondrogenic differentiation of ATDC5 cells.FEBS Lett., 469, 83-87.

Siebler, T., Robson, H., Shalet, S.M. and Wikkiams,G.R. (2002): Dexamethasone inhibits and thy-roid hormone promotes differentiation of mousechondrogenic ATDC5 cells. Bone, 31, 457-464.

Silvestrini, G., Ballanti, P., Patacchioli, F.R., Mocetti,P., Di Grezia, R., Wedard, B.M., Angelucci, L.and Bonucci, E. (2000): Evaluation of apoptosisand the glucocorticoid receptor in the cartilagegrowth plate and metaphyseal bone cells of ratsafter high-dose treatment with corticosterone.Bone, 26, 33-42.

Sklans, S., Taylor, R.G. and Shklar, G. (1967): Effectof diphenylhydantoin sodium on healing ofexperimentally produced fractures in rabbitmandibles. J. Oral Surg., 25, 310-319.

Stevens, D.A., Hasserjian, R.P., Robson, H., Siebler,T., Shalet, S.M. and Williams, G.R. (2000):Thyroid hormones regulate hypertrophic chon-drocyte differentiation and expression of par-athyroid hormone-related peptide and itsreceptor during endochondral bone formation. J.Bone Miner. Res., 15, 2431-2442.

Sudo, H., Kodama, H., Amagai, Y., Yamamoto, S. andKasai, S. (1983): In vitro differentiation and cal-cification in a new clonal osteogenic cell linederived from newborn mouse calvaria. J. CellBiol., 96, 191-198.

Vortkamp, A., Lee, K., Lanske, B., Segre, G.V.,Kronenberg, H.M. and Tabin, C.J. (1996): Reg-ulation of rate of cartilage differentiation byindian hedgehog and PTH-related protein. Sci-ence, 273, 613-622.

Waes, J.G.V., Bennett, G.D. and Finnell, R. (1999):Phenytoin-induced alterations in craiofacialgene expression. Teratology, 59, 23-34.