Available online at www.sciencedirect.com

ta 1783 (2008) 12–22www.elsevier.com/locate/bbamcr

Biochimica et Biophysica Ac

Thrombin peptide TP508 prevents nitric oxide mediated apoptosis inchondrocytes in the endochondral developmental pathway

M. Zhong a, L.J. Wike a, J.T. Ryaby b, D.H. Carney c, B.D. Boyan a,⁎, Z. Schwartz a,d,e

a Wallace H. Coulter Department of Biomedical Engineering at Georgia Tech and Emory University, Georgia Institute of Technology, Atlanta, Georgia, USAb OrthoLogic Corporation, Tempe, Arizona, USA

c Department of Biochemistry and Molecular Biology, University of Texas Medical Branch, Galveston, Texas, USAd Department of Periodontics, University of Texas Health Science Center at San Antonio, San Antonio, Texas, USA

e Department of Periodontics, Hebrew University Hadassah, Jerusalem, Israel

Received 7 March 2007; received in revised form 12 October 2007; accepted 16 October 2007Available online 20 November 2007

Abstract

TP508 is a 23-amino acid peptide derived from human prothrombin that increases cartilage matrix production and reduces alkaline phosphataseactivity without changing chondrocyte proliferation. This study tested the hypothesis that TP508 acts by blocking the onset of apoptosis associatedwith hypertrophy. Rat costochondral resting zone chondrocytes and human auricular chondrocytes were cultured in DMEM containing 50μMascorbic acid and 10% FBS. Apoptosis was induced by treatment of confluent cultures with chelerythrine, tamoxifen, or inorganic phosphate (Pi)for 24h. One half of the cultures received TP508 (0, 0.7, or 7μg/ml). Apoptosis was assessed as a function of DNA fragmentation ([3H]-thymidinelabeled DNA fragments), TUNEL staining, and cell viability using the MTT assay, as well as by assessing the Bcl-2/Bax mRNA and protein ratiosand caspase-3 activity. The universal NO synthase inhibitor L-NMMAwas used to assess the effect of NO production on chondrocyte apoptosisand specific NO synthase subspecies were identified using iNOS inhibitor 1400W and nNOS inhibitor vinyl-L-NIO, as well as L-NAME, whichinhibits both iNOS and eNOS. Finally, we assessed if TP508 would block NO production induced by the apoptogens. Chelerythrine, tamoxifenand Pi-induced apoptosis and this was reversed by TP508. All apoptogens increased NO production and this was reduced by TP508. TP508reduced NO levels to the same extent as 1400W but not to the same extent as L-NAME, suggesting that its effects are mediated primarily by iNOS.In addition, TP508 reduced the effect of chelerythrine to the same extent as 1400W and L-NAME, again indicating that it acts via inhibition of aniNOS pathway. TP508 also regulated Bcl-2/Bax mRNA in a time and dose-dependent manner. The Bcl-2/Bax mRNA ratio was 0.11 in theabsence of TP508 at 1h and 4.95 at 7μg/ml TP508; by 3h the ratio was approximately 1 in both groups. The Bcl-2/Bax protein ratio also increasedby 63% at 1h. TP508 did not affect caspase-3 activity. TP508 also caused a dose-dependent increase in protein kinase C (PKC) activity within9min that was maximal at 270min. These results show that TP508 prevents apoptosis in growth plate chondrocytes via inhibition of iNOS-dependent NO and suggest a possible role for PKC in the mechanism.© 2007 Elsevier B.V. All rights reserved.

Keywords: Chondrocytes; TP508; Thrombin peptide; Apoptosis; Growth plate; Endochondral ossification

1. Introduction

Bone repair occurs via endochondral bone formation arounda fracture site. Its efficiency depends on the balance betweenproliferation, maturation and death of cells that form the car-

⁎ Corresponding author. Department of Biomedical Engineering, GeorgiaInstitute of Technology, 315 Ferst Drive NW, Atlanta, Georgia 30332-0363,USA. Tel.: +1 404 385 4108; fax: +1 404 894 2291.

E-mail address: Barbara.Boyan@bme.gatech.edu (B.D. Boyan).

0167-4889/$ - see front matter © 2007 Elsevier B.V. All rights reserved.doi:10.1016/j.bbamcr.2007.10.009

tilage template preceding bone formation, all of which areregulated by multiple systemic and local regulators. The processof bone repair recapitulates the events that occur duringembryonic bone formation and in many respects, the eventsthat occur during post-natal long bone growth [1,2]. Duringlong bone growth, cells in the resting zone, which are embeddedin an extracellular matrix rich in type II collagen and pro-teoglycan aggregates containing sulfated glycosaminoglycans,are induced to proliferate, followed by a shift in phenotypeto a hypertrophic cell that ultimately calcifies its extracellular

13M. Zhong et al. / Biochimica et Biophysica Acta 1783 (2008) 12–22

matrix. Chondrocytes in the hypertrophic zone of the growthplate routinely undergo physiological apoptosis, especially duringterminal differentiation [3–6], and studies indicate that elevatedextracellular phosphate and nitric oxide (NO) production areinvolved [7,8].

Apoptosis in the resting zone occurs less frequently [9]. It isnot known whether it is induced by pathways that act in thehypertrophic zone or if factors exist that can rescue resting zonecells from apoptosis once it has been induced. To better under-stand the mechanisms involved in regulation of apoptosis inthese cells, we took advantage of recent observations concerningthrombin peptide TP508 (Chrysalin®, OrthoLogic Corp, Tempe,AZ). TP508 is a 23-amino acid peptide representing the humanthrombin amino acid sequence originally identified as thereceptor binding domain responsible for thrombin binding andactivation of a high-affinity class of receptors on fibroblasts [10].It has been shown to accelerate fracture repair in a rat closeddiaphyseal fracture model, which heals through an endochondralprocess [1,11]. One effect of the peptide was to increase thecallus mass, a tissue that consists of proliferating and prehyper-trophic chondrocytes. Ultimately, cells in the callus undergohypertrophy and calcify their matrix, leading to vascular inva-sion and bone formation.

The purpose of the callus is to stabilize the bony ends of thefracture so agents that cause this to happen more rapidly arelikely to affect cells at early stages of endochondral maturation.Studies using rat costochondral growth plate chondrocytes,which are similar to the prehypertrophic cells in the fracturecallus, support this hypothesis [12]. These cells were differen-tially affected by TP508 in a manner that promoted prolonga-tion of the chondrocytic lineage. Although TP508 had no effecton proliferation, it stimulated production of sulfated glycosa-minoglycans, a hallmark of cartilage extracellular matrix, and itinhibited growth plate chondrocyte maturation based on de-creased alkaline phosphatase activity. Increased alkalinephosphatase is associated with chondrocyte hypertrophy andapoptosis [13,14]. Therefore, it is important to elucidatewhether TP508 is able to block chondrocyte apoptosis, and ifso, its mechanism of action.

The apoptotic pathway in growth plate chondrocytes can beactivated by high concentrations of inorganic phosphate (Pi) [7],and is mediated through signals that include a rapid increase innitric oxide [8]. Chick sternal chondrocytes possess induciblenitric oxide synthase (iNOS), neuronal NOS (nNOS), andendothelial NOS (eNOS) and treatment of the cells with NOSinhibitors decreases alkaline phosphatase [15]. This suggeststhat peptides like TP508, which reduce alkaline phosphatase,may also act as anti-apoptotic agents via a mechanism thatinvolves reduction in NO. To determine if TP508 can preventapoptosis in these cells, we first examined whether TP508regulates expression of Bcl-2 or Bax, since previous studieshave shown that an increase in the Bcl-2/Bax ratio favorsprotection against apoptosis [16,17]. In addition, we assessedthe effects of TP508 on protein kinase C (PKC) activity sinceactivation of this enzyme has been associated with cellproliferation in other systems [18–20]. Apoptosis was inducedby treating confluent cultures with three different apoptotic

agents. Whether TP508 could reverse their ability to stimulateNO production or apoptosis was determined.

2. Materials and methods

2.1. Materials

The 23-amino acid synthetic thrombin peptide TP508, also known asChrysalin®, comprising amino acids 508–530 of human prothrombin(AGYKPDEGKRGDACEGDSGGPFV, mol wt of 2311.5) synthesized byBachem California (Torrance, CA), was provided for these studies byOrthoLogic Corp. (Tempe, AZ). The peptide was shown by HPLC to begreater than 97% pure. It was stored as a lyophilized powder and reconstitutedwith medium prior to addition to cells.

2.2. Cell culture

Chondrocytes were isolated from the resting zone of the costochondralcartilage of 125-g Sprague–Dawley rats, as described previously [21]. Initialexperiments compared cells from male and female donors to rule out any sex-specific differences in mechanism. At third passage confluence, the chondro-cytes were subpassaged into 24-well plates such that there were six separatecultures per variable per experiment. TP508 was dissolved in culture media andthen added to the cultures at concentrations of 0.07, 0.7 and 7.0μg/ml. Previousstudies validated the specificity of the response of chondrocytes to TP508 byusing a scrambled peptide containing the same amino acids as TP508, but in adifferent order (AVDPEGRAGCGGDPKYGDFSKEG) [12].

To determine if TP508 also had effects on human cartilage cells, humanauricular chondrocytes were used. Human cartilage was obtained via an IRB-approved protocol, cleaned of adjacent connective tissue, minced and washedwith 10ml Hank's Balanced Salt Solution (Invitrogen, Carlsbad, CA) containing10% penicillin/streptomycin (Mediatech, Herndon, VA) twice for 20min.Following treatment with 5ml 0.25% trypsin–1mM EDTA (Invitrogen) for 1h,the samples were treated for 6h with 12.5ml 0.2% collagenase type II(Worthington Biochem, Lakewood, NJ) containing 1% penicillin/streptomycin.The collagenase treated samples were sieved through a cell strainer (BDBiosciences, San Jose, CA), collected in a 50ml centrifuge tube, and centrifugedat 2000rpm for 10min. The supernatant was removed, and the pellet wassuspended in 10ml media containing 10% fetal bovine serum (FBS), 1%penicillin/streptomycin (Mediatech, Herndon, VA) and 1% ascorbic acid(Sigma-Aldrich, St. Louis, MO). After determining cell yield using a BeckmanCoulter Z1 Cell Counter (Beckman Coulter, Fullerton, CA), the cells were platedin 24-well plates at a density of 10,000cells/cm2.

2.3. Regulation of Bcl-2 and Bax

Effects of TP508 on Bcl-2 and Bax expression were assessed by RT-PCR ofRNA isolated from female resting zone chondrocytes that were treated with 0,0.07, 0.7 or 7μg/ml TP508 for 1, 3, 6 and 12h. RNAwas isolated after lysis of thecells with TRIzol (Invitrogen, Carlsbad, CA). Total RNAwas reverse-transcribedusing specific primers described below and the Omniscript RT kit fromQIAGEN(Valencia, CA). Following the initial reverse transcription reaction, the cDNAwas amplified using a polychain amplification reaction (PCR) by adding a pair ofspecific primers and Taq DNA polymerase (Fisher Scientific, Hampton, NH).The PCR products were separated by electrophoresis and measured semi-quantitatively by densitometry using VersaDoc Imaging System (Bio-Rad,Hercules, CA). Primer sequences were designed using Beacon Designer 5(Premier Biosoft International, Palo Alto, CA) according to the publishedsequence of rat Bcl-2 (GenBank accession number L14680) and Bax (GenBankaccession number AF235993) mRNA. The primers used in this study are: Bcl-2sense primer, 5′-CTCCTGGCTGTCTCTGAAG-3′ and antisense primer, 5′-TCTGCTGACCTCACTTGTG-3′; Bax sense primer, 5′-CCATCAGAAGCAG-TAGCC-3′ and antisense primer, 5′-TGGAGAGAGGAGGAAAGG-3′. Theexpected product sizes were 95 base pairs (Bcl-2) and 134 base pairs (Bax).

Effects on Bcl-2 and Bax protein levels were determined by Western blot ofcell culture lysates prepared from female rat resting zone chondrocytes.

14 M. Zhong et al. / Biochimica et Biophysica Acta 1783 (2008) 12–22

Samples (30μg protein) were resolved on 10% SDS-polyacrylamide gels. Blotsof the gels were probed with mouse monoclonal antibodies to Bcl-2 (cat#610539, purified mouse anti-Bcl-2, BD Biosciences, Franklyn Lakes, NJ) andBax (sc-7480, Bax(B9), Santa Cruz Biotechnology, Santa Cruz, CA, USA).Two other antibodies against Bcl-2 [sc-7382, Bcl-2(C-2) and sc-783, Bcl-2(DC21) (Santa Cruz Biotechnology)], and two other antibodies against Bax [sc-6236, Bax(δ21) (Santa Cruz Biotechnology) and a purified mouse anti-Baxmonoclonal antibody (cat#556467, BD Biosciences)] were also tested, but nopositive bands were obtained. Immunoreactive bands were detected using1:5000 dilutions of horseradish peroxidase-conjugated goat anti-mouse IgG(Santa Cruz Biotechnology), and visualized using enhanced chemilumines-cence (ECL; Amersham Biosciences, Piscataway, NJ). Controls wereperformed with anti-GAPDH antibodies (Mab374, Chemicon International,Temecula, CA). Immunoreactive bands were scanned and their relativeintensities determined.

2.4. Protein kinase C activity

PKC specific activity was determined in cell layer lysates as a function ofdose and time using an assay kit (Protein Kinase C Enzyme Biotrak Assay

Fig. 1. Panel A: Effect of TP508 on Bcl-2 and Bax expression in rat costochondralTP508 for 1, 3, 6, and 12 h. Relative levels of each mRNAwere determined by denslysates were separated by SDS-PAGE and Bcl-2 and Bax detected byWestern blot. Daof TP508 on PKC specific activity. Cultures were treated with 0.7 to 14 μg/ml TP50presented are from one of at least two separate experiments, all of which had compTP508 v. all other treatments at 270 min.

System, Amersham Biosciences, Buckinghamshire, England), as previouslydescribed [22]. NO production was determined using a 2,3-diaminonaphthlene-based assay (DAN) of the conditioned media as described by Misko et al. [23].

2.5. Apoptosis assays

Apoptosis was induced using three different methods. Confluent cultures ofresting zone chondrocytes were treated for 24h with chelerythrine (0.1, 1, or10μM), which has been shown to induce apoptosis in leukocytes, cardiacmyocytes and some cancer cells [24–26] or tamoxifen (10− 10, 10− 9, or 10− 8M),which has been shown to induce apoptosis in multiple human cancer cells [27–29]. Both of these agents inhibit PKC in resting zone cells in a dose-dependentmanner [30,31]. In addition, we tested the effects of 10μM chelerythrine and 10−9M tamoxifen in combination. Alternatively, confluent cultures were treated for24h with Pi (0, 2.5, 5, and 7.5mM), which has been shown to induce apoptosis inchick growth plate chondrocytes [7]. To determine if TP508 could block theapoptotic effects of these agents, cultures were treated with 10μM chelerythrine,10− 9M tamoxifen, or 7.5mM Pi; one half of the cultures also received TP508 (0,0.7 or 7μg/ml). Similarly, human auricular chondrocytes were treated withcontrol media or 10μM chelerythrine with and without 7μg/ml TP508.

resting zone chondrocytes. Cultures were treated with 0, 0.07, 0.7 and 7 μg/mlitometry. Panel B: Effect of TP508 on Bcl-2 and Bax protein levels. Cell cultureta are from one of three experiments, all with comparable results. Panel C: Effect8 for 9 to 720 min. Data are means±SEM for N=6 independent cultures. Dataarable results. ⁎Pb0.05, treatment v. vehicle at each time; #Pb0.05, 0.7 μg/ml

Fig. 2. Effect of TP508 on cell number (A), cell viability (B) and DNAfragmentation (C) of confluent resting zone costochondral growth platechondrocyte cultures isolated from female rats. Cultures were treated with 0,0.7 or 7 μg/ml for 24 h in the presence or absence of 10 μM chelerythrine. Dataare means±SEM for N=6 independent cultures. Data presented are from one ofat least two separate experiments, all of which had comparable results. ⁎Pb0.05,TP508 treatment v. vehicle; #Pb0.05, chelerythrine v. control at each con-centration of TP508.

15M. Zhong et al. / Biochimica et Biophysica Acta 1783 (2008) 12–22

Apoptosis in rat and human chondrocytes was quantified as a function ofDNA fragmentation [32]. Cells were labeled with [3H]-thymidine for 4h beforethe treatment. After the treatment, the cells were lysed and centrifuged toseparate the fragmented DNA and the intact DNA. [3H] in the supernatant andthe pellet was measured by scintillation counting, and the percentage ofapoptotic cells was determined by dividing the amount of fragmented DNAwiththe total DNA. In addition, cells treated with chelerythrine or tamoxifen wereexamined by TUNEL staining using an In Situ Cell Death Detection Kit (RocheApplied Science, Indianapolis, IN) in order to verify that DNA fragmentationhad occurred [33]. Cell viability was determined using an MTT assay kit(Promega, Madison, WI).

Caspase-3 activity was assayed using a Caspase-3 Assay Kit (Sigma-Aldrich, St. Louis, MO) following the manufacturer's instructions. Confluentcultures were treated for 24h with 10μM chelerythrine or 10− 9M tamoxifen.Lysates of these cultures were incubated with acetyl Asp-Glu-Val-Asp 7-amido-4-methylcoumarin, a substrate of caspase-3, for 1h. At that time, fluorescencewas measured at excitation and emission wavelengths of 360 and 460nm,respectively, to determine the amount of 7-amino-4-methylcoumarin (AMC)produced. Results were normalized to cell number.

2.6. Role of nitric oxide

The universal NO synthase inhibitor L-NMMA (Sigma-Aldrich, St. Louis,MO) was used to assess the effect of NO production on chondrocyte apoptosis[34]. To determine the specific NO synthase subspecies that was involved,inducible NO synthase (iNOS) inhibitor 1400W (Sigma-Aldrich, St. Louis, MO)[35] and neuronal NO synthase (nNOS) inhibitor vinyl-L-NIO (Sigma-Aldrich,St. Louis, MO) [36], as well as L-NAME (Sigma-Aldrich, St. Louis, MO), whichinhibits both iNOS and eNOS [37], were used together with TP508 to block NOproduction.

2.7. Statistical analysis

All cell culture experiments were repeated at least twice. For eachexperiment, there were six independent cultures for each variable. Statisticalsignificance was determined using ANOVA and post-hoc testing withBonferroni's modification of Student's t-test. The threshold of significancewas set as P-value b 0.05.

3. Results

3.1. Bcl-2/Bax

The ratio of Bcl-2 to Bax mRNAs varied with TP508 doseand treatment time (Fig. 1A). At 1h, TP508 caused a dose-dependent increase in Bcl-2 and a decrease in Bax expression.By 3h of treatment, Bax levels increased and the ratio ofmRNAs for the two proteins was essentially 1. At 6 and 12h,Bcl-2 mRNAs were relatively constant and independent ofTP508 treatment. At 6h, levels of Bax varied with dose but thelevels of Bcl-2 were constant up to 7μg/ml TP508. At 12h, therewas little evidence of Bax expression in control cultures but allTP508 doses caused a comparable increase in Bax, resulting in aBcl-2/Bax ratio approximating 1. Changes in Bcl-2 and Baxprotein levels reflected the mRNA ratios. At 1h, there was a62.9% increase in the Bcl-2/Bax protein ratio (Fig. 1B).

3.2. Protein kinase C activity

TP508 caused a dose-dependent increase in PKC activity infemale resting zone cells that was evident within 9min andremained elevated over 6h (Fig. 1C). At 9min, the stimulatory

effects were seen at concentrations of 7 and 14μg/ml TP508. By90min, even the low dose of TP508 was effective.

3.3. Regulation of apoptosis

Chelerythrine caused a decrease in cell number (Fig. 2A) andcell viability (Fig. 2B) in female rat chondrocytes, indicating

Table 1The effect of TP508 on chelerythrine-induced DNA fragmentation of humanauricular chondrocytes

Control 10 μM chelerythrine

Control 3.74±0.95 8.01±1.32 a, b

7 μg/ml TP508 4.62±1.34 4.67±0.49a pb0.05 against control group.b pb0.05 against 0 μg/ml TP508 groups.

16 M. Zhong et al. / Biochimica et Biophysica Acta 1783 (2008) 12–22

that chelerythrine induced apoptosis. The inhibitory effects ofchelerythrine on cell number and cell viability were partiallyreversed by TP508 in a dose-dependent manner. Chelerythrinealso caused an increase in DNA fragmentation in female resting

Fig. 3. Effect of chelerythrine and tamoxifen on the number of TUNEL-labeled celgrowth plate chondrocyte cultures isolated from female rats. Cultures were treated wpresented are from one of at least two separate experiments, all of which had comv. tamoxifen alone.

zone chondrocytes (Fig. 2C) and human auricular chondrocytes(Table 1), which was blocked by TP508 in a dose-dependentmanner. Similar effects on cell number, MTT activity and DNAfragmentation were observed in male chondrocytes (data notshown). TUNEL staining confirmed that chelerythrine inducedDNA fragmentation in rat chondrocytes (Fig. 3B) comparedwith control cultures (Fig. 3A). TUNEL staining showed thattamoxifen also induced DNA fragmentation (Fig. 3C), and thetwo agents in combination increased the number of TUNEL-labeled cells over the amount seen in response to either agentalone (Fig. 3D). DNA fragmentation in the rat chondrocytes inresponse to tamoxifen was comparable to that seen in culturestreated with chelerythrine, and the effects of the two agents were

ls (A–D) and DNA fragmentation (E) of confluent resting zone costochondralith 10−9 M tamoxifen, 10 μM chelerythrine or their combination for 24 h. Dataparable results. ⁎Pb0.05, treatment v. 1.0; #Pb0.05, tamoxifen+chelerythrine

Table 2The dose-dependent effect of Pi on DNA fragmentation

Pi concentration Percent fragmented Viability

17M. Zhong et al. / Biochimica et Biophysica Acta 1783 (2008) 12–22

additive (Fig. 3E). No changes in caspase-3 activity weredetected as a function of chelerythrine, tamoxifen, or Pi treat-ment (data not shown).

Fig. 4. Effect of TP508 on tamoxifen and Pi-induced cell death of confluentresting zone costochondral growth plate chondrocyte cultures isolated fromfemale rats. Cultures were treated with 10−9 M tamoxifen (A, B) or 7.5 mM Pi(C) for 24 h in the presence or absence of 7 μg/ml TP508. Cell death wasmeasured using MTT assay to assess viability (A) and by DNA fragmentationassay (B,C). Data are means±SEM for N=6 independent cultures. Datapresented are from one of at least two separate experiments, all of which hadcomparable results. ⁎Pb0.05, treatment v. vehicle; #Pb0.05, tamoxifen aloneand tamoxifen+TP508 v. TP508 alone (A,B); #Pb0.05, Pi+TP508 v. Pi alone(C); •Pb0.05, tamoxifen+TP508 v. tamoxifen alone (A).

(mM) DNA (%) (treatment/control)

0 4.67±0.63 12.5 8.42±1.08 a 0.85±0.045 9.91±1.11 a 0.84±0.047.5 16.06±0.10 a, b 0.72±0.03a pb0.05 against control group.b pb0.05 against 2.5 mM Pi group.

TP508 reduced the effect of tamoxifen on chondrocyteviability (Fig. 4A) and on DNA fragmentation (Fig. 4B). Theapoptogen Pi also induced a dose-dependent decrease inchondrocyte viability and an increase in DNA fragmentation(Table 2). TP508 reduced the effects of Pi on DNA fragmenta-tion (Fig. 4C). In the tamoxifen data set shown, TP508 alone hada small inhibitory effect onMTTactivity and a small stimulatoryeffect on DNA fragmentation. This effect of TP508 on viabilitywas seen in four of six experiments, but the effect on DNAfragmentation was only observed in three of ten experiments.The Pi example is typical of experiments showing a lack ofTP508 effect on DNA fragmentation.

3.4. Role of nitric oxide

Chelerythrine (Fig. 5A), tamoxifen (Fig. 5B), and Pi(Fig. 5C), all caused a dose-dependent increase in NO pro-duction in rat resting zone chondrocytes. On a comparativebasis, chelerythrine was the most potent stimulator and Pi theleast potent. The effects of tamoxifen were bimodal with peakstimulation at 10− 9M. TP508 significantly reduced chelerythr-ine-dependent NO production in a dose-dependent manneralthough levels remained greater than in control cultures even atthe highest concentration of peptide (Fig. 5D). TP508 blockedthe stimulatory effects of tamoxifen (Fig. 5E) and Pi (Fig. 5F).

When the rat cartilage cells were treated with chelerythrinetogether with the NO synthase inhibitor L-NMMA, L-NMMAblocked NO production induced by chelerythrine (Fig. 6A).Moreover, L-NMMA also blocked loss of cell viability causedby chelerythrine (Table 3). At least part of the NO produced inresponse to chelerythrine was via an iNOS mechanism based onthe reduction in NO by the iNOS/eNOS inhibitor L-NAME(Fig. 6B) and the iNOS inhibitor 1400W (Fig. 6C). TP508reduced NO levels in control cultures to the same extent as1400W but not to the same extent as L-NAME. TP508 and L-

NAME both reduced NO production induced by chelerythrinebut the peptide was less effective. In contrast, TP508 and1400W had comparable inhibitory effects on NO production innon-stimulated cultures and both had comparable inhibitoryeffects on cultures stimulated with chelerythrine. The nNOSinhibitor had no effect (Fig. 6D).

The stimulatory effect of chelerythrine on NO production byrat resting zone chondrocytes was time-dependent; it wasevident at 90min and had maximal increases at 12h (Fig. 6E).TP508 reduced NO levels in control cultures at 9 and 90min.In addition, TP508 caused a greater than 90% reduction in

Fig. 5. Effect of TP508 on NO production in response to chelerythrine, tamoxifen and Pi by confluent resting zone costochondral growth plate chondrocyte cultures isolated from female rats. Cultures were treated with0.1 to 10 μM chelerythrine (A), 10−11 to 10−9 M tamoxifen (B) or 2.5 to 7.5 mM Pi for 24 h (C). Alternatively, cultures were treated with 0.7 and 7 μg/ml TP508 in the presence or absence of 10 μM chelerythrine (D),10−9 M tamoxifen (E) or 7.5 mM Pi (F) for 24 h. Data are means±SEM for N=6 independent cultures. Data presented are from one of at least two separate experiments, all of which had comparable results. ⁎Pb0.05,treatment v. vehicle controls; #Pb0.05, 10−9 and 10−8 M tamoxifen v. 10−10 M tamoxifen (B); #Pb0.05, treatment v. control at each concentration of TP508 (D,E,F); •Pb0.05, 10 μm chelerythrine v. 0.1 and 1 μmchelerythrine (A); •Pb0.05, 10−8 M tamoxifen v. 10−10 M tamoxifen (B).

18M.Zhong

etal.

/Biochim

icaet

Biophysica

Acta

1783(2008)

12–22

Fig. 6. Effect of NOS inhibitors on chelerythrine-induced NO production (A–D) and the time course of chelerythrine-induced NO production (E) of confluent restingzone costochondral growth plate chondrocyte cultures isolated from female rats. Cultures were treated with 7 μg/ml TP508, 10 μM chelerythrine or their combination,in the presence or absence of 1 mM L-NMMA (A), 50 μM L-NAME (B), 5 μM1400W (C) and 1 μMvinyl-L-NIO (D). Alternatively, cultures were treated with 7 μg/mlTP508, 10 μM chelerythrine, or their combination for 9 to 1440 min (E). Data are means±SEM for N=6 independent cultures. Data presented are from one of at leasttwo separate experiments, all of which had comparable results. ⁎Pb0.05, treatment v. vehicle; #Pb0.05, inhibitor v. control groups for each treatment; •Pb0.05,chelerythrine+TP508 v. chelerythrine alone.

19M. Zhong et al. / Biochimica et Biophysica Acta 1783 (2008) 12–22

chelerythrine-induced NO production at all time pointsexamined. Moreover, the general NO inhibitor L-NMMAreduced the effect of chelerythrine (Fig. 7A) and Pi (Fig. 7B)on DNA fragmentation, indicating that the inhibitory effect ofchelerythrine on cell viability was mediated by NO and thatTP508 blocked this effect by reducing NO production.

4. Discussion

In this study, we first established chelerythrine, tamoxifenand inorganic phosphate as apoptosis inducers in rat growth zonechondrocytes as indicated by increased DNA fragmentation,TUNEL staining and decreased cell viability. Although other

Table 3The effect of L-NMMA on chelerythrine-induced NO production

Chelerythrine (μM) NO production (nmol NO2−/cell×10−5)

Control L-NMMA

0 0.79±0.11 0.92±0.070.1 3.02±0.19 a 0.82±0.16 b

1 3.78±0.59 a 1.92±0.40 a, b, c

10 9.28±0.29 a, d 3.33±0.84 a, b, c, d

a pb0.05 against 0 μM chelerythrine groups.b pb0.05 against control groups.c pb0.05 against 0.1 μM chelerythrine groups.d pb0.05 against 1 μM chelerythrine groups.

Fig. 7. Effect of L-NMMA on chelerythrine (A) or Pi (B) induced apoptosis ofconfluent resting zone costochondral growth plate chondrocyte cultures isolatedfrom female rats. Cultures were treated with 10 μMchelerythrine (A) or 7.5M Pi(B) in the presence or absence of 7 μg/ml TP508 for 24 h. Data are means±SEMfor N=6 independent cultures. Data presented are from one of at least twoseparate experiments, all of which had comparable results. ⁎Pb0.05, treatmentv. vehicle; #Pb0.05, apoptogen+L-NMMA v. apoptogen alone.

20 M. Zhong et al. / Biochimica et Biophysica Acta 1783 (2008) 12–22

types of cell death may also result in DNA fragmentation, theDNA fragmentation assay we used is specific to apoptotic cells.However, we did not observe an increase in caspase-3 activitycaused by these apoptosis inducers. It is possible that ourcaspase-3 assay was not sensitive enough, or we had not foundthe exact time when caspase-3 is activated. Alternatively, theseapoptotic pathways might be caspase-independent, as pre-viously reported in other cases of chondrocyte apoptosis [38,39].

We demonstrated that TP508 blocks apoptosis induced bychelerythrine and tamoxifen in resting zone chondrocytes. Bothof the apoptogens inhibit PKC in these cells [30,31]. AlthoughPKC activation has been associated with proliferation in manycell types [18,19,40], TP508 does not increase proliferation inresting zone chondrocytes, either as a function of cell number oras a function of [3H]-thymidine incorporation [12]. We did notspecifically test whether TP508 blocked the decrease in PKCcaused by chelerythrine and tamoxifen; however, it is possiblethat the ability of TP508 to increase PKC activity in resting zonecells may play a role in its anti-apoptotic action.

Inorganic phosphate also induced apoptosis in the rat growthplate chondrocytes, as has been noted by others examiningchick growth plate chondrocytes [7]. Interestingly, the effects ofPi in the present study were on resting zone cells, which undernormal physiological conditions would not be exposed to highconcentrations of free phosphate typical of the hypertrophiczone of the growth plate [41,42]. This supports the hypothesisthat regulation of Pi concentration in the extracellular matrix isan important feature of growth plate development [7,43]. Aswas seen with chelerythrine and tamoxifen, TP508 was able toreduce Pi's effects on DNA fragmentation and MTT activity.

The sensitivity of the rat growth plate chondrocytes tochelerythrine and Pi was not sex-specific. Both female and malecells exhibited similar loss of viability and similar increases inDNA fragmentation. Tamoxifen also induced apoptosis in cellsderived from both female and male rats, but there weredifferences in sensitivity; tamoxifen's effect on male cells wasless than on female cells. Tamoxifen is used clinically to blockthe effects of estrogen and has been reported to do so both by itsinhibitory effects on estrogen-dependent PKC signaling [31], aswell as by binding to estrogen receptors [44]. We have shownthat its ability to inhibit PKC in rat growth plate chondrocytes isnot sex-specific [31], but male chondrocytes possess fewerestrogen receptors than female cells [45], which may accountfor the difference in tamoxifen-dependent apoptosis. The fact

that the effects of chelerythrine and tamoxifen were additiveregardless of the sex of the cells indicates that the two PKCinhibitors worked via different mechanisms but the twomechanisms were the same in both sexes.

Our results show that TP508′s anti-apoptotic effect alsoworks for human chondrocytes. Moreover, our results supportthose of others using chick growth plate chondrocyte as a modelthat Pi induces apoptosis by an NO-mediated pathway [15].Here, we show that NO production was also increased incultures treated with chelerythrine and tamoxifen. Treatment ofchelerythrine-stimulated cells with the general NO synthaseinhibitor L-NMMA reduced the apoptotic effects, indicating thatNO production also mediated induction of apoptosis in theresting zone cells. Importantly, TP508 was able to reduce NOproduction caused by all three agents that induced apoptosis inthese cells. In addition, TP508 was able to reduce the increase inDNA fragmentation and the decrease in MTT activity caused bythese three apoptogens, suggesting that its anti-apoptotic effectwas via inhibition of NO production.

Our results indicate that TP508 reduces NO that is producedby iNOS. Specific inhibition of iNOS with 1400W reducedchelerythrine-dependent NO production to a similar extent asTP508. L-NAME, which inhibits both iNOS and eNOS, was

21M. Zhong et al. / Biochimica et Biophysica Acta 1783 (2008) 12–22

more effective than 1400W, suggesting the possibility thateNOS might also play a role. eNOS is present in chick sternalchondrocytes that are undergoing hypertrophy [46] and othershave reported eNOS in rat osteoblasts [47], suggesting that ratchondrocytes may also possess this form of the enzyme.Osteoblasts from eNOS knockout mice exhibit reduced alkalinephosphatase [48] and inhibition of eNOS with L-NAME reducesalkaline phosphatase in the chick cartilage cells [15]. We havepreviously noted that TP508 reduces alkaline phosphataseactivity in rat growth plate chondrocytes [12], further support-ing a role for eNOS. Although nNOS has also been found inchondrocytes [15], the nNOS inhibitor had no effect, indicatingthat nNOS was not involved.

These results indicate that TP508, besides its effect onretaining the rat resting zone chondrocyte phenotype, is alsoable to directly protect it from undergoing apoptosis. Interest-ingly, the effect of TP508 on PKC activation and NO inhibitionwas rapid and sustained over 12 h, indicating nongenomicsignaling and suggesting downstream genomic effects. Expres-sion of Bcl-2 and Bax mRNA and protein was regulated byTP508 in a dose and time-dependent manner, with the effectsoccurring within 1 h. Upregulation of Bcl-2 and down-regulation of Bax are consistent with the hypothesis thatTP508 reduces apoptosis through the Bcl-2/Bax mechanismearly in the response pathway. Our results suggest that the maineffect of TP508 is to expand the reservoir of chondrocytes thatserve to populate the template for endochondral bone formation.This study also confirmed that NO produced by iNOS is acrucial mediator in chondrocyte apoptosis, and suggests a rolefor eNOS in the apoptotic mechanism.

Acknowledgements

This work was supported by grants from OrthoLogic Corp(Tempe, AZ) to Drs. Boyan, Schwartz and Carney; and NSFEEC 9731643 to BDB. TP508 (also known as Chrysalin®) wassupplied by OrthoLogic and is in clinical trials as anexperimental drug not yet approved for human use. Dr. Carneyand Dr. Ryaby have a financial interest in OrthoLogic.

References

[1] D. Simmons, J. Yang, S. Yang, Acceleration of Rat Femoral Fracture Healingby a Synthetic Thrombin Peptide., CalciumMetabolism, BioScientifica Ltd.,1998.

[2] A. Vortkamp, S. Pathi, G.M. Peretti, E.M. Caruso, D.J. Zaleske, C.J. Tabin,Recapitulation of signals regulating embryonic bone formation duringpostnatal growth and in fracture repair, Mech. Dev. 71 (1998) 65.

[3] C.S. Adams, I.M. Shapiro, The fate of the terminally differentiatedchondrocyte: evidence for microenvironmental regulation of chondrocyteapoptosis, Crit Rev. Oral Biol. Med. 13 (2002) 465.

[4] G. Gibson, Active role of chondrocyte apoptosis in endochondralossification, Microsc. Res. Tech. 43 (1998) 191.

[5] M. Hatori, K.J. Klatte, C.C. Teixeira, I.M. Shapiro, End labeling studies offragmented DNA in the avian growth plate: evidence of apoptosis interminally differentiated chondrocytes, J. BoneMiner. Res. 10 (1995) 1960.

[6] K. Ohyama, C. Farquharson, C.C. Whitehead, I.M. Shapiro, Furtherobservations on programmed cell death in the epiphyseal growth plate:comparison of normal and dyschondroplastic epiphyses, J. Bone Miner.Res. 12 (1997) 1647.

[7] K. Mansfield, R. Rajpurohit, I.M. Shapiro, Extracellular phosphate ionscause apoptosis of terminally differentiated epiphyseal chondrocytes,J. Cell Physiol. 179 (1999) 276.

[8] C.C. Teixeira, K. Mansfield, C. Hertkorn, H. Ischiropoulos, I.M. Shapiro,Phosphate-induced chondrocyte apoptosis is linked to nitric oxidegeneration, Am. J. Physiol. Cell Physiol 281 (2001) C833–C839.

[9] H.I. Roach, New aspects of endochondral ossification in the chick:chondrocyte apoptosis, bone formation by former chondrocytes, and acidphosphatase activity in the endochondral bone matrix, J. Bone Miner. Res.12 (1997) 795.

[10] K.C. Glenn, G.H. Frost, J.S. Bergmann, D.H. Carney, Synthetic peptidesbind to high-affinity thrombin receptors and modulate thrombin mitogen-esis, Pept. Res. 1 (1988) 65.

[11] M.R. Sheller, R.S. Crowther, J.H. Kinney, J. Yang, S. Di Jorio, T. Breunig,D.H. Carney, J.T. Ryaby, Repair of rabbit segmental defects with thethrombin peptide, TP508, J. Orthop. Res. 22 (2004) 1094.

[12] Z. Schwartz, D.H. Carney, R.S. Crowther, J.T. Ryaby, B.D. Boyan,Thrombin peptide (TP508) treatment of rat growth plate cartilage cellspromotes proliferation and retention of the chondrocytic phenotype whileblocking terminal endochondral differentiation, J. Cell Physiol. 202 (2005)336.

[13] H.I. Roach, J. Erenpreisa, T. Aigner, Osteogenic differentiation of hyper-trophic chondrocytes involves asymmetric cell divisions and apoptosis,J. Cell Biol. 131 (1995) 483.

[14] H.I. Roach, J. Erenpreisa, The phenotypic switch from chondrocytes tobone-forming cells involves asymmetric cell division and apoptosis,Connect. Tissue Res. 35 (1996) 85.

[15] C.C. Teixeira, H. Ischiropoulos, P.S. Leboy, S.L. Adams, I.M. Shapiro,Nitric oxide-nitric oxide synthase regulates key maturational events duringchondrocyte terminal differentiation, Bone 37 (2005) 37.

[16] S.J. Korsmeyer, J.R. Shutter, D.J. Veis, D.E. Merry, Z.N. Oltvai, Bcl-2/Bax: a rheostat that regulates an anti-oxidant pathway and cell death,Semin. Cancer Biol. 4 (1993) 327.

[17] D. Magne, G. Bluteau, C. Faucheux, G. Palmer, C. Vignes-Colombeix, P.Pilet, T. Rouillon, J. Caverzasio, P. Weiss, G. Daculsi, J. Guicheux,Phosphate is a specific signal for ATDC5 chondrocyte maturation andapoptosis-associated mineralization: possible implication of apoptosis inthe regulation of endochondral ossification, J. Bone Miner. Res. 18 (2003)1430.

[18] S. Carlin, K.X. Yang, R. Donnelly, J.L. Black, Protein kinase C isoforms inhuman airway smooth muscle cells: activation of PKC-zeta duringproliferation, Am. J. Physiol 276 (1999) L506–L512.

[19] L. Piacentini, M. Gray, N.Y. Honbo, J. Chentoufi, M. Bergman, J.S.Karliner, Endothelin-1 stimulates cardiac fibroblast proliferation throughactivation of protein kinase C, J. Mol. Cell Cardiol. 32 (2000) 565.

[20] Y. Zhou, E. Dziak, M. Opas, Adhesiveness and proliferation of epithelialcells are differentially modulated by activation and inhibition of proteinkinase C in a substratum-dependent manner, J. Cell Physiol. 155 (1993)14.

[21] B.D. Boyan, Z. Schwartz, L.D. Swain, D.L. Carnes Jr., T. Zislis,Differential expression of phenotype by resting zone and growth regioncostochondral chondrocytes in vitro, Bone 9 (1988) 185.

[22] R.C. Kinney, Z. Schwartz, K. Week, M.K. Lotz, B.D. Boyan, Humanarticular chondrocytes exhibit sexual dimorphism in their responses to17beta-estradiol, Osteoarthritis. Cartilage. 13 (2005) 330.

[23] T.P. Misko, R.J. Schilling, D. Salvemini, W.M. Moore, M.G. Currie, Afluorometric assay for the measurement of nitrite in biological samples,Anal. Biochem. 214 (1993) 11.

[24] S. Yamamoto, K. Seta, C. Morisco, S.F. Vatner, J. Sadoshima, Chele-rythrine rapidly induces apoptosis through generation of reactive oxygenspecies in cardiac myocytes, J. Mol. Cell Cardiol. 33 (2001) 1829.

[25] J.F. Sweeney, P.K. Nguyen, K.B. Atkins, D.B. Hinshaw, Chelerythrinechloride induces rapid polymorphonuclear leukocyte apoptosis throughactivation of caspase-3, Shock 13 (2000) 464.

[26] S.J. Chmura, M.E. Dolan, A. Cha, H.J. Mauceri, D.W. Kufe, R.R.Weichselbaum, In vitro and in vivo activity of protein kinase C inhibitorchelerythrine chloride induces tumor cell toxicity and growth delay invivo, Clin. Cancer Res. 6 (2000) 737.

22 M. Zhong et al. / Biochimica et Biophysica Acta 1783 (2008) 12–22

[27] R.R. Perry, Y. Kang, B. Greaves, Effects of tamoxifen on growth andapoptosis of estrogen-dependent and -independent human breast cancercells, Ann. Surg. Oncol. 2 (1995) 238.

[28] M.F. El Etreby, Y. Liang, R.W. Lewis, Induction of apoptosis bymifepristone and tamoxifen in human LNCaP prostate cancer cells inculture, Prostate 43 (2000) 31.

[29] W.Z. Gu, Z. Chen, S.K. Tahir, S.H. Rosenberg, S.C. Ng, Synergistic effectof paclitaxel and 4-hydroxytamoxifen on estrogen receptor-negative coloncancer and lung cancer cell lines, Anticancer Drugs 10 (1999) 895.

[30] S. Helm, V.L. Sylvia, T. Harmon, D.D. Dean, B.D. Boyan, Z. Schwartz,24,25-(OH)2D3 regulates protein kinase C through two distinct phospho-lipid-dependent mechanisms, J. Cell Physiol 169 (1996) 509.

[31] Z. Schwartz, V.L. Sylvia, T. Guinee, D.D. Dean, B.D. Boyan, Tamoxifenelicits its anti-estrogen effects in growth plate chondrocytes by inhibitingprotein kinase C, J. Steroid Biochem. Mol. Biol. 80 (2002) 401.

[32] A.Grey, Q. Chen, K. Callon,X.Xu, I.R. Reid, J. Cornish, The phospholipidssphingosine-1-phosphate and lysophosphatidic acid prevent apoptosis inosteoblastic cells via a signaling pathway involving G(i) proteins andphosphatidylinositol-3 kinase, Endocrinology 143 (2002) 4755.

[33] N. Kasagi, Y. Gomyo, H. Shirai, S. Tsujitani, H. Ito, Apoptotic cell death inhumangastric carcinoma: analysis by terminal deoxynucleotidyl transferase-mediated dUTP-biotin nick end labeling, Jpn. J. Cancer Res. 85 (1994) 939.

[34] N.M. Olken, M.A. Marletta, NG-methyl-L-arginine functions as analternate substrate and mechanism-based inhibitor of nitric oxide synthase,Biochemistry 32 (1993) 9677.

[35] E.P. Garvey, J.A. Oplinger, E.S. Furfine, R.J. Kiff, F. Laszlo, B.J. Whittle,R.G. Knowles, 1400W is a slow, tight binding, and highly selectiveinhibitor of inducible nitric-oxide synthase in vitro and in vivo, J. Biol.Chem. 272 (1997) 4959.

[36] B.R. Babu, O.W. Griffith, N5-(1-Imino-3-butenyl)-L-ornithine. A neuronalisoform selective mechanism-based inactivator of nitric oxide synthase,J. Biol. Chem. 273 (1998) 8882.

[37] D.D. Rees, R.M. Palmer, R. Schulz, H.F. Hodson, S. Moncada,Characterization of three inhibitors of endothelial nitric oxide synthasein vitro and in vivo, Br. J. Pharmacol. 101 (1990) 746.

[38] D. Mistry, Y. Oue, M.G. Chambers, M.V. Kayser, R.M. Mason, Chon-drocyte death during murine osteoarthritis, Osteoarthr. Cartil. 12 (2004)131.

[39] M. Whiteman, J.S. Armstrong, N.S. Cheung, J.L. Siau, P. Rose, J.T.Schantz, D.P. Jones, B. Halliwell, Peroxynitrite mediates calcium-dependent mitochondrial dysfunction and cell death via activation ofcalpains, FASEB J. 18 (2004) 1395.

[40] Y. Zhou, E. Dziak, M. Opas, Adhesiveness and proliferation of epithelialcells are differentially modulated by activation and inhibition of proteinkinase C in a substratum-dependent manner, J. Cell Physiol. 155 (1993)14.

[41] D.S. Howell, J.C. Pita, J.F. Marquez, R.A. Gatter, Demonstration ofmacromolecular inhibitors of calcification and nucleational factors in fluidfrom calcifying sites in cartilage, J. Clin. Invest. 48 (1969) 630.

[42] S. Kakuta, E.E. Golub, I.M. Shapiro, Morphochemical analysis ofphosphorus pools in calcifying cartilage, Calcif. Tissue Int. 37 (1985) 293.

[43] D.S. Howell, J.C. Pita, Calcificaiton of growth plate cartilage with specialreference to studies on micropuncture fluids, Clin. Orthop. Relat. Res.(1976) 208.

[44] V.C. Jordan, Antiestrogenic and antitumor properties of tamoxifen inlaboratory animals, Cancer Treat. Rep. 60 (1976) 1409.

[45] E. Nasatzky, Z. Schwartz, W.A. Soskolne, B.P. Brooks, D.D. Dean, B.D.Boyan, A. Ornoy, Evidence for receptors specific for 17 beta-estradioland testosterone in chondrocyte cultures, Connect. Tissue Res. 30 (1994)277.

[46] M.V. Hukkanen, L.A. Platts, D.M. Fernandez, I.M. O'Shaughnessy, I.MacIntyre, J.M. Polak, Developmental regulation of nitric oxide synthaseexpression in rat skeletal bone, J. Bone Miner. Res. 14 (1999) 868.

[47] S.W. Fox, J.W. Chow, Nitric oxide synthase expression in bone cells, Bone23 (1998) 1.

[48] J. Aguirre, L. Buttery, M. O'Shaughnessy, F. Afzal, D.M. Fernandez, I.M.Hukkanen, P. Huang, I. MacIntyre, J. Polak, Endothelial nitric oxidesynthase gene-deficient mice demonstrate marked retardation in postnatalbone formation, reduced bone volume, and defects in osteoblast maturationand activity, Am. J. Pathol. 158 (2001) 247.