[CANCER RESEARCH 60, 3470–3476, July 1, 2000]

A Plasmin-derived Hexapeptide from the Carboxyl End of Osteocalcin CounteractsOxytocin-mediated Growth of Inhibition of Osteosarcoma Cells1

J. F. Novak,2 M. B. Judkins, M. I. Chernin, P. Cassoni, G. Bussolati, J. A. Nitche, and S. K. NishimotoDepartment of Biology, Bucknell University, Lewisburg, Pennsylvania 17837 [J. F. N., M. B. J., M. I. C., J. A. N.]; Department of Biomedical Sciences and Oncology, University ofTurin, Turin, Italy 10126 [P. C., G. B.]; and Department of Biochemistry, University of Tennessee, Memphis, Tennessee 38163 [S. K. N.]

ABSTRACT

We have previously described the presence of the functional plasmin-ogen activator system on the surfaces of bone neoplastic cells and the factthat plasmin specifically cleaves bone matrix protein osteocalcin (OC).The cleavage of OC to NH2-midterminal (1–44) and COOH-terminalRFYGPV hexapeptide (44–49) proceeds with detachment of both prod-ucts from bone mineral. Because the sequence of OC-derived hexapeptide(HP) is nearly identical to the E2 region of the oxytocin receptor (OTR),we set out to ascertain whether the HP interferes with the osteosarcoma(OS)-associated oxytocin (OT) system. We documented the presence andfunctional activity of OTRs in several OS cells by means of (a) OT-mediated inhibition of OS growth; (b) expression of OTR mRNA bymeans of reverse transcription-PCR; (c) immunofluorescence stainingwith IF3 monoclonal antibody specific for human OTR; and (d) saturationbinding and Scatchard analysis of OT binding to the receptors of isolatedmembranes or intact OS cells. Although we could not demonstrate directbinding of HP to OT, the presence of HP in cultures of OS cells antago-nizes the inhibitory effect of OT on these cells. Additionally, in competitivebinding assays, the HP effectively competes with binding of OT to itscognate receptors. The results indicate the existence of an OTR/OT systemin tumor cells of bone origin. Suggested plasminogen activator-OC-OTR/OT interactions may have an effect on the regulation of cell prolif-eration within the bone tissue as well as properties of the extracellularmatrix surrounding the tumor foci in the bone.

INTRODUCTION

The mechanism of bone metastasis from extraskeletal tumors aswell as the progression of primary bone tumorsin situ through thebone tissue is an active area of research (1, 2). The PA3 system hasbeen implicated in extracellular matrix degradation, facilitation oftumor proliferation, local invasion, and metastasis (3). Plasminogen, asubstrate of PAs, may be readily available in the bone marrow/boneinterface via bone marrow sinusoids that exhibit large fenestrationsthat facilitate the exchange of plasma proteins. Once in the interstitialtissue space, the plasminogen may be retained by high-capacity/low-affinity receptors present on osteoblast cells (4). The conformation ofthe cell-bound plasminogen renders this molecule amenable to cleav-age and activation by PAs (5), resulting in a plasmin that remainsbound to the same receptor with an increased affinity (6). Plasmin isa serine protease acting on extracellular matrix proteins or activatinglatent matrix metalloproteases or growth factors (7, 8). A bone-specific effect of plasmin could be mediated by the bone protein OC.OC is a highly conserved protein produced predominantly by osteo-blasts late in the mineralization process. The precise function of OC

in bone metabolism is not known. A relationship between serum OCand bone formation led to the hypothesis that OC is a marker forenhanced bone activity (9, 10).

We found that plasmin can cleave OC at a single site within itsCOOH end. The cleavage creates a NH2-midterminal 1–43 peptideand a short COOH-terminal 44–49 HP (RFYGPV) (11). The NH2-midterminal peptide is the most abundant fragment in serum (12).Plasmin cleaves OC both in solution and when bound to hydroxyap-atite. When treated with plasmin, both OC cleavage products detachfrom the hydroxyapatite. We hypothesize that the plasmin-mediatedlysis of the free and hydroxyapatite-bound OC could be responsiblefor the abundant NH2-midterminal peptide and COOH-terminal HP inserum. If so, plasmin cleavage of OC could play a role in OCmetabolism as well as bone homeostasis. The COOH-terminal pen-tapeptide was previously shown to function as a cellular chemoattrac-tant (13). The objective of this investigation was to ascertain whetherthe COOH-terminal HP RFYGPV produced by plasmin cleavage ofOC has biological activity. The COOH-terminal RRFYGPV se-quence, a plasmin cleavage site in OC, is evolutionarily conserved intetrapod vertebrates from frog to man (14–15). A BLASTP4 searchfor proteins including the sequence . . . R-F-Y-G-P-V. . . matched theOTR in which the RFYGPD sequence constitutes a part of the highlyconserved second extracellular E2 region (Fig. 1). The E2 region isessential for OT binding (16). Here we present evidence for theinteraction between plasmin action, the OC HP, and OTR in modu-lating the growth of OS cells.

MATERIALS AND METHODS

Materials. OC HP (NH2-Arg-Phe-Tyr-Gly-Pro-Val-COOH) was synthe-sized by Bio-Synthesis, Inc. (Lewisville, TX). OT acetate was purchased fromSigma (St. Louis, MO). The peptides were reconstituted in PBS (pH 8) andstored at280°C. [3H]OT was obtained from New England Nuclear (Boston,MA); [tyrosyl-2,6-3H]OT, 43.5 Ci/mmol].

Cells. Human OS cell lines U2OS and MG-63 were purchased from Amer-ican Type Culture Collection (Manassas, VA). Established human OS celllines OS9 and OS15 were generously provided by Dr. Nicola Baldini (IstitutiOrtopedici Rizzoli, Bologna, Italy). OS9 cells were established from bonemetastasis, whereas OS15 cells are from primary osteoblastic OS. MCF-7human breast carcinoma cell line was provided by Dr. David Beidler (Univer-sity of Michigan, Ann Arbor, MI). The cells were maintained under thefollowing conditions:a-MEM/F12 (1:1 mixture of MEM and F-12 nutrientmixture) media were used for MG-63 and U2OS cells; Iscove’s modifiedessential medium was used for maintaining OS9, OS15, and MCF-7 cells. Allmedia (Life Technologies, Inc., Grand Island, NY) were enriched with 10%heat-inactivated fetal bovine serum (PAA Laboratories, Inc., Parker Ford, PA),1 mg/ml streptomycin, and 1000 units/ml penicillin (Sigma). All cells weremaintained at 37°C in a humidified 5% CO2 atmosphere.

Cell Growth Studies. U2OS and MG-63 cells were seeded into 6-welltissue culture plates (2.53 104 cells/well; Corning, Cambridge, MA) using theserum-rich media as described above (2 ml/well). After 16 h, the media werereplaced with fresh media containing appropriate peptide concentrations. Thecontrol and experimental media were replaced every 48 h until the time ofharvest [120 h (U2OS cells) and 144 h (MG-63 cells) after peptide treatment].At the end of culture periods, the cells were harvested by trypsinization and

Received 12/3/99; accepted 5/3/00.The costs of publication of this article were defrayed in part by the payment of page

charges. This article must therefore be hereby markedadvertisementin accordance with18 U.S.C. Section 1734 solely to indicate this fact.

1 Preliminary results were reported at “Towards the Eradication of OsteosarcomaMetastases” meeting held in Oslo, August 1998, and the 90th AACR Meeting, Philadel-phia, March 1999.

2 To whom requests for reprints should be addressed, at Department of Biology,Bucknell University, Lewisburg, PA 17837. Phone: (570) 577-1286; Fax: (570) 577-3537;E-mail: novak@bucknell.edu.

3 The abbreviations used are: PA, plasminogen activator; OC, osteocalcin; HP,hexapeptide; OT, oxytocin; OTR, oxytocin receptor; OS, osteosarcoma; RT-PCR, reversetranscription-PCR; MAb, monoclonal antibody. 4 www.ncbi.nlm.nih.gov/BLAST/.

3470

Research. on January 1, 2016. © 2000 American Association for Cancercancerres.aacrjournals.org Downloaded from

counted with a hemocytometer. Each condition was repeated with eight indi-vidual samples, and statistical analysis was performed with SPSS.

RNA Extraction. Total RNA was extracted by the method described byChomczynski and Sacchi (17) using 4M guanidinium thiocyanate, 0.75Msodium citrate (pH 7), 10% sarcosyl, and 0.1M 2-mercaptoethanol. All solu-tions were prepared in diethylpyrocarbonate-treated water.

RT-PCR. The first-strand cDNA synthesis was performed with 1mg oftotal RNA using Moloney murine leukemia virus reverse transcriptase pro-vided with Amplimer Sets for RT-PCR (Clontech, Palo Alto, CA). Oligo(dT)18

primer (20 mM) was added to the RNA preparation, and the samples weredenatured at 70°C for 2 min and quenched on ice. To complete cDNAsynthesis, the following components were added to each reaction: 53 reactionbuffer (4 ml); deoxynucleotide triphosphate mix (10 mM each); recombinantRNAse inhibitor (20 units); and Maloney murine leukemia virus reversetranscriptase (200 units). The reaction was incubated at 42°C for 1 h and thenheated at 94°C for 5 min to stop cDNA synthesis and destroy DNase activity.The following PCR protocol was revised from that provided along withAdvatage RT-for-PCR kit (Clontech). Each reaction contained the followingreagents (volume/final concentration): 103buffer KCl-Tris-HCl (5 ml/Tris-HCl, 10 mM; KCl, 50 mM); MgCl2 (3 ml/1.5 mM); deoxynucleotide triphos-phate mix (1ml/0.2 mM each); cDNA [5ml; (1:20)]; primers (1.25ml of 59,1.25 ml of 39; 12.5 pM); and AmpliTaq DNA polymerase (0.4ml/2.0 units).Oligonucleotide primers (OTR-sense, 59-CCTTCATCGTGTGCTGGACG-39;OTR-antisense, 59-CTAGGAGCAGAGCACTTATG-39) were synthesized byGenosys Biotechnologies, Inc. (Woodlands, TX). These primers were designedto amplify a 391-bp fragment of the OTR cDNA and modeled after those usedby Ito et al. (18). A PvuII digest of the PCR product produced the expected118- and 273-bp fragments. Target mRNA was amplified using the followingtemperatures and intervals: critical denaturing (94°C, 1 min), annealing (59°C,1 min), and strand elongation (72°C, 2 min) temperatures for 30 cycles in athermometer (ERICOMP Single Block System). PCR reaction products werestored at 4°C until analyzed by gel electrophoresis.

Immunofluorescence and Flow Cytometry.The MG-63 OS cells weregrown on glass coverslips to subconfluent density. The cells were washed withPBS, fixed in paraformaldehyde for 5 min at room temperature, and incubatedfor 30 min with a primary monoclonal antibody (MAb) to human OTR, cloneIF3 (19), diluted to 1:20 in PBS. Fluorescein-labeled secondary antiserumSera-Lab Ltd., Sussex, England) diluted 1:10 in PBS was used for 30 min atroom temperature.

The MG-63 cells (53 106 cells ml) were resuspended in PBS and fixed inparaformaldehyde for 30 min at 4°C with anti-OTR MAb IF3. After twowashes, the cells were incubated at 4°C for 30 min with fluorescein-labeledsecondary antiserum. Fluorescence was analyzed by means of a FACSort(Becton Dickson, San Jose, CA).

Isolation of Plasma Membrane. Membranes were isolated from cellscollected by rubber policeman into a cell collection buffer [250 mM sucrose, 50

mM Tris-HCl (pH 7.4), 5 mM MgCl2, 1 mM EDTA, and 1 mM phenylmethyl-fulfonyl fluoride]. The cells were collected by centrifugation and homogenizedin a homogenization buffer [50 mM Tris-HCl (pH 7.4), 5 mM MgCl2, 1 mM

EDTA, and protease inhibitors mixture (Sigma)] using a glass/Teflon Elve-hjem homogenizer (five cycles each 10 s using a Wheaton Instruments over-head stirrer). Large debris was removed by centrifugation (1,5003 g, 10 min,4°C), and the supernatant was centrifuged for 30 min at 160,0003 g. Thecrude membrane preparation was resuspended in binding buffer (see below),and its protein concentration was determined by Bradford assay (Bio-Rad,Hercules, CA).

Radioligand Binding Assays.The binding to the membrane receptors wasconducted with isolated membranes and intact cells. The intact cells, whenused for binding studies, were scraped into a Locke’s solution [154 mM NaCl,5.6 mM KCl, 5.6 mM glucose, 5 mM 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid, 40 mg/liter gentamycin, 200,000 units/liter penicillin, 2 g/literBSA, 1 mM MgCl2, and 2.3 mM CaCl2] containing 1 mM phenylmethylsulfonyl fluoride, collected by centrifugation, and enumerated by counting ina hemocytometer. The cells were resuspended into binding buffer or Locke’ssolution without Ca21 and Mg21 ions.

The saturation binding was conducted with increasing concentrations of[3H]OT in the presence or absence of 1mM unlabeled OT at 4°C for 4 h. Thebinding was conducted in duplicate or triplicate assays. The membranes orcells were retained on the Whatman GF/B filters (Clifton, NJ). The filters werepresoaked in the binding buffer for 1 h, and filtration was conducted withLocke’s solution without divalent ions. The filters were placed into scintilla-tion vials, and radioactivity was determined with a Packard scintillation coun-ter fitted with the quenching correction curves. The saturation binding datawere determined using the Prism (GraphPad Software, Inc., San Diego, CA)nonlinear least square analysis program.

Competition Binding Assays. Competition binding studies were con-ducted with one concentration of [3H]OT and increasing concentrations of OCHP with and without the presence of 1mM OT. The incubations and analyseswere performed as described above, and IC50 values were derived fromnonlinear least square analysis.

RESULTS

Growth of OS Cells Is Modulated by OT and the OC COOH-terminal HP

OT Inhibits OS Cell Growth. Previously, we found that thegrowth of cells derived from neuroectodermal and mammary tumorsis inhibited by OT (20, 21). To determine the growth effect of OT andthat of OC COOH-terminal HP, the 2.53 104 OS cells were seededinto 6-well plates and treated with peptides as described in “Materialsand Methods.” Fig. 2 shows the effect of OT on the growth of twoosteoblast-like OS cell lines. OT significantly depresses the cellnumber when present in the complete medium at 1027

M (MG-63, Fig2A, P , 0.02; U2OS, Fig. 2B,P , 0.001). The inhibition of growthwas concentration dependent because OT at 1028

M had only a minoreffect on the growth of the cells. The growth inhibition becamesignificant after more than 100 h of growth in the presence of OT.Preliminary tests for hormone-induced apoptosis did not indicateevidence of DNA ladder formation. A similar OT-mediated inhibitionof growth was observed for OS15 and OS9 OS cell lines. Thus, OTinhibits all tested OS cell lines (MG-63, U2OS, OS9, and OS15) in thepresence of 10% heat-inactivated FCS.

The OT-mediated Growth Inhibition Is Reversed by OCCOOH-terminal HP. Realization that the COOH-terminal peptide ofOC matches the essential and conserved sequence of the E2 region ofthe OTR led us to experiments that tested the effects of HP on thegrowth of OS cells. Fig. 2,C andD, shows that the HP alone at 1026

M to 1029M has a somewhat promoting but not significant effect on

OS cell growth.Fig. 3 shows that the simultaneous addition of OT and OC HP at

equimolar concentrations negates the growth-inhibitory activity of OT

Fig. 1. Schematic representation of OTR and its common sequence with the COOH-terminal of OC. The four extracellular regions labeledE1 throughE4are composed of theNH2-terminal and three loops between the seven transmembrane domains. The binding ofOT to the receptor requires intact E1 and E2 regions for high-affinity binding as indicatedby the arrow. The similarity between the OTR E2 and COOH-terminal OC sequences hasbeen found with BLASTP program.

3471

INTERACTION BETWEEN OC AND OT IN OS

Research. on January 1, 2016. © 2000 American Association for Cancercancerres.aacrjournals.org Downloaded from

alone for MG-63 cells. Addition of oxytocin at 1027M decreased cell

number compared to control cultures in full media (P , 0.002), butthe presence of both the HP and OT at 1027

M recovered control levelsof cells. Identical responses were seen for U2OS cells (data notshown). Lower doses of HP 1028 and 1029

M did not reverse thegrowth-inhibitory activity of OT (data not shown).

OS Cells Synthesize and Express OTR

RT-PCR Evidence That OS Cells Express OTR mRNA.Growth inhibition of OS cell lines by OT implies the presence ofOTRs in OS cells. Synthesis of OTR mRNA was verified by PCRamplification of the expected 391-bp product from cDNA mixturesreverse-transcribed from OS cell mRNA. RT-PCR was conductedwith 10 mg of total RNA and oligonucleotide primers specific forOTR cDNA sequences (18). Fig. 4 shows that MG-63 and U2OS OScells express the OTR, as does the breast carcinoma cell line MCF-7used here as a positive control (21). The OS15 and OS9 OS cells alsoexpress the OTR (data not shown). In parallel experiments, we found

that HT29 colon cancer cells do not express OTR, and these cells weretherefore used as a negative control.

OTRs Are Present on Cell Surface of OS Cells.The presence ofOTR on cell membranes of the OS cells was studied by means of theOTR-specific IF3 MAb, followed by visualization via fluorescencesecondary antibody. Fig. 5 shows that OTR is localized on the cellsurfaces and exhibits either a uniformly stained pattern all along thecell periphery or manifests itself as bright heavily stained spots.MCF-7 cells that were previously shown to express OTR were used asa positive control. A replacement of IF3 with a normal mouse serum

Fig. 2. OT inhibits OS cell proliferation. Osteoblast-like OS cell lines MG-63 andU2OS exhibit a dose-dependent decrease in cell number when grown in the presence ofOT. The asterisk indicatesP , 0.02 (t test for paired samples), thedouble asteriskindicatesP , 0.001;n 5 8 for all data.A andC, MG-63 mean cell number after 144-hincubation with OT at 1028 and 1027 M (dark bars,A), or with 1029 and 1026 M OCCOOH-terminal HP (hatched bars,C) compared to control cells cultured in controlmedium (h).B andD, U2OS mean cell number after 120-h incubation with OT (o),B)or OC COOH-terminal HP (hatched bars,D) compared to the control cells cultured incontrol medium.Error bars, SD.

Fig. 3. OC COOH-terminal HP blocks the growth-inhibitory effect of OT. MG-63 OScells were cultured in control medium (Control), medium containing 1027 M OT, ormedium with 1027 OT and 1027 M OC COOH-terminal HP (OT1 HX); n 5 8 for allconditions. Theasteriskindicates that the OT-treated cultures were significantly different(P , 0.02) from control or OT1 HP-treated cells.Error bars, SD. Note that 1027 M HPalone has no effect on cell number, as indicated by the experiment in Fig. 2.

Fig. 4. Proof of OTR gene expression in OS cells by RT-PCR. Isolated mRNA fromOS cell lines MG-63 and U2OS or MCF-7 breast carcinoma cells produced the expected391-bp PCR product after reverse transcription using oligo(dT) primers and amplificationby intron-spanning OTR-specific PCR primers. The internal positive controls were PCRamplification of the OTR plasmid (data not shown) and RT-PCR or mRNA and primersprovided with the kit (Clontech), which resulted in a band of 900 bp as expected. Thenegative control was RT-PCR of HT29 colon carcinoma cells. Theleft lane showsmolecular weight markers and the estimated size (bp). RT-PCR of OS mRNAs wasperformed as described in the “Materials and Methods.” Twoml of the PCR reactionswere mixed with glycerol-bromophenol blue containing sample buffer, loaded onto 2%agarose gel, electrophoresed, and visualized by UV fluorescence after ethidium bromidestaining.

Fig. 5. Immunofluorescence of the OTR. Immunofluorescence microscopy was con-duced as described in the “Materials and Methods” using IF3 MAb. The antibody stainingrevealed the presence of OTR on the surfaces of the majority of MG63 OS cells.

3472

INTERACTION BETWEEN OC AND OT IN OS

Research. on January 1, 2016. © 2000 American Association for Cancercancerres.aacrjournals.org Downloaded from

and the use of IF3 with HT29 cells colon carcinoma cells that do notexpress OTR constituted negative controls.

The fluorescence intensity and the percentage of OTR-positive cellswere also determined by fluorescence-activated cell-sorting analyses.As reported in Fig. 6, the large majority of MG-63 cells (;90%)showed a strong reactivity to OTR antibodies. The results wereidentical in unfixed (data not shown) and paraformaldehyde-fixedcells. An identical pattern was observed in U2OS cells (data notshown).

OTRs of OS Cells Exhibit Saturation Binding that Can BeInhibited by OC COOH-terminal HP

Intact OS Cells or Cell Membranes Exhibit Saturation Bindingof [3H]OT Ligand. Fig. 7 shows that [3H]OT binding to intact U2OScells is saturable with a calculatedKd of 7.1 3 1028

M by Scatchardanalysis. Fig. 8 shows that [3H]OT binds to the membranes isolatedfrom MG-63 cells in a saturation manner (Kd 5 1.5 3 1028

M).IndicatedKd values are significantly lower than the values reportedfor OT binding to endogenous OTR in OT target tissues (22).

The COOH-terminal HP of OC Inhibits Binding of OT to CellMembrane Receptors.Fig. 9 shows that when increasing concen-trations of the OC COOH-terminal HP are present, there is a dose-dependent inhibition of [3H]OT binding by isolated MG-63 cellmembranes. The calculated IC50 for inhibition is 1.53 1027

M. There

was no demonstrable binding of iodinated HP to either the cellmembranes or OT (data not shown). It is possible that iodination ofthe tyrosine abrogated binding because the tyrosine is a conservedfeature of both the HP and the OTR E2 domain sequence RFYGPD.

DISCUSSION

Until recently, the OT/OTR system has been studied exclusively inthe context of its functions in traditional target tissues and cells:epithelial cells in the breast-smooth muscle of the uterus; and hypo-thalamic neurons of the brain (23). Consequently, milk ejection,initiation of parturition, and behavioral effects are well known phys-iological processes controlled in part by OT. Recently, however, theOTR has been found in a number of primary tumor cells of bothepithelial [breast and endometrium (19, 24)] and neural origin [neu-roblastoma and glioblastoma (25)]. Here we report on the expressionof the OTR by OS cells, a tumor cell of mesenchymal origin. TheOTR has been localized on OS cell membranes by immunofluores-cence, and OT displays saturation binding to both isolated membranesand intact cells. The finding of OT-mediated inhibition of cell growthsuggests that the OTR is connected to the growth-regulatory path-ways. The functionality of the OT/OTR system in other tumor cells

Fig. 6. One-parameter flow-cytometric analysis of the OTR reactivity with MG-63cells. Anti-OTR IF3 MAb shows a strong reaction with the majority (;90%) of MG-63OS cells.C-axis, fluorescence intensity/cell;y-axis, number of cells registered/channel. Atotal of 20,000 of cells was analyzed.

Fig. 7. Specific binding of OT to isolated U2OS cells.Increasing concentrations of [3H]OT added to intact U2OScells exhibited saturation binding. Nonspecific binding esti-mated in the presence of a 1000-fold excess of unlabeled OTwas subtracted from total binding as described in “Materialsand Methods.” Scatchard analysis (inset) reveals a calculatedKd of 7.1 3 1028 M.

Fig. 8. Specific binding of OT by isolated cell membranes of MG-63 OS cells.Increasing concentrations of [3H]OT added to isolated membranes of MG-63 cellsexhibited saturation binding. Nonspecific binding estimated in the presence of 1000-foldexcess of unlabeled OT was subtracted from total binding as described in “Materials andMethods.” Calculations from Scatchard analysis yield aKd of 1.5 3 1028 M.

3473

INTERACTION BETWEEN OC AND OT IN OS

Research. on January 1, 2016. © 2000 American Association for Cancercancerres.aacrjournals.org Downloaded from

has been previously ascertained because OT inhibited proliferation ofsome breast cancer cell lines (20) but promoted their growth underdifferent cell culture conditions (26). The differences in responses arelikely to depend on the cell subclone, its steroid responsivity, specificexperimental conditions, and the presence of appropriate ligands intheir surroundings. The steroid hormone progesterone has been foundrecently to bind the OTR and to decrease OT binding (27). Otherconditions may change the affinities and/or specificity of ligands fortheir receptors, including changes in the amino acid sequences andinteractions between different receptor subunits, neighboring mole-cules at the cell membrane, or the presence of similar ligands (28).Thus, both biological and experimental conditions may pertain to theOT/OTR system in OS cells (Kd 5 15 nM, membranes; 71 nM, intactcells) and may explain the difference inKd values obtained with targettissues [;2–3 nM (29)] and those with model cells [0.6 nM (30)].

OC is known to affect several processes that impact on bone

remodeling. This is supported by its ability to bind osteopontin (31),collagen type I (32), and bone mineral with high affinity and ste-reospecificity (33). In addition, OC inhibits cross-linking of osteopon-tin by transglutaminase (34). OC was demonstrated to alter adhesive,chemotactic, and gene expression properties of giant tumor cells, aneoplastic counterpart of osteoclasts (35). Thus, the plasmin-mediatedlysis of OC may involve the release of OC not only from the hy-droxyapatite (11) but also from other components of the extracellularmatrix. The action of cell-bound plasmin on extracellular proteins hasspecific consequences in neural tissues. Deprivation of cell attachmentsubstrates via plasmin digestion of laminin in normal neural tissuesleads to anoikis of neurons (36). Plasmin cleavage of OC and its lossfrom the bone mineral should lead to profound changes in the extra-cellular matrix surrounding bone cells and should therefore affectbone remodeling. Both bone density and cortical bone thickness areincreased in OC-null mice (37), suggesting that either the intact OC orone of its fragments may be a negative regulator of bone formation.Plasmin cleavage of the conserved site within OC and subsequentrelease of active peptides with novel properties would constituteproteolytic activation of OC, not unlike that of other vitamin K-dependent proteins involved in blood homeostasis (38). Thus, inaddition to the effect of intact OC, released OC peptides, a COOH-terminal HP (11) and NH2-mid peptide (39), could also have newbiological activities. This may be substantiated by the fact that theplasmin-mediated cleavage occurs at a highly evolutionary conservedRR site of the OC COOH-terminal sequence. The similarity betweenthe COOH-terminal OC HP NH2-RFYGPV-COOH and E2 region ofOTR-RFYGPD- might extend to the sixth position valine becauseboth COOH-terminal valine and aspartic acid have negatively chargedcarboxylate groups. Our results support the interaction between HP,the OTR, and OT. The OC-derived HP reverses the growth-inhibitoryeffect of OT, a phenomenon with potential significance forin situregulatory bone mechanisms. This presumably could occur via bind-ing to the OT itself, a possibility that we could not confirm by any ofthe following methods: gel chromatography; native gel polyacryl-amide electrophoresis; and capillary electrophoresis (results notshown). These negative data are reminiscent of the recent report that

Fig. 9. The OC COOH-terminal HP inhibits binding of OT to cell membrane receptors.Increasing concentrations of OC COOH-terminal HP inhibit binding of OT to MG-63 cellmembranes. At 1.53 1027 M HP, one half of the binding of OT to cell membranes wasblocked (IC50).

Fig. 10. The model depicting interrelationsamong the PA, OC-derived active HP and OTsystem in bone-derived cells. The model starts atthe upper left, with an osteoblast (OB), OS cell(OS), or metastatic bone tumor cell (CA)with PA(uPA) and plasminogen (Plg) bound to its surface.The activated plasmin (Pl) that results from theactivation of plasminogen remains bound to thesurface of the cell. Plasmin cleaves OC, whichcauses detachment of OC from the bone matrix andthe release of OCNH2 midterminal fragment (N-mid) and COOH-terminal HP (HP). The HP canblock the inhibitory effect of circulating OT to itsreceptor (OTR). The presence of both HP and OTin the bone environment enhances cell growth oftumor cells within bone or may modulate remod-eling processes in the normal bone environment(inset).

3474

INTERACTION BETWEEN OC AND OT IN OS

Research. on January 1, 2016. © 2000 American Association for Cancercancerres.aacrjournals.org Downloaded from

peptides representing various extracellular regions of the vasopressinreceptor interfere with vasopressin binding without measurable bind-ing to the ligand itself (40). We conclude that the interference of HPmust occur at the site of the receptor because HP competes for OTbinding with IC50 of 1.7 3 1027

M.The enhanced expression of a cell surface plasmin system is one of

the hallmarks of the neoplastic and metastatic states (41) and apredictor of poor prognosis in mammary and other tumors (42). Wehave shown that such a system is also active on OS cells (4, 43). Theproduction of COOH-terminal OC by a cell surface plasmin system inbone may have relevance to thein situ development of OS and theability of other tumors to metastasize into bone. We suggest (Fig. 10)that the presence of HP could enhance tumor growth by relievingOT-mediated growth inhibition of OS cells. By virtue of the nearlyexclusive occurrence of OC in skeletal tissues, the plasmin-OC OTsystem would be specific for the bone environment and could there-fore determine the success of OS cells or metastatic bone tumors toproduce localized growth in skeletal tissues. Interactions among theOTR/OT system, OC, and steroid hormones may also reflect on thepossible role of these mechanisms in the development of bone met-abolic diseases such as osteoporosis. Similar interactions could alsooccur in normal or metabolic disease-affected bone tissue where allthree components of the PA-OT/OTR system were already docu-mented: (a) hormonally regulated PA system (44); (b) presence of OCfragments (12) resembling those derived by plasmin (11); and (c)high-affinity OTRs in normal osteoblasts (45).

In conclusion, we have shown that OS cells express the OTR andthat the binding of this receptor by its ligand results in the inhibitionof growth. Our results with osteoblast-like cells in culture add to agrowing list of tumor systems that express endogenous OTR and inwhich the OT/OTR systems have been shown to modulate either thegrowth or properties of the tumor cells. The participation of theOT/OTR system in a number of novel physiological mechanisms hasbeen documented (46, 47), and in addition, normal human osteoblasts(45) and endothelial cells (48) have also been recently shown toexpress OTRs. These results suggest that OT possesses much broaderphysiological significance for both females and males than previouslythought. The present findings may be of particular significance forunderstanding primary bone tumors and those neoplasms with pred-ilection for metastatic growth in the bone. It may be of importance thatOS cells can express OC in both pre- and postproliferative stages ofgrowth (49), whereas the normal osteoblasts do so only during min-eralization phase (50). Nearly without exception, the tumors andparticularly their metastatic variants are known to express a cellsurface plasmin system. We suggest that the ability to invade bonemay be due in part to release of the OC COOH-terminal peptide thatin turn acts as a potent antagonist of OT binding to the OTR. We planto evaluate the proposed sequence of events in a suitablein vivomodelof bone neoplastic disease. It could be of therapeutic value to identifycompounds that could interfere with degradation of OC. Alternatively,potent ligands for OTRs and strong competitors for OC HP could alsolead to the inhibition of tumor growth within the bone environment.

REFERENCES

1. Krempien, B. Morphological findings in bone metastasis, tumorosteopathy and anti-osteolytic therapy.In: I. J. Diel, M. Kaufmann, and G. Bastert (eds.), Metastatic-BoneDisease. Fundamental and Clinical Aspects, pp. 59–85. Berlin: Springer-Verlag,1994.

2. Arguello, F., and Cohen, H. J. Structural and microenvironmental bone marrowinvolved in the pathogenesis of bone metastasis.In: F. W. Orr and G. Singh (eds.),Bone Metastasis—Mechanisms and Pathophysiology, pp. 31–45. Austin, TX: R. G.Landes Company, 1996.

3. Kwaan, H. C. The plasminogen-plasmin system in malignancy. Cancer MetastasisRev.,11: 291–311, 1992.

4. Campbell, P. G., Wines, K., Yanosick, T. B., and Novak, J. F. Binding and activationof plasminogen on the surface of osteosarcoma cells. J. Cell. Physiol.,159: 1–10,1994.

5. Gonzalez-Gronow, M., Stack, S., and Pizzo, S. V. Plasmin binding to the plasmin-ogen receptor enhances catalytic efficiency and activates the receptor for subsequentligand binding. Arch. Biochem. Biophys.,286: 625–628, 1991.

6. Correc, P., Zhang, S., Komano, O., Laurent, M., and Burtin, P. Visualization of theplasmin receptor on carcinoma cells. Int. J. Cancer,50: 767–771, 1992.

7. Alexander, C. M. and Werb, Z. Proteinases and extracellular matrix remodeling. Curr.Opin. Cell Biol.,1: 974–982, 1989.

8. Novak, J. F., McMaster, J. H., and Campbell, P. G. Regulation of transforminggrowth factor-b binding to type III receptor in osteosarcoma by plasmin.In: J. F.Novak, and J. H. McMaster (eds.), Frontiers in Osteosarcoma Research, pp. 525–529.Seattle, WA: Hogrefe and Huber, 1993.

9. Nishimoto, S. K., Chang, C. H., Gendler, E., Stryker, W. F., Nimmi, M. E. The effectof aging on bone formation in rats: biochemical and biological evidence for decreasedbone formation capacity. Calcif. Tiss. Int.,37: 617–624, 1985.

10. Seibel, M. J., and Pols, H. A. P. Clinical application of bone biochemical markers ofbone metabolism.In: J. P. Bilezikian, L. G. Raisz, and G. A. Rodan (eds.), Principlesof Bone Biology, pp. 1293–1311. San Diego, CA: Academic Press, 1996.

11. Novak, J. F., Hayes, J. D., and Nishimoto, S. K. Plasma-mediated proteolysis ofosteocalcin. J. Bone Mineral Res.,12: 1035–1042, 1997.

12. Garnero, P., Grimaux, M., Demiaux, B., Preaudat, C., Seguin, P., and Delmas, P. D.Measurement of serum osteocalcin with a human-specific two-site immunoradiomet-ric assay. J. Bone Mineral Res.,7: 1389–1398, 1992.

13. Mundy, G. R., and Poser, J. W. Chemotactic activity of theg-carboxyglutamic acidcontaining protein in bone. Calcif. Tissue Int.,35: 164–168, 1983.

14. Hauschka, P. V., Lian, J. B., Cole, D. E. C., and Gundberg, C. M. Osteocalcin andmatrix Gla protein: vitamin K-dependent proteins in bone. Physiol. Rev.,69: 990–1047, 1989.

15. Cancela, M. L., Williamson, M. K., and Price, P. A. Amino acid sequence of bone Glaprotein from the african toadXenopus laevisand the fishSparus aurata. Int. J. PeptideProtein Res.,46: 419–423, 1995.

16. Postina, R., Kojro, E., and Fahrenholz, F. Separate agonist and peptide antagonistbinding to sites of the oxytocin receptor defined by their transfer into V2 vasopressinreceptor. J. Biol. Chem.,271: 31593–31601, 1996.

17. Chomczynski, P., and Sacchi, N. Single-step method of RNA isolation by acidguanidium thiocyanate-phenol-chloroform extraction. Anal. Biochem.,162: 156–159, 1987.

18. Ito, Y., Kobayashi, T., Kimura, T., Matsuura, N., Wakasugi, E., Takeda, T., Shimano,T., Kubota, Y., Nobunaga, T., Makino, Y., Azuma, C., Saji, F., and Monden, M.Investigation of the oxytocin receptor expression in human breast cancer tissue usingnewly established monoclonal antibodies. Endocrinology,137: 773–779, 1996.

19. Bussolati, G., Cassoni, P., Ghisolfi, G., Negro, F., and Sapino, A. Immunolocalizationand gene expression of oxytocin receptors in carcinomas and non-neoplastic tissuesof the breast. Am. J. Pathol.,148: 1895–1903, 1996.

20. Cassoni, P., Sapino, A., Negro, F., and Bussolati, G. Oxytocin inhibits proliferationof human breast cancer cell lines. Virchows Arch.,425: 467–472, 1994.

21. Cassoni, P., Sapino, A., Papotti, M., and Bussolati, G. Oxytocin and oxytocin-analogue F314 inhibit cell proliferation and tumor growth of rat and mouse mammarycarcinomas. Int. J. Cancer,66: 817–820, 1996.

22. Soloff, M., S., and Wieder, M. H. Oxytocin receptors in rat involuting mammarygland. Can. J. Biochem. Cell Biol.,61: 631–635, 1983.

23. Mohr, E., Meyerhof, W., and Richter, D. Vasopressin and oxytocin: molecularbiology and evolution of the peptide hormones and their receptors. Vitamins Hor-mones,51: 235–266, 1995.

24. Cassoni, P., Fulcheri, E., Carcangiu, M. L., Deaglio, Stella A., Deaglio, S., andBussolati, G. Oxytocin receptors in human adenocarcinomas of the endometrium:presence and biological significance. J. Pathol., in press, 2000.

25. Cassoni, P., Sapino, A., Stella, A., Fortunati, N., and Bussolati, G. Presence andsignificance of oxytocin receptors in human neuroblastomas and glial tumors. Int. J.Cancer,77: 695–700, 1998.

26. Taylor, A. H., Ang, V. T. Y., Jenkins, J., Silverlight, J., Coombes, R. C., andLuqmani, Y. A. Interaction of vasopressin and oxytocin with human breast carcinomacells. Cancer Res.,50: 7882–7886, 1990.

27. Grazzini, E., Guillon, G., Mouillac, B., and Zingg, H. H. Inhibition of oxytocinreceptor function by direct binding of progesterone. Nature, Lond.392: 509–512,1988.

28. Szwajkajzer, D., and Carey, J. Molecular and biological constraints on ligand-bindingaffinity and specificity. Biopolymers,44: 181–198, 1997.

29. Kojro, E., Hackenberg, M., Zsigo, J., and Fahrenholz, F. Identification and enzymaticdeglycosylation of the myometrial oxytocin receptor using a radioiodinated photore-active antagonist. J. Biol. Chem.,266: 21416–21421, 1991.

30. Thibonnier, M., Conarty, D., M., Preston, J. A., Wilkins, P. L., Berti-Mattera, L. N.,and Mattera, R. Molecular pharmacology of human vasopressin receptors.In: H. H.Zingg, C. W., Bourque, and D. G. Bichet (eds.), Vasopressin and Oxytocin, pp.251–276. New York: Plenum Press, 1998.

31. Ritter, N. M., Farach-Carson, M. C., and Butler, W. T. Evidence for the formation ofa complex between osteopontin and osteocalcin. J. Bone Mineral Res.,7: 877–884,1992.

32. Prigodich, R. V., and Vesely, M. R. Characterization of a complex between bovineosteocalcin and type I collagen. Arch. Biochem. Biophys.,345: 339–341, 1997.

33. Fujisawa, R., and Kuboki, Y. Preferential adsorption of dentin and bone acidicproteins on the (100) face of hydroxyapatite crystals. Biochim. Biophys. Acta,1075:56–60, 1991.

3475

INTERACTION BETWEEN OC AND OT IN OS

Research. on January 1, 2016. © 2000 American Association for Cancercancerres.aacrjournals.org Downloaded from

34. Kaartinen, M. T., Pirhonen, A., Linnala-Kankkunen, A., and Maenpaa, P. H. Trans-glutaminase-catalyzed crosslinking of osteopontin is inhibited by osteocalcin. J. Biol.Chem.,272: 22736–22741, 1997.

35. Chenu, C., Colucci, S., Grano, M., Zigrino, P., Zambonin, G., Balsini, N., Vergnaud,P., Delmas, P. D., and Zallone, A. Z. Osteocalcin induces chemotaxis, secretion ofmatrix proteins, and calcium-mediated intracellular signaling in human osteoclast-likecells. J. Cell Biol.,127: 1149–1158, 1994.

36. Chen, Z-L., and Strickland, S. Neuronal death in the hippocampus is promoted byplasmin-catalyzed degradation of laminin. Cell,91: 917–925, 1997.

37. Ducy, P., Desbois, C., Boyce, B., Pinero, G., Syory, B., Dunstan, C., Smith, E.,Bonadio, J., Goldstein, S., Gundberg, C., Bradley, A., and Karsenty, G. Increasedbone formation in osteocalcin-deficient mice. Nature (Lond.)382: 448–452, 1996.

38. Tulinsky, A. The structures of domains of blood proteins. Thrombosis Haemostasis,66: 16–31, 1991.

39. Chen, J-T., Hosoda, K., Hasumi, K., Ogata, E., and Shiraki, M. Serum N-terminalosteocalcin is a good indicator for estimating responders to hormone replacementtherapy in postmenopausal women. J. Bone Miner. Res.,11: 1784–1792, 1996.

40. Mendre, C., Dufour, M. N., Le Roux, S., Seyer, R., Guillou, L., Calas, B., andGuillon, G. Synthetic rat V1a vasopressin receptor fragments interfere with vasopres-sin binding via specific interaction with the receptor. J. Biol. Chem.,272: 21027–21036, 1997.

41. Johnsen, M., Lund, L. R., Romer, J., Almholt, K., and Dano, K. Cancer invasion andtissue remodeling: common themes in proteolytic matrix degradation. Curr. Opin.Cell Biol., 10: 667–671, 1998.

42. Solomayer, E-F., Diel, I. J., Wallwiener, D., Bode, S., Meyberg, G., Sillem, M.,Gollan, C., Kramer, M. D., Krainick, U., and Bastert, G. Prognostic relevance ofurokinase plasminogen activator detection in micrometastatic cells in the bone mar-row of patients with primary breast cancer. Br. J. Cancer,76: 812–818, 1997.

43. Campbell, P. G., Novak, J. F., Wines, K., and Walton, P. E. Localization of plasminactivity on osteosarcoma cells: cell surface proteolysis of insulin-like growth factorbinding proteins. Growth Regulation,3: 95–98, 1993.

44. Catherwood, B. D., Titus, L., Evans, C-O., Rubin, J., Boden, S. D., and Nanes, M. S.Increased expression of tissue plasminogen activator messenger ribonucleic acid is animmediate response to parathyroid hormone ion neonatal rat osteoblasts. Endocrinol-ogy, 134: 1429–1436, 1994.

45. Copland, J. A., Ives, K. L., Simmons, D. J., and Soloff, M. S. Functional oxytocinreceptors discovered in human osteoblasts. Endocrinology,140: 4371–4374, 1999.

46. Gutkowska, J., Jankowski, M., Lambert, C., Mukaddam-Daher, S., Zingg, H. H., andMcCann, S. M. Oxytocin releases atrial natriuretic peptide by combining withoxytocin receptors in the heart. Proc. Natl. Acad. Sci. USA,94: 11704–11709, 1997.

47. Pedersen, C. A., Caldwell, J. D., Jirikowski, G. F., and Insel, T. R. (eds.). Oxytocinin maternal, sexual, and social behaviors. Ann. N. Y. Acad. Sci.,652:253–270, 1992.

48. Thibonnier, M., Conarty, D. M., Preston, J. A., Plesnicher, C. L., Dweik, R. A., andErzurum, S. C. Human vascular endothelial cells express oxytocin receptors. Endo-crinology,140: 1301–1309, 1999.

49. Lian, J. B., Stein, G. S., Stein, J. L., Bortell, R., Owen, T. A., Shalhoub, V., Heinrichs,A., Van Oordt, C. W., and Van Wijnen, A. J. Implications of abrogated growthcontrol on expression of a cell cycle-regulated histone and the osteoblast-specificosteocalcin genes in osteosarcoma cells.In: J. F. Novak and J. H. McMaster (eds.),Frontiers of Osteosarcoma Research, pp. 227–287. Seattle, WA: Hogrefe and Huber,1993.

50. Owen, T. A., Aronow, M., Shalhoub, V., Barone, L. M., Wilming, L., Tassinari,M. S., Kennedy, M. B., Pockwinse, S., Lina, J. B., and Stein, G. S. Progressivedevelopment of the rat osteoblast phenotypein vitro: reciprocal relationships inexpression of genes associated with osteoblast proliferation and differentiation duringformation of the bone extracellular matrix. J. Cell. Physiol.,143: 420–430, 1990.

3476

INTERACTION BETWEEN OC AND OT IN OS

Research. on January 1, 2016. © 2000 American Association for Cancercancerres.aacrjournals.org Downloaded from

2000;60:3470-3476. Cancer Res   J. F. Novak, M. B. Judkins, M. I. Chernin, et al.   Inhibition of Osteosarcoma CellsOsteocalcin Counteracts Oxytocin-mediated Growth of A Plasmin-derived Hexapeptide from the Carboxyl End of

  Updated version

  http://cancerres.aacrjournals.org/content/60/13/3470

Access the most recent version of this article at:

   

   

  Cited articles

  http://cancerres.aacrjournals.org/content/60/13/3470.full.html#ref-list-1

This article cites 41 articles, 8 of which you can access for free at:

  Citing articles

  http://cancerres.aacrjournals.org/content/60/13/3470.full.html#related-urls

This article has been cited by 2 HighWire-hosted articles. Access the articles at:

   

  E-mail alerts related to this article or journal.Sign up to receive free email-alerts

  Subscriptions

Reprints and

  .pubs@aacr.orgDepartment at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

  Permissions

  .permissions@aacr.orgDepartment at

To request permission to re-use all or part of this article, contact the AACR Publications

Research. on January 1, 2016. © 2000 American Association for Cancercancerres.aacrjournals.org Downloaded from