phenotypic comparison of periodontal ligament cells in vivo and in vitro

9
The periodontium consists of soft and mineralized connective tissues arranged in cellular domains of remarkable coherence and spatially preserved dimensions (1, 2). As the periodontium contains cell populations that rapidly synthesize and remodel soft and mineralizing connective tissues, it is a useful model for studies of connective tissue homeostasis (3, 4). In spite of the application of Phenotypic comparison of periodontal ligament cells in vivo and in vitro Lekic P, Rojas J, Birek C, Tenenbaum H, McCulloch CAG: Phenotypic comparison of periodontal ligament cells in vivo and in vitro. J Periodont Res 2001; 36: 71– 79. # Munksgaard, 2001. The mammalian periodontal ligament contains heterogeneous populations of connective tissue cells, the precise function of which is poorly understood. Despite close proximity to bone and the application of high amplitude physical forces, cells in the periodontal ligament (PL) are capable of expressing regulatory factors that maintain PL width during adult life. The study of PL homeostasis and PL cell dierentiation requires culture and phenotypic methods for precise characterization of PL cell populations, in particular those cells with an inherently osteogenic program. Currently it is unknown if cells cultured from the PL are phenotypically similar to the parental cells that are present in the tissues. We have compared the phenotype of cells in vivo with cells derived from the PL and expanded in vitro to assess the general validity of in vitro models for the study of phenotypic regulation in vivo. Rat PL cells were isolated by either scraping the root of the extracted first mandibular molars (Group A), or by scraping the alveolar socket following extraction of first mandibular molars (Group B), or by obtaining a mixture of cells after disaggregating a block of tissue consisting of first mandibular molar, PL and the surrounding alveolar bone (Group C). Cultured cells at confluence were fixed and immunostained for a-smooth muscle actin (a-SMA), osteopontin (OPN), alkaline phosphatase (AP), or bone sialoprotein (BSP). For in vivo assessments, frontal sections of rat first mandibular molar were immunostained for a-SMA, OPN, AP and BSP. We examined osteogenic dierentiation of cultured PL cell cultures by bone nodule-forming assays. In vivo and at all examined sites, 468% of PL cells were immunostained for AP; *50% and *51% for OPN and a-SMA ( p=0.3), respectively, while only *8% were positively stained for BSP ( p50.01). Analysis of cultured PL cells in Groups A, B and C showed 54%, 53% and 56% positive staining for a-SMA respectively; 51%, 56%, 54% for OPN; 66%, 70%, 69% for AP and 2.2%, 1.4% and 2.8% for BSP. The mean percentage of PL cells in situ stained for the dierent markers was similar to that of cultured PL cells (Group A*Group B*Group C in situ for p40.2) except for BSP which was 3 to 4 fold higher in vivo ( p50.01). PL cell cultures treated with dexamethasone showed mineralized tissue formation for all groups (A, B, C), but no mineralized tissue formation was detected in the absence of dexamethasone. As PL cells express quantitatively similar phenotypes in vitro and in vivo, we conclude that the in vitro models used here for assessment of PL cell dierentiation appear to be appropriate and are independent of the cell sampling method. Further, dexamethasone-dependent progenitors are present both on the root and bone-related sides of the PL. P. Lekic 1 , J. Rojas 2 , C. Birek 1 , H. Tenenbaum 3 , C. A. G. McCulloch 3 1 Department of Dental Diagnostics and Surgical Sciences, Faculty of Dentistry, University of Manitoba, 2 Division of Undergraduate Orthodontics and Pediatric Dentistry, School of Dentistry, Faculty of Medicine and Dentistry, University of Western Ontario, 3 MRC Group in Periodontal Physiology, Faculty of Dentistry, University of Toronto, Canada Dr P. C. Lekic, Faculty of Dentistry, University of Manitoba, 780 Bannatyne Av., R3E 0W2, Winnipeg, Manitoba, Canada Tel: z1204 789 3507 Fax: z1204 789 3913 Key words: phenotypic stability; periodontal ligament cells; in vivo; in vitro Accepted for publication June 12, 2000 J Periodont Res 2001; 36: 71–79 Printed in UK. All rights reserved

Upload: p-lekic

Post on 06-Jul-2016

215 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: Phenotypic comparison of periodontal ligament cells in vivo and in vitro

The periodontium consists of soft and mineralizedconnective tissues arranged in cellular domains ofremarkable coherence and spatially preserveddimensions (1, 2). As the periodontium contains

cell populations that rapidly synthesize andremodel soft and mineralizing connective tissues,it is a useful model for studies of connective tissuehomeostasis (3, 4). In spite of the application of

Phenotypic comparisonof periodontal ligamentcells in vivo and in vitroLekic P, Rojas J, Birek C, Tenenbaum H, McCulloch CAG:Phenotypic comparison of periodontal ligament cells in vivo and in vitro.J Periodont Res 2001; 36: 71±79. # Munksgaard, 2001.

The mammalian periodontal ligament contains heterogeneous populationsof connective tissue cells, the precise function of which is poorly understood.Despite close proximity to bone and the application of high amplitude physicalforces, cells in the periodontal ligament (PL) are capable of expressing regulatoryfactors that maintain PL width during adult life. The study of PL homeostasisand PL cell di�erentiation requires culture and phenotypic methods for precisecharacterization of PL cell populations, in particular those cells with an inherentlyosteogenic program. Currently it is unknown if cells cultured from the PL arephenotypically similar to the parental cells that are present in the tissues. We havecompared the phenotype of cells in vivo with cells derived from the PL andexpanded in vitro to assess the general validity of in vitro models for the study ofphenotypic regulation in vivo. Rat PL cells were isolated by either scraping the rootof the extracted ®rst mandibular molars (Group A), or by scraping the alveolarsocket following extraction of ®rst mandibular molars (Group B), or by obtainingamixture of cells after disaggregating a block of tissue consisting of ®rstmandibularmolar, PL and the surrounding alveolar bone (Group C). Cultured cells atcon¯uence were ®xed and immunostained for a-smooth muscle actin (a-SMA),osteopontin (OPN), alkaline phosphatase (AP), or bone sialoprotein (BSP).For in vivo assessments, frontal sections of rat ®rst mandibular molar wereimmunostained for a-SMA, OPN, AP and BSP. We examined osteogenicdi�erentiation of cultured PL cell cultures by bone nodule-forming assays. In vivoand at all examined sites, 468% of PL cells were immunostained for AP;*50%and*51% for OPN and a-SMA (p=0.3), respectively, while only *8% werepositively stained for BSP (p50.01). Analysis of cultured PL cells in Groups A, Band C showed 54%, 53% and 56% positive staining for a-SMA respectively;51%, 56%, 54% for OPN; 66%, 70%, 69% for AP and 2.2%, 1.4% and 2.8%for BSP. The mean percentage of PL cells in situ stained for the di�erent markerswas similar to that of cultured PL cells (Group A*Group B*Group C in situ forp40.2) except forBSPwhichwas 3 to 4 fold higher in vivo (p50.01). PL cell culturestreated with dexamethasone showed mineralized tissue formation for all groups(A, B, C), but no mineralized tissue formation was detected in the absence ofdexamethasone. As PL cells express quantitatively similar phenotypes in vitro andin vivo, we conclude that the in vitro models used here for assessment of PL celldi�erentiation appear to be appropriate and are independent of the cell samplingmethod. Further, dexamethasone-dependent progenitors are present both on theroot and bone-related sides of the PL.

P. Lekic1, J. Rojas2, C. Birek1,H. Tenenbaum3, C. A. G. McCulloch3

1Department of Dental Diagnostics andSurgical Sciences, Faculty of Dentistry,University of Manitoba, 2Division ofUndergraduate Orthodonticsand Pediatric Dentistry, School of Dentistry,Faculty of Medicine and Dentistry, Universityof Western Ontario, 3MRC Group in PeriodontalPhysiology, Faculty of Dentistry, University ofToronto, Canada

Dr P. C. Lekic, Faculty of Dentistry, Universityof Manitoba, 780 Bannatyne Av., R3E 0W2,Winnipeg, Manitoba, CanadaTel: z1204 789 3507Fax: z1204 789 3913

Key words: phenotypic stability; periodontalligament cells; in vivo; in vitro

Accepted for publication June 12, 2000

J Periodont Res 2001; 36: 71±79Printed in UK. All rights reserved

Page 2: Phenotypic comparison of periodontal ligament cells in vivo and in vitro

high amplitude physical forces followed by con-stant formation and remodeling of alveolar boneand cementum, cells in the periodontal ligament(PL) are capable of expressing regulatory factorsthat maintain PL width.PL cell populations comprise a renewal cell

system in steady state (5, 6). The renewing cellpopulations undergo extensive turnover and areprincipally located around the blood vessels in thePL (7) and in the endosteal spaces of the alveolarbone (8). These cells proliferate and migrate toproduce more di�erentiated cells that can synthes-ize bone, cementum and the extracellular matrix ofthe PL (9). As these processes are the bases forremodeling and healing in periodontal tissues, it islikely that some forms of signaling systems haveevolved to orchestrate, in a temporally and spatiallyappropriate manner, repopulation and di�erenti-ation responses (10). However, a detailed under-standing of remodeling processes in periodontiumis complicated by the diversity of the cells and tissuetypes. Methods that can identify di�erent PL cellpopulations and recognize discrete stages of celldi�erentiation are of central importance for ourunderstanding of remodeling and healing processesin periodontal tissues. However, as there is node®ned marker for periodontal ligament cell popu-lations, some secreted matrix proteins and cyto-skeletal proteins have been used to characterizethese cells (11±13). For example, some PL®broblastsexpress a-smooth muscle actin (a-SMA, Ref. 11) orosteopontin (OPN, Ref. 12). Further, other cellscan express proteins that are characteristic ofparticular lineages, notably, alkaline phosphatase(AP, Ref. 13) which is expressed predominantlyby mineralized tissue-forming PL cells. How-ever, another marker, bone sialoprotein (BSP) isused as a nucleator of mineralization to charac-terize mineralized tissue forming cells (14). As onlya small portion of PL cells is associated withmineralized tissue formation (osteoblasts, cemento-blasts or odontoblasts), consequently only few PLcells will be expressing this marker (15±17).Notably, a-SMA, OPN, AP and BSP could beuseful markers to identify ®broblastic and osteo-genic cell lineages in the PL with a restricted di�er-entiation repertoire. Currently, however, there is nodetailed knowledge of the relative importance ofthese proteins in regenerating PL cell populationsor the relationship of these markers in vitro to thecells that express these proteins in vivo.The use of appropriate model systems is useful

for obtaining a detailed understanding of the originof cells, their regulation and di�erentiation. Valida-tion of in vitro models requires phenotypic char-acterization of PL cell populations in vitro andin vivo, particularly for their application to actual

physiological conditions. We have compared thephenotype of cells in vivo with cells derived fromthe PL and expanded in vitro. We have tested thehypothesis that cultured PL cells preserve thephenotypic characteristics of the original parentalcell populations.

Material and methods

Phenotypic markers of PL cells in vitro

PL cells from Sprague Dawley male rats weighing110±130 g were isolated by either scraping theroot of the extracted left ®rst mandibular molar(Group A), or by scraping the alveolar socketfollowing extraction of the left ®rst mandibularmolar (Group B), or by obtaining a mixture of cellsafter excision of a block of tissue consisting of left®rst mandibular molar, PL and the surroundingalveolar bone (Group C). The rationale for thesemethods was that cells obtained in Groups A or Bwould be expected to include precursor cellsrestricted to the root-related or bone-related com-partments of the PL, whereas cells from the blockof tissue (alveolar bone, PL, cementum of the root;Group C) would be expected to include also theprecursor cells located in the endosteal spaces ofthe alveolar bone. After isolation from the explantcultures, cells were grown in T-25 ¯asks containingDulbecco's modi®ed Eagle's medium (DMEM)with 10% fetal bovine serum at 37³C in ahumidi®ed atmosphere containing 5% CO2. Uponreaching con¯uences, cells were trypsinized, trans-ferred into 8-well chamber slides overnight, ®xedand prepared for immunostaining as described indetail earlier (18). We used immunohistochemistrymethods as these were established previously (4, 12,19) and could facilitate assessment of the functionalactivity of cell proteins (20). To prevent non-speci®c binding, cells were incubated with 1% caseinin phosphate bu�ered blocking solution (12, 19).Cells were then incubated with primary antibodyfor 3 hours in a moist chamber. Immunolocaliza-tion for a-smooth muscle actin was performed witha mouse monoclonal anti-a-SMA antibody (Clone1A4, MPIII10, Sigma Chemical Co., St Louis,MO, USA; 1:50 dilution) and for osteopontin witha mouse monoclonal anti-OPN antibody (Hybri-doma Bank, University of Baltimore, MD, USA;1:600 dilution; Ref. 12). Detection of alkaline pho-sphatase was done with a rabbit polyclonal anti-APantibody (Provided by Dr Lary Fisher, Bethesda,Maryland, USA; 1:200 dilution), and for bonesialoprotein with a rabbit polyclonal anti-BSPantibody (provided by Dr J. Sodek, MRC Groupin Periodontal Physiology, University of Toronto,ON, Canada; 1:200 dilution). Previous studies haveshown that these antibodies (a-SMA ± Ref. 11;

72 Lekic et al.

Page 3: Phenotypic comparison of periodontal ligament cells in vivo and in vitro

OPN ± Ref. 12; AP ± Ref. 13; BSP ± Ref. 19) maybe used as phenotypic markers for the PL cellpopulation. However, some other soft and miner-alized connective tissue markers (osteocalcin, osteo-nectin) were not well identi®ed in the PL (21) andfor this reason failed to prove their usefulness asPL cell phenotypic markers. Negative control forpolyclonal antibodies has been described previously(22), whereas for monoclonal antibodies an irrelev-antmonoclonal antibody (anti-humanCD4 lympho-cyte antigen; Coulter Electronics, Burlington, ON,Canada) was used. Following this rationale andafter incubating for 3 hours with OPN and a-SMA,primary cells were washed in phosphate bu�eredsaline (PBS) (265 min) and left with the secondaryantibody (biotinylated goat anti-mouse; VectorLaboratories, Burlingame, CA, USA; 1:200 dilu-tion). For BSP and AP, cells were stained withthe appropriate primary antibody, washed withPBS and incubated with secondary antibody(biotinylated goat anti-rabbit, Vector Laboratories;1:200 dilution). Cells were then washed in PBS(265 min), incubated with streptavidin ABC com-plex/HRP (PK-6100, Vectastain, Vector Laborat-ories) for 30 min, washed again in PBS (265 min),and incubated with diaminobenzidine (SK-4100,Vector Laboratories) for 15 min. Cells were thenwashed, counterstained with hematoxylin and eosin,dehydrated and cover-slipped.

Tissue preparation

The right mandibles of the same animals used inthe in vitro studies were ®xed in periodate-lysine-paraformaldehyde (23) at pH 7.4 for 24 h at 4³C,demineralized for 24 h in 0.2 N HCl, and washed inPBS for 20 h. Specimens were dehydrated in gradedethanol, cleared in toluene, and embedded in par-a�n. Serial frontal sections (5 mm) through theperiodontium surrounding the ®rst mandibularmolar were attached to glass slides for immuno-histochemical analyses. Sections were immuno-stained using the same reagents and the sameconcentrations of primary and secondary antibo-dies as described above.

Immunostaining assessment

Control slides were for a-SMA and OPN treatedwith an irrelevant antibody (anti-human CD4lymphocyte antigen; Coulter Electronics), and forAP and BSP as described (22). The localization ofproteins for the in vivo specimens was studied onfrontal sections of the ®rst mandibular molar usingpreviously described quantitative methods (4). Thepercentages of cells with intracellular a-SMA,OPN, AP, and BSP staining were computed in the

central compartment of the PL (at the cellular/acellular cementum junction, zone 2, Fig. 1), thecervical part of the PL (400 mm coronally fromthe central compartment, zone 1, Fig. 1) and in theapical part of the PL (400 mm apically from thecentral compartment, zone 3, Fig 1). These zoneswere chosen to study the phenotypic characteristicsof periodontal ligament cells within the functionalzones (cervical ± tension; central ± tension andpressure; apical ± pressure). Intracellular stainingwas computed on 300±400 cells for each compart-ment of the PL. Staining data were obtained fromassessments of at least 9 sections from 3 di�erentanimals under a light microscope (Laborlux K,Leica, Wetzlar, Germany) with an intraoculargrid (2506250 mm) containing 100 squares of625 mm each. The estimate of sample size for therequired number of animals was based on pre-vious studies and the relative variance terms anddi�erences between mean values of immunostainingassessment (19).

Fig. 1. A drawing of a frontal section through a rat molar rootto illustrate the locations for immunostaining assessments.Mandible (M), alveolar bone (AB), acellular cementum (AC),cellular cementum (CC), dentin (D), pulp (P). ``Z1'' depictscervical compartment of the PL (zone 1), ``Z2'' depicts centralpart of the PL (zone 2) and ``Z3'' depicts apical compartment ofthe PL (zone 3).

Phenotypic comparison of PL cells in vivo and in vitro 73

Page 4: Phenotypic comparison of periodontal ligament cells in vivo and in vitro

The percentage of cells with positive stainingfor a-SMA, OPN, AP, and BSP was obtained inthe same manner as from PL cell cultures. As thecell count was lower in vitro (80±100) than in vivo(300±400), each culture dish was assessed at 4 dif-ferent sites within the culture to provide a balancednumber of cells for comparisons.

Osteogenic differentiation of PL cells in vitro

We assessed the potential for osteogenic di�erentia-tion in PL cell cultures by treating each of thegroups (A±C) with DMEM supplemented with10% fetal serum, 50 mg/ml ascorbic acid, 10 mM

sodium b-glycerophosphate, 1078M dexametha-

sone (Sigma Chemical Co.) and 0.3 mg/ml ampho-tericin B (fungizoneTM, GibcoBRL, Grand Island,NY, USA). Control cultures were treated with thesame mixture but without dexamethasone. Themediumwas changed every 2 days, and after 21 dayscells were ®xed for 15 min in 1% methanol andstained by the Von Kossa method as described (24).

Statistical methods

Raw data from each animal or culture were keptseparate and the mean for protein staining in eachzone was calculated. The mean result from eachzone per animal was considered as an independentsample. Data were assessed by analysis of variance(ANOVA) to evaluate the di�erences betweenthe model systems with respect to immunostainingassessments at the di�erent sites. Di�erencesbetween zones were examined by ANOVA andwere considered signi®cant at p50.05. Data wereexpressed as mean¡SEM (standard error of themean). Post hoc comparisons were made usingDuncan's multiple range test.

Results

Expression of differentiation markers in PL cellpopulation in vivo and in vitro

We used a-SMA, OPN, AP and BSP as di�erentia-tion markers of connective tissue cell populations inthe PL (11±13). Expression of these markers wasobserved throughout the periodontal ligament andin PL cell cultures (Figs 2 to 6). Notably, in vivo,PL cells showed high percentages (50±65%) of cellspositively stained for a-SMA at all examined sites(zone 1, 2, 3; Fig. 2A). PL cells in the apical area(zone 3) exhibited slightly higher percentages ofcells stained for a-SMA compared to the PL cells inzones 1 and 2 (Figs 2A, 4C), but this di�erence wasnot statistically signi®cant (p40.2). a-SMA stain-ing in the PL cell population in vitro was also high(50±55%) and was similar for all experimentalgroups (Group A*Group B*Group C; p40.2;Fig. 2B).

PL cells in vivo exhibited, in the central and inthe cervical portions of the ligament, the same per-centages of cells stained for OPN (zone 1 andzone 2*50%; Fig. 3A), whereas staining for OPNin the apical part of the PL (zone 3) was slightlyhigher (p40.2; Figs 3A, 4A).Notably,OPNstainingin PL cell cultures was similar to that of PL cells invivo (*50%) and there were no di�erences betweenthe experimental groups (Group A*Group B*Group C*50%; p40.2; Figs 3B, 4B).

For all the markers examined here, PL cellsexhibited the highest percentages of AP positivelystained cells (460%; Figs 5A, 5B; signi®cantlyhigher than OPN, p50.05). Notably, PL cells in vivoshowed similar percentages of cells stained for APin the central and cervical portions of the ligament(zone 1 and zone 2*68%; Fig. 5A), whereas therewas a slightly higher percentage of cells stained

Fig. 2. Histograms showing percentages of a-smooth muscle actin (a-SMA) positive stained cells in vivo (A) and in vitro (B). Data aremean¡SEM % positive cells.

74 Lekic et al.

Page 5: Phenotypic comparison of periodontal ligament cells in vivo and in vitro

for AP in the apical part of the PL (zone 3; 480%;Fig. 5A). AP staining of PL cells in vitro wasvery similar among the examined groups (GroupA*Group B*Group C*65±70%; p40.2;Figs 4D, 5B).PL cells in vivo were rarely stained for BSP

(Figs 4E, 6A) compared to the staining for the othermarkers (a-SMA, OPN, and AP; p50.001; Figs 2,3, 5). Similar to OPN and AP (Figs 3A, 5A), PLcells in the apical portion of the ligament exhibitedslightly higher percentages of cells stained forBSP (zone 3 : zone 1, p50.05; Fig. 6A). Notably,BSP staining of PL cells in vivo was much largerthan the percentages of BSP stained cells in vitro(2±3 fold higher; p50.01; Figs 4F, 6A, 6B).

Dexamethasone treatment of PL cells in vitro

We treated cultured PL cells (Groups A, B, C)for 21 days with dexamethasone-supplementedmedium to assess osteogenic di�erentiation of cellpopulations within di�erent PL compartments.PL cell cultures treated with the supplementedmedium exhibited nodules after staining with VonKossa, indicating mineralized tissue formation forall three groups (Fig. 4G). Interestingly, miner-alized nodules were evenly distributed throughoutthe dexamethasone-treated PL cultures. ControlPL cells cultured for the same period of time(21 days) but without dexamethasone did notform mineralized tissue (Fig. 4H).

Discussion

An interesting feature of the PL is the remarkablemaintenance of PL width over time, an importantmeasure of PL homeostasis (25). This preservationof PL width provides insight into a biologicalsystem in which soft (PL) and surrounding miner-alizing tissue domains (cementum and alveolarbone) are maintained in a dynamic equilibrium.

In spite of the constant physiological tooth driftand the presence of cells bearing osteoblasticmarkers in the PL (26, 27), resident PL cells con-stitutively express regulatory factors that maintainPL width by inhibiting osteogenesis (28±30). Tofacilitate studies on the regulatory mechanisms ofPL cell di�erentiation, we determined the immuno-staining of cells cultured from di�erent PL sitesand compared these data to the staining character-istics of PL cells in vivo. Notably, the central ®ndingof this study was that PL cells cultured from di�er-ent sites within the periodontium showed similarpercentages of positively stained cells as PL cellsin vivo. Notably, after treatment of PL cells withdexamethasone, osteoblastic di�erentiation of cellswas promoted, indicating the potential for at leastsome PL cells associated with the root or thealveolar bone to enter the osteogenic pathway (30).

Phenotypic stability of PL

Renewing PL cells undergo extensive turnover tomaintain the steady state (5) and, in spite of con-stant bone remodeling and cellular cementumformation, the cellular domains of the PL arepreserved. Paracrine inhibition of osteogenesis byPL ®broblasts has been suggested as a possibleregulatory mechanism that preserves PL homeo-stasis (28, 29). On the basis of the data shownhere, we suggest that the phenotypic uniformity ofpopulations across the PL may also contribute tothe preservation of periodontal tissue domains.Indeed, the results from this study indicate thatstaining of cells for most of the phenotypic markers(a-SMA, OPN, AP) is relatively constant in theroot and bone-associated compartments of the PL.Further support for the phenotypic stability of PLcell populations was the similar staining for thesemarkers reported in previous in vitro and in vivostudies (11, 31, 32). Not surprisingly and indepen-dent of the isolation technique (Groups A±C),cultured precursor cells di�erentiated to form cell

Fig. 3. Histograms showing osteopontin (OPN) staining of PL cell population in vivo (A) and in vitro (B). Data are mean¡SEM %positive cells.

Phenotypic comparison of PL cells in vivo and in vitro 75

Page 6: Phenotypic comparison of periodontal ligament cells in vivo and in vitro

76 Lekic et al.

Page 7: Phenotypic comparison of periodontal ligament cells in vivo and in vitro

populations that expressed the same percentage ofcells stained for a-SMA, OPN and AP as PL cellsin vivo. This staining similarity indicates that cellscultured from di�erent PL sites (root-associatedPL ± Group A; alveolar bone associated PL ±Group B; or PLzendosteal spaces ± Group C)can maintain similar phenotypic characteristics

under widely di�erent growth conditions (i.e.in vivo compared to in vivo). Further, this equival-ence of phenotypic characteristics of cultured cellswith the in vivo situation indicates that theregulatory mechanisms that prevent continuousosteoblastic di�erentiation of PL cells may be aninherent characteristic of PL cells and under

Fig. 4. A. Photomicrograph of the apical compartment of the PL (Zone 3, Z3) showing positive staining for osteopontin (arrow heads).Intense staining for osteopontin was noted in the PL as well as in the alveolar bone (AB) and the cementum of the root (C). Bar 20 mm.B. Immunostaining for osteopontin (arrow heads) of cultured PL cells explanted from the block of tissue consisting of left ®rstmandibular molar, PL and the surrounding alveolar bone. Note the presence of staining for osteopontin in cultured PL cells. Bar10 mm. C. Photomicrograph of the cervical compartment of the PL (Zone 1, Z1) immunostained for a-smooth muscle actin. Note thepresence of positively stained PL cells for a-smooth muscle actin (arrow heads) located throughout the ligament. Cementum of the root(C) and the alveolar bone (AB) are not stained for a-smooth muscle actin. Bar 20 mm. D. Immunostaining for alkaline phosphatase(arrow heads) of cultured PL cells explanted from the root of the ®rst mandibular molar. Bar 5 mm. E. Photomicrograph of the ®rstmandibular molar immunostained for bone sialoprotein. Note the presence of few cells in the central compartment of the PL (Zone 2,Z2) positively stained for bone sialoprotein (arrow heads). PL cells positively stained for bone sialoprotein are located in the proximityof the alveolar bone (AB) and the cementum of the root (C). Bar 20 mm. F. Immunostaining for bone sialoprotein of cultured PL cellsexplanted from the block of tissue consisting of left ®rst mandibular molar, PL and the surrounding alveolar bone. Note the presence offew cultured PL cells (arrow heads) with positive staining for bone sialoprotein. Bar 10 mm. G. Von Kossa staining of cultured PL cellsexplanted from an extracted root of a ®rst mandibular molar and treated with dexamethasone-supplemented medium. Note bonenodule formation (arrow heads) throughout the culture. Bar 50 mm. H. Von Kossa staining of cultured PL cells explanted from theblock of tissue consisting of left ®rst mandibular molar, PL and the surrounding alveolar bone, and cultured with regular DMEMmedium as in Fig. 4G but without dexamethasone. Note the absence of bone nodule formation throughout the culture. Bar 50 mm.

Fig. 5. Histograms showing alkaline phosphatase (AP) staining for in vivo (A) and in vitro (B) PL cell populations. Data aremean¡SEM % positive cells.

Fig. 6. Histograms showing bone sialoprotein (BSP) staining for in vivo (A) and in vitro (B) PL cell populations. Data are mean¡SEM% positive cells.

Phenotypic comparison of PL cells in vivo and in vitro 77

Page 8: Phenotypic comparison of periodontal ligament cells in vivo and in vitro

physiological conditions is independent of the e�ectof systemic factors.BSP was the only phenotypic marker to exhibit

greater percentages of positively stained cells in vivocompared to in vitro PL cell cultures. PL cells in vivoalso exhibited higher percentages of cells stainedfor BSP in the apical portion than in the centralor cervical portions of the PL. However, as BSPstained cells were only found in the proximity of thealveolar bone and cementum of the root and as thismarker is detected in mineralized tissue-formingcells (15, 16), we suggest that some of the PLcells have di�erentiated into osteoblasts or cemento-blasts (12). Indeed, increased BSP staining of PLcells in vivo correlates to the zone of continuousbone remodeling and cellular cementum formation.Notably, application of high physical forces duringmastication stimulates remodeling of mineralizedperiodontal tissues, thus increasing the expressionof mineralized tissue markers in the residing PLcells (i.e. AP; Ref. 31). As apical and centralportions of the PL showed increased staining ofcells for BSP, it seems that masticatory forces mayhave promoted osteoblastic and cementoblasticdi�erentiation of PL cells in these compartments.Indeed, assessment of stress distribution withinthe PL, following the application of physical forces,has shown increased compressive stresses in thecervical and apical portion of the ligament (33).These compressive stresses may lead to relativeavascularity of the PL and induce osteoclast activitywhich may be followed by enhanced osteoblasticand cementoblastic di�erentiation of PL cells.

Osteogenic recruitment of PL cells

Cultured PL cells, as well as cells in vivo, exhibitedhigh percentages of cells stained for AP and OPN.Previous studies have also shown the expression ofthese proteins by PL cells in vivo (4, 13) and sinceAP and OPN are predominantly expressed by osteo-genic cells (33, 34), these studies support the notionthat most of the resident PL cells have osteogenicpotential. In contrast, cell cultures and cells in vivoexhibited low percentages of BSP cells. As theexpression of BSP is closely associated with boneformation and with nucleation of mineralization(14), the lack of BSP staining by PL cells indicatesthat these cells do not normally express the fullrepertoire of di�erentiation-associated proteins inosteogenesis. This low percentage of BSP stainingin PL cells may explain the limited expression ofthe osteogenic potential of these cells, as shownby the absence of mineralization in the body ofthe PL during normal function. However, treat-ment of PL cell cultures with dexamethasonepromoted osteoblastic di�erentiation of cells and

led to the production of mineralized tissue. Inter-estingly, in our study mineralized tissue was formedin all PL cell cultures (Groups A, B, C) that weretreated with dexamethasone and this suggeststhat throughout the PL there are populations ofdexamethasone-dependent osteogenic populations.The diversion of the normal di�erentiation reper-toire of PL cells to an osteogenic phenotype hasbeen reported in other in vitro studies (29, 35), aswell as in vivo following systemic treatment withbisphosphonates (36, 19). Treatment with bisphos-phonates leads to a signi®cant increase in BSPexpression of PL cells, presumably by modulatingthe di�erentiation of these cells to a bone-formingphenotype (36). This recruitment of PL cells toa bone-forming phenotype promotes formation ofalveolar bone and cementum at the expense of thePL, thereby disrupting PL width homeostasis (19).The ability to induce mineralized tissue formationby PL cells in vivo as well as in vitro indicates thatthe di�erentiation repertoire of some PL cells caninclude osteoblastic di�erentiation. Our ®ndings ofstable phenotypic characteristics in primary PL cellcultures support the application of in vitro modelsfor assessment of PL cell di�erentiation and couldhave signi®cant implications for the interpretationof in vitro data.

References

1. Vignery A, Baron A. Dynamic histomorphometry ofalveolar bone remodeling in the adult rat. Anat Rec 1991;196:191±200.

2. Pitaru S, McCulloch CAG, Narayanan SA. Cellular originsand di�erentiation control mechanisms during periodontaldevelopment and wound healing. J Periodont Res 1994;29:81±94.

3. Gould TRL, Melcher AH, Brunette DM. Migration anddivision of progenitor cell populations in periodontalligament after wounding. J Periodont Res 1980;15:20±42.

4. Lekic P, Sodek J, McCulloch CAG. Osteopontin and bonesialoprotein expression in regenerating rat periodontalligament and alveolar bone. Anat Rec 1996;244:50±58.

5. McCulloch CAG, Melcher AH. Cell migration in theperiodontal ligament of mice. J Periodont Res 1983;18:339±352.

6. McCulloch CAG, Barghava U, Melcher AH. Cell deathand the regulation of population of cells in the periodontalligament. Cell Tissue Res 1989;255:129±138.

7. Gould TRL, Melcher AH, Brunette DM. Location ofprogenitor cells in periodontal ligament stimulated bywounding. Anat Rec 1977;188:133±141.

8. McCulloch CAG, Nemeth E, Lowenberg B, Melcher AH.Paravascular cells in endosteal spaces of alveolar bonecontribute to periodontal ligament cell populations. AnatRec 1987;219:2233±2242.

9. Palmer RM, Lumsden AGS. Development of periodontalligament and alveolar bone in homografted recombinationsof enamel organs and papillary, pulpal and follicularmesenchyme in the mouse. Arch Oral Biol 1987;32:281±289.

10. Melcher AH. On the repair potential of periodontal tissues.J Periodontol 1976;47:256±260.

78 Lekic et al.

Page 9: Phenotypic comparison of periodontal ligament cells in vivo and in vitro

11. Arora P, McCulloch CAG. Dependence of collagenremodeling on a a-smooth muscle actin expression by®broblasts. J Cell Physiol 1994;159:161±175.

12. Lekic P, Sodek J, McCulloch CAG. Relationship of cellularproliferation to expression of osteopontin and bonesialoprotein in regenerating rat periodontium. Cell TissueRes 1996;285:491±500.

13. Groenveld MC, Everts V, Beertsen W. A quantitativeenzyme histochemical analysis of the distribution of alkal-ine phosphatase activity in the periodontal ligament of therat incisor. J Dent Res 1993;72:1344±1350.

14. Hunter GK, Goldberg HA. Nucleation of hydroxyapatiteby bone sialoprotein. Proc Natl Acad Sci USA 1993;90:8561±8565.

15. Chen J, Shapiro HS, Wrana JL, Reimers S, Heersche JNM.Localization of bone sialoprotein (BSP) expression to thesite of mineralized tissue formation in fetal rat tissues byin situ hybridization. Matrix 1991;11:133±143.

16. Chen J, Shapiro HS, Sodek J. Developmental expression ofbone sialoprotein mRNA in rat mineralized connectivetisues. J Bone Miner Res 1992;8:987±997.

17. MacNeil RL, Sheng N, Strayhorn C, Fisher LW,Somerman MJ. Bone sialoprotein is localized to the rootsurface. J Bone Miner Res 1994;9:1597±1606.

18. Zohar R, Lee W, Arora P, Cheifetz S, McCulloch CAG,Sodek J. Single cell analysis of intracellular osteopontin inosteogenic cultures of fetal rat calvarial cells. J Cell Physiol1997;170:88±100.

19. Lekic P, Rubbino I, Krasnoshtein F, Cheifetz S, McCullochCAG, Tenenbaum HC. Bisphosphonate modulates prolif-eration and di�erentiation of rat periodontal ligament cellsduring wound healing. Anat Rec 1997;247:329±340.

20. Kuru L, Parkar MH, Gri�ths GS, Newman HN, Olsen I.Flow cytometry analysis of gingival and periodontalligament cells. J Dent Res 1998;77:555±564.

21. Tenorio D, Cruchley A, Hughes FJ. Immunohistochemicalinvestigation of the rat cementoblast phenotype. J PeriodontRes 1993;28:411±419.

22. Chen J, Zhang Q, McCulloch CAG, Sodek J. Immuno-histochemical localization of bone sialoprotein in foetalporcine bone tissues: comparisons with secreted phospho-protein 1 (SPP-1, osteopontin) and SPARC (osteonectin).Histochem J 1991;23:281±289.

23. McLean IW, Nakane PK. Periodate-lysine-paraformalde-hyde ®xative: a new ®xative for immunoelectron micro-scopy. J Histochem Cytochem 1974;22:1077±1083.

24. Lillie RD. Histopathological technique and practical histo-chemistry, 3rd edn. New York: McGraw Hill Book Co,1965.

25. McCulloch CAG, Melcher AH. Cell density and cellgeneration in the periodontal ligament of mice. Am J Anat1983;167:43±58.

26. Nojima N, Kobayashi M, Shionone M, Takahashi N,Suda T, Hasegawa K. Fibroblastic cells derived from bovineperiodontal ligaments have the phenotypes of osteoblasts.J Periodont Res 1990;25:179±185.

27. Basdra EK, Komposch G. Osteoblast-like properties ofhuman periodontal ligament cells: an in vitro analysis.Eur J Orthodont 1997;19:615±621.

28. Ogiso B, Hughes FJ, Melcher AH, McCulloch CAG.Fibroblasts inhibit mineralized bone nodule formation byrat bone marrow stromal cells in vitro. J Cell Physiol 1991;146:442±450.

29. Ogiso B, Hughes FJ, Davies JE, McCulloch CAG.Fibroblastic regulation of osteoblast function by prosta-glandins. Cell Signal 1992;4:627±639.

30. Arceo N, Sauk JJ, Moehring J, Foster RA, Somerman MJ.Human periodontal cells initiate mineral-like nodulesin vitro. J Periodontol 1991;62:499±503.

31. Zohar R, Sodek J, McCulloch CAG. Characterization ofstromal progenitor cells enriched by ¯ow cytometry. Blood1997;90:3471±3481.

32. Groenveld MC, Everts V, Beertsen W. Alkaline phosphat-ase activity in the periodontal ligament and gingiva of therat molar: its relation to cementum formation. J Dent Res1995;74:1374±1381.

33. Wilson AN, Middleton J, Jones MI, McGuiness NJ. The®nite analysis of stress in the periodontal ligament whensubject to vertical orthodontic forces. Br J Orthodont 1994;21:161±167.

34. Genge BR, Sauer GR, Wu LNY, McLean FM, WuthierRE. Correlation between loss of alkaline phosphataseactivity and accumulation of calcium during matrixvesicle-mediated mineralization. J Biol Chem 1988;263:18513±18519.

35. Cho M-I, Lin W-L, Moshier A, Ramakrishnan PR. In vitroformation of mineralized nodules by periodontal ligamentcell from the rat. Calcif Tissue Int 1992;50:459±467.

36. Wesselink PR, Beertsen W. Ankylosis of the mouse molarafter systemic administration of 1-hydroxyethylindene-1,1-bisphosphonate (HEBP). J Clin Periodontol 1994;21:465±471.

Phenotypic comparison of PL cells in vivo and in vitro 79