201801.500 031512.300 030118.900 040522.666 091313.875 010709.500160907*500

Type II collagen and TGF-βs in developingand aging porcine mandibular condylar cartilage:immunohistochemical studiesJ.R. Moroco1, R. Hinton2, P. Buschang1, S.B. Milam3, A.M. Iacopino2

1 Department of Orthodontics, Baylor College of Dentistry, P.O. Box 660677, Dallas, TX 75266-0677, USA2 Department of Biomedical Sciences, Baylor College of Dentistry, P.O. Box 660677, Dallas, 75266-0677, USA3 Department of Oral and Maxillofacial Surgery, University of Texas Health Science Center at San Antonio,7703 Floyd Curl Drive, San Antonio, TX 78284-7908, USA

&misc:Received: 11 July 1996 / Accepted: 30 December 1996

&p.1:Abstract. Transforming growth factor-betas (TGF-βs)have been associated with the development and mainte-nance of articular cartilage. However, no studies haveaddressed their role in the postnatal development ofmandibular condylar cartilage. This investigation repre-sents the first immunohistochemical characterization ofTGF-β isoforms and type II collagen in porcine mandib-ular condylar cartilage from various age groups. Further-more, it is the first description of possible age-relatedchanges in the expression of these proteins during post-natal development of this tissue. Condylar cartilage wasdissected from freshly harvested temporomandibularjoints of newborn, 6-, 12-, 24-, and 36-month-old farmswine. TGF-β1, TGF-β2, TGF-β3, and type II collagenwere localized via standard immunohistochemical pro-cedures. An immunoblot technique was employed tocompare the relative amount of each protein present inthe various age groups. Immunoreactivity was detectedin mandibular condylar cartilage for all three isoforms ofTGF-β and for Type II collagen. All age groups demon-strated some evidence of immunostaining, primarily inthe cytoplasm of cells from most zones of the cartilage.Immunoblot results indicated that TGF-β isoforms hadindividualized patterns of expression. When newbornprotein levels were taken as the baseline, TGF-β1 dem-onstrated a significant increase at ages 24 and 36months. TGF-β2 significantly increased at 6, 12, 24, and36 months (peak levels at 24 months; similar levels at 6,12, and 36 months), whereas TGF-β3 remained stable atall ages. Type II collagen demonstrated increases thatparalleled the increased levels of TGF-β1 and TGF-β2 at24 and 36 months.

&kwd:Key words: TGF-β – Condylar cartilage – Postnatal de-velopment – Immunohistochemistry – Pig

Introduction

The role of cytokines in the regulation of matrix andstructural changes is not well understood for limb articu-lar cartilage and has yet to be explored for the condylarcartilage of the mandible. In limb articular cartilage, thetransforming growth factor-beta (TGF-β) family of cyto-kines has received considerable attention, because of anextremely broad spectrum of biological activity in nu-merous cell types. In humans, three isoforms, viz., TGF-β1, TGF-β2, and TGF-β3, have been shown to be associ-ated with cell replication, fibrosis, and bone/cartilage for-mation. Moreover, these isoforms each have unique spa-tial and temporal patterns of expression (Millan et al.1991), suggesting that each has a distinct biological role,despite a 60%–80% sequence homology. Considerableevidence exists showing that TGF-β1 and TGF-β2 are in-timately involved in the control of proteoglycan metabo-lism in articular cartilage (Malemud et al. 1991; Moraleset al. 1991), specifically by increasing hyaluronic ac-id/proteoglycan monomer synthesis and by promoting theformation of link protein stabilized aggregates.

The growth and development of the mandibular con-dyle, and specifically the condylar cartilage, have beensubjects of intense interest for many years (Petrovic etal. 1975; Carlson 1994). Experimental and clinical stud-ies have been undertaken to examine the histological andmetabolic changes of mandibular condylar cartilage inresponse to physical and chemical stimuli (Carlson et al.1980; Copray et al. 1988; Hinton 1992). However, fewinvestigators have attempted to explore, at the molecularlevel, the role of local growth factors that have beenshown to be involved in the developmental mechanismsof articular cartilage. This study represents an initialcharacterization of TGF-βs in postnatal mandibular con-dylar cartilage.

This work was supported by grants from the Texas Association ofOrthodontists and the Baylor College of Dentistry Intramural Re-search Program, and by P.J. Polverini, University of MichiganSchool of Dentistry.

Correspondence to:R.J. Hinton (Tel.: +1-214-828-8272; Fax: +1-214-828-8951)&/fn-block:

Cell Tissue Res (1997) 289:119–124

© Springer-Verlag 1997

120

Materials and methods

Isolation and identification of cartilage specimens

Freshly harvested temporomandibular joint (TMJs) from farmswine of five age groups (newborn 6, 12, 24, and 36 months;n=12 per group) were purchased from Hedgemoor Farms (Aledo,Tex.). Each specimen consisted of approximately a 3 cm3 blockresection containing the condylar head, articular disc, and man-dibular fossa, with the exception of newborn specimens, whichincluded one half of the entire head. All right joints were placedimmediately in dry ice. All left joints were placed immediately in3.7% formalin. An incision into the inferior joint space of the leftTMJs was made to ensure contact of the fixative with the condy-lar head.

Immunohistochemistry

Antibodies specific for TGF-β1, TGF-β2, and TGF-β3 were pur-chased from R&D Systems (Minneapolis, Minn.). The antibodyspecific for TGF-β1 was produced in chickens by injection ofhighly purified, native porcine platelet TGF-β1; characterizationstudies demonstrated that this antibody had 5% cross-reactivitywith porcine TGF-β2 and TGF-β3. Antibodies specific for TGF-β2 and TGF-β3 were produced in goats by injection of purifiednative porcine TGF-β2 and purified recombinant chicken TGF-β3, respectively. The antibody of TGF-β2 did not neutralize TGF-β1 or TGF-β3, although it did neutralize TGF-β1.2 at a 50-foldhigher concentration. The antibody to TGF-β3 did not cross-reactwith TGF-β1, TGF-β1.2, or TGF-β2. Each antibody was dis-solved in sterile phosphate-buffered saline (PBS, pH 7.4) to give aworking stock concentration of 1 mg/ml. A 1:2000 dilution (0.5mg/ml) of each antibody in 0.05% Tween/PBS buffer was freshlyprepared for use in immunohistochemical procedures. Antibodyspecific for Type II collagen was purchased from Chemicon Inter-national (Temecula Calif.). The antibody was prepared in rabbitsby injection of purified bovine Type II collagen. Prior to use, theantibody was dissolved in 500 ml distilled water and diluted 1:100in 0.05% Tween/PBS buffer for use in immunohistochemical pro-cedures.

The condyle of each left joint was dissected from the tissueblock with the aid of an electric saw and PBS irrigation. Once iso-lated, a sagittal cut was made through the condyle to enable grossidentification of the bone/cartilage interface. This interface washistologically confirmed in decalcified hematoxylin/eosin-stainedsections. Cartilage was blunt-dissected from the condyle at this in-terface and placed immediately into 3.7% formalin fixative for 24h at 4° C. Following fixation, the tissue was dehydrated in gradedethanol/anyl acetate washes and embedded in low-melting-pointparaffin (<55° C). Paraffin-embedded cartilage was sectioned at 6µm, and the sections were mounted on positively charged glassslides (Fisher Scientific, Dallas, Tex.). Slides were placed on awarmer overnight (42° C for approximately 12 h) and stored inslide boxes at room temperature.

ImmunoPure Ultra-Sensitive ABC Staining Kits (Pierce Labo-ratories, Rockford, Ill.) were used for the detection of TGF-β1,TGF-β2, TGF-β3, and Type II collagen in sections of condylarcartilage. Slides were deparaffinized in toluene (2×10 min), fol-lowed by a decreasing gradient wash in ethanol (5 min, 50/50 tol-uene absolute ethanol; 2 min each, 2×absolute ethanol, 95%, 76%,and 50% ethanol/PBS). Slides were washed for 5 min in distilledwater. Endogenous peroxidase activity was quenched by incuba-tion of the tissue in 3% H2O2 in methanol for 30 min followed bya PBS wash for 20 min. Slides were then incubated in 1 mg/mlhyaluronidase with 0.1 M sodium acetate (pH 5.5) in 0.15 MNaCl for 30 min, followed by washes in iced PBS (3×5 min). Toreduce non-specific staining, sections were blocked with normal

rabbit serum in the case of TGF-β1, TGF-β2, and TGF-β3, andnormal goat serum in the case of Type II collagen. Sections wereincubated for 24 h at 4° C with primary antibody, washed exten-sively in PBS, incubated with an appropriate biotinylated second-ary antibody (rabbit anti-chicken for TGF-β1; rabbit anti-goat forTGF-β2, and TGF-β3; goat anti-rabbit for Type II collagen) for 30min, washed, and stained for 30 min with an avidin/biotinylatedenzyme complex. Detection of the immunoreaction was achievedby the addition of 3,3′-diaminobenzidine and hydrogen peroxidesubstrate (Vector Laboratories, Burlingame, Calif.) for 7 min. Oneexperimental slide from each group was counterstained with he-matoxylin. Primary antibody preabsorbed with the relevant pro-tein (R&D Systems) was used as a negative control in adjacentsections.

Immunoblot assay

Frozen tissue blocks containing the right joint were thawed at 4°C for 36–48 h. The condyle of each joint was dissected, and thebone/cartilage interface was identified as previously described.Cartilage was blunt-dissected from each half of the condyle andplaced immediately into 10 ml PBS (pH 7.4). Samples were ho-mogenized with a Tissue-Mizer (Beckman Instruments, Palo Alto,Calif.) and centrifuged for 30 min (5000 rpm, 4° C). The superna-tants were collected and stored at 4° C.

To determine the concentration of total protein (mg/ml) ineach supernatant, a spectrophotometric (DU64, Beckman Instru-ments) protein assay was performed (Lowry et al. 1951). Threedilutions from each sample containing 1, 5, and 10 µg total pro-tein were loaded onto pretreated nitrocellulose membranes (nor-mal goat or rabbit serum, 24 h, 4° C) via the 96-well Easy-TiterELIFA Unit (Pierce Laboratories). An internal control was runwith each membrane by loading 1, 5, and 10 mg bovine serumalbumin (BSA) in one row of the unit. Following the applicationof samples, membranes were removed from the unit and allowedto air-dry. The control row was separated from the membranesand washed (2×5 min) in 0.4% Tween/PBS. It was then incubat-ed for 30 min with 100 ml India ink in 100 ml 0.3% Tween/PBSto visualize the optical density of the BSA standards and washedin multiple changes of double-distilled H2O. In order to quenchendogenous peroxidases, the remaining membranes were incu-bated for 10 min with 3% H2O2 in PBS, followed by multiplewashes with PBS. To reduce non-specific staining, membraneswere blocked for 2 h with normal rabbit serum in the case ofTGF-β1, TGF-β2, and TGF-β3, and normal goat serum in thecase of Type II collagen. Membranes were incubated for 24 h at4° C with primary antibody (same dilutions as for immunohisto-chemistry), washed extensively in PBS, incubated with an ap-propriate biotinylated secondary antibody for 15 min, washed,and stained for 15 min with an avidin/biotinylated enzyme com-plex. Immunoreaction was detected as previously described.Densitometric scanning (Shimadzu CS 9000U; Tokyo, Japan)was used to compare relative amounts of protein present in thesamples. Protein levels were expressed in units of signal intensi-ty.

Statistical analysis

Statistical analysis was performed on data obtained from the sam-ple dilutions for each antibody. Background readings obtainedfrom BSA controls were subtracted from the raw data prior to sta-tistical analysis. Statistical significance was evaluated by the non-parametric Kruskal-Wallis test at the P<0.05 level. This test wasselected, because the raw data for each antibody were not normal-ly distributed, and because Bartlett’s test for homogeneity of vari-ance performed on mean data indicated that differences amongstandard deviations were highly significant.

Results

Immunohistochemistry

Immunoreactivity for Type II collagen, included primari-ly for comparative purposes in the immunoblot studies,was observed in all age groups. Immunoreactivity waspredominantly detected in the cytoplasm and extracellu-lar matrix of cells from the chondroblastic zone of thecondylar cartilage (Fig. 1B). In contrast, there was littleor no reactivity in the matrix surrounding the hypertro-phic chondrocytes, despite the positive reaction of thecells (Fig. 1C). The quality of staining was intense in allzones for all ages.

Immunoreactivity for all three TGF-β isoforms wasobserved in every age group and almost exclusively inthe cytoplasm. Variability in intensity of staining pre-cluded generalizations regarding age changes, althoughstaining was generally weaker in newborns. Typically,immunoreactivity was present mainly in the chondro-blastic and hypertrophic zones (Fig. 2), although someimmunoreactivity was observed in the prechondroblasticzone in the neonatal and 6-month age groups.

Immunoblot assay

Immunoreactivity to Type II collagen was detected at allages. When taking newborn protein levels as a baseline,there was a significant elevation in immunoreactivity at24 months and 36 months of age (P<0.05). Immunoreac-

tivity to TGF-β1 was detected at all ages. Like Type IIcollagen, a highly significant increase occurred at 24 and36 months of age. Immunoreactivity to TGF-β2 was de-tected at all ages, peak levels of TGF-β2 being demon-strated at 24 months, with similar levels of TGF-β2 be-ing observed at 6, 12, and 36 months. Immunoreactivityto TGF-β3 was detected at all ages. However, relative tonewborns, no significant differences existed between anyof the age groups.

Graphical representation of Type II collagen, TGF-β1, TGF-β2, and TGF-β3 levels is presented in Fig. 3.Additionally, immunoblot data for TGF-βs is presentedin a normalized format (ratiometric table) in Table 1.Each TGF-β isoform is expressed as a percentage of thetotal TGF-β immunoreactivity for each timepoint of thestudy. The percentage of TGF-β1 immunoreactivity ishighest early (newborn) and late (24 and 36 months) indevelopment, whereas levels of this isoform are low at 6and 12 months. The percentage of TGF-β2 is extremely

121

Table 1.Ratiometric presentation of TGF-β isoforms expressed asa percentage of the total TGF-β immunoactivity&/tbl.c:&tbl.b:

Age (months) TGF-β1 TGF-β2 TGF-β3

Newborn 40 5 556 21 50 29

12 19 46 3524 45 33 2236 47 28 25

&/tbl.b:

Fig. 1A–C. Representative photomicrographs of the immunohistochemical localization of Type II collagen in mandibular condylar carti-lage from a 6-months-old pig. A Control, B chondroblastic zone, C hypertrophic zone. Bar: 25 µm&/fig.c:

122

Fig. 2A–F. Representative photomicrographs of the immunohisto-chemical localization of TGF-β1 and TGF-β2 isoforms in man-dibular condylar cartilage. A TGF-β1 control, neonate, B TGF-β1in upper chondroblastic zone, neonate, C TGF-β1 in hypertrophic

zone, neonate, D TGF-β2 control, neonate, E TGF-β2 in chon-droblastic zone, neonate, F TGF-β2 in hypertrophic zone, 6-month-old pig. Bar: 25 µm&/fig.c:

Fig. 3. Graphical presentation (mean±SEM) of immu-noblot data for Type II collagen(COL), TGF-β1 (B1),TGF-β2 (B2), and TGF-β3 (B3) in porcine mandibularcondylar cartilage from various age groups (NB new-born). Significant differences are present for TGF-β1between the newborn and 24-months and 36-monthslevels, and for TGF-β2 between the newborn and allother levels; no differences are present for TGF-β3. ForType II collagen, significant differences are present be-tween the newborn, 24-month, and 36-months levels

TGF-β2 immunoreactivity. The temporal and spatial re-lationship between these TGF-βs and Type II collagenprovides circumstantial evidence for a role of TGF-βs inthe development of the cartilaginous matrix of the con-dyle, a notion supported by previous studies that haveimplicated specific TGF-β isoforms in chondrogenesis(Morales and Roberts 1988; Morales et al. 1991; Mo-rales 1991). Specifically, the observed changes in immu-noreactive Type II collagen, TGF-β1, and TGF-β2 at 24and 36 months of age may reflect increased functionaldemands on the pig TMJ, since dental attrition is mostnoticeable during these ages (Tonge and McCane 1973).

An alternative explanation for the observed age-relat-ed changes in immunoreactive Type II collagen, TGF-β1, and TGF-β2 may involve an unmasking process.Newly synthesized TGF-β molecules are deposited inthe extracellular matrix and may bind specific extracel-lular molecules (Sporn and Roberts 1991). These bind-ing events may mask immunoreactive isotopes. Loss ofspecific binding molecules or changes in the molecularcomposition of the extracellular matrix may affect epi-topic exposure, thereby influencing the amounts of TypeII collagen, TGF-β1, and TGF-β2 that are detectable byimmunological techniques.

One must exercise caution when interpreting immu-nological data. TGF-β isoforms have been identified inthis study by using immunohistochemistry and immuno-blotting techniques. TGF-βs are secreted in biologicallyinactive pro-isoforms and subsequently activated by pro-teolytic cleavage (Kryceve-Martinerie et al. 1985; Law-rence et al. 1985; Keski-Oja et al. 1987). Liberated ac-tive forms bind to specific cell surface receptors elicitingbiological responses from affected cell populations(Wakefield et al. 1987). Therefore, the mere presence ofa TGF-β isoform, as detected by immunologically basedmethods, does not necessarily signify biological activity.

In summary, the results of this investigation demon-strate that TGF-βs are present in the condylar cartilagethroughout postnatal development. Co-expression ofsome isoforms with Type II collagen suggests that one ormore isoforms play a role in the postnatal developmentof porcine condylar cartilage.

123

low initially, becomes the predominant isoform at 6 and12 months, and then decreases later. Levels of TGF-β3are initially high, accounting for more than half of allTGF-β immunoreactivity, with levels decreasing to25%–35% of total immunoreactivity for the subsequentage groups.

Discussion

Immunoreactivity for TGF-β1, TGF-β2, and TGF-β3 inpostnatal porcine condylar cartilage, as characterized byimmunoblot, demonstrates distinct patterns of expres-sion up to 36 months of age. Such differential expressionof TGF-β isoforms in bone and cartilage has been estab-lished in previous studies (Carrington et al. 1988;Gatherer et al. 1990; Millan et al. 1991; Villiger andLotz 1992; Thorp et al. 1992). Furthermore, some cellsmay be responsive to ratios of TGF-β isoforms (Bascomet al. 1989). We have found distinct age-related changesin the ratio of TGF-β isoforms in porcine condylar carti-lage. Newborn tissues exhibit a predominance of TGF-β1 and TGF-β3, tissues at 6–12 months demonstrate apredominance of TGF-β2, and at 24–36 months exhibit apredominance of TGF-β1.

Although it is tempting to propose a linkage of age-related changes in TGF-β immunoreactivity with func-tional changes occurring in the pig TMJ at specificpoints of development, such speculation must be under-taken with caution. For example, weaning of the pig isinitiated at 6–8 weeks of age (Mount 1968). The currentinvestigation has detected a significant increase in TGF-β2 immunoreactivity from newborn to 6 months of age.Although this relationship cannot be assessed more pre-cisely from our data, the increased expression of thisprotein is supported by the observation that TGF-β2 ismore effective at stimulating chondrocyte prolifera-tion/chondrogenesis than is TGF-β1 (Hiraki et al. 1988;Joyce et al. 1990). Similarly, a significant increase in-Type II collagen immunoreactivity is observed at 24months of age and again at 36 months of age, and is ac-companied by a significant increase in both TGF-β1 and

124

&p.2:Acknowledgements.Special thanks to Peter J. Polverini of theUniversity of Michigan School of Dentistry for additional techni-cal assistance.

References

Bascom CC, Wolfshohl JR, Coffey RJ, Madisen L, Webb NR,Purchio AR, Derynck R, Moses HL (1989) complex regula-tion of transforming growth factor beta 1, beta 2, and beta 3mRNA expression in mouse fibroblasts and keratinocytes bytransforming growth factors beta 1 and beta 2. Mol Cell Biol9:5508–5515

Carlson DS (1994) Growth of the temporomandibular joint. In:Zarb GA, Carlson DS, Sessle BJ, Mohl ND (eds) Temporo-mandibular joint and masticatory muscle disorders. Munksga-ard, Copenhagen, pp 128–155

Carlson DS, McNamara JA, Graber LW, Hoffman DL (1980) Ex-perimental studies of growth and adaptation of TMJ. In: IrbyWB (ed) Current advances in oral surgery. Mosby, St. Louis,pp 28–77

Carrington JL, Roberts AR, Flanders KC, Roche NS, Reddi AH(1988) Accumulation, localization, and compartmentation oftransforming growth factor β during endochondral bone devel-opment. J Cell Biol 107:1969–1975

Copray JCVM, Dibbets JMH, Kantomaa T (1988) The role ofcondylar cartilage in the development of the temporomandibu-lar joint. Angle Orthod 125:369–380

Gatherer DP, Ten Dijke P, Baird DT, Akhurst RJ (1990) Expres-sion of TGF-β isoforms during first trimester human embryo-genesis. Development 110:445–460

Hinton RJ (1992) Alterations in rat condylar cartilage followingdiscectomy. J Dent Res 71:1292–1297

Hiraki Y, Inoue H, Hirai R, Kato Y, Suzuki F (1988) Effect oftransforming growth factor β on cell proliferation and glycos-aminoglycan synthesis by rabbit growth-plate chondrocytes inculture. Biochim Biophys Acta 969:91–99

Joyce ME, Terek RM, Jingushi S, Bolander ME (1990) Role oftransforming growth factor-β in fracture repair. Ann N YAcad Sci 593:107–111

Keski Oja J, Lyons RM, Moses HL (1987) Inactive secreted form(s)of TGF-β: activation by proteolysis. J Cell Biochem 11:60–65

Kryceve-Martinerie C, Lawrence DA, Crochet J, Jullien P, VigierP (1985) Further study of beta-TGFs released by virallytransformed and non-transformed cells. Int J Cancer 35:553–558

Lawrence DA, Pircher R, Jullien P (1985) Conversion of a highmolecular weight latent TGF-β from chicken embryo fibro-blasts into a low molecular weight active TGF-β under acidicconditions. Biochem Biophys Res Commun 133:1026–1034

Lowry OH, Rosebrough NJ, Farr AL, Randall RJ (1951) Proteinmeasurement with the Folin-phenol reagent. J Biol Chem193:265–275

Malemud CJ, Killeen W, Hering TM, Purchio AF (1991) Enhancdsulfated-proteoglycan core protein synthesis by incubation ofrabbit chondrocytes with recombinant transforming growthfactor-β1. J Cell Physiol 149:152–159

Millan FA, Denhenz F, Kondaiah P, Akhurst RJ (1991) Embryonicgene expression patterns of TGF-β1, β2, and β3 suggest differ-ent development functions in vivo. Development 111:131–144

Morales TI (1991) Transforming growth factor-β1 stimulates syn-thesis of proteoglycan aggregates in calf articular cartilage or-gan cultures. Arch Biochem Biophys 286:99–106

Morales TI (1988) Transforming growth factor β regulates themetabolism of proteoglycans in bovine cartilage organ cul-tures. J Biol Chem 263:12828–12831

Morales TI, Joyce MB, Sobel ME, Danielpour D (1991) Trans-forming growth factor-β in calf articular cartilage organ cul-tures: synthesis and distribution. Arch Biochem Biophys 288:397–405

Mount LE (1968) The climatic physiology of the pig. Arnold,London

Petrovic AG, Stutzmann JJ, Oudet CL (1975) Control processes inthe postnatal growth of the condylar cartilage of the mandible.In: McNamara JA (ed) Determinants of mandibular form andgrowth. University of Michigan Press, Ann Arbor, pp 101–154

Sporn MB, Roberts AB (1991) What is TGF-β? In: Bock GR,Marsh J (eds) Clinical applications of TGF-β. Wiley, Chiches-ter,pp 176–198

Thorp BH, Anderson I, Jakowlew SB (1992) TGF-β1, -β2, and β3in cartilage and bone cells during endochondral ossification inthe chick. Development 114:907–911

Tonge CH, McCane RA (1973) Normal development of the jawsand teeth in pigs, and the delay and malocclusion produced bycalorie deficiencies. J Anat 115:3–22

Villiger PM, Lotz M (1992) Differential expression of TGF-β iso-forms by human articular chondrocytes in response to growthfactors. J Cell Physiol 151:318–325

Wakefield LM, Smith DM, Masui T, Harris CC, Sporn MB (1987)Distribution and modulation of the cellular receptor for trans-forming growth factor-beta. J Cell Biol 105:965–975