Search tips
Search criteria 


Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
J Dent Res. Author manuscript; available in PMC 2010 July 23.
Published in final edited form as:
PMCID: PMC2909184

Clinically-Relevant Cell Sources for TMJ Disc Engineering


Tissue-engineering of the temporomandibular joint (TMJ) disc aims to provide patients with TMJ disorders an option to replace diseased tissue with autologous, functional tissue. This study examined clinically-relevant cell sources by comparing costal chondrocytes, dermal fibroblasts, a mixture of the two, and TMJ disc cells in a scaffoldless tissue-engineering approach. It was hypothesized that all constructs would produce matrix relevant to the TMJ disc, but the mixture constructs were expected to appear most like the TMJ disc constructs. Costal chondrocyte and mixture constructs were morphologically and biochemically superior to the TMJ disc and dermal fibroblast constructs, and their compressive properties were not significantly different. Costal chondrocyte constructs produced almost 40 times more collagen and 800 times more glycosaminoglycans than TMJ constructs. This study demonstrates the ability of costal chondrocytes to produce extracellular matrix that may function in a TMJ disc replacement.


There are approximately 10 million patients in the United States that suffer from TMJ disorders.1 In severely diseased joints, surgical options are met with varying amounts of success, as reviewed elsewhere.2,3 Efforts to engineer the TMJ disc may create a viable alternative to current treatment options.

While previous work using TMJ disc cells characterized their in vitro behavior for the purpose of tissue-engineering, a clinical solution will likely not involve these cells. This conclusion was reached following the experience of our laboratory and others, demonstrating the difficulty in culturing these cells, failure of constructs to achieve sufficient morphology, and poor mechanical strength of the engineered tissue.410

A clinically-feasible cell source should be abundant, healthy, and leave little donor site morbidity. Selection of an alternative source must also consider the functionality of the cells. Previous characterization data guide this selection; the TMJ disc has properties of both fibrous tissue and cartilage, indicating the need for a fibrochondrocytic cell source.11 Specifically, the cells should produce tissue containing collagen type I, type II, and glycosaminoglycans and should support both tensile and compressive loads. Dermal fibroblasts exhibit chondrogenic potential despite being inherently fibrogenic in nature.12,13 Costal cartilage contains both collagen types II and I (in a ratio of 5:1) and glycosaminoglycans suggesting its potential to function as a fibrocartilage replacement.14 Indeed, costochondral grafts are already used in mandibular reconstruction.15 In addition to their functional potential, costochondral cartilage and dermal fibroblasts are easily harvested and more abundant than TMJ disc cells, making them a more clinically-feasible source of cells for tissue-engineering.

Constructs of costal chondrocytes, dermal fibroblasts, TMJ disc cells, and a 50/50 costal chondrocyte/dermal fibroblast mixture were examined in this study. Based on known tissue characteristics, it was hypothesized that all cell types would produce extracellular matrix similar to the TMJ disc cell construct. However, the costal chondrocyte/dermal fibroblast mixture was expected to produce the most TMJ disc cell-like construct. Specifically, mechanical and biochemical properties that are most similar to the TMJ would be observed via glycosaminoglycan and collagen II production by the costal chondrocytes and collagen I production by the dermal fibroblasts.

Materials and methods

Cell isolations

Cells were isolated from three skeletally-mature, Spanish, female goats. TMJ disc cells were isolated as previously described and cultured until 70–90% confluent.10 They were passaged with trypsin-EDTA (Gibco, Carlsbad, CA) until passage 2. Culture medium was Dulbecco’s Modified Eagle’s Medium (DMEM) with L-glutamine and 4.5g/L glucose (Biowhittaker, Walkersville, MS), 10% fetal bovine serum (Gemini Bio-Products, Woodland, CA), 1% Penicillin-Streptomycin-Fungizone (Cambrex, Walkersville, MD), 1% non-essential amino acids (Gibco), 25μL/mL L-ascorbic acid (Sigma, St. Louis, MO), and 1μL/mL insulin (Sigma).

Skin was cut into 1cm2 squares, digested in 0.5% dispase (Gibco) at 4°C overnight, and then epidermis and adipose layers were removed. The remaining dermis was placed in 0.05% type II collagenase (Worthington, Lakewood, NJ). After 24hrs, samples were passed through a 70μm cell strainer, yielding a single-cell suspension, which was plated and fed with DMEM containing Glutamax, 10% fetal bovine serum, 1% Penicillin-Streptomycin-Fungizone, and 1% non-essential amino acids. When confluent, cultures were exposed to 0.5% dispase for 30min to remove keratinocytes. The purified dermis cells were allowed to expand, and passage 2 cells were used.

Cartilage was scraped from non-floating ribs, minced into cubes of approximately 1mm3, and digested overnight with 0.2% collagenase in DMEM. After isolation, cells were frozen in DMEM with 10% dimethyl sulfoxide, 20% fetal bovine serum, 1% Penicillin-Streptomycin-Fungizone, and 1% non-essential amino acids to allow for concurrent seeding of costal cartilage constructs with the other groups.

Construct culture

Constructs were formed using a scaffoldless method described previously.16 Two million cells (TMJ, costal chondrocytes, dermal fibroblast or a 50:50 costal chondrocyte:dermal fibroblast mix) were seeded in 3mm wells formed with 2% agarose (Fisher Scientific). Media changes occurred everyday using DMEM with 1% Penicillin-Streptomycin-Fungizone, 1% NEAA, 1% insulin-transferrin-selenium+ premix (BD Biosciences, San Jose, CA), 0.1μM dexamethasone, 40μg/mL L-proline (EMD Chemicals, Gibbstown, NJ), 50μg/mL ascorbate 2-phosphate (Sigma), and 100μg/mL sodium pyruvate (Fisher). After 2 wks, constructs were transferred to agarose-coated plates. Samples were removed for biochemistry and histology at 3 wks and 6 wks. Additionally, mechanical testing was performed at 6 wks.


Samples were frozen in HistoPrep (Fisher), and 14μm sections were prepared. Slides were stained with picrosirius red for collagen or safranin-O/fast green for glycosaminoglycans. Immunohistochemistry slides were stained for collagen types I and II, as described previously.17


Four samples per group were lyophilized for 2 days and digested at 4°C with constant agitation for 7 days with 125μg/mL papain (Sigma) digest, followed by 2 days of 1mg/mL elastase (Sigma) digestion. Samples were stored at −20°C.

Cell numbers were determined using PicoGreen® (Molecular probes) with a conversion factor of 7.7pg DNA/cell.18 Total collagen was measured using a modified hydroxyproline assay.19 Sulfated glycosaminoglycans were quantified with a dimethylmethlylene blue Blyscan kit (Biocolor, Newtownabbey, Ireland). Type I collagen was quantified with an indirect ELISA, described previously.20

Mechanical testing

At least five samples per group were tested in tension and compression. Tensile testing was performed on an Instron 5565 (Norwood, MA) to determine ultimate tensile strength and elastic modulus. Samples were cut with into a dog bone shape using a scalpel blade and biopsy punch and tested at 10% strain rate/min until failure.

Specimens were tested in unconfined compression with an indentation apparatus.21 Each sample was tare-loaded with 0.00196mN until equilibrium was reached (deformation less than 10−6mm/s) or loading time reached 10min. A step load of 0.00686N was then applied until equilibrium or 1hr elapsed. Creep data were analyzed with Matlab’s (The Math Works, Inc) curve fitting tool using the viscoelastic model.22


Data were analyzed with a 2-way analysis of variance where time and cell type were factors with two and four levels, respectively. When an F-test indicated significance (p<0.05), a Tukey’s post hoc test was performed.


Morphology and histology

Morphologically, dermal fibroblast and TMJ disc constructs contracted, yielding mostly spherical constructs measuring about 1mm in diameter. Mixture constructs contracted to a diameter just less than 2mm and maintained a more cylindrical shape. Costal chondrocyte constructs expanded to a diameter and height just over 3mm (Fig. 1 and Table 1).

Figure 1
Gross morphology and histology at wk6 for tissue-engineered constructs with various cell sources: TMJ (a column), dermal fibroblast (b column), costal chondrocyte/dermal fibroblast (c column), and costal chondrocyte (d column). a–h illustrate ...
Table 1
Quantitative size data (mean ± standard deviation) of all groups. Groups separated by different letters are considered significantly different (p<0.05). Volumes were significantly greater at wk6 than at wk3 as indicated by the *. Time ...

TMJ and dermal fibroblast groups did not stain with safranin-O (Fig. 1m–p). All groups stained positive for collagen (Fig. 1i–l). The costal chondrocyte group stained intensely for glycosaminoglycans throughout the construct, whereas the costal chondrocyte/dermal fibroblast group stained only around the periphery of the construct. Immunohistochemistry demonstrated positive collagen I staining for all groups. Only groups which contained costal chondrocytes, however, stained positive for collagen II (Fig. 2).

Figure 2
Immunohistochemistry staining for constructs made from TMJ disc cells (a, f), dermal fibroblasts (b, g), costal chondrocyte/dermal fibroblast co-culture (c, h), and costal chondrocytes (d, i). a–e show the collagen I staining which is present ...


Costal chondrocyte constructs had significantly more cells than any other group, and costal chondrocyte/dermal fibroblast constructs had significantly more than dermal fibroblast or TMJ constructs (Fig. 3a). Cell numbers were not significantly affected by time. The initial seeding was 2 million cells/construct, which only the costal chondrocyte constructs maintained.

Figure 3
Biochemical quantities for cells (a), collagen per construct (b), and percent increase in collagen type I (c). All data are shown as mean + standard deviation with a sample size equal to four for all groups. Groups separated by different letters are considered ...

Costal chondrocyte constructs had significantly more total collagen than any other group (Fig. 3b). Collagen type I was normalized to the amount in the TMJ constructs, and there was significantly more collagen I for costal chondrocyte and costal chondrocyte/dermal fibroblast over the other constructs (Fig. 3c). Both total collagen and collagen I increased significantly at wk6 from wk3.

Glycosaminoglycan per construct also increased significantly at the later time point. For costal chondrocyte constructs it was orders of magnitudes higher than the other groups with values at 3 and 6 wks of 640±17μg and 1700±94μg, respectively. TMJ constructs made 1.7±0.6μg at 3 wks and 1.1±0.4μg at 6 wks. At 3 wks the dermal fibroblast constructs contained 2.6±1.0μg and 1.1±0.3μg at 6 wks. The co-culture constructs produced 1.8±0.7μg and 2.0±0.8μg at 3 and 6 wks, respectively. Despite cell quantity changes, matrix normalization to cell number resulted in similar trends to the per construct normalization.

Mechanical properties

There were no statistical differences between any of the measured compressive properties. For all three tensile properties, the TMJ constructs were significantly higher (Table 2).

Table 2
Mechanical properties (mean ± standard deviation) of all groups at wk6. TMJ disc cell constructs were significantly stronger in tension than any other group as indicated by the * (p<0.05). No other statistical differences were observed. ...


Due to the prevalence of TMJ disorders and limited treatment options, it is essential to examine possible alternatives to current surgical techniques, such as engineering a replacement disc.3 Using TMJ disc cells in this approach has numerous drawbacks, such as a limited population of healthy cells and donor site morbidity. The current study examines the potential of dermal fibroblasts and costal chondrocytes as alternative cell sources for fibrocartilage tissue-engineering. The results show significant increases in extracellular matrix produced by costal chondrocytes from that produced by dermal fibroblasts or TMJ disc cells, while co-culture of costal chondrocytes and dermal fibroblasts made extracellular matrix in quantities between either individual cell type. These distinctions were also apparent in construct size and weight, where costal chondrocyte constructs were significantly larger. The hypothesis that a costal chondrocyte/dermal fibroblast co-culture would produce constructs most similar to those produced by the TMJ cells was not supported by this work. Instead, dermal fibroblast constructs were most like TMJ constructs. Costal chondrocyte/dermal fibroblasts trended toward improved biochemical content in addition to improved morphology over TMJ cells and dermal fibroblasts. However, costal chondrocyte constructs exceeded expectations by producing significantly more glycosaminoglycans, total collagen, and collagen type I than any other construct.

The extracellular matrix results obtained with the TMJ disc cells are representative of previous studies, which suggest that these cells do not exhibit a robust synthetic ability. In contrast, costal chondrocyte constructs demonstrate that cells derived from this source are highly productive relative to the others tested. The most collagen produced previously with TMJ disc cells was approximately 60μg total collagen per construct—made with over 6 million cells.7 At the same time point (6 wks), the costal chondrocyte constructs produced over 450μg of total collagen with an initial seeding density of 2 million cells. Additionally, the costal chondrocyte constructs yielded almost 100 times more glycosaminoglycans than previous TMJ disc cell constructs.7

The 200–300% increase in collagen I of costal chondrocyte constructs over TMJ constructs further illustrates the productive capacity of the costal chondrocytes. Since rib cartilage contains both collagen types I and II, and skin contains primarily type I, the dermal fibroblast constructs were expected to produce the most collagen type I followed by costal chondrocyte/dermal fibroblast constructs, and, finally, costal chondrocyte constructs. However, like total collagen, dermal fibroblast constructs contained the least collagen I, and no statistical difference was seen between the costal chondrocyte and costal chondrocyte/dermal fibroblast group. This reinforces the total extracellular matrix data that suggest costal chondrocytes alone may be a viable cell source for functional tissue-engineering of the TMJ disc.

While extracellular matrix data indicate that costal chondrocytes are most likely to succeed in fibrocartilage tissue-engineering, mechanical data did not correspond to extracellular matrix changes. Generally, an increase in glycosaminoglycans increases compressive resilience, while more collagen improves tensile strength. In this experiment, the TMJ disc cell constructs had significantly higher tensile properties. This could be due to tighter cell packing or better organization of the extracellular matrix. However, even the largest of any of the constructs’ tensile properties were still orders of magnitude below the native values for the TMJ disc.23 Despite the lower tensile strength, the high quantities of extracellular matrix suggest that with the proper stimuli (biochemical or mechanical) the costal chondrocytes can produce a more mechanically robust construct, perhaps through better organization of the collagen fibers. Mechanical stimuli are particularly well-suited for altering organizational changes, as seen with a variety of tissue engineering studies, and will be an important area of future research for TMJ disc tissue engineering.2427

While the scaffoldless approach used in this study has clear advantages, like avoiding immune responses due to biomaterials, many of the TMJ disc cells and dermal fibroblasts were not retained at even the first time point. Low cell retention was also seen previously with TMJ disc cells on scaffolds.7 At both time points, only one-eighth of the original cells were measured for the TMJ and dermal fibroblast constructs, and one-fourth remained in the costal chondrocyte/dermal fibroblast constructs. Dermal fibroblast and TMJ constructs also contracted significantly from the initial well diameter, while the costal chondrocyte constructs retained their initial size or grew slightly. A reduction in size makes it more difficult to engineer a replacement tissue with functional dimensions. Considering the retention of cells, most overall extracellular matrix production, and ability to create a replacement tissue with clinically-relevant dimensions, the costal chondrocytes appear to be the most likely cell source candidate, of those studied here, for TMJ disc replacement, particularly with this scaffoldless method.

Finally, the costal chondrocyte constructs offer several other advantages as a cell source for TMJ disc reconstruction. Large quantities of costal cartilage can be obtained from almost any patient with a minimally-invasive harvest technique producing limited morbidity and complications.28 While costal cartilage is a relatively acellular tissue,29 the protocol could be optimized to limit the amount of needed tissue, for example, by expanding the cells before construct formation. However, patients requiring a tissue-engineered disc would not have sufficient quantity of nonpathological tissue to provide an adequate number of TMJ disc cells, even with passaging, particularly considering the size reduction discussed previously.30 While using costal cartilage without in vitro manipulation is appealing, previous work reveals complications, like tissue overgrowth31,32 and undesirable calcification. These concerns can be addressed by controlling the in vitro environment used in a tissue-engineering approach. By influencing growth conditions and applied stimuli, constructs can be engineered to produce the appropriate dimensions, mechanical properties, and biochemical properties. After examining the integrative capacities of engineered neotissue at different maturities, grafts may be more readily integrated with the native joint.

Although costal chondrocytes are clearly superior in this experiment as a highly productive and feasible cell source for tissue-engineering, the simplicity of dermal fibroblast harvest warrants its continued examination. Further optimization is also needed to improve the costal chondrocyte constructs’ mechanical properties. With the application of external stimuli, like growth factors or mechanical forces, scaffoldless costal chondrocyte constructs may produce sufficient quantities of organized matrix to function as a TMJ disc replacement and serve as a feasible alternative for patients.


We acknowledge funding from NIDCR#R01DE015038-01A2 and NIAMS#R01AR47839-2.


1. NIDCR. TMJ Disorders. 2006.
2. Dimitroulis G. The role of surgery in the management of disorders of the temporomandibular joint: a critical review of the literature. Part 2. Int J Oral Maxillofac Surg. 2005;34:231–7. [PubMed]
3. Wong ME, Allen KD, Athanasiou KA. Tissue engineering of the temporomandibular joint. In: Bronzino JD, editor. Biomedical Engineering Handbook Third Edition: Tissue Engineering and Artificial Organs. CRC Press; 2006. pp. 51-1–52-22.
4. Puelacher WC, Wisser J, Vacanti CA, Ferraro NF, Jaramillo D, Vacanti JP. Temporomandibular joint disc replacement made by tissue-engineered growth of cartilage. J Oral Maxillofac Surg. 1994;52:1172–7. discussion 1177–8. [PubMed]
5. Springer IN, Fleiner B, Jepsen S, Acil Y. Culture of cells gained from temporomandibular joint cartilage on non-absorbable scaffolds. Biomaterials. 2001;22:2569–77. [PubMed]
6. Allen KD, Athanasiou KA. Growth factor effects on passaged TMJ disk cells in monolayer and pellet cultures. Orthod Craniofac Res. 2006;9:143–52. [PubMed]
7. Almarza AJ, Athanasiou KA. Effects of initial cell seeding density for the tissue engineering of the temporomandibular joint disc. Ann Biomec Eng. 2005;33:943–950. [PubMed]
8. Almarza AJ, Athanasiou KA. Effects of hydrostatic pressure on TMJ disc cells. Tissue Eng. 2006;12:1285–94. [PubMed]
9. Detamore MS, Athanasiou KA. Evaluation of three growth factors for TMJ disc tissue engineering. Ann Biomed Eng. 2005;33:383–90. [PubMed]
10. Johns DE, Athanasiou KA. Improving culture conditions for temporomandibular joint disc tissue engineering. Cells Tissues Organs. 2007;185:246–57. [PubMed]
11. Almarza AJ, Athanasiou KA. Design characteristics for the tissue engineering of cartilaginous tissues. Ann Biomed Eng. 2004;32:2–17. [PubMed]
12. Deng Y, Hu JC, Athanasiou KA. Isolation and chondroinduction of a dermis-isolated, aggrecan-sensitive subpopulation with high chondrogenic potential. Arthritis Rheum. 2007;56:168–76. [PubMed]
13. French MM, Rose S, Canseco J, Athanasiou KA. Chondrogenic differentiation of adult dermal fibroblasts. Ann Biomed Eng. 2004;32:50–6. [PubMed]
14. Safronova EE, Borisova NV, Mezentseva SV, Krasnopol’skaya KD. Characteristics of the macromolecular components of the extracellular matrix in human hyaline cartilage at different stages of ontogenesis. Biomed Sci. 1991;2:162–8. [PubMed]
15. Lindqvist C, Jokinen J, Paukku P, Tasanen A. Adaptation of autogenous costochondral grafts used for temporomandibular joint reconstruction: a long-term clinical and radiologic follow-up. J Oral Maxillofac Surg. 1988;46:465–70. [PubMed]
16. Hu JC, Athanasiou KA. A self-assembling process in articular cartilage tissue engineering. Tissue Eng. 2006;12:969–79. [PubMed]
17. Detamore MS, Orfanos JG, Almarza AJ, French MM, Wong ME, Athanasiou KA. Quantitative analysis and comparative regional investigation of the extracellular matrix of the porcine temporomandibular joint disc. Matrix Biol. 2005;24:45–57. [PubMed]
18. Kim YJ, Sah RL, Doong JY, Grodzinsky AJ. Fluorometric assay of DNA in cartilage explants using Hoechst 33258. Anal Biochem. 1988;174:168–76. [PubMed]
19. Woessner JF. The determination of hydroxyproline in tissue and protein samples containing small proportions of this imino acid. Arch Biochem Biophys. 1961;93:440–7. [PubMed]
20. Darling EM, Athanasiou KA. Growth factor impact on articular cartilage subpopulations. Cell Tissue Res. 2005;322:463–73. [PubMed]
21. Athanasiou KA, Agarwal A, Dzida FJ. Comparative study of the intrinsic mechanical properties of the human acetabular and femoral head cartilage. J Orthop Res. 1994;12:340–9. [PubMed]
22. Leipzig ND, Athanasiou KA. Unconfined creep compression of chondrocytes. J Biomech. 2005;38:77–85. [PubMed]
23. Beatty MW, Bruno MJ, Iwasaki LR, Nickel JC. Strain rate dependent orthotropic properties of pristine and impulsively loaded porcine temporomandibular joint disk. J Biomed Mater Res. 2001;57:25–34. [PubMed]
24. Aufderheide AC, Athanasiou KA. A direct compression stimulator for articular cartilage and meniscal explants. Ann Biomed Eng. 2006;34:1463–74. [PubMed]
25. Eastwood M, V, Mudera C, McGrouther DA, Brown RA. Effect of precise mechanical loading on fibroblast populated collagen lattices: morphological changes. Cell Motil Cytoskeleton. 1998;40:13–21. [PubMed]
26. Huang D, Chang TR, Aggarwal A, Lee RC, Ehrlich HP. Mechanisms and dynamics of mechanical strengthening in ligament-equivalent fibroblast-populated collagen matrices. Ann Biomed Eng. 1993;21:289–305. [PubMed]
27. Seliktar D, Black RA, Vito RP, Nerem RM. Dynamic mechanical conditioning of collagen-gel blood vessel constructs induces remodeling in vitro. Ann Biomed Eng. 2000;28:351–62. [PubMed]
28. Yotsuyanagi T, Mikami M, Yamauchi M, Higuma Y, Urushidate S, Ezoe K. A new technique for harvesting costal cartilage with minimum sacrifice at the donor site. J Plast Reconstr Aesthet Surg. 2006;59:352–9. [PubMed]
29. Stockwell RA. The cell density of human articular and costal cartilage. J Anat. 1967;101:753–63. [PubMed]
30. Detamore MS, Hegde JN, Wagle RR, Almarza AJ, Montufar-Solis D, Duke PJ, Athanasiou KA. Cell type and distribution in the porcine temporomandibular joint disc. J Oral Maxillofac Surg. 2006;64:243–8. [PubMed]
31. Baek RM, Song YT. Overgrowth of a costochondral graft in reconstruction of the temporomandibular joint. Scand J Plast Reconstr Surg Hand Surg. 2006;40:179–85. [PubMed]
32. Samman N, Cheung LK, Tideman H. Overgrowth of a costochondral graft in an adult male. Int J Oral Maxillofac Surg. 1995;24:333–5. [PubMed]